Developmentally Restricted Actin-Regulatory Molecules Control Morphogenetic Cell Movements in the Zebrafish Gastrula

Current Biology, Vol. 14, 1632–1638, September 21, 2004, 2004 Elsevier Ltd. All rights reserved.

DOI 10.1016/j .c ub . 20 04 . 08 .0 2 4

Developmentally Restricted Actin-Regulatory Molecules Control Morphogenetic Cell Movements in the Zebrafish Gastrula David F. Daggett,1,3,* Catherine A. Boyd,1 Philippe Gautier,1 Robert J. Bryson-Richardson,1,4 Christine Thisse,2 Bernard Thisse,2 Sharon L. Amacher,3 and Peter D. Currie1,4 1 Comparative and Developmental Genetics Section Medical Research Council Human Genetics Unit Western General Hospital Edinburgh EH4 2XU United Kingdom 2 Institut de Ge´ne´tique et de Biologie Mole´culaire et Cellulaire Centre National de la Recherche Scientifique Institut National de la Sante´ et de la Recherche Me´dicale Universite´ Louis Pasteur 1 rue Laurent Fries BP 163 Cite´ Universitaire de Strasbourg 67404 Illkirch Cedex France 3 Department of Molecular and Cellular Biology University of California, Berkeley Berkeley, California 94720-3200

Summary Although our understanding of the regulation of cellular actin and its control during the development of invertebrates is increasing [1–9], the question as to how such actin dynamics are regulated differentially across the vertebrate embryo to effect its relatively complex morphogenetic cell movements remains poorly understood. Intercellular signaling that provides spatial and temporal cues to modulate the subcellular localization and activity of actin regulatory molecules represents one important mechanism [10– 13]. Here we explore whether the localized gene expression of specific actin regulatory molecules represents another developmental mechanism. We have identified a cap1 homolog and a novel guanine nucleotide exchange factor (GEF), quattro (quo), that share a restricted gene expression domain in the anterior mesendoderm of the zebrafish gastrula. Each gene is required for specific cellular behaviors during the anterior migration of this tissue; furthermore, cap1 regulates cortical actin distribution specifically in these cells. Finally, although cap1 and quo are autonomously required for the normal behaviors of these cells, they are also nonautonomously required for convergence and extension movements of posterior tissues. Our results provide direct evidence for the deployment of developmentally restricted actin-regu*Correspondence: [email protected] 4 Present address: Victor Chang Institute of Medical Research, 384 Victoria Street, Darlinghurst 2010, New South Wales, Australia.

latory molecules in the control of morphogenetic cell movements during vertebrate development.

Results and Discussion In an in situ hybridization-based screening approach to identify spatially and temporally regulated genes positioned to effect developmental events, we isolated two genes encoding putative regulators of actin dynamics with restricted expression in the developing anterior mesendoderm. The gene encoding zebrafish cyclaseassociated protein-1 (Cap1) is a homolog of Drosophila cap. cap is required for the apical regulation of actin dynamics in cells during morphogenetic furrow movement in the eye disc and in the developing follicular epithelium [8, 9]. Zebrafish Cap1 shares 63%, 63%, and 45% overall identity with human, Xenopus, and Drosophila orthologs, respectively, with highly conserved regions such as the G-actin binding domains shown to directly modulate actin filament growth by regulating the availability of monomeric actin (Figure 1A) [14, 15]. Zebrafish cap1 expression is first detected in anterior mesendodermal cells between 60% and 70% epiboly and remains prominent in these cells throughout the anterior extension of the embryo until early somitogenesis (Figures 1B and 1C; data not shown). As anterior mesoderm expression diminishes, cap1 transcripts become specifically prominent in the adaxial slow-muscle precursors adjacent to the notochord and increasingly so in non-neural ectoderm, defining further developmentally restricted expression domains (data not shown). A second gene, with a remarkably similar expression profile throughout early development (Figures 1D and 1E; not shown), was identified by in situ hybridizationbased cDNA screening and bioinformatic analysis of the emerging zebrafish genome. It encodes a novel GEF for Rho family GTPases, as defined by the presence of highly conserved, tandem double homology (DH)-pleckstrin homology (PH) domains (Figure 1A). We have named this protein Quattro (Quo) to reflect its strong RhoGEF domain homology to Dbl family GEFs such as Dbl and Trio [16]. Phylogenetic analysis of metazoan genomes with both the DH/PH and a novel N-terminal domain revealed that Quo is the founding member of a new family of previously unidentified RhoGEFs (Figure 1F; also see Figure S1 in the Supplemental Data available with this article online). The role of RhoGEFs in catalyzing GTP/GDP exchange on Rho family small GTPases, leading to the nucleation and modulation of actin filaments underlying changes in cell motility and morphology, has been documented in many developmental contexts [2, 4, 5]. The identification of two putative actin-regulatory genes with remarkably similar expression restricted to the anterior mesendoderm is consistent with this group of cells uniquely possessing coordinated cell motility behaviors. Examination of live, late epiboly stage embryos reveals a compact population of cells protruding

Actin Regulation during Zebrafish Gastrulation 1633

Figure 1. Identification of Developmentally Regulated cap1 and quo Genes (A) Conserved domains of zebrafish Cap1 and Quo proteins. Abbreviations are as follows: AC, adenylate cyclase; Act-Cof, actin-cofilin; DH, double homology; and PH, pleckstrin homology. cap1 (B and C) and quo (D and E) expression in the anterior mesendoderm at 90% epiboly (B and D) and 3-5 somite stage (C and E) is shown. Anterior is up. (F) Phylogenetic analysis of RhoGEF domains demonstrates that Quo is the first characterized member of a novel family of RhoGEF domain-containing proteins (red).

from the anterior hypoblast toward the animal pole (Figure 2A) [17, 18]. These cells display dynamic cellular behaviors characteristic of active motility. Extension and retraction of filopodia and lamellipodia toward the animal pole is accompanied by cellular extrusions out of the compacted group into the remarkably cell-free anterior space, along with simultaneous forward advancement of the anterior mesendoderm (Figures 2A and 2B; Movie 1). By the tailbud stage, this group of motile cells is recognizable as the polster (Figures 2C and 2D), the mesendodermal derivative marked by hatching gland-1 (hgg1) expression [19]. Analysis of hgg1 expression illus-

trates discrete aspects of polster development. After the initial anterior extension of shield derivatives, including polster precursors [20, 21], individual and small aggregates of hgg1-expressing cells move anteriorly and become consolidated into a dense group by the end of epiboly (Figures S2A and S2B). This coordinated group moves forward to occupy a symmetrical crescentshaped position at the front of the prechordal plate mesoderm ahead of the anterior edge of the neural plate by bud stage (Figures 3A and S2C) [22]; it maintains this relative position as the polster continues moving anteriorly until the early segmentation stages. These observations indicate that the early formation and anterior movement of the polster involves the coordinated dynamic behavior of individual cells. To determine a role for Cap1 and Quo in polster cell behaviors, we performed loss-of-function experiments with morpholino antisense oligonucleotides [23]. To control for specificity, two morpholino sequences each were designed against the 5⬘ untranslated and start codon regions of both cap1 and quo predicted mRNAs, along with corresponding mismatch control morpholinos (see Supplemental Data). For each gene, injection of both gene-specific morpholinos led to reproducible defects in particular aspects of polster cell behaviors (below), ruling out mistargeting of an unrelated gene. Injection of the Cap1 or Quo mismatch morpholino sequences had no effect on polster morphology or behavior (n ⫽ 150; Figures 3C and 3E), reducing the likelihood that nonspecific morpholino effects influenced these phenotypes. In Cap1 morphants, polster precursor cells were able to form a relatively consolidated group that protruded toward the animal pole by late epiboly and displayed filopodial and lamellipodial processes. However, they failed to advance as a group beyond the animal pole or to acquire the coordinated crescent shape anterior to the neural plate by tailbud stage (Figures 2E–2H, 3B, and S2D–S2F; n ⫽ 250). The defects within Quo morphants were more severe, as a morphologically distinct polster could not be identified (Figures 2I–2L). The anterior mesendodermal cells failed to form a consolidated group, with cells abnormally dispersed both anteroposteriorly and mediolaterally in a thin, twodimensional layer throughout the prechordal plate as small groups or individuals, less distinguishable from neighboring mesendoderm (Figures 2I and 2J, 3D, and S2G–S2I; n ⫽ 250). In both cases, the failure of anterior movement or consolidation of the anterior mesendoderm also prevented the anterior resolution of the overlying neural plate, illustrated by more diffuse dlx3 expression (Figures 3B and 3D). The polster defects are consistent with the onset of cap1 and quo expression in anterior mesendoderm by 70% epiboly. Accordingly, at the shield stage, the ventral marker gata2 is expressed across the ventral half of the embryo in both wild-type embryos and morpholino-injected embryos whose siblings develop later polster defects, indicating that early dorsal-ventral specification has not been affected by the morpholinos (Figure S3). Similarly, at 65% epiboly, the dorsal mesoderm marker goosecoid is reproducibly expressed within a comparable anteroposterior and mediolateral domain in wild-type and morpholinoinjected embryos, demonstrating that dorsal mesoder-

Current Biology 1634

Figure 2. Cap1 and Quo Are Required for the Normal Formation and Movement of the Mesendodermal Polster (A–D) Wild-type; (E–H) cap1 morphants; (I–L) Quo morphants. (A, B, E, F, I, and J) Mesendodermal cells at the leading edge of the hypoblast layer at 90% epiboly; arrows denote axial domain of the hypoblast, and arrowheads indicate filopodial/lamellipodial extensions. (C, D, G, H, K, and L) Tailbud stage. (C, G, and K) Relationship between mesendodermal polster and overlying neurectoderm; arrowheads denote the anterior edge of mesendoderm, and arrows indicate the anterior limit of neurectoderm. Quo morphants lack a morphologically distinct polster (K). (D, H, and L) Anterior embryonic extension. Unlike wild-type embryos (D), Cap1 morphants fail to extend anteriorly beyond the animal pole ([H]; arrowheads), whereas Quo morphants, though lacking discernible polster, display a more-elongated axis (L).

mal derivatives are both properly specified and positioned prior to the strong expression of cap1 and quo (Figure S3). This analysis reveals that polster development consists of at least two distinct, yet integrated phases of migratory behavior. Quo is required for the anteriorly directed convergence, aggregation, and compaction of individual precursor cells into a morphologically distinct polster, whereas Cap1 is required for the migration of the consolidated cells of the polster toward and beyond the animal pole. Through in vitro, cell culture, and invertebrate studies, Cap1 orthologs across species have been shown to directly control actin filament growth, with functional consequences for cell shape changes in development [8, 9, 14, 15, 24]. RhoGEFs act with varied specificity

directly upon the Rho family GTPases, such as Rho, Rac, and Cdc42, which themselves lie upstream of multiple signaling pathways that have been most heavily implicated in the modulation of the actin cytoskeleton [2, 5]. Thus, the distribution of actin within polster cells in wildtype, morpholino-injected, and mismatch morpholinoinjected embryos was examined. We combined fluorescently conjugated phalloidin staining with Forkhead-2 (Fkd2) immunofluorescent labeling of polster cells [25] to visualize actin distributions within the polster and surrounding tissues. In wild-type embryos, actin is enriched cortically and distributed continuously around the perimeter of the densely packed polster cells (Figure 4A; n ⫽ 35). In Cap1 morphants, this cortical actin staining is markedly reduced, discontinuous, and patchy (Fig-

Figure 3. Cap1 and Quo Morphants Display Defects in the Anterior Movement and Consolidation of the Polster along with Simultaneous Abnormal Convergent Extension of Posterior Tissue Wild-type embryos(A), Cap1 morphants (B), Cap1 mismatch morpholino-injected embryos (C), Quo morphants (D), and Quo mismatch morpholino-injected embryos (E) are shown at early tailbud stage. In wild-type embryos, individual hgg1-expressing polster cells consolidate and coordinately move anterior to the anterior edge of the neural plate (dlx3 expression) as the tissue advances beyond the animal pole (A). The posterior notochord (ntl staining [A⬘]) has converged medially (arrowheads) and extended anteroposteriorly. Cap1 morphants display a general consolidation of polster cells, but they fail to move anterior to the neural plate, which is itself abnormally resolved (B). This is accompanied by a lack of normal convergence and extension of the posterior notochord (B⬘). Cap1 mismatch morpholinoinjected embryos display normal polster formation and movement (C), as well as posterior convergence and extension of the notochord (C⬘). In Quo morphants, polster precursor cells are more dramatically dispersed, and fail to properly aggregate within the anterior-most mesoderm (D); this failure is accompanied by a poorly resolved overlying neural plate (lack of clear dlx3 staining). A strong defect in the convergence of the posterior notochord (D⬘) correlates with the lack of convergence of the anterior mesendoderm (D). Quo mismatch morpholino-injected embryos appear to be normal (E and E⬘).

Actin Regulation during Zebrafish Gastrulation 1635

Figure 4. Cap1 Is Required to Maintain a Cortical Actin Distribution within Polster Cells Abbreviations are as follows: WT, wild-type; MO, morpholino-injected; and MM, mismatch morpholino-injected. Phalloidin staining of the actin cytoskeleton (A–G) and Fkd2 immunolabeling of polster cells (A⬘–G⬘) reveals a cortical enrichment of actin within wild-type polster cells at 95% epiboly (A). Cap1 morphants (B and F) display reduced, discontinuous cortical actin in polster cells (arrows in [B]), whereas that of neighboring tissue is unaffected (F). Cap1 mismatch morpholinoinjected embryos display normal actin levels (C). Quo morphants, while displaying abnormally dispersed polster cells (G), display relatively normal cortical actin in both polster cells (D and G) and neighboring tissue (G). Cortical actin in Quo mismatch morpholinoinjected embryos is unaffected (E).

ures 4B and 4F; n ⫽ 27), whereas actin in neighboring cells appears to be unaffected (Figure 4F). Although cells appeared to be less densely packed in Quo morphants (Figure 2A), the cortical actin distribution was relatively normal (Figure 4D; n ⫽ 29). The actin distributions in polster cells of Cap1 and Quo mismatch morpholinoinjected embryos were indistinguishable from those in the wild-type (Figures 4A, 4C, and 4E). These results demonstrate that the restricted expression of cap1 is required for proper regulation of cortical actin distribution within polster cells. Combined with the requirement of Cap1 for the proper coordinated movements of these cells, this strongly suggests that Cap1 regulation of cortical actin is required for the morphogenetic event of anterior polster migration. Although we did not observe gross changes in the actin distribution in Quo morphant polster cells, the striking requirement of Quo for the proper migration of these cells is consistent with in-

volvement in a pathway leading to modulation of cell behavior via the actin cytoskeleton. Whether this putative RhoGEF mediates polster cell behaviors by regulating aspects of actin dynamics other than cortical enrichment or through means independent of actin specializations remains to be seen. Finer analyses of subcellular processes and actin dynamics, and a further characterization of Quo signaling, including RhoGTPase specificity, may yet reveal requirements for Quo in actinbased mechanisms of polster cell signaling. Interestingly, in the case of both morphants, the compromised polster phenotypes were accompanied by apparent defects in convergent extension, the set of gastrulation cell movements leading to a narrowing of the mediolateral axis and lengthening of the anteroposterior axis. In Cap1 morphants the notochord, marked by ntl expression, was less extended in the anterior-posterior direction and broader along the mediolateral axis (Figure

Current Biology 1636

2B⬘; n ⫽ 180). The lack of convergence within axial mesoderm was more pronounced in the case of Quo morphants (Figure 3D⬘), although axial extension was less compromised (Figure 3D⬘; Figure 2L; n ⫽ 225). As in the zebrafish convergent-extension mutant silberblick (wnt11) [26], the distance between polster cells marked by hgg1 and notochord cells marked by ntl is reduced in Cap1 morphants by bud stage, indicating a lack of extension behavior in prechordal plate mesoderm (data not shown). A similar reduction is seen in Quo morphants, although hgg1-expressing polster cells have failed to fully advance out of the prechordal plate region, so it is unclear whether posterior prechordal plate extension is impaired. Together, these observations are consistent with convergent-extension defects during late gastrulation [27]. Thus, Cap1 and Quo are specifically required for the formation and anterior movement of the polster, and surprisingly, this process may be required for the normal, coincident convergence and extension of posterior tissues of the embryo. The localized expression and putative function of Cap1 and Quo in modulating actin filament dynamics suggested that they would act autonomously in the developing polster and nonautonomously during convergent extension. In situ analysis of Cap1 and Quo gene expression in early blastula stage embryos revealed an absence of maternal mRNA transcripts consistent with this notion (data not shown). To further test this, we performed cell transplantation experiments in which donor cells from Cap1, Quo, or Cap1/Quo double morphants were introduced into wild-type hosts and their ability to undergo characteristic movements analyzed. Cap1/Quo double morphant cells transplanted into wildtype hosts at the lateral margin, which gives rise to lateral, posterior tissues, consistently converged toward the midline and became distributed in the A-P axis (Figures 5A–5C; n ⫽ 15), demonstrating that neither of these molecules is autonomously required for the convergentextension behaviors of these cells. However, when groups of Cap1 or Quo single morphant cells, transplanted into wild-type shield tissue, contributed to the polster lineage, the donor cells failed to move anteriorly in the proper manner. This failure led to the appearance of lagging or asymmetrical polsters, the severity of which reflected the levels of morphant cell contribution (Figures 5E–5G and 5M–5O; n ⫽ 17). In contrast, when morphant cells contributed primarily to more-posterior axial tissues derived from the shield, the polster and posterior tissues behaved normally (Figures 5I–5K and 5Q–5S; n ⫽ 9). Collectively, these results are consistent with Cap1 and Quo acting autonomously in the development and anterior movement of the polster, and nonautonomously in the simultaneous convergent extension of more-posterior tissues. Zebrafish convergent extension is best understood in terms of the regulation of cell polarity. Noncanonical Wnt signaling through the vertebrate equivalent of the Drosophila planar cell polarity (PCP) pathway is essential for the establishment of the cell polarity that is required for both the mediolateral cellular intercalation and directed cell migration that occur during convergent extension (reviewed in [28, 29]). Recent work also demonstrates that the dorsal-ventral gradient of bone mor-

Figure 5. Cap1 and Quo Act Autonomously in Polster Cell Movements and Nonautonomously in the Convergent Extension of Posterior Tissues (A–C) Cap1/Quo double morphant cells were transplanted 45⬚ lateral to the shield of wild-type, shield stage embryos (A). At the 3–5 somite stage, the morphant cells have converged to, and extended along, the midline (B and C). (D–S) Cap1 (D–K) or Quo (L–S) morphant cells containing rhodamine/biotin-dextran were transplanted into the shield of shield stage wild-type embryos (D, H, L, and P). At early tailbud stages, embryos were fluorescently imaged (E, I, M, and Q), labeled for hgg1 (F, J, N, and R), and stained for biotin (brown; [G, K, O, and S]), revealing the localization of the transplanted cells. Polsters containing Cap1 (E–G) or Quo (M–O) morphant cells show a lack of forward advancement in the morphant region (arrows), whereas polsters containing negligible numbers of morphant cells behave normally (I–K and Q–S). Red arrowheads provide comparative orientation.

phogenetic protein (BMP) activity in the zebrafish gastrula influences the deployment of the PCP pathway and thus regulates the relative contributions of cell migration and cell intercalation to the process [30]. Our finding that Cap1-/Quo-mediated polster migration may also influence global convergent-extension movements through a nonautonomous mechanism is consistent with recent observations that zebrafish embryos lacking Stat-3 activity in the prechordal plate display autonomous loss of prechordal plate migration along with nonautonomous defects in convergent extension [31]. Although yet-to-be-characterized properties of anterior mesendoderm may affect more global cell polarity, the probable direct role of Cap1 and Quo in mediating actindriven cell movements suggests that the migration of anterior tissues per se is responsible for any secondary effect on more-widespread convergent-extension movements. In many zebrafish convergent-extension mutants

Actin Regulation during Zebrafish Gastrulation 1637

that lack the function of widely expressed PCP components, the polster is anteroposteriorly extended concomitantly with convergent-extension defects throughout the embryo [26, 27, 32]. In cell transplant experiments in which cells from the PCP mutant slb contributed to polster cells in wild-type embryos, polster morphology was relatively normal, suggesting that the polster abnormalities in PCP mutants may be a result of more-global cell polarity and convergent extension defects, as opposed to a requirement for these genes in specific cell behaviors during polster morphogenesis [22]. However, the extent to which polster defects in PCP mutants contribute to global convergent-extension defects, as suggested by this work, remains unresolved. Interestingly, the Quo morphant is novel in that its polster cells are also dispersed mediolaterally, in addition to anteroposteriorly, representing the first example of such a dramatic lack of anterior-mesoderm convergence. The corresponding lack of mediolateral consolidation of the posterior notochord provides evidence suggesting a role for anterior mesoderm convergence in the convergence of posterior mesoderm. The presence of this strong convergence defect without a severe posterior-extension defect in Quo morphants is consistent with recent zebrafish work demonstrating that convergence and extension are separable events [33, 34]. Taken together, these works describe a growing diversity of both polster and convergent-extension phenotypes and illustrate how the morphogenetic cell movements within anterior mesoderm and posterior tissues are highly integrated, yet differentially regulated events. A direct role for the migration of anterior mesoderm acting in concert with the convergent extension of posterior mesoderm driven by mediolateral cellular intercalation has been described during the gastrulation of the anuran amphibians, Xenopus and Rana [35–37]. In addition, similar to that of zebrafish and unlike that of anurans, the embryo of the urodele amphibian Pleurodeles does not undergo such widespread mesodermal mediolateral intercalation, and the active migration of anterior mesoderm plays a necessary role in overall extension movements [38]. Indeed, the polster cell movements described here are reminiscent of anterior-mesoderm migration in the amphibian gastrula, where cells have been shown to migrate toward the animal pole on aligned networks of ectodermal fibronectin deposits that act as directional guidance cues [35, 39]. It is tempting to speculate that a conserved molecular mechanism similar to that described here exists in amphibian mesoderm. Recent work found a Xenopus cap1 ortholog to be expressed in the mesoderm of the gastrula [40], and it will be interesting to determine whether it functions in anterior migration. Quo in particular, as a large, multidomain RhoGEF, would be potentially well placed within signal transduction pathways to interpret guidance cues during these events.

Conclusions This work provides genetic and cellular evidence indicating that the regional expression of cap1 and quo is required for coordinated cell behaviors in the developing anterior mesendoderm of the zebrafish embryo. In addi-

tion, cap1 is required for the proper maintenance of cortical actin in these cells. These results demonstrate for the first time that a mechanism involving tissue restriction of actin-regulatory molecules is employed in the control of morphogenetic cell movements in vertebrate development. In addition to characterizing further details of the cell biology mediated by this regionally restricted regulation of cellular actin dynamics, it will be of great interest to determine how these localized events are integrated with the global control of cell polarity to effect morphogenetic movements in a broad range of architecturally varied vertebrate embryos. Supplemental Data Supplemental Data including three figures, a movie, and detailed Experimental Procedures used in this work are available at www. current-biology.com/cgi/content/full/14/18/1632/DC1/ Acknowledgments We thank J.B. Wallingford, L. Solnica-Krezel, C.P. Heisenberg, C. Henry, and T. Han for thoughtful discussion of the data, R.E. Hill and D.I. Bassett for critical reading of the manuscript, and the Currie group, S. Bruce, and B. Mellone for efforts and support. We are grateful to R. Warga for generously providing anti-Fkd antibodies and M. Ekker for dlx-3 cDNA. Work supported by Burroughs Wellcome Fund Research Travel Grant (D.F.D.), National Institutes of Health National Research Service Award (D.F.D.), National Institutes of Health grant GM61952-01 (S.L.A.), Pew Scholar Award (S.L.A.), and Medical Research Council Career Development Award (P.D.C.). B.T. and C.T. supported by Institut National de la Sante´ et de la Recherche Me´dicale, Centre National de la Recherche Scientifique, Hoˆpital Universitaire de Strasbourg, Association pour la Recherche sur le Cancer, Ligue National Contre le Cancer, and National Institutes of Health (R01 RR15402). Received: June 17, 2004 Revised: July 27, 2004 Accepted: July 27, 2004 Published: September 21, 2004 References 1. Mitchison, T.J., and Cramer, L.P. (1996). Actin-based cell motility and cell locomotion. Cell 84, 371–379. 2. Etienne-Manneville, S., and Hall, A. (2002). Rho GTPases in cell biology. Nature 420, 629–635. 3. Burridge, K., and Wennerberg, K. (2004). Rho and Rac take center stage. Cell 116, 167–179. 4. Settleman, J. (2001). Rac ‘n Rho: The music that shapes a developing embryo. Dev. Cell 1, 321–331. 5. Zheng, Y. (2001). Dbl family guanine nucleotide exchange factors. Trends Biochem. Sci. 26, 724–732. 6. Jacinto, A., Wood, W., Woolner, S., Hiley, C., Turner, L., Wilson, C., Martinez-Arias, A., and Martin, P. (2002). Dynamic analysis of actin cable function during Drosophila dorsal closure. Curr. Biol. 12, 1245–1250. 7. Zhou, Z., Caron, E., Hartwieg, E., Hall, A., and Horvitz, H.R. (2001). The C. elegans PH domain protein CED-12 regulates cytoskeletal reorganization via a Rho/Rac GTPase signaling pathway. Dev. Cell 1, 477–489. 8. Benlali, A., Draskovic, I., Hazelett, D.J., and Treisman, J.E. (2000). act up controls actin polymerization to alter cell shape and restrict Hedgehog signaling in the Drosophila eye disc. Cell 101, 271–281. 9. Baum, B., and Perrimon, N. (2001). Spatial control of the actin cytoskeleton in Drosophila epithelial cells. Nat. Cell Biol. 3, 883–890. 10. Keller, R. (2002). Shaping the vertebrate body plan by polarized embryonic cell movements. Science 298, 1950–1954.

Current Biology 1638

11. Wodarz, A. (2002). Establishing cell polarity in development. Nat. Cell Biol. 4, 39–44. 12. Habas, R., Yoichi, K., and He, X. (2001). Wnt/Frizzled activation of Rho regulates vertebrate gastrulation and requires a novel Formin homology protein Daam1. Cell 107, 843–854. 13. Iijima, M., Huang, Y.E., and Devreotes, P. (2002). Temporal and spatial regulation of chemotaxis. Dev. Cell 3, 469–478. 14. Hubberstey, A.V., and Mottillo, E.P. (2002). Cyclase-associated proteins: CAP1acity for linking signal transduction and actin polymerization. FASEB J. 16, 487–499. 15. Moriyama, K., and Yahara, I. (2002). Human CAP1 is a key factor in the recycling of cofilin and actin for rapid actin turnover. J. Cell Sci. 115, 1591–1601. 16. Debant, A., Serra-Pages, C., Seipel, K., O’Brien, S., Tang, M., Park, S.H., and Streuli, M. (1996). The multidomain protein Trio binds the LAR transmembrane tyrosine phosphatase, contains a protein kinase domain, and has separate rac-specific and rhospecific guanine nucleotide exchange factor domains. Proc. Natl. Acad. Sci. USA 93, 5466–5471. 17. Kane, D.A., and Warga, R.M. (1994). Domains of movement in the zebrafish gastrula. Semin. Dev. Biol. 5, 101–109. 18. Solnica-Krezel, L., Stemple, D.L., and Driever, W. (1995). Transparent things: Cell fates and cell movements during early embryogenesis of zebrafish. Bioessays 17, 931–939. 19. Vogel, A., and Gerster, T. (1997). Expression of a zebrafish cathepsin L gene in anterior mesendoderm and hatching gland. Dev. Genes Evol. 206, 477–479. 20. Kimmel, C.B., Warga, R.M., and Schilling, T.F. (1990). Origin and organization of the zebrafish fate map. Development 108, 581–594. 21. Saude, L., Woolley, K., Martin, P., Driever, W., and Stemple, D.L. (2000). Axis-inducing activities and cell fates of the zebrafish organizer. Development 127, 3407–3417. 22. Heisenberg, C.P., and Nusslein-Volhard, C. (1997). The function of silberblick in the positioning of the eye anlage in the zebrafish embryo. Dev. Biol. 184, 85–94. 23. Nasevicius, A., and Ekker, S.C. (2000). Effective targeted gene ‘knockdown’ in zebrafish. Nat. Genet. 26, 216–220. 24. Freeman, N.L., and Field, J. (2000). Mammalian homolog of the yeast cyclase associated protein, CAP1/Srv2p, regulates actin filament assembly. Cell Motil. Cytoskeleton 45l, 106–120. 25. Warga, R.M., and Nusslein-Volhard, C. (1999). Origin and development of the zebrafish endoderm. Development 126, 827–838. 26. 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. 27. Solnica-Krezel, L., Stemple, D.L., Mountcastle-Shah, E., Rangini, Z., Neuhauss, S.C.F., Malicki, J., Schier, A.F., Stainer, D.Y.R., Zwartkruis, S.A., and Driever, W. (1996). Mutations affecting cell fates and cellular rearrangements during gastrulation in zebrafish. Development 123, 67–80. 28. 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. 29. Tada, M., Concha, M.L., and Heisenberg, C.-P. (2002). Noncanonical Wnt signaling and regulation of gastrulation movements. Semin. Cell Dev. Biol. 13, 251–260. 30. Myers, D.C., Sepich, D.S., and Solnica-Krezel, L. (2002). Bmp activity gradient regulates convergent extension during zebrafish gastrulation. Dev. Biol. 243, 81–98. 31. 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. 32. Hannus, M., Feiguin, F., Heisenberg, C.-P., and Eaton, S. (2002). Planar cell polarization requires Widerborst, a B regulatory subunit of protein phosphatase 2A. Development 129, 3493–3503. 33. Glickman, N.S., Kimmel, C.B., Jones, M.A., and Adams, R.J. (2003). Shaping the zebrafish notochord. Development 130, 873–887. 34. Bakkers, J., Kramer, C., Pothof, J., Quaedvlieg, N.E.M., 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.

35. Winklbauer, R., and Nagel, M. (1991). Directional mesoderm cell migration in the Xenopus gastrula. Dev. Biol. 148, 573–589. 36. Johnson, K.E., Darribere, T., and Boucaut, J.C. (1993). Mesodermal cell adhesion to fibronectin-rich fibrillar extracellular matrix is required for normal Rana pipiens gastrulation. J. Exp. Zool. 265, 40–53. 37. Keller, R., Davidson, L., Edlund, A., Elul, T., Ezin, M., Shook, D., and Skoglund, P. (2000). Mechanisms of convergence and extension by cell intercalation. Philos. Trans. R. Soc. Lond. B Biol. Sci. 355, 897–922. 38. Shi, D.-L., Delarue, M., Darribe`re, T., Riou, J.-F., and Boucaut, J.-C. (1987). Experimental analysis of the dorsal marginal zone in Pleurodeles waltl gastrulae. Development 100, 147–161. 39. Nakatsuji, N., and Johnson, K.E. (1983). Conditioning of a culture substratum by the ectodermal layer promotes attachment and oriented locomotion by amphibian gastrula mesodermal cells. J. Cell Sci. 59, 43–60. 40. KhosrowShahian, F., Hubberstey, A.V., and Crawford, M.J. (2002). CAP1 expression is developmentally regulated in Xenopus. Mech. Dev. 113, 211–214. Accession Numbers The accession numbers AY162326 and AY654285 have been assigned to the GenBank database.