Gastrulation Movements: the Logic and the Nuts and Bolts

Developmental Cell, Vol. 8, 305–320, March, 2005, Copyright ©2005 by Elsevier Inc. DOI 10.1016/j.devcel.2005.02.007

Gastrulation Movements: the Logic and the Nuts and Bolts Maria Leptin* Institut für Genetik Universität zu Köln Zülpicherstrasse 47 D-50674 Köln Germany

Gastrulation, the period during the early development of animals when major cell and tissue movements remodel an initially unstructured group of cells, requires coordinated control of different types of cellular activities in different cell populations. A hierarchy of genetic control mechanisms, involving cell signaling and transcriptional regulation, sets up the embryonic axes and specify the territories of the future germ layers. Cells in these territories modulate their cytoskeleton and their adhesive behavior, resulting in shape changes and movement. Similarities among different species in patterning and cell biological mechanisms are beginning to allow us to recognize general, conserved principles and speculate on possible ancestral mechanisms of gastrulation. Introduction: What Is Gastrulation, and Is It the Same in All Animals? The word gastrulation is derived from the Greek word “gaster,” meaning belly or gut (as does the Latin “venter,” also pertinent here). The gastrula was first defined by Haeckel as the embryonic stage at which an infolding that gives rise to the digestive system and musculature can be seen as a distinct cell layer on the inside of the embryo. As it turns out, not all animals go through a stage of precisely this description. It is therefore more sensible to define gastrulation as the transition from a simple, not very highly organized group of cells, usually a hollow epithelial sphere or “blastula” but sometimes a compact ball or even a sheet of cells, to a more complex, organized, and multilayered embryo with the distinguishable germ layers of ectoderm, mesoderm, and endoderm. Even by this definition, a gastrula stage is difficult to define for some organisms, for example, C. elegans; it has therefore been discussed whether all animals do, in fact, gastrulate. The example of certain cnidarians, which create germ layers simply by making cells on the inside of a ball of cells develop according to a different program from those on the outside, shows that layers can, in fact, be created without any apparent cell movement (see Byrum and Martindale, 2004). However, we will see that the processes of gastrulation are very flexible and appear to have changed rapidly during evolution, responding to changes in the environment and the architecture of the egg. It seems reasonable to state as a general rule, therefore, that any animal that ends up with a gut, musculature, and a circulatory system, and a skin and nervous system, i.e., the derivatives of the *Correspondence: [email protected]

Review

three germ layers (or, in the case of diploblastic animals, the two germ layers that make gut and epidermis) must have gastrulated. Gastrulation should be regarded as the sum of the processes that lead to the initial establishment of the germ layers. Gastrulation is clearly not the same in all animals. This is frustrating for the biologist who wishes to discover unifying principles that allow us to understand not only how one given animal develops, but also the underlying rules that explain why it develops in this way and how it relates to other animals, or, more importantly, to a postulated common ancestor. It is not yet clear what the unifying principles or mechanisms of gastrulation may be or whether they even exist. This review will attempt to discuss where such principles may have been identified and might be found in the future and which cell biological mechanisms operating during gastrulation are shared between different species. It cannot cover all possible modes of gastrulation and will therefore concentrate mainly on three species, Xenopus, Drosophila, and the sea urchin, to represent different phyla (Figure 1). However, the variations during gastrulation are such that to say that these animals “represent” their phyla may be inappropriate, as the model organisms chosen for the study of development turn out not to be typical in all aspects. Thus, Drosophila develops as a long germ embryo in which all segments are determined simultaneously and early development is extremely rapid, in contrast to short germ embryos, which develop more slowly, with segments arising sequentially during embryogenesis. The rapid long germ mode of development puts strong constraints on gastrulation, which has to be more efficient and reliable in detail than in the more leisurely developing short germ insects. Xenopus deserves its place because of its historical importance and prevalence as an experimental organism, but unlike most other vertebrates, it has almost no epithelial-mesenchymal transitions during gastrulation; furthermore, it has a yolk-rich egg in which the yolk is partitioned into a subset of cells, whereas birds and fish have yolkrich eggs in which the yolk is not partitioned into the cells that will produce the embryo, and mammals have no yolk at all. Adaptations to the distinct architectures of the eggs have entailed great differences in the geometry of the early embryo and different modes of bringing cells into its inside. The essence of gastrulation is that different cells in the embryo move in different ways and directions. For this to happen, different cells have to be given different instructions, or cell fates. Since disruption of cell fates usually also results in the disruption of gastrulation movements, it is necessary to understand how the cell fates are determined. It has not been easy to disentangle the roles of genes determining fate from those controlling movement, or to establish the genetic cascade from “master regulators” to genes directly controlling cell behavior. Since fate-determining genes are often referred to as genes “controlling gastrulation,” they will

Developmental Cell 306

Figure 1. Fate Maps and Gastrulation Movements of the Sea Urchin, Drosophila, and Xenopus The germ layers are shown in the same colors for all three species (endoderm: green; mesoderm: blue; ectoderm: gray). Note that the scales are not the same for the different organisms. The body axes are marked by A (anterior) and P (posterior), D (dorsal) and V (ventral), AN (animal) and VEG (vegetal). An excellent and more detailed set of descriptions of amphibian and sea urchin gastrulation with many photographs, diagrams and movies can be found at J. Hardin’s sites http://worms.zoology.wisc.edu/urchins/SUgast_intro.html and http://worms.zoology.wisc.edu/frogs/welcome. html. Movies of fly gastrulation and early development are at http://sdb.bio.purdue.edu/dbcinema/kaufman/kaufman.html and http://flymove. uni-muenster.de/Processes/Gastrulation/GastrulAdditional.html. (A–D) Four stages of sea urchin gastrulation (modified from Wolpert, 2001). The first cells to enter the blastocoel are the primary mesenchyme cells (PMC), which delaminate from the vegetal plate of the blastoderm epithelium. The epithelium then forms a thickened placode, which eventually invaginates, probably by apical constriction, to form the archenteron, which lengthens into the blastocoel by convergent extension. The PMC extend long filopodia throughout the blastocoel, which are stabilized when they attach to the animal pole. By contracting, they contribute to extending the archenteron. Finally, the tip of the archenteron fuses with the overlying oral ectoderm to create a new opening, the mouth of the larva. (E–L) Modified from Leptin, 1995. Four stages of Drosophila gastrulation. The left column shows whole embryos, the right shows diagrams of cross-sections. Cells next to each cross-section represent a mesodermal cell undergoing its characteristic shape changes. The first cells to enter the inside of the blastoderm stage embryo are the prospective mesodermal cells on the ventral side of the embryo. They form a deep invagination by apical constriction and baso-apical shortening. Almost simultaneously, the posterior endoderm invaginates by the same type of cell shape changes, carrying the primordial germ cells (pole cells, PC) with it. Cell intercalation in a dorso-ventral direction lengthens the germ band and pushes its posterior end onto the dorsal side. The invaginated mesodermal tube disperses into mesenchymal cells, which spread out to form a single cell layer. (M–T) Modified from Keller and Shook, 2004. Cell rearrangements within the vegetal cell mass (vegetal rotation) initiate the internalization of the endoderm and lead to the subsequent involution movements at the blastopore. At this point, the bottle cells undergo apical constriction, further supporting involution movements. The prospective mesoderm is thereby positioned onto the inside of the blastocoel, where it can begin its fibronectin-dependent, anteriorly directed migration, aided by convergent extension both in the mesoderm and the ectoderm. The boxed region in panel (O) and (S) is shown at higher magnification in (T) to show the overlapping mesodermal cells migrating on the blastocoel roof.

be described here and their roles in different animals compared. Part 1. The Specification of Cells That Participate in Gastrulation: Establishment of the Primordia of the Germ Layers The Determination of the Embryonic Axes The identification of genes responsible for the determination of cell fates in the early embryo has led to the

surprising discovery of homology of genetic pathways in the development of widely separated species. Two axes are used to describe the spatial organization of adult animals: the antero-posterior axis and the dorsoventral axis (the left-right axis is directly defined by these two). “Anterior” is the side with the mouth and the highest concentration of neural tissue; posterior, the opposite end. The Hox gene complexes found in all animals control the organization of cell fates between

Review 307

these poles, but homeotic differences in cell fate along this axis are not a deciding factor for gastrulation. The ventral side in the adult vertebrate body is, as the name indicates, the belly-side, thus, the side facing downward and containing the abdominal organs; the dorsal side is the side with the spine or notochord and the central nervous system, usually the side facing away from the ground. In an adult animal and larva the anterior-posterior and dorso-ventral axes are usually more or less perpendicular to each other. Primates and birds, moving about on their hind legs only, illustrate how confusing these definitions can be: in our case, our “front” sides correspond exactly to our ventral sides. Similarly, it is not necessarily the case that the axes between the primordia of the future anterior and posterior or dorsal and ventral structures in the embryo are perpendicular. Additional axes are therefore often defined, such as the animal-vegetal in amphibians or the oral-aboral in sea urchins. To make matters even more complex, the various axes move relative to each other during gastrulation in ways that often make it difficult to understand possible homologies between distant phyla. The definitions of the dorsal and ventral sides of vertebrates as the sides facing upwards and downwards were simply transferred to invertebrates. Hence, the central nervous system (located dorsally in vertebrates), which lies on the side of the animal facing the ground in most invertebrates, became the “ventral nerve chord,” and the heart, considered a ventral structure in vertebrates, became the “dorsal vessel” in worms and insects. As we now know, this nomenclature completely disregards evolutionary relationships, and it would be more appropriate to think of invertebrates moving about with their dorsal side facing downward and their belly up (Arendt and Nübler-Jung, 1997). This insight came from the discovery of the genes that determine cell fates along the dorso-ventral axis in vertebrates and insects, specifically from the study of mutants in Drosophila and the search for the molecular basis of Spemann’s Organizer in frog embryos. The study of gastrulation in vertebrates is closely linked with that of the induction of the mesodermal cell fate. This is mainly because of the property of the Organizer, which, when transplanted into a recipient Xenopus embryo, can induce gastrulation movements as well as mesoderm respecification at ectopic sites. The effect of the Organizer on frog gastrulation movements is not direct (although in zebrafish some possible direct involvement is emerging [Yamashita et al., 2002]), and signals emitted from the Organizer mainly subdivide the previously specified mesoderm primordium into regions with different fates (reviewed in Lemaire and Yasuo, 1998). Nevertheless, the discovery of these signals, following the identification of genes affecting patterning of the dorso-ventral axis in Drosophila, which encode similar molecules, have shown that early embryonic axis determination uses a molecular mechanism conserved from a common ancestor. Cell fates along the dorso-ventral axis in both organisms are determined by a gradient of the TGF-β family member (Dpp in flies and its homolog BMP-2/4 in amphibians [De Robertis and Sasai, 1996; Holley and Ferguson, 1997]) (Figure 2). In both cases, the gradient is established with the help of the BMP-2/4 antagonist Chordin or Sog, as well as other modulators, and this

cassette of morphogens and modulators is conserved in other invertebrates as well. For example, in sea urchins, BMP-2/4 is involved in differential fate determination along the oral-aboral axis (Duboc et al., 2004). Whereas the high point of BMP2/4 activity is on the ventral side in vertebrates, in Drosophila it is on the “dorsal” side (Figure 2). This is the basis for the view that insects develop and live upside-down relative to vertebrates. It is important to remember that there is a snag in this apparently perfect unification of the description of at least one aspect of early development: the patterning under the control of Dpp and Sog in Drosophila concerns exclusively the ectoderm, as it does in the sea urchin, whereas in chordates, it is responsible for the subdivision of the endomesoderm as well as the ectoderm. This does not invalidate the unification, but illustrates how very much the deployment of regulatory cassettes has been varied during the evolution of the species we study. Although the BMP-chordin axis was the first to be discovered as a conserved molecular principle of embryonic development, it is not the first to be established in the vertebrate embryo. BMP-2/4 subdivides a region that has been set up under the control of earlier patterning events, as have the expression of BMP2/4 and chordin themselves. These patterning systems, too, are conserved, at least in the deuterostomes. They are mediated by two further signaling molecules: Wnt and Nodal (reviewed in Heasman, 1997; Schier, 2003) (Figure 2). In the case of Wnt signaling, it may be more appropriate to speak of stabilization and nuclear localization of β-catenin, as this is the conserved component of the pathway, and Wnt-dependent receptor activation has not been shown in all cases to be the trigger for β-catenin stabilization. Xenopus will serve as the prototype to discuss these two systems. In the Xenopus embryo, nodal-related factors (Xnrs), also members of the BMP family, and β-catenin stabilization act in almost perpendicular axes (Figure 2). Xnrs originate from the vegetal pole, where they are transcribed under the control of the maternal transcription factor VegT (Clements et al., 1999; Kofron et al., 1999) to induce mesoderm in the ring of cells above. β-catenin stabilization defines the dorsal side. It modifies the effect of the Xnrs signal received by the mesoderm in this region, which is thereby induced to become the Spemann-Mangold Organizer (Kimelman et al., 1992). In addition, Xnrs are direct targets of β-catenin, such that a dorso-ventral gradient of Nodal signaling is established. Nodal and canonical Wnt signaling also pattern the sea urchin embryo, an example of a nonchordate deuterostome (Duboc et al., 2004; Logan et al., 1999; Wikramanayake et al., 1998). Stabilization of β-catenin is seen in the vegetal micromeres, which give rise to the region that invaginates during gastrulation and later forms the gut. Nodal sets up distinct cell fates along the perpendicular, oral-aboral axis (Figure 2). This is different from the situation in Xenopus in that Nodal acts along the same axis as BMP-2/4 signaling in the sea urchin. Furthermore, the determination of the oralaboral axis is not independent of the vegetal-animal signaling system. If stabilization of β-catenin is induced ubiquitously or is inhibited, this results not only in an expansion or reduction of vegetal fates, respectively, but also the loss of oral-aboral asymmetry (Angerer and

Developmental Cell 308

Figure 2. Patterning Molecules and Germ Layer Determinants in the Sea Urchin, Drosophila, and Xenopus Arrows denote the activity gradients (not necessarily expression patterns!) of β-catenin (black), Xenopus nodal-related factors (white), BPM2/4 or dpp (solid pink), and chordin/sog (open pink). (A) Sea urchin. β-catenin is stabilized in the vegetal micromeres (top) and determines the endoderm as well as the patterning of the adjacent prospective ectoderm. The endoderm expresses GATAe (green). Nodal and BMP2/4 (along with a number of other genes, including goosecoid) are expressed in the future oral ectoderm (purple) and pattern the oral-aboral axis. Brachyury (yellow) is expressed on the oral side as well as in a ring around the endoderm primordium, similar to the situtation at the posterior end of the Drosophila embryo. (B) Drosophila. β-catenin or Nodal signaling has not been found to participate in primary axis determination in the Drosophila embryo. The ectoderm of the embryo is patterned by opposing activities of Dpp (BMP2/4) and Sog (chordin). The Brachyury-homolog brachyenteron (yellow) determines the ectodermal hindgut; the endoderm is marked by the GATA factor Serpent (green), the mesoderm by Snail and Twist (light and dark blue), both of which are also expressed in the anterior endoderm. (C) Xenopus. The early frog egg is patterned along two axes by Nodal related factors (Xnrs) and β−catenin stabilized at a position opposite the sperm entry site. Xnrs are responsible for defining the future mesoderm, which expresses brachyury (yellow). Xnrs and β-catenin together determine the Spemann-Mangold Organizer region, which expresses nodal, goosecoid, siamois, and other genes (purple) and serves as a source of chordin to counteract BMP2/4 in patterning the dorso-ventral axis (see also De Robertis and Kuroda, 2004).

Angerer, 2003). Wnt signaling seems to be necessary to set up a zone of competence in which the oral-aboral patterning system can operate, and nodal expression itself is part of the system that depends on Wnt signaling. This also differs from the situation in vertebrates, where β-catenin and Nodal are at least initially independent of each other. However, vertebrates do seem to retain a trace of this congruency in that β-catenin does influence the level of Nodal in Xenopus (Agius et al., 2000; Hyde and Old, 2000; Takahashi et al., 2000; Xanthos et al., 2002; Yang et al., 2002a) and is also needed for proper expression of the Nodal-homolog Squint in zebrafish (Solnica-Krezel and Driever, 2001). Nodal is the earliest asymmetrically expressed gene in the sea urchin, but the primary source of asymmetry along the oral-aboral axis is not known. The oldest axis is likely the one determined by β-catenin stabilization, since it exists even in radially symmetric, diploblastic animals like the sea anemone (Wikramanayake et al., 2003), which have only one axis and only two germ layers. Here β-catenin stabilization specifies the endoderm, which may be its true ancestral function. This is also consistent with the role in ascidians and C. elegans, where endoderm specification depends on stabilization of β-catenin (Imai et al., 2000; Rocheleau et al., 1997; Thorpe et al., 1997). Surprisingly, a role for β-catenin stabilization during early axis formation or germ layer determination has not been found in Drosophila or other higher invertebrates, nor is an equivalent of Nodal signaling known. Transcription Factors as Determinants for the Germ Layers The result of axis determination is the definition of regions within the embryo that will give rise to the germ layers. These can be recognized by the expression of specific transcription factors, which then direct the further development of the germ layer (often in multiply backed-up pathways, so that loss of a single gene does

not necessarily lead to complete loss of the germ layer). In Drosophila, the mesodermal primordium coincides largely with the region expressing Snail and Twist (Kosman et al., 1991; Leptin, 1991), while the GATA-factor Serpent defines the endoderm (Reuter, 1994) (Figure 2). Brachyury is expressed in a ring around the posterior endoderm that will give rise to the ectodermal hindgut (Kispert et al., 1994). In vertebrates, the expression domain of Brachyury, set up under the control of Nodal, encompasses the mesodermal primordium (SchulteMerker et al., 1992; Smith et al., 1991; Wilkinson et al., 1990). The subset of mesodermal cells that constitute the Organizer express Goosecoid (Cho et al., 1991), activated by Nodal together with nuclear β-catenin. Snail and Twist, the mesodermal determinants in Drosophila, have important morphogenetic functions in the vertebrate mesoderm only at later stages. The future endoderm cannot be unambiguously distinguished from the mesoderm early. In sea urchins, the situation is different yet again. The homologs of the mesodermal determinants Goosecoid and Brachyury are expressed in the oral ectoderm (Angerer et al., 2001; Croce et al., 2003; Gross and McClay, 2001), which seems to function as a signaling center similar to the Organizer but here patterns the ectoderm. Brachyury is, in addition, expressed in the endoderm, a ring at the periphery of the future archenteron, in a pattern analogous to brachyury expression in Drosophila. As in Drosophila, the sea urchin endoderm also expresses a GATA-factor (Figure 2). All of these transcription factors participate in directing the morphogenetic behavior of the regions in which they are expressed. It has even been proposed that the ancestral role of Brachyury was to control the cell movements that internalize cells, rather than the control of cell differentiation (Gross and McClay, 2001). Some of the factors expressed in the primordia of the different germ layers are shared between different species, but the only recognizable general principle of

Review 309

germ layer determination seems to be to make certain regions of the embryo distinct from the ectoderm. This might indicate that the distinction between endoderm and mesoderm is not conserved. An ancestral endomesoderm with absorptive properties characteristic for endodermal tissue and contractile properties typical of mesoderm, still found, for example, in the sea anemone, might have become subdivided in different ways during the evolution of new phyla (also discussed more extensively by others [Ball et al., 2004; Martindale et al., 2004; Rodaway and Patient, 2001]). In many species, endoderm and mesoderm arise as common primordia and are separated late (e.g., frog and fish) or share regulators (e.g., Snail and Twist in the mesoderm and the anterior endoderm in Drosophila). This may also mean that we cannot expect to find cell behaviors that are restricted to one germ layer. Thus, convergent extension movements occur in the endoderm in sea urchins, the ectoderm in Drosophila, and in both mesoderm and ectoderm in Xenopus. Nevertheless, in the germ layers of extant species, transcription factors control cell behavior — perhaps by specific modulation of an underlying ancestral behavior — and are responsible for the movements of cells. Part 2: Gastrulation Movements and Their Control Types of Cell Movement The central issue in gastrulation is movement. Excellent descriptions of cell movement during gastrulation have been generated, using the whole range of available microscopic techniques, but surprisingly little is known about the generation of force and the molecular mechanisms driving and directing movement. Four major modes of movement contribute to the rearrangement of cell groups during gastrulation: bending of epithelial cell sheets, rearrangement of cells within the plane of epithelial or pseudo-epithelial sheets, dissociation of cells from epithelial structures (this includes delamination of single cells as well as epithelial-mesenchymal transitions [EMTs] of whole epithelia), and cell migration of individual cells and groups of cells. Epithelial Bending. The bending of epithelial sheets, often one of the earliest gastrulation movements to be seen, provides a way of translocating large groups of cells from the surface into the interior of an epithelial sphere, creating a two-layered structure. In sea urchins, it occurs during the invagination of the archenteron; in Drosophila, it represents the first step of mesoderm invagination (formation of the ventral furrow) (Figures 1C and 1J). The initiation of involution in amphibians, itself a combination of various types of cell movements, involves a small group of cells, the bottle cells, undergoing cell shape changes typical of bending cell sheets (Figures 1N and 1R). None of these processes are understood in detail, but they share a number of characteristics. Typically, the apical circumferences of the invaginating cells constrict, probably induced by contractile acto-myosin networks, causing the cells to become wedge-shaped or bottle-shaped. This shape change is often accompanied by lengthening of the cells along the apical-basal axis, which creates a thick epithelial placode and displacement of nuclei away from the apical sides of the cells (for example, in the

Drosophila ventral furrow; see Figures 1I–M). Modeling shows that apical constrictions can be sufficient to cause invagination. In one type of model, a wave of constrictions is triggered by a single cell at the center of a prospective pit, making the initial constriction (Odell et al., 1981). This causes stretching in neighboring cells, which acts as a trigger for their constriction, which in turn stretches the next row of neighbors, etc. Attractive as it is, this model has not found experimental support. Instead, it appears that cells constrict autonomously under the direct control of a program dictated by their fate-determining transcription factors. For example, the cell shape changes associated with ventral furrow formation in Drosophila depend on the transcription factors Snail and Twist, and single wildtype cells in a snail twist mutant mesoderm are able to undergo their typical shape changes while the surrounding mutant tissue remains inert (Leptin and Roth, 1994). Also, in sea urchins, the shape changes are local and cell autonomous and no global coordination is necessary: half invaginations can be made in embryos in which the developmental program of one blastomere has been disturbed (Logan et al., 1999). Paradoxically, it appears that epithelial bending, although it is the initiating step of mesoderm invagination in Xenopus and Drosophila, may not be an essential step for mesoderm internalization. In Xenopus, microablation of bottle cells does not prevent the internalization of the mesodermal germ layer (Keller, 1981), and in Drosophila, the cells of the mesoderm primordium appear to be able to become internalized even if the orderly sequence of early cell shape changes is disrupted, for example, by untimely cell divisions (Seher and Leptin, 2000). Thus, at least in these cases, apical constriction and epithelial bending may serve primarily to place the mesodermal cells in an advantageous starting position for their later migration on the inner surface of the ectoderm. Cell Rearrangement within Sheets. Rearrangement of cells within epithelia or layers of mesenchymal cells is used to change the dimensions of cell sheets. Cell intercalation can turn dome-shaped invaginations into long tubes, as during the formation of the archenteron in sea urchins (Hardin, 1989), and short, wide regions into long narrow areas, as during convergence and extension in vertebrate gastrulation and neurulation (Shih and Keller, 1992a; Shih and Keller, 1992b; Wilson and Keller, 1991) or during “germ band extension,” the lengthening of the segmented part of the Drosophila embryo (Irvine and Wieschaus, 1994). During convergence and extension, cells within a mesenchymal sheet change their positions relative to each other, such that they intercalate between each other with a preferred direction. Convergence and extension movements in frogs and fish continue beyond the stage at which the germ layers have been established. Cell intercalation in Drosophila does not affect germ layer formation at all, but the movements are involved in the overall shaping of the embryo and are therefore relevant for gastrulation. Cell intercalation typically uses interactions between neighboring cells within the cell sheet, as opposed to interactions with an underlying substratum, to generate the force and directionality of the movement. This is most clearly demonstrated by the fact that explants of

Developmental Cell 310

Figure 3. Mechanisms of Cell Intercalation during Convergence and Extension Movements: Two Ways of Changing Positions of Cells (A) During convergent extension movements in Xenopus, the initially undirected protrusive activity of filopodia becomes polarized in a dorso-ventral direction. Force exerted along this axis draws cells between each other. (B) During germ band extension in the Drosophila embryo, myosin becomes enriched at the anterior and posterior borders of the blastoderm epithelial cells. These sides then contract to a point, bringing previously separate, dorsally and ventrally located cells near each other. The initial anterior-posterior neighbors lose contact, allowing a new area of contact to be established between the dorsal and ventral cells.

mesodermal cells from a Xenopus embryo can undergo intercalation movements in the absence of a substratum (Elul and Keller, 2000; Keller et al., 1989). The processes of cell intercalation within cell sheets in different animals and tissues look similar, but it is not clear whether they are based on the same cell biological mechanisms. They can occur in proper epithelia, as in the Drosophila ectoderm, in which case the integrity of the epithelium is maintained during the movements. They can also occur in more loosely associated mesenchymal masses, as in the Xenopus mesoderm during convergent extension, a very dynamic process with attachments between neighboring cells being made, dissolved, and reestablished rapidly. The elucidation of the underlying mechanisms has benefited tremendously from in vivo time-lapse microscopy. In convergence and extension during early vertebrate gastrulation, motility is associated with the extension of lamellipodia (Wilson and Keller, 1991). Cells that initially produce lamellipodia at random positions of their circumference begin to protrude them in a polarized fashion, preferentially in the medio-lateral axis (Shih and Keller, 1992a; Shih and Keller, 1992b). Tensile forces can then bring about polarized intercalation of cells (Figure 3A). By contrast, in the ectoderm of the Drosophila embryo where tall columnar epithelial cells are packed in a tight hexagonal pattern, no protrusions are seen. Instead, cell rearrangement is mediated by cell shape changes and concomitant changes in the contacting surfaces between neighboring cells (Lecuit, 2004). Cells change from a hexagonal shape, where they have on average six neighbors (Figure 3B), through a diamondshape with four neighbors in which they have minimized their contact with their anterior and posterior neighbors to a point, and then back to a differently oriented hexagonal shape in which the contact point with dorsal and ventral neighbors is expanded to a large area. Many of the molecules involved in convergence and extension have been identified (see below), but one of the main open questions is how direction and order are imposed on cell intercalation. Signals are needed that determine in which direction the bipolar lamellipodia

are extended and when convergence and extension behavior stops. A solution may have been found for the latter problem in the mesoderm in Xenopus. Cells whose lamellipodia have reached the notochord boundary strive to expand their contact with this boundary, remain attached to it, and cease their protrusive activity, a process Keller named “boundary capture” (Shih and Keller, 1992b). Thus, they become immobile. It is possible that this behavior might use a similar subcellular mechanism as MDCK epithelial cells in culture, which, once they have made contact with each other, use rac-dependent cytoskeletal remodeling to maximize their contact (Ehrlich et al., 2002). Almost nothing is known about the signals that determine the direction of intercalation. There are hints in Drosophila that it is imposed by the anterior-posterior patterning system in the embryo (Blankenship and Wieschaus, 2001; Zallen and Wieschaus, 2004). MyosinII is enriched at the anterior and posterior contacts of the cells in the blastoderm epithelium, where it presumably mediates contraction of the contact area (Lecuit, 2004), while PAR-3, a molecule involved in establishing polarity in many cell types, is enriched at the dorsal and ventral contacts (Zallen and Wieschaus, 2004). This distribution, as well as dorso-ventral intercalation behavior, is abolished in mutant embryos in which segmentation is disrupted. It remains to be seen whether this type of intercalation is related to that seen in vertebrate convergent extension, but the tantalizing discovery that antero-posterior tissue polarity affects mesoderm convergent extension in frogs suggests more parallels between the two processes than were initally apparent (Ninomiya et al., 2004). Epithelial to Mesenchymal Transitions. The third type of movement allows cells to move out of epithelia. Two cases can be distinguished, the delamination of single cells and the transition of part or all of an epithelium to a mesenchymal state. Both can be considered to be epithelial-mesenchymal transitions (EMT; comprehensively reviewed in (Shook and Keller, 2003)). Delamination of single cells occurs, for example, during the ingression of the PMC in sea urchin embryos, the ingression of endodermal precursors in C. elegans, and the ingression of mesodermal cells from the epiblast in the avian embryo or the blastocoel roof in amphibians.

Review 311

Epithelial cells are connected to each other near their apical surfaces (which in the case of the blastula stage embryo faces outward) via septate or tight junctions and adherens junctions. They are often supported by an extracellular matrix on their basal sides and may be in contact with protective layers on the apical outer side. These junctions must be dissolved in the cells, leaving the epithelium but maintained or reestablished between the cells that are left behind (except in cases where the complete epithelium is transformed to mesenchyme). One of the best-described examples of delamination of single cells is the ingression of the primary mesenchyme cells in sea urchins (reviewed in McClay et al., 2004). The delaminating cell first constricts at its apical side to from a “bottle cell.” This reduces the size of the hole that will be left behind in the epithelium, often further helped by the neighboring cells sending out extensions toward each other. At the same time the PMCs lose affinity for the outer protective hyaline layer, which is contacted by the apical side of the cell, and acquire affinity for fibronectin found on the inside in the basal lamina (Fink and McClay, 1985). Disassembly of adherens junctions and loss of adhesion with neighboring cells allows PMCs to leave the epithelium and become mesenchymal, migratory cells (summarized with further citations in Shook and Keller, 2003). This sequence of events can also occur in groups of adjacent cells. In the extreme case of a large part of an epithelium following this sequence, the first step — apical constriction — leads to the bending and invagination of the epithelium. Events like the internalisation of the mesoderm in Drosophila could perhaps be viewed as a group of cells undergoing the changes typical for ingression; the formation of the ventral furrow corresponds to the formation of bottle cells, which in this case leads to the internalisation of the whole epithelial primordium. Only then are the junctions between the cells dissolved and the mesoderm disperses into individual, mesenchymal cells, which move away from the site of invagination. One may speculate whether ingression of single cells via bottle cell formation and loss of adhesion is the oldest mechanism for cells to form a second layer within a blastula-like ancestral organism. The grouping of such cells in one place in the embryo may then have enabled epithelial invaginations, leading to the formation of an archenteron. Certainly bottle cell formation and shifting of adhesion sites appears to be a mechanism for morphogenesis extending beyond the animal kingdom (Nishii et al., 2003). Cell Migration. The fourth type of cell movement, cell migration, involves cells moving across a substratum, which they use for translocation. It occurs mainly at later stages of gastrulation, when the early gastrulation movements have placed cells of one germ layer onto the substratum of another. Examples include spreading of the anterior mesoderm on the blastocoel roof in Xenopus, spreading of the mesoderm on the ectoderm in Drosophila, migration of neural crest cells on somites in vertebrates, and many others. Cell migration also makes a contribution to extension movements during convergence and extension in the fish, and some of the involution movements at the Xenopus blastopore may qualify as migration.

Cells may migrate as individuals, but in many cases they move as groups, with the group acting together as a community rather than as a collection of independent cells on the way to the same target. This was specifically shown in the case of the mesoderm migrating on the blastocoel roof of the Xenopus embryo. Explants can move efficiently and rapidly on substrates in vitro, but if single cells or small sectors of mesodermal tissue are explanted, they advance more slowly than large sheets of cells with an intact migrating front (Davidson et al., 2002b). The leading cells maintain contact with those following, their posterior parts overlapping the lamellipodia of the cells behind. When these contacts are lost, the migrating cells also lose polarity and the ability to follow guidance cues (Winklbauer et al., 1992). Far from being a specialty of Xenopus cell migration during gastrulation, movement of cells in groups may rather be the rule. It has also been observed for germ cells in vertebrates (Gomperts et al., 1994), and, more recently, even for metastatic tumor cells migrating in three dimensional matrices in vitro (Friedl, 2004). Cell migration usually is thought to be guided by chemotaxis. Whereas chemotaxis has been well established as a mechanism guiding fibroblasts and cells of the blood and lymphoid system and operates during a number of late morphogenetic processes (e.g., migration of vertebrate limb mesenchyme, Drosophila border cells in the ovary and tracheal cells in the embryo), few clear cases have been demonstrated during gastrulation, although the presence in the embryo of typical chemotactic ligands such as Slit and Ephrin is suggestive (Oates et al., 1999; Yeo et al., 2001). One important example of chemotactic migration appears to the behavior of cells from the primitive streak in avian embryos, after they have undergone EMT to leave the streak. Experiments in chicks show that members of the FGF family have the ability to provide instructive, directional signals (Yang et al., 2002b). By contrast, the effects of FGF on gastrulation movements in the mouse are consistent with a permissive role of FGF-signaling, that of inducing the ability of cells to leave the streak by triggering EMT via downregulation of cadherin levels (see below). However, directionality of movement need not be caused by distant chemotactic signals. Simple spatial constraints can lead to an apparently directed migration. Both in Xenopus and Drosophila, the mass of mesodermal cells initially lies near the site of invagination, and the only direction available for migration is away from this site. Thus, any signal inducing motility would automatically induce directional movement. If, in addition, motility continued until each cell had reached a target tissue that it “liked” better than its mesodermal neighbors — for example, the ectoderm in the case of the Drosophila embryo — and only then ceased (as in the case of boundary capture), this would lead to the dispersal away from the site of ingression and the establishment of a single cell layer. More interestingly, local properties of the substratum can influence the direction of migration. The extracellular matrix of the basal side of the Xenopus blastocoel roof can be transferred to a glass surface and act as substratum for the migration of explanted mesodermal cells. Remarkably, mesodermal cells plated on such a substratum migrate

Developmental Cell 312

toward the point of the ECM that was derived from the region closest to the animal pole, i.e., in the correct direction (Winklbauer and Nagel, 1991). The recent finding that PDGF-A and its receptor are needed for this directionality to be established in the ectoderm and interpreted by the migrating mesoderm may point to a chemotactic input, but it may also be that PDGF provides the local polarity cue fixed in the matrix (Montero et al., 2003; Nagel et al., 2004). Combining and Integrating Different Cell Movements I have divided the types of cell movements during gastrulation into four groups. Obviously, it is the eye and the mind of the scientist that desire such classifications, whereas the distinctions may not always be so clear in the developing embryo. The example of bending and invagination of an epithelial sheet as being interpretable as the first stage of EMT within a coherent group of cells has been mentioned. Cells can switch from one type of morphogenetic behavior to another: a balance of cell migration and convergent extensionstyle neighbor exchange shapes the nascent mesodermal cell layer in zebrafish, with different degrees of migration occurring in different parts of the mesoderm. For example, lateral cells initially migrate dorsally and only later begin to show polarized, mediolateral intercalation behavior (Myers et al., 2002b). Parallel or overlapping types of movement can sometimes be separated experimentally (vindicating the scientist’s reductionist view), for example, by mutations or inhibitors affecting only one type of movement. In Drosophila, mutations affecting anterior-posterior patterning of the embryo affect germ band extension but not the invagination of the mesoderm or endoderm. Conversely, mutations of snail or twist completely abolish the mesoderm, but posterior endoderm formation and germ band extension are unaffected. In fish and frogs, the contributions of migration and convergent extension are beginning to be distinguishable by the molecules required for only one or the other. The zebrafish mutants in the gene no tail, which codes for brachuryhomolog, have defects only in convergence, but not in extension movements (Glickman et al., 2003; Myers et al., 2002b). Interfering with the production by mesodermal cells of hyaluronan, a cell surface polysaccharide previously implicated in cell adhesion and migration, has the same effect (Bakkers et al., 2004). Conversely, activation of the PDGF-A receptor in mesodermal cells is required for the migration of anterior mesodermal cells toward the animal pole in the Xenopus embryo, but not for convergence (Nagel et al., 2004). Intercellular and Subcellular Events What is the cell biological basis for the various types of movement, and do similar processes in different groups of animals use the same mechanisms? Obvious candidates for mediators of cell movement are the extracellular matrix and cellular receptors for its components, the cytoskeleton, cell adhesion systems, and chemoattractants and their signal transmission systems. All of these have been studied in many cell types in vivo and in vitro, and their contributions to cell behavior are well understood. Their roles in gastrulation will be listed briefly here with an emphasis on recent new insights. The Extracellular Matrix. The extracellular matrix serves many functions. It protects and supports epithe-

lia and participates in the delivery of secreted signaling molecules; its own components, recognized by cells via integrins, have signaling functions as well. It also acts as a barrier to cell movement. In Drosophila, the matrix and its receptors have been found to play no role in early gastrulation movement, consistent with the fact that the ECM becomes detectable only after gastrulation is well under way. In Xenopus embryos, the migration of the anterior mesoderm and convergent extension movements depend on fibronectin secreted by the blastocoel roof (Marsden and DeSimone, 2001). Purified fibronectin supports random mesodermal cell motility in vitro, whereas the fibrils of the matrix mediate directional cell movement (Winklbauer and Nagel, 1991). It is not yet known how polarity is imposed on the fibronectin matrix of the blastocoel roof, but PDGF may have a critical role (Nagel et al., 2004). In the sea urchin, a basal lamina underlies the blastula epithelium, which must be penetrated by the ingressing primary mesenchyme cells (PMCs). These cells acquire adhesive properties for fibronectin during ingression (Fink and McClay, 1985), suggesting that the ECM supports their further movement, for example, their migration toward the animal pole, but this has not been shown experimentally. The matrix covering the outside of the embryo has been implicated in triggering or supporting the invagination of the archenteron: the swelling of secreted hygroscopic matrix material may push the underlying cells toward the interior of the blastula (Davidson et al., 1999; Lane et al., 1993). Interactions of the prospective endodermal cells with the matrix may contribute to bottle cell formation (Kimberly and Hardin, 1998; Marsden and Burke, 1998). The secondary mesenchyme cells that draw the archenteron inwards attach their cytoplasmic extensions to the basal lamina of the ectoderm, but the precise molecular interactions have not been determined. Cytoskeleton. The cytoskeleton is involved in nearly all aspects of cellular morphogenetic behavior and, precisely for this reason, cannot be analyzed easily in specific processes in vivo. Disrupting the actin cytoskeleton with drugs leads to drastic effects in Drosophila (mainly because cellularization is blocked), but, surprisingly, does not interfere with the early steps of gastrulation in sea urchins (Lane et al., 1993). Later events, like the pulling of the archenteron by the filopodia of the secondary mesenchyme cells, fail to occur in the presence of cytoskeletal inhibitors (Miller et al., 1995). Because of the global effects of disrupting the actin cytoskeleton and the difficulties in interpreting the resulting phenotypes, the role of actin has mostly been studied by the analysis of its modifiers and regulators, such as the rho family GTPases, for which numerous functions during gastrulation have been described. One of the more dramatic cases is DRhoGEF2, which is essential specifically for the cell shape changes that drive the invagination of the mesoderm and the endoderm in Drosophila (Barrett et al., 1997; Häcker and Perrimon, 1998). DRhoGEF2 is present throughout the egg, and it is not yet clear how it is activated in mesodermal and endodermal cells. MyosinII, one of the most important modulators of the actin cytoskeleton, is assumed to be involved in

Review 313

apical constrictions and perhaps in generating the tensile force during convergent extension in vertebrates, but no experimental evidence has been provided to support this. Recent studies in Drosophila have shown myosinII to be involved in force generation during the cell intercalation movements that drive germ band extension movements (Lecuit, 2004). MyosinII is asymmetrically distributed within the plane of the ectodermal epithelium, such that it is enriched at the sides of the cells that shorten to reduce contact between neighboring cells. If myosin function is impaired, so is the process of shortening. Signaling Systems and Chemoattractants. Cell signaling molecules, especially those of the BMP, Wnt, and FGF families, play major roles in gastrulation. One important goal will be to disentangle their functions in fate determination from those that directly affect morphogenesis. BMP family members are mostly responsible for cell fate determination, whereas Wnt signaling and FGF signaling have many functions both in cell fate determination and in morphogenesis. BMPs clearly affect the behavior of mesodermal cells in vertebrates (Graff et al., 1994). They are involved in the subdivision of the mesoderm in Xenopus and zebrafish, and different subpopulations of mesodermal cells along the dorso-ventral axis show different morphogenetic activities (Keller and Danilchik, 1988; Myers et al., 2002b), but there is no evidence that BMP signaling affects these behaviors directly rather than by controlling the expression of other genes. The events downstream of FGF signaling that allow mesodermal cell migration have been worked out for cells ingressing through the primitive streak in the mouse embryo. In FGF or FGF-receptor loss-of-function mutants, these cells fail to migrate away from their site of ingression, but when compared with wild-type cells for their ability to migrate in vitro, no differences were seen (Ciruna and Rossant, 2001). The defects in gastrulation turned out to result from the cells’ inability to deepithelialize and leave the primitive streak because they could not downregulate cadherin. This, in turn, was due to the transcription factor Snail not being turned on (Ciruna and Rossant, 2001), which is known to control EMT by downregulation of E-cadherin (Cano et al., 2000). A more direct effect of FGF on cellular behavior is likely to be responsible for chemotactic functions, but these mechanisms remain to be elucidated. In the Drosophila embryo, FGF signaling directs the invaginated mesodermal cells toward the ectoderm, enabling them to spread efficiently over the inner surface of the ectoderm (Beiman et al., 1996; Gryzik and Müller, 2004; Shishido et al., 1997; Stathopoulos et al., 2004). The timing of this process is such that it cannot be regulated via fate changes, and, indeed, it does not act via MAPK and, therefore, probably not by gene activation (Wilson et al., 2005). Instead, a requirement for the rhoGEF Pebble (Schumacher et al., 2004; Smallhorn et al., 2004) suggests that IGF signaling directly mediates changes in the cytoskeleton. Wnt signaling also has roles both in cell fate determination and morphogenesis. The divide between the two functions roughly parallels the divide between the two intracellular branches of the pathway: the canonical Wnt signaling pathway is mainly associated with cell

fate determination, while the noncanonical or planar cell polarity (PCP) pathway is associated with morphogenetic activities. It is important to remember that although the pathways are usually referred to as Wnt signaling pathways, it is not necessarily the case that they are triggered by localized extracellular signals in all situations. For example, β-catenin stabilization on the future dorsal side of the Xenopus embryo is achieved by cortical rotation after sperm entry, and components of the PCP pathway have been found to play roles in morphogenetic events for which no involvement of Wnt-like ligands can be shown. The components of the PCP pathway, their epistatic relationships, and their requirement for planar cell polarity have been worked out in Drosophila, Xenopus, and zebrafish and have been the subject of recent excellent reviews (Fanto and McNeill, 2004; McEwen and Peifer, 2000; Veeman et al., 2003). Briefly, reception and interpretation of the Wnt signal requires the receptor Frizzled along with the extracellular heparan sulfate proteoglycan Knypek and the transmembrane protein Van Gogh/Strabismus/Trilobite. The signal is transmitted by the large docking protein Dishevelled and can act through or in parallel with various cytoskeletal regulators including cdc42; Rac; Rho and Rho-kinase, MARCKS; jnk; and daam (Bakkers et al., 2004; Choi and Han, 2002; Habas et al., 2003; Habas et al., 2001; Iioka et al., 2004; Marlow et al., 2002; Yamanaka et al., 2002). Calcium signaling has also been shown to play a role in PCP signaling in vertebrates (Wallingford et al., 2001). The best-studied role of the PCP pathway during gastrulation is during convergent extension movements, where it is needed for the polar extension of membrane protrusions. Analysis of mutants in zebrafish and morpholino studies in Xenopus have shown that the disruption of components of the PCP pathway, such as Wnt-11 (silberblick) and Wnt5 (pipetail), knypek, frz, strabismus/vangogh/trilobite, dvl, and prickle disrupts gastrulation movements without affecting cell fate (Carreira-Barbosa et al., 2003; Djiane et al., 2000, Wallingford et al., 2000 Heisenberg et al., 2000; Jessen et al., 2002; Kilian et al., 2003; Sokol, 1996; Topczewski et al., 2001). The cellular basis of this defect is the disruption of cell polarity in intercalating cells. The formation of membrane protrusions as such is not impaired, but the protrusions are no longer aligned along a medial-lateral axis, nor does the cell body elongate along a medio-lateral axis. Thus, once a polarity signal has been received, the components of the PCP pathway are clearly necessary to implement the polarity of the rac-dependent membrane protrusions, but no satisfactory explanation for movement mediated by PCP can yet be derived from its function. It is neither completely clear what cell biological events act downstream of the PCP pathway, nor, more importantly, what provides the asymmetry in the first place, especially in the case of convergence and extension movements. It is puzzling in this context that the movements are apparently cell autonomous; Xenopus explants suggest that the region by itself is responsible for convergent extension behavior with no need for external signals (Wilson and Keller, 1991). Furthermore, constitutively active or ubiquitously expressed components of the planar cell polarity pathway can rescue defects caused by loss of the Wnt sig-

Developmental Cell 314

nal (Heisenberg et al., 2000; Kilian et al., 2003; Marlow et al., 2002), suggesting that directionality of polarity is mediated by a different signal and function of the PCP pathway is necessary for the interpretation of the signal. This indicates that the directionality was imposed on this region at an earlier point and is not needed at the time when the PCP pathway mediates polar movement. There is now good evidence that a major contribution comes from the anterior-posterior patterning of the notochord primordium by activin-like factors (Ninomiya et al., 2004). The PCP pathway and Wnt signaling also operate in morphogenetic mechanisms other than convergent extension. Fz controls the separation of the mesodermal and ectodermal cell layers during mesoderm migration on the blastocoel roof (Winklbauer et al., 2001) and Wnt signaling also plays a role in the crawling of hypoblast cells on the epiblast immediately after involution (Ulrich et al., 2003), a movement distinct from that during convergent extension. In this case, the ligand Wnt11 is made by cells in the epiblast and acts on cells in the hypoblast. While this might sound like a case of chemotaxis, this doesn’t appear to be how Wnt11 functions in this case. Instead, it affects the adhesive properties of the hypoblast cells after they have made contact with the epiblast (C.P. Heisenberg, personal communication). Indeed, Wnt11 is unlikely to act as a chemotactic agent, as ubiquitous expression rescues the silberblick mutant phenotype in zebrafish (Heisenberg et al., 2000; Kilian et al., 2003). This function, too, is more likely to be permissive than instructive. However, overexpression of components of the noncanonical Wnt pathway can cause similar defects as their loss (e.g., Moon et al., 1993), suggesting that some aspect of directionality of movement may be due to specific spatial distributions of Wnts. Cell-Cell Adhesion Molecules. The establishment, maintenance, and severing of contacts between cells is important during all morphogenetic events. During gastrulation, modulation of adhesion plays major roles in cell rearrangements within cell sheets and in epithelial mesenchymal transitions (reviewed in Shook and Keller, 2003). During gastrulation in amniotes, cells ingressing through the streak switch from E-cadherin to N-cadherin expression (Hatta and Takeichi, 1986), and it is indeed necessary for E-cadherin to be downregulated to allow ingressing cells to leave the primitive streak. In the mouse, the downregulation has been shown to be mediated by Snail, and Snail transcription in turn is activated by FGF-signaling (Ciruna and Rossant, 2001). A similar switch from E-cadherin to N-cadherin occurs in the invaginating mesoderm in Drosophila (Oda et al., 1998) but has not yet been shown to have any function. Modulation of cadherin is also seen in the ingressing primary mesenchyme cells in sea urchins as they undergo bottle cell formation and delamination; convergence and extension in the sea urchin archenteron is accompanied by a loss of E-cadherin (Miller and McClay, 1997a). While the molecular basis of adhesive changes in convergence and extension movements in vertebrates has not been fully determined, modulation of the adhesive function of C-cadherin may play a role (Marsden and DeSimone, 2003),

and cadherins are involved in the separation of germ layers in Xenopus (Wacker et al., 2000). Endocytosis. A perhaps unexpected cell biological mechanism that is turning out to have an important role in enabling various cellular processes during gastrulation is endocytosis. In sea urchin embryos stained with antibodies against components of apical junctions such as cadherin, the immunofluorescence signal is seen at the site of these junctions in epithelial cells, but not in PMCs that are getting ready to delaminate. In these cells, the staining is seen in intracellular spots, which has been interpreted as the result of endocytosis of junction material (Miller and McClay, 1997b). Whether this is the cause of the loss in adhesiveness or simply due to removal of excessive membrane material in response to apical constriction is not known. However, other cell surface molecules are transported from intracellular stores to the cell surface at the same time, indicating a regulated exchange of adhesion molecules by endocytosis and exocytosis (Miller and McClay, 1997b). Whether apical constriction is also associated with endocytosis in other situations remains to be tested. Hierarchy of Control Mechanisms One of the great challenges for developmental cell biologists is to understand how the determination of cell fate, manifested in the expression of specific transcription factors, is translated into the modulation of cell behavior. A correlation between cell fates and particular types of morphogenetic behavior has been observed in many instances. For example, in Xenopus, only anterior mesodermal and endodermal cells are able to spread on fibronectin when tested in vitro, consistent with their role of migrating across the blastocoel roof. More generally, loss of the function of a transcription factor determining cell fate usually also causes the loss of typical morphogenetic behavior of cells. In Drosophila, loss of either of the mesodermal cell fate regulators Snail or Twist leads to failure of mesodermal cells to differentiate and to undergo cell shape changes typical for the prospective mesoderm. The same is true for Brachyury and other transcription factors in vertebrates. What then are the cascades of gene activity downstream of these transcription factors that lead to the cell behavior that shapes the embryo? What are the target genes of transcription factors that are directly responsible for changing a cell’s adhesion, shape, and motility? The problem of elucidating such cascades is not made easier by the fact that many factors can have effects both on cell fate via the transcription of differentiation genes and on morphogenesis via modulation of the cytoskeleton or adhesion or other properties. In several instances, it has been possible to distinguish these effects experimentally. Convergence and extension in the zebrafish and the frog is affected by all factors that control mesodermal cell fate and dorsoventral patterning, including FGF8. In the case of FGF signaling, one of the negative modulators of the signal transduction pathway, Sprouty, can suppress cell movement without affecting MAPK activation and mesodermal gene expression (Nutt et al., 2001). Similarly, in Drosophila, the MAPK-independent morphogenetic effect of FGF signaling can be separated from its MAPK-dependent function in heart cell determination

Review 315

(Wilson et al., 2005). In the zebrafish, genes are beginning to be identified that affect only cell movement and not fate (Bakkers et al., 2004; Topczewski et al., 2001; Yamashita et al., 2002). Whether these genes act under the control of or in parallel with fate determining genes is an interesting question worth further debate and analysis (see also Myers et al., 2002a). Surprisingly, the homolog of the transcription factor Brachyury, one of the major mesodermal fate determinants in vertebrates, can influence gastrulation movements in the sea urchin independent of cell fate, as deduced from experiments in which a fusion protein of Brachyury with the repressor domain of Drosophila Engrailed blocked gastrulation but not expression of several endodermal and mesodermal markers (Gross and McClay, 2001). However, this effect of Brachyury, too, must ultimately be mediated by transcriptional regulation, and the targets responsible for the observed effects remain to be identified. In spite of such complications, some cassettes of genetic hierarchies from fate determination to molecules directly modulating cell movement are beginning to be worked out. A search for target genes of Brachyury identified Wnt11 (Tada and Smith, 2000), which, like Bra, is required for involution and convergence and extension movements in fish and frogs. To prove that Brachyury acts through Wnt11, it would be necessary to show that the defects resulting from lack of Brachyury can be alleviated or abolished by expressing Wnt11. However, in tissue explants, a constitutively active form of Dishevelled could not rescue the defects (Tada and Smith, 2000), indicating that other targets of Brachury must also participate. In Drosophila, the genes acting downstream of the mesodermal transcriptional activator Twist have been tested for their ability to direct morphogenetic behavior. Surprisingly, the major role in mediating cell shape changes and invagination is played by Snail, a transcriptional repressor (T. Seher and M.L., unpublished data). Finally, the effort to establish the “gene regulatory network” for the micromeres in sea urchin embryos aims, in a more encompassing way, at the solution of the same problem (Davidson et al., 2002a). In some cases, transcription factors have been shown to act as switches between alternative cell behaviors. The example of Snail as regulator between an epithelial and a mesenchymal state has been discussed above. Brachyury may control the difference in the types of movement in mesodermal cells in frogs and fish. In tissue explants in culture, Brachyury is required for convergence and extension movements, but not for cell migration (Conlon and Smith, 1999). To qualify as a “switch,” Brachyury needs to be able not only to turn convergence and extension behavior on, but also to turn migration off, which is precisely what it does (Kwan and Kirschner, 2003). A transcription factor acting as a switch in the interpretation of the FGF signal in the chick has been identified recently. The Zinc finger transcriptional activator Churchill downregulates mesodermal gene activity and movement downstream of FGF signaling (via activating the Smad-interacting-protein1), simultaneously sensitizing cells to neural induction factors from the node (Sheng et al., 2003). Conservation of Morphogenetic and Regulatory

Mechanisms. We have seen that there are recurrent themes in the repertoire of mechanisms used during gastrulation by different animals. Bottle cell formation, de-epithelialization, migration, and BMP, FGF, and Wnt signaling occur in all organisms. But can they be viewed as homologous, and are cassettes of regulatory pathways exchangeable? Whereas it is generally accepted that the underlying logic and geometry of gastrulation is conserved in vertebrates (Beddington and Robertson, 1998), this is not so clear for the cellular mechanisms. Whether gastrulation movements in fish and amphibians are the same has been discussed extensively (Myers et al., 2002a; Wallingford et al., 2002). Because of certain differences (more migration — apparently directed — in the fish than the frog, larger dependence of frog gastrulation on mediolateral intercalation), the conclusion has been that they are quite different. However, it has to be kept in mind that large differences can be found even within the taxa of the amphibians or fish, often between closely related groups. Surely this does not mean that the underlying logic and principles are not conserved, especially as those movements seen in the fish but not Xenopus (directed dorsal migration) depend on many of the same molecules as the “true” convergent extension behaviors. More likely, different species make use of the constituent mechanisms to different degrees, resulting in perhaps large apparent differences. Often, these are probably dictated by the architecture of the egg and the speed of development. An example for the latter is found in the invagination of the mesoderm in insect embryos. In dipterans like Drosophila, in which the primordia of all segments are present before gastrulation begins, the ventral furrow invaginates rapidly along the entire length of the embryo and the mesoderm becomes completely internalized within less than an hour. The cells only begin to proliferate after this point; a mitotic block in the ventral furrow ensures the orderly progression of cell shape changes necessary for the rapid and efficient invagination (Grosshans and Wieschaus, 2000; Mata et al., 2000; Seher and Leptin, 2000). By contrast, in short germ insects like Tribolium, in which segments arise sequentially during embryogenesis, mesoderm invagination is slow and less orderly. Individual cells show cell shape changes similar to those seen in Drosophila, but proliferation occurs at the same time, so that apically constricting cells are interspersed with dividing cells (Handel et al., 2005). Cross-sections through the invaginating mesoderm of a Tribolium embryo look strikingly similar to Drosophila tribbles mutant embryos in which the mitotic block in the mesoderm has been released. This suggests that the mitotic block may have evolved to ensure the rapid, reliable invagination of the mesoderm needed to allow the fast development of the dipteran embryo. A mitotic block has also been observed to contribute to proper mesodermal morphogenesis in frogs (Leise and Mueller, 2004), cell proliferation is regulated by Snail in the mouse (Ciruna and Rossant, 2001), and Xenopus Tribbles participates in the control of mitosis (Saka and Smith, 2004), another set of tantalizing similarities which do not quite yet add up to a coherent picture. Just as new cell biological mechanisms can apparently be coopted during evolution to serve the chang-

Developmental Cell 316

Figure 4. Modulation of Genetic Hierarchies around a Conserved Cassette Activation of the FGF-receptor expressed in the mesoderm of the mouse embryo results in the transcription of Snail. Snail is required for the downregulation of E-cadheirn, to allow mesodermal cells to lose their epithelial coherence and move into the embryo. It is not clear to what extent FGF might have other, direct morphogenetic effects on cell behavior at this point (for example, by acting as a motogen and increasing protrusive activity), but after ingression, migratory mesodermal cells in the chick show chemotactic behavior in response to FGF. In Drosophila, the genes for both Snail and the FGF-receptor are activated in the mesoderm primordium by Twist, and Snail represses E-cadherin. Snail expression does not depend on FGFsignaling. FGF induces protrusive activity of mesodermal cells after they have invaginated, which allows them to spread out on the ectoderm (cf. Figure 1H).

ing requirements of the gastrulating embryo, so is there a suggestion that gene regulatory hierarchies are flexible and can easily be rewired. A particularly striking case concerns the roles of Snail, E-cadherin, and FGF signaling in mouse versus Drosophila (Figure 4). The mesoderm in both organisms express Snail, responds to FGF signaling, and downregulates E-cadherin as it begins to spread out after ingression or invagination. However, the relationships between the three proteins are completely different. In the mouse, FGF signaling activates Snail transcription, and Snail then downregulates E-cadherin expression, which allows the mesoderm to become migratory. In Drosophila, Snail also represses E-cadherin expression (Leptin, 1991), suggesting a conserved regulatory cassette, but there is no evidence that downregulation of E-cadherin is important. Furthermore, although FGF signaling is active in the mesoderm during EMT in Drosophila, it is required for the spreading of the mesoderm on the ectoderm, but not for EMT. Thus, although all three genes are active during gastrulation both in Drosophila and in vertebrates, the hierarchy of their activities has apparently not been conserved. Another mesoderm-specific transcription factor, Twist, is important for myogenesis in invertebrates and vertebrates, but in Drosophila also acts at the top of the cascade of mesoderm development, controlling all aspects of mesoderm morphogenesis and differentiation. Open Questions, Conclusion Many aspects of gastrulation have now been explained in molecular detail, and we can expect further rapid

progress in many organisms, but it is not obvious that a unified understanding of gastrulation will arise from these studies. Among the cellular processes that require further investigation are the processes of apical constriction, for which only two molecules have been ascribed specific functions (Costa et al., 1994; Haigo et al., 2003; Hildebrand and Soriano, 1999), and as mentioned above, the source of the directionality during PCP-mediated movements. A further interesting unsolved question concerns the interpretation of gradients: how is a graded distribution of a molecule translated into qualitatively distinct cell responses? In the case of the gradients of transcription factors in the Drosophila embryo, the principles are well understood (number and affinities of binding sites for transcription factors in the promoter regions of downstream genes, feedback loops). It is less clear how different concentrations of ligand lead to qualitatively different responses of events downstream of activated transmembrane receptors, for example, in the case of the signal from the Organizer in vertebrates, the dorsoventral dpp gradient in Drosophila. Furthermore, gradients by definition have no sharp borders, whereas cells do eventually have to be one thing or another; thus, fuzzy transitions have to be converted to sharp boundaries with on-off responses. Finally, the coordination of the various cell behaviors and their place in the hierarchy of gene activities are only beginning to be understood. To reach an integrated understanding, it will be necessary in this field, more than in any other, to keep the evolutionary aspect in mind. What is the ancestral subdivision of germ layers, and what is the role of canonical Wnt signaling in the ancestral protostome? A good guess is the definition of the gut or endoderm, because this is a likely ancestral distinction: absorptive tissues for taking up nutrients versus an outer protective tissue containing sensory cells to interpret the environment. If endoderm induction by β-catenin is the first axis/germ layer differentiation, then what is the second? Mesoderm induction by Nodal-like factors might be the suggestion from the viewpoint of the amphibians and fish, but how does that square with Nodal function in sea urchins, or BMPs in Drosophila? Perhaps endoderm and mesoderm were initially set aside as one unit and an ancestral endomesoderm found diverse ways of sorting itself into two distinct germ layers (see also (Ball et al., 2004) (Arendt, 2004)). With a wider range of animals now being analyzed genetically and molecularly, aided by the rapid increase in the availability of new whole genome sequences, these questions should soon find their answers.

Acknowledgments I am very grateful to Juliane Hancke for her expert artwork that provided the templates for the figures. Among the many friends and colleagues who have helped with suggestions, corrections, and discussions, I want to thank especially Lilianna Solnica-Krezel, John Wallingford, Rudi Winklbauer, and Carl-Philipp Heisenberg for their critical reading of the manuscript and their helpful comments.

Review 317

References

cally along the oral-aboral axis of the embryo and is involved in skeletogenesis. Mech. Dev. 120, 561–572.

Agius, E., Oelgeschlager, M., Wessely, O., Kemp, C., and De Robertis, E.M. (2000). Endodermal Nodal-related signals and mesoderm induction in Xenopus. Development 127, 1173–1183.

Davidson, L.A., Oster, G.F., Keller, R.E., and Koehl, M.A. (1999). Measurements of mechanical properties of the blastula wall reveal which hypothesized mechanisms of primary invagination are physically plausible in the sea urchin Strongylocentrotus purpuratus. Dev. Biol. 209, 221–238.

Angerer, L.M., and Angerer, R.C. (2003). Patterning the sea urchin embryo: gene regulatory networks, signaling pathways, and cellular interactions. Curr. Top. Dev. Biol. 53, 159–198. Angerer, L.M., Oleksyn, D.W., Levine, A.M., Li, X., Klein, W.H., and Angerer, R.C. (2001). Sea urchin goosecoid function links fate specification along the animal-vegetal and oral-aboral embryonic axes. Development 128, 4393–4404. Arendt, D. (2004). Comparative Aspects of Gastrulation. In Gastrulation: From Cells to Embryos, C.D. Stern, ed. (New York: Cold Spring Harbor Press), pp. 679–694. Arendt, D., and Nübler-Jung, K. (1997). Dorsal or ventral: similarities in fate maps and gastrulation patterns in annelids, arthropods and chordates. Mech. Dev. 61, 7–21. Bakkers, J., Kramer, C., Pothof, J., Quaedvlieg, N.E., Spaink, H.P., and Hammerschmidt, M. (2004). Has2 is required upstream of Rac1 to govern dorsal migration of lateral cells during zebrafish gastrulation. Development 131, 525–537. Ball, E.E., Hayward, D.C., Saint, R., and Miller, D.J. (2004). A simple plan - cnidarians and the origins of developmental mechanisms. Nat. Rev. Genet. 5, 567–577. Barrett, K., Leptin, M., and Settleman, J. (1997). The Rho GTPase and a putative RhoGEF mediate a signaling pathway for the cell shape changes in Drosophila gastrulation. Cell 91, 905–915.

Davidson, E.H., Rast, J.P., Oliveri, P., Ransick, A., Calestani, C., Yuh, C.H., Minokawa, T., Amore, G., Hinman, V., Arenas-Mena, C., et al. (2002a). A genomic regulatory network for development. Science 295, 1669–1678. Davidson, L.A., Hoffstrom, B.G., Keller, R., and DeSimone, D.W. (2002b). Mesendoderm extension and mantle closure in Xenopus laevis gastrulation: combined roles for integrin alpha(5)beta(1), fibronectin, and tissue geometry. Dev. Biol. 242, 109–129. De Robertis, E.M., and Sasai, Y. (1996). A common plan for dorsoventral patterning in Bilateria. Nature 380, 37–40. De Robertis, E.M., and Kuroda, H. (2004). Dorsal-ventral patterning and neural induction in Xenopus embryos. Annu. Rev. Cell Dev. Biol. 20, 285–308. Djiane, A., Riou, J., Umbhauer, M., Boucaut, J., and Shi, D. (2000). Role of frizzled 7 in the regulation of convergent extension movements during gastrulation in Xenopus laevis. Development 127, 3091–3100. Duboc, V., Rottinger, E., Besnardeau, L., and Lepage, T. (2004). Nodal and BMP2/4 signaling organizes the oral-aboral axis of the sea urchin embryo. Dev. Cell 6, 397–410.

Beddington, R.S., and Robertson, E.J. (1998). Anterior patterning in mouse. Trends Genet. 14, 277–284.

Ehrlich, J.S., Hansen, M.D., and Nelson, W.J. (2002). Spatio-temporal regulation of Rac1 localization and lamellipodia dynamics during epithelial cell-cell adhesion. Dev. Cell 3, 259–270.

Beiman, M., Shilo, B.Z., and Volk, T. (1996). Heartless, a Drosophila FGF receptor homolog, is essential for cell migration and establishment of several mesodermal lineages. Genes Dev. 10, 2993–3002.

Elul, T., and Keller, R. (2000). Monopolar protrusive activity: a new morphogenic cell behavior in the neural plate dependent on vertical interactions with the mesoderm in Xenopus. Dev. Biol. 224, 3–19.

Blankenship, J.T., and Wieschaus, E. (2001). Two new roles for the Drosophila AP patterning system in early morphogenesis. Development 128, 5129–5138.

Fanto, M., and McNeill, H. (2004). Planar polarity from flies to vertebrates. J. Cell Sci. 117, 527–533.

Byrum, C.A., and Martindale, M.Q. (2004). Gastrulation in the Cnidaria and Ctenophora. In Gastrulation: From Cells to Embryos., C.D. Stern, ed. (New York: Cold Spring Harbor Press), pp. 33–50. Cano, A., Perez-Moreno, M.A., Rodrigo, I., Locascio, A., Blanco, M.J., del Barrio, M.G., Portillo, F., and Nieto, M.A. (2000). The transcription factor snail controls epithelial-mesenchymal transitions by repressing E-cadherin expression. Nat. Cell Biol. 2, 76–83. Carreira-Barbosa, F., Concha, M.L., Takeuchi, M., Ueno, N., Wilson, S.W., and Tada, M. (2003). Prickle 1 regulates cell movements during gastrulation and neuronal migration in zebrafish. Development 130, 4037–4046. Cho, K.W., Blumberg, B., Steinbeisser, H., and De Robertis, E.M. (1991). Molecular nature of Spemann's organizer: the role of the Xenopus homeobox gene goosecoid. Cell 67, 1111–1120. Choi, S.C., and Han, J.K. (2002). Xenopus Cdc42 regulates convergent extension movements during gastrulation through Wnt/ Ca2+ signaling pathway. Dev. Biol. 244, 342–357. Ciruna, B., and Rossant, J. (2001). FGF signaling regulates mesoderm cell fate specification and morphogenetic movement at the primitive streak. Dev. Cell 1, 37–49. Clements, D., Friday, R.V., and Woodland, H.R. (1999). Mode of action of VegT in mesoderm and endoderm formation. Development 126, 4903–4911. Conlon, F.L., and Smith, J.C. (1999). Interference with brachyury function inhibits convergent extension, causes apoptosis, and reveals separate requirements in the FGF and activin signalling pathways. Dev. Biol. 213, 85–100. Costa, M., Wilson, E., and Wieschaus, E. (1994). A putative cell signal encoded by the folded gastrulation gene coordinates cell shape changes during Drosophila gastrulation. Cell 76, 1075–1089. Croce, J., Lhomond, G., and Gache, C. (2003). Coquillette, a sea urchin T-box gene of the Tbx2 subfamily, is expressed asymmetri-

Fink, R.D., and McClay, D.R. (1985). Three cell recognition changes accompany the ingression of sea urchin primary mesenchyme cells. Dev. Biol. 107, 66–74. Friedl, P. (2004). Prespecification and plasticity: shifting mechanisms of cell migration. Curr. Opin. Cell Biol. 16, 14–23. Glickman, N.S., Kimmel, C.B., Jones, M.A., and Adams, R.J. (2003). Shaping the zebrafish notochord. Development 130, 873–887. Gomperts, M., Garcia-Castro, M., Wylie, C., and Heasman, J. (1994). Interactions between primordial germ cells play a role in their migration in mouse embryos. Development 120, 135–141. Graff, J.M., Thies, R.S., Song, J.J., Celeste, A.J., and Melton, D.A. (1994). Studies with a Xenopus BMP receptor suggest that ventral mesoderm-inducing signals override dorsal signals in vivo. Cell 79, 169–179. Gross, J.M., and McClay, D.R. (2001). The role of Brachyury (T) during gastrulation movements in the sea urchin Lytechinus variegatus. Dev. Biol. 239, 132–147. Grosshans, J., and Wieschaus, E. (2000). A genetic link between morphogenesis and cell division during formation of the ventral furrow in Drosophila. Cell 101, 523–531. Gryzik, T., and Müller, H.A. (2004). FGF8-like1 and FGF8-like2 Encode Putative Ligands of the FGF Receptor Htl and Are Required for Mesoderm Migration in the Drosophila Gastrula. Curr. Biol. 14, 659–667. Habas, R., Kato, Y., and He, X. (2001). Wnt/Frizzled activation of Rho regulates vertebrate gastrulation and requires a novel Formin homology protein Daam1. Cell 107, 843–854. Habas, R., Dawid, I.B., and He, X. (2003). Coactivation of Rac and Rho by Wnt/Frizzled signaling is required for vertebrate gastrulation. Genes Dev. 17, 295–309. Häcker, U., and Perrimon, N. (1998). DRhoGEF2 encodes a member of the Dbl family of oncogenes and controls cell shape changes during gastrulation in Drosophila. Genes Dev. 12, 274–284.

Developmental Cell 318

Haigo, S.L., Hildebrand, J.D., Harland, R.M., and Wallingford, J.B. (2003). Shroom induces apical constriction and is required for hingepoint formation during neural tube closure. Curr. Biol. 13, 2125–2137. Handel, K., Basal, A., Fan, X., and Roth, S. (2005). Tribolium castaneum twist: gastrulation and mesoderm formation in a short-germ beetle. Dev. Genes Evol. 215, 13–31. Hardin, J. (1989). Local shifts in position and polarized motility drive cell rearrangement during sea urchin gastrulation. Dev. Biol. 136, 430–445. Hatta, K., and Takeichi, M. (1986). Expression of N-cadherin adhesion molecules associated with early morphogenetic events in chick development. Nature 320, 447–449. Heasman, J. (1997). Patterning the Xenopus blastula. Development 124, 4179–4191. Heisenberg, C.P., Tada, M., Rauch, G.J., Saude, L., Concha, M.L., Geisler, R., Stemple, D.L., Smith, J.C., and Wilson, S.W. (2000). Silberblick/Wnt11 mediates convergent extension movements during zebrafish gastrulation. Nature 405, 76–81. Hildebrand, J.D., and Soriano, P. (1999). Shroom, a PDZ domaincontaining actin-binding protein, is required for neural tube morphogenesis in mice. Cell 99, 485–497. Holley, S.A., and Ferguson, E.L. (1997). Fish are like flies are like frogs: conservation of dorsal-ventral patterning mechanisms. Bioessays 19, 281–284. Hyde, C.E., and Old, R.W. (2000). Regulation of the early expression of the Xenopus nodal-related 1 gene, Xnr1. Development 127, 1221–1229. Iioka, H., Ueno, N., and Kinoshita, N. (2004). Essential role of MARCKS in cortical actin dynamics during gastrulation movements. J. Cell Biol. 164, 169–174. Imai, K., Takada, N., Satoh, N., and Satou, Y. (2000). (beta)-catenin mediates the specification of endoderm cells in ascidian embryos. Development 127, 3009–3020. Irvine, K.D., and Wieschaus, E. (1994). Cell intercalation during Drosophila germband extension and its regulation by pair-rule segmentation genes. Development 120, 827–842. Jessen, J.R., Topczewski, J., Bingham, S., Sepich, D.S., Marlow, F., Chandrasekhar, A., and Solnica-Krezel, L. (2002). Zebrafish trilobite identifies new roles for Strabismus in gastrulation and neuronal movements. Nat. Cell Biol. 4, 610–615. Keller, R.E. (1981). An experimental analysis of the role of bottle cells and the deep marginal zone in gastrulation of Xenopus laevis. J. Exp. Zool. 216, 81–101. Keller, R., and Danilchik, M. (1988). Regional expression, pattern and timing of convergence and extension during gastrulation of Xenopus laevis. Development 103, 193–209. Keller, R., and Shook, D. (2004). Gastrulation in amphibians. In Gastrulation: From Cells to Embryos, C.D. Stern, ed. (New York: Cold Spring Harbor Press), pp. 171–204. Keller, R., Cooper, M.S., Danilchik, M., Tibbetts, P., and Wilson, P.A. (1989). Cell intercalation during notochord development in Xenopus laevis. J. Exp. Zool. 251, 134–154. Kilian, B., Mansukoski, H., Barbosa, F.C., Ulrich, F., Tada, M., and Heisenberg, C.P. (2003). The role of Ppt/Wnt5 in regulating cell shape and movement during zebrafish gastrulation. Mech. Dev. 120, 467–476. Kimberly, E.L., and Hardin, J. (1998). Bottle cells are required for the initiation of primary invagination in the sea urchin embryo. Dev. Biol. 204, 235–250. Kimelman, D., Christian, J.L., and Moon, R.T. (1992). Synergistic principles of development: overlapping patterning systems in Xenopus mesoderm induction. Development 116, 1–9. Kispert, A., Herrmann, B.G., Leptin, M., and Reuter, R. (1994). Homologs of the mouse Brachyury gene are involved in the specification of posterior terminal structures in Drosophila, Tribolium, and Locusta. Genes Dev. 8, 2137–2150. Kofron, M., Demel, T., Xanthos, J., Lohr, J., Sun, B., Sive, H., Osada, S., Wright, C., Wylie, C., and Heasman, J. (1999). Mesoderm induc-

tion in Xenopus is a zygotic event regulated by maternal VegT via TGFbeta growth factors. Development 126, 5759–5770. Kosman, D., Ip, Y.T., Levine, M., and Arora, K. (1991). Establishment of the mesoderm-neuroectoderm boundary in the Drosophila embryo. Science 254, 118–122. Kwan, K.M., and Kirschner, M.W. (2003). Xbra functions as a switch between cell migration and convergent extension in the Xenopus gastrula. Development 130, 1961–1972. Lane, M.C., Koehl, M.A., Wilt, F., and Keller, R. (1993). A role for regulated secretion of apical extracellular matrix during epithelial invagination in the sea urchin. Development 117, 1049–1060. Lecuit, T. (2004). Junctions and vesicular trafficking during Drosophila cellularization. J. Cell Sci. 117, 3427–3433. Leise, W.F., 3rd, and Mueller, P.R. (2004). Inhibition of the cell cycle is required for convergent extension of the paraxial mesoderm during Xenopus neurulation. Development 131, 1703–1715. Lemaire, P., and Yasuo, H. (1998). Developmental signalling: a careful balancing act. Curr. Biol. 8, R228–R231. Leptin, M. (1991). twist and snail as positive and negative regulators during Drosophila mesoderm development. Genes Dev. 5, 1568–1576. Leptin, M. (1995). Drosophila gastrulation: From pattern formation to morphogenesis. Annu. Rev. Cell Biol. 11, 189–212. Leptin, M., and Roth, S. (1994). Autonomy and non-autonomy in Drosophila mesoderm differentiation and morphogenesis. Development 120, 853–859. Logan, C.Y., Miller, J.R., Ferkowicz, M.J., and McClay, D.R. (1999). Nuclear beta-catenin is required to specify vegetal cell fates in the sea urchin embryo. Development 126, 345–357. Marlow, F., Topczewski, J., Sepich, D., and Solnica-Krezel, L. (2002). Zebrafish Rho kinase 2 acts downstream of Wnt11 to mediate cell polarity and effective convergence and extension movements. Curr. Biol. 12, 876–884. Marsden, M., and Burke, R.D. (1998). The betaL integrin subunit is necessary for gastrulation in sea urchin embryos. Dev. Biol. 203, 134–148. Marsden, M., and DeSimone, D.W. (2001). Regulation of cell polarity, radial intercalation and epiboly in Xenopus: novel roles for integrin and fibronectin. Development 128, 3635–3647. Marsden, M., and DeSimone, D.W. (2003). Integrin-ECM interactions regulate cadherin-dependent cell adhesion and are required for convergent extension in Xenopus. Curr. Biol. 13, 1182–1191. Martindale, M.Q., Pang, K., and Finnerty, J.R. (2004). Investigating the origins of triploblasty: 'mesodermal' gene expression in a diploblastic animal, the sea anemone Nematostella vectensis (phylum, Cnidaria; class, Anthozoa). Development 131, 2463–2474. Mata, J., Curado, S., Ephrussi, A., and Rørth, P. (2000). Tribbles coordinates mitosis and morphogenesis in Drosophila by regulating string/CDC25 proteolysis. Cell 101, 511–522. McClay, D.R., Gross, J.M., Range, R., Peterson, R.E., and Bradham, C. (2004). Sea Urchin Gastrulation. In Gastrulation: From Cells to Embryos., C.D. Stern, ed. (New York: Cold Spring Harbor Press), pp. 123–128. McEwen, D.G., and Peifer, M. (2000). Wnt signaling: Moving in a new direction. Curr. Biol. 10, R562–R564. Miller, J.R., and McClay, D.R. (1997a). Changes in the pattern of adherens junction-associated beta-catenin accompany morphogenesis in the sea urchin embryo. Dev. Biol. 192, 310–322. Miller, J.R., and McClay, D.R. (1997b). Characterization of the role of cadherin in regulating cell adhesion during sea urchin development. Dev. Biol. 192, 323–339. Miller, J., Fraser, S.E., and McClay, D. (1995). Dynamics of thin filopodia during sea urchin gastrulation. Development 121, 2501–2511. Montero, J.A., Kilian, B., Chan, J., Bayliss, P.E., and Heisenberg, C.P. (2003). Phosphoinositide 3-kinase is required for process outgrowth and cell polarization of gastrulating mesendodermal cells. Curr. Biol. 13, 1279–1289. Moon, R.T., Campbell, R.M., Christian, J.L., McGrew, L.L., Shih, J.,

Review 319

and Fraser, S. (1993). Xwnt-5A: a maternal Wnt that affects morphogenetic movements after overexpression in embryos of Xenopus laevis. Development 119, 97–111. Myers, D.C., Sepich, D.S., and Solnica-Kreyel, L. (2002a). Convergence and extension in vertebrate gastrulae: cell movements according to or in search of identity? Trends Genet. 18, 447–455. Myers, D.C., Sepich, D.S., and Solnica-Krezel, L. (2002b). Bmp activity gradient regulates convergent extension during zebrafish gastrulation. Dev. Biol. 243, 81–98. Nagel, M., Tahinci, E., Symes, K., and Winklbauer, R. (2004). Guidance of mesoderm cell migration in the Xenopus gastrula requires PDGF signaling. Development 131, 2727–2736. Ninomiya, H., Elinson, R.P., and Winklbauer, R. (2004). Antero-posterior tissue polarity links mesoderm convergent extension to axial patterning. Nature 430, 364–367. Nishii, I., Ogihara, S., and Kirk, D.L. (2003). A kinesin, invA, plays an essential role in volvox morphogenesis. Cell 113, 743–753. Nutt, S.L., Dingwell, K.S., Holt, C.E., and Amaya, E. (2001). Xenopus Sprouty2 inhibits FGF-mediated gastrulation movements but does not affect mesoderm induction and patterning. Genes Dev. 15, 1152–1166. Oates, A.C., Lackmann, M., Power, M.A., Brennan, C., Down, L.M., Do, C., Evans, B., Holder, N., and Boyd, A.W. (1999). An early developmental role for eph-ephrin interaction during vertebrate gastrulation. Mech. Dev. 83, 77–94.

and ensheathing of longitudinal axon tracts in CNS. Development 124, 2119–2128. Shook, D., and Keller, R. (2003). Mechanisms, mechanics and function of epithelial-mesenchymal transitions in early development. Mech. Dev. 120, 1351–1383. Smallhorn, M., Murray, M.J., and Saint, R. (2004). The epithelialmesenchymal transition of the Drosophila mesoderm requires the Rho GTP exchange factor Pebble. Development 131, 2641–2651. Smith, J.C., Price, B.M., Green, J.B., Weigel, D., and Herrmann, B.G. (1991). Expression of a Xenopus homolog of Brachyury (T) is an immediate-early response to mesoderm induction. Cell 67, 79–87. Sokol, S.Y. (1996). Analysis of Dishevelled signalling pathways during Xenopus development. Curr. Biol. 6, 1456–1467. Solnica-Krezel, L., and Driever, W. (2001). The role of the homeodomain protein Bozozok in zebrafish axis formation. Int. J. Dev. Biol. 45, 299–310. Stathopoulos, A., Tam, B., Ronshaugen, M., Frasch, M., and Levine, M. (2004). pyramus and thisbe: FGF genes that pattern the mesoderm of Drosophila embryos. Genes Dev. 18, 687–699. Tada, M., and Smith, J.C. (2000). Xwnt11 is a target of Xenopus Brachyury: regulation of gastrulation movements via Dishevelled, but not through the canonical Wnt pathway. Development 127, 2227–2238.

Oda, H., Tsukita, S., and Takeichi, M. (1998). Dynamic behavior of the cadherin-based cell-cell adhesion system during Drosophila gastrulation. Dev. Biol. 203, 435–450.

Takahashi, N., Ishihara, S., Takada, S., Tsukita, S., and Nagafuchi, A. (2000). Posttranscriptional regulation of alpha-catenin expression is required for Wnt signaling in L cells. Biochem. Biophys. Res. Commun. 277, 691–698.

Odell, G.M., Oster, G., Alberch, P., and Burnside, B. (1981). The Mechanical Basis of Morphogenesis. Epithelial Folding and Invagination. Dev. Biol. 85, 446–462.

Thorpe, C.J., Schlesinger, A., Carter, J.C., and Bowerman, B. (1997). Wnt signaling polarizes an early C. elegans blastomere to distinguish endoderm from mesoderm. Cell 90, 695–705.

Reuter, R. (1994). The gene serpent has homeotic properties and specifies endoderm versus ectoderm within the Drosophila gut. Development 120, 1123–1135.

Topczewski, J., Sepich, D.S., Myers, D.C., Walker, C., Amores, A., Lele, Z., Hammerschmidt, M., Postlethwait, J., and Solnica-Krezel, L. (2001). The zebrafish glypican knypek controls cell polarity during gastrulation movements of convergent extension. Dev. Cell 1, 251–264.

Rocheleau, C.E., Downs, W.D., Lin, R., Wittmann, C., Bei, Y., Cha, Y.H., Ali, M., Priess, J.R., and Mello, C.C. (1997). Wnt signaling and an APC-related gene specify endoderm in early C. elegans embryos. Cell 90, 707–716. Rodaway, A., and Patient, R. (2001). Mesendoderm. an ancient germ layer? Cell 105, 169–172. Saka, Y., and Smith, J.C. (2004). A Xenopus tribbles orthologue is required for the progression of mitosis and for development of the nervous system. Dev. Biol. 273, 210–225. Schier, A.F. (2003). Nodal signaling in vertebrate development. Annu. Rev. Cell Dev. Biol. 19, 589–621. Schulte-Merker, S., Ho, R.K., Herrmann, B.G., and Nusslein-Volhard, C. (1992). The protein product of the zebrafish homologue of the mouse T gene is expressed in nuclei of the germ ring and the notochord of the early embryo. Development 116, 1021–1032. Schumacher, S., Gryzik, T., Tannebaum, S., and Muller, H.A. (2004). The RhoGEF Pebble is required for cell shape changes during cell migration triggered by the Drosophila FGF receptor Heartless. Development 131, 2631–2640.

Ulrich, F., Concha, M.L., Heid, P.J., Voss, E., Witzel, S., Roehl, H., Tada, M., Wilson, S.W., Adams, R.J., Soll, D.R., and Heisenberg, C.P. (2003). Slb/Wnt11 controls hypoblast cell migration and morphogenesis at the onset of zebrafish gastrulation. Development 130, 5375–5384. Veeman, M.T., Axelrod, J.D., and Moon, R.T. (2003). A second canon. Functions and mechanisms of beta-catenin-independent Wnt signaling. Dev. Cell 5, 367–377. Wacker, S., Grimm, K., Joos, T., and Winklbauer, R. (2000). Development and control of tissue separation at gastrulation in Xenopus. Dev. Biol. 224, 428–439. Wallingford, J.B., Rowning, B.A., Vogeli, K.M., Rothbacher, U., Fraser, S.E., and Harland, R.M. (2000). Dishevelled controls cell polarity during Xenopus gastrulation. Nature 405, 81–85. Wallingford, J.B., Ewald, A.J., Harland, R.M., and Fraser, S.E. (2001). Calcium signaling during convergent extension in Xenopus. Curr. Biol. 11, 652–661.

Seher, T.C., and Leptin, M. (2000). Tribbles, a cell-cycle brake that coordinates proliferation and morphogenesis during Drosophila gastrulation. Curr. Biol. 10, 623–629.

Wallingford, J.B., Fraser, S.E., and Harland, R.M. (2002). Convergent extension: the molecular control of polarized cell movement during embryonic development. Dev. Cell 2, 695–706.

Sheng, G., dos Reis, M., and Stern, C.D. (2003). Churchill, a zinc finger transcriptional activator, regulates the transition between gastrulation and neurulation. Cell 115, 603–613.

Wikramanayake, A.H., Huang, L., and Klein, W.H. (1998). betaCatenin is essential for patterning the maternally specified animalvegetal axis in the sea urchin embryo. Proc. Natl. Acad. Sci. USA 95, 9343–9348.

Shih, J., and Keller, R. (1992a). Cell motility driving mediolateral intercalation in explants of Xenopus laevis. Development 116, 901–914. Shih, J., and Keller, R. (1992b). Patterns of cell motility in the organizer and dorsal mesoderm of Xenopus laevis. Development 116, 915–930. Shishido, E., Ono, N., Kojima, T., and Saigo, K. (1997). Requirements of DFR1/Heartless, a mesoderm-specific Drosophila FGFreceptor, for the formation of heart, visceral and somatic muscles,

Wikramanayake, A.H., Hong, M., Lee, P.N., Pang, K., Byrum, C.A., Bince, J.M., Xu, R., and Martindale, M.Q. (2003). An ancient role for nuclear beta-catenin in the evolution of axial polarity and germ layer segregation. Nature 426, 446–450. Wilkinson, D.G., Bhatt, S., and Herrmann, B.G. (1990). Expression pattern of the mouse T gene and its role in mesoderm formation. Nature 343, 657–659. Wilson, P., and Keller, R. (1991). Cell rearrangement during gastrula-

Developmental Cell 320

tion of Xenopus: direct observation of cultured explants. Development 112, 289–300. Wilson, R., Vogelsang, E., and Leptin, M. (2005). FGF signalling and the mechanism of mesoderm spreading in Drosophila embryos. Development 132, 491–501. Winklbauer, R., and Nagel, M. (1991). Directional mesoderm cell migration in the Xenopus gastrula. Dev. Biol. 148, 573–589. Winklbauer, R., Selchow, A., Nagel, M., and Angres, B. (1992). Cell interaction and its role in mesoderm cell migration during Xenopus gastrulation. Dev. Dyn. 195, 290–302. Winklbauer, R., Medina, A., Swain, R.K., and Steinbeisser, H. (2001). Frizzled-7 signalling controls tissue separation during Xenopus gastrulation. Nature 413, 856–860. Wolpert, L. (2001). Principles of Development (Oxford: Oxford University Press). Xanthos, J.B., Kofron, M., Tao, Q., Schaible, K., Wylie, C., and Heasman, J. (2002). The roles of three signaling pathways in the formation and function of the Spemann Organizer. Development 129, 4027–4043. Yamanaka, H., Moriguchi, T., Masuyama, N., Kusakabe, M., Hanafusa, H., Takada, R., Takada, S., and Nishida, E. (2002). JNK functions in the non-canonical Wnt pathway to regulate convergent extension movements in vertebrates. EMBO Rep. 3, 69–75. Yamashita, S., Miyagi, C., Carmany-Rampey, A., Shimizu, T., Fujii, R., Schier, A.F., and Hirano, T. (2002). Stat3 Controls Cell Movements during Zebrafish Gastrulation. Dev. Cell 2, 363–375. Yang, J., Tan, C., Darken, R.S., Wilson, P.A., and Klein, P.S. (2002a). Beta-catenin/Tcf-regulated transcription prior to the midblastula transition. Development 129, 5743–5752. Yang, X., Dormann, D., Munsterberg, A.E., and Weijer, C.J. (2002b). Cell movement patterns during gastrulation in the chick are controlled by positive and negative chemotaxis mediated by FGF4 and FGF8. Dev. Cell 3, 425–437. Yeo, S.Y., Little, M.H., Yamada, T., Miyashita, T., Halloran, M.C., Kuwada, J.Y., Huh, T.L., and Okamoto, H. (2001). Overexpression of a slit homologue impairs convergent extension of the mesoderm and causes cyclopia in embryonic zebrafish. Dev. Biol. 230, 1–17. Zallen, J.A., and Wieschaus, E. (2004). Patterned gene expression directs bipolar planar polarity in Drosophila. Dev. Cell 6, 343–355.