Daam1 Regulates the Endocytosis of Ephb During the Convergent Extension of the Zebrafish Notochord

Daam1 regulates the endocytosis of EphB during the convergent extension of the zebrafish notochord

Yasuyuki S. Kida*†, Takayuki Sato†, Kota Y. Miyasaka†, Asami Suto†, and Toshihiko Ogura*†‡

*Department of Developmental Neurobiology, Institute of Development, Aging, and Cancer and †Graduate School of Life Sciences, Tohoku University, 4-1, Seiryo, Aoba, Sendai, Miyagi 980-8575, Japan

Edited by Kathryn V. Anderson, Sloan–Kettering Institute, New York, NY, and approved February 21, 2007 (received for review October 10, 2006) Convergent extension (CE) movement of cells is one of the funda- distinct responses. EphB/EphrinB-mediated repulsion is known mental processes that control the organized morphogenesis of to be controlled by a mechanism involving endocytosis of Eph tissues and organs. The molecular events connecting the nonca- receptors (17, 18), yet upstream signaling cascades controlling nonical Wnt pathway and CE movement, however, are not well this endocytic event remain to be solved. understood. We show that subcellular localization of Daam1, an In this study, we show that Daam1 regulates endocytosis of essential component of noncanonical Wnt signaling, changes dy- EphB molecules and cytoskeletal remodeling to coordinate cell namically during notochord formation. In the early phases, Daam1 behavior during notochord development. complexes with EphB receptors and Disheveled 2. This complex is incorporated into endocytic vesicles in a dynamin-dependent man- Results ner, thereby resulting in the removal of EphB from the cell surface Subcellular Localization of Daam1. Using an anti-Daam1 antibody with subsequent switching of cell adhesiveness. In the next step, to label HEK293 cells, we found Daam1 in vesicle-like dot Daam1 colocalizes with the actin cytoskeleton to induce morpho- structures where early endosomal autoantigen 1 (EEA1) colo- logical extension of cells. We elucidate the molecular mechanism calized (Fig. 1A). In addition, Daam1 and the fluorescent styryl underlying the CE movement of notochord cells with Daam1 as a dye (FM1–43FX) were both located in the vesicles, indicating dynamic coordinator of endocytosis and cytoskeletal remodeling. incorporation of FM1–43FX into Daam1-positive endocytic vesicles (Fig. 1B). When stimulated by Wnt1-conditioned me- Daam ͉ Eph ͉ Wnt ͉ planar cell polarity dium, cells changed their shape and developed cytoplasmic protrusions in which the Daam1-positive vesicles accumulated he establishment of cell polarity is one of the fundamental (yellow arrowheads, Fig. 1C). Colocalization with Dvl2 and Tprocesses critical for cell division, migration, and convergent incorporated transferrin was also confirmed [see supporting extension (CE) cell movement during development (1, 2). Po- information (SI) Fig. 6]. larity within the cells in the epithelial plane, planar cell polarity Endocytosis serves numerous functions in cells, including internalization of ligand-receptor molecules, signal transduction, (PCP), has been characterized extensively in Drosophila, high- and cell–cell communication. It has been shown that small Rho lighting two core PCP genes, frizzled (Dfz) and disheveled (dsh) GTPases, such as RhoB, RhoD, and Rab5, play pivotal roles (1, 3). Both genes are highly conserved in vertebrates, playing in cell migration (19), which is also regulated by endocytosis. essential roles in the establishment of cell polarity and cell When the EGFP-tagged RhoB protein was expressed in movement (4, 5). In invertebrates and vertebrates, the Fz and HEK293 cells, the Daam1 and RhoB proteins colocalized in the Dsh (Dvl) families control cell fate and proliferation as part of vesicles and at the cell membrane (Fig. 1D). When RhoB(N19), the canonical Wnt signaling pathway involving ␤-catenin (6). In a dominant-negative mutant RhoB (20), was expressed, the contrast, Fz and Dvl are involved in noncanonical Wnt signaling number of endocytic vesicles was decreased. Instead, a large in the PCP/CE pathway. Fz and Dv1 activate another core PCP vacuole that was both Daam1- and RhoB-positive was formed molecule, RhoA, a member of the Rho family of GTPases (2, 7). (Fig. 1E). These results confirm that Daam1 and RhoB colo- Its activation requires a formin-homology (FH) protein, Daam1 calize in the endocytic vesicles even when endocytosis is abnor- (disheveled-associated activator of morphogenesis) (8). Daam1 mally disrupted. binds to both RhoA and Dvl, making a pivotal complex that is The repulsion of cells is controlled by dynamin, a large essential for PCP/CE signaling during Xenopus gastrulation. guanosine triphosphatase that functions during endocytic traf- Hence, Daam1 relays the noncanonical Wnt signal to the ficking of membrane molecules (21). When a dominant-negative cytoskeleton, which causes it to initiate its remodeling to change mutant of dynamin [Dnm1(K44A)], defective in GTP binding cell behaviors. and hydrolysis (21), was expressed, Daam1 was not distributed in During gastrulation, cells migrate away from their original the vesicles, but rather localized at the cell membrane (yellow position in a concerted fashion specifically through modulation arrows, Fig. 1F). Daam1 was present in the vesicles of a cell not of cell adhesion (2). Wnt11, which acts in the noncanonical Wnt expressing Dnm1(K44A) (yellow arrowheads, Fig. 1F). These cascade, is involved in this process, controlling the cohesion of observations indicate that Daam1 colocalizes in the endocytic mesoendodermal cells through the Rab5c-mediated endocytosis vesicles with dynamin. In addition, Daam1 interacts with RhoA of E-cadherin (9). In addition to the noncanonical Wnt cascade, multiple signaling pathways, such as BMP (10), Stat3 (11), and Eph/Ephrin Author contributions: Y.S.K. designed research; Y.S.K., T.S., K.Y.M., and A.S. performed (12–14), regulate key processes of gastrulation. Eph receptor research; Y.S.K. and T.O. wrote the paper. tyrosine kinases and Ephrins play important roles in cell migra- The authors declare no conﬂict of interest. tion and adhesion during development. Activation of Eph leads This article is a PNAS Direct Submission. to opposite effects, deadhesion/repulsion or adhesion/attraction Abbreviations: CE, convergent extension; MO, morpholino oligonucleotide; PCP, planar of cells, depending on the cell type and signaling contexts (15). cell polarity. Consequently, this cascade regulates diverse developmental ‡To whom correspondence should be addressed. E-mail: [email protected]. processes, including gastrulation, segmentation, angiogenesis, This article contains supporting information online at www.pnas.org/cgi/content/full/ migration of neural crest cells, and axonal pathfinding (16). It is 0608946104/DC1. unclear which mechanisms enable switching between these © 2007 by The National Academy of Sciences of the USA

6708–6713 ͉ PNAS ͉ April 17, 2007 ͉ vol. 104 ͉ no. 16 www.pnas.org͞cgi͞doi͞10.1073͞pnas.0608946104 Downloaded by guest on September 29, 2021 A A Lane: 1 2 EphB2-HA+ + Myc-Daam1 + + Daam1 EEA1 ephrinB1-FC -+ B I.P IB:HA Ly sate Daam1/DAPI FM1-43FX C control CM Wnt1 CM I.P: Myc B Lane: 1 2 3 4 Daam1/DAPI ephrinB1-Fc+ + ++ D EphB2-HA+ + ++ Myc-Daam1 + +

GFP-RhoB Daam1 FL-Dvl2 + + GFP-Rho Daam1 E Wnt 1 Wnt 3a (N19) I.P: Myc Flag Myc Flag IB:HA GFP-Dnm1 F (K44A) I.P EphB2 IB:pTy Daam1 Lysate Fig. 1. Localization of Daam1 in the endocytic vesicles. (A) In HEK293 cells, endogenous Daam1 was observed in the vesicles, which were visualized by staining with anti-Daam1 and anti-EEA1 antibodies. (B) After exposure to Fig. 2. Association of phosphorylated EphB2 with Daam1 and Wnt signaling. fluorescent FM1–43FX, cells were fixed and stained with the anti-Daam1 (A) HA-tagged EphB2 and Myc-tagged Daam1 were coexpressed in HEK293 antibody and DAPI. Three spots of incorporated FM1–43FX (arrows) colocal- cells. When EphrinB1-FC was added, Daam1 and EphB2 associated, resulting in ized with the Daam1-positive vesicles. A small area indicated by white dotted coprecipitation with Daam1 by an anti-Myc tag antibody. (B) The EphB2 lines was magnified and shown (Inset). (C) When stimulated by Wnt1- proteins associated with Daam1 were phosphorylated when coprecipitated conditioned medium, HEK293 cells changed cell shape to form protrusions. with Daam1 and detected with an anti-phospho-tyrosine antibody (lane 1). In The Daam1-positive endocytic vesicles accumulated in the protrusion (arrows). contrast, phosphorylation of EphB2 molecules associated with Dvl2 was low (D) After GFP-tagged RhoB was expressed in HEK293 cells, cells were stained (lane 2). In the presence of Daam1 and Wnt1, phosphorylation of EphB2 with the anti-Daam1 antibody. Colocalization of Daam1 and RhoB was evi- associated with Daam1 was enhanced (lane 3) compared with that of Dvl2- dent in the vesicles (arrows). (E) Introduction of a dominant-negative RhoB associated EphB2 stimulated by Wnt3a (lane 4). (N19) tagged with GFP reduced the number of vesicles and resulted in the formation of a large vacuole, in which Daam1 colocalized (arrows). (F) Expres- sion of a kinase-dead mutant of GFP-Dnm1 [Dnm1(K44A)] disrupted the Daam1 and Dvl2 participate. To confirm this possibility, we vesicle-like distribution of Daam1. Instead, Daam1 localized at the cell mem- performed an immunoprecipitation assay using HEK293 cells. In brane (arrows). In a Dnm1(K44A)-negative cell, Daam1 localized in the endo- the absence of EphrinB1-FC stimulation, HA-tagged EphB2 and cytic vesicles (arrowheads). Myc-Daam1 proteins associated weakly, whereas most of EphB2 protein coprecipitated along with Daam1 from cells stimulated (8), RhoB, and Dnm1 (data not shown). These data again by EphrinB1-FC (Fig. 2A). This indicates that ligand-stimulated support the involvement of Daam1 in the trafficking of the EphB2 preferentially associates with Daam1. The same result endocytic vesicle. Localization of overexpressed Myc-tagged was obtained when EphB1 was used (data not shown). Finally, BIOLOGY Daam1 was also confirmed (SI Fig. 6). we confirmed that the N-terminal of Daam1 interacted with the DEVELOPMENTAL C-terminal domain of EphB1 (SI Fig. 7). EphB1 Interacts with Daam1. Ephrins and Eph receptors play a pivotal role in the regulation of cell migration, with the endo- Noncanonical Wnt Signal Induces Association of Phosphorylated EphB cytosis of the EphB/EphrinB complex being a key event for and Daam1. In HEK293 cells, the EphB2 molecules that copre- termination of cell–cell adhesion and subsequent repulsive cipitated with Daam1 were phosphorylated, as evidenced by movement (17, 18). It has been shown that EphB receptors anti-phospho-tyrosine antibody binding (Fig. 2B, lane 1). In interact with Dvl2 in the Xenopus eye (22). Thus, it is likely that contrast, the EphB2 proteins that coprecipitated with Dvl2 were EphB receptors are directly involved in endocytosis in which phosphorylated at a lower level (Fig. 2B, lane 2). To address the

Kida et al. PNAS ͉ April 17, 2007 ͉ vol. 104 ͉ no. 16 ͉ 6709 Downloaded by guest on September 29, 2021 putative effects of Wnt signaling on the phosphorylation of EphB2, we analyzed its phosphorylation level in cells stimulated ABC D by either Wnt1 or Wnt3a. Wnt1 activates both the canonical and noncanonical Wnt cascades (8). Nonetheless, canonical signaling, as measured with the TOP-flash reporter plasmid, was not stimulated in Wnt1-stimulated HEK293 cells expressing Daam1. In contrast, Wnt3a activated the TOP-flash reporter activity in wt embryo zNDaam1a injected Wnt3a-stimulated cells expressing Dvl2 (22) (SI Fig. 8). E EphB2 proteins were extensively phosphorylated in the pres- 11 hpf ence of noncanonical Wnt signaling, compared with those stimulated by the canonical Wnt pathway (Fig. 2B, lanes 3 and 4). These data suggest that the noncanonical Wnt signal induces the preferential association of Daam1 and phosphorylated EphB. Consistent with this, the nonphosphorylated mutant of FGHIJK EphB1 (K631F/K660F) does not bind to Daam1 (SI Fig. 7).

Disruption of the Noncanonical Wnt or the EphB Signaling Cascades Resulted in a Phenotype Consistent with CE. To explore the Wnt and EphB signaling cascades further, we performed loss-of-function experiments with zebrafish embryos. Injection of mRNA encoding zNDaam1a resulted in a short body axis, a kinked tail, and cyclopia (Fig. 3B). NDaam1, lacking its C-terminal half, acted as a dominant-negative form (8). The same phenotype was induced by injection of an mRNA encoding Xdd1, a mutant Dvl2 lacking the C-terminal part of its PDZ domain (data not shown). wt zNDaam1a Sol zDnm1 Dnm1- EphB2/4 mEphB1 MO Injection of another mRNA encoding a soluble form of EphB1 14 hpf (K44A) MO comprising only its extracellular domain again resulted in a similar phenotype (Fig. 3C) (12, 14). The phenotypes of L CE phenotype (%) M CE phenotype (%) Daam1a- and EphB4-morphants were milder than those produced by the zNDaam1a and soluble EphB1 injection experiments (SI Fig. 9). These results may be explained by the redundancy of the Daam and EphB genes. Indeed, not only zNDaam1a 19.2 Sol mEphB1 28.8 Daam1b, Daam2, EphB2, and EphB3 but also EphA4b and Sol mEphB1 Sol mEphB1 11.4 EphA3 are expressed in the notochord (ZFIN web). Injection of 11.6 + zCDaam1a Dnm1(K44A) mRNA induced a similar phenotype (Fig. 3D). zNDaam1a + To visualize notochord cell shape, EGFP protein was ex- 49.4 zCDaam1a 9.0 pressed with the promoter of the floating head (flh) gene (22) Sol mEphB1 (Fig. 3E). In the wild-type embryo, notochord cells elongated to myod expression at 14 hpf show the CE cell movement. In contrast, when the flh promoter– EGFP construct was coinjected with zNDaam1a mRNA, cells Fig. 3. Disruption of the noncanonical Wnt pathway or the EphB signaling became round and failed to show either polarized cell shape or cascade induces the same CE phenotype. (A–D) Morphology [72 hours post- the typical CE movement. This indicates that inactivation of fertilization (hpf)] of the wild-type embryo (A), embryo injected with ze- Daam1a results in an abnormal cell shape and a loss of polarized brafish NDaam1a mRNA (B), mRNA encoding a soluble form of mouse EphB1 (C), and mRNA encoding a K44A mutant of zebrafish Dnm1 (D). These embryos cell movement. share a short and bent body axis (red arrowheads), as well as cyclopia (blue Next, we examined the expression patterns of myoD in em- arrowheads). (E) EGFP protein was expressed in the notochord cells to visualize bryos injected with the mutant mRNAs and morpholino oligo- their shape. In the wild-type embryo, the notochord cells elongate to show the nucleotides (MOs). This gene is normally expressed in adaxial intercalation movement. When zNDaam1 was expressed, the notochord cells cells and segmented somites (Fig. 3F). In the zNDaam1-injected became round and did not develop a polarized shape. Dorsal views at 11 hpf embryo, expression of myoD in the somites became compressed are shown. Rostral to the left, caudal to the right. (F) At 14 hpf, myoD is (Fig. 3G). Expression of the soluble form of EphB1 and the expressed in the somites and the adaxial cells in the wild type. (G) Injection of Dnm1 mutant K44A resulted in the same phenotype (Fig. 3 H zNDaam1 mRNA resulted in a short body axis with compressed somites. (H and and I). Injection of MO against Dnm1 also produced a similar I) Injection of mRNA encoding the soluble form of mouse EphB1 (H)orthe dominant-negative dynamin1 [Dnm1(K44A)] (I) induced a similar phenotype. short and bent notochord (Fig. 3J). Coinjection of MOs against (J) Injection of Dnm1-MO resulted in similar but more severe defects. (K) EphB2 and EphB4, both of which are expressed in the notochord, Coinjection of EphB4-MO and EphB2-MO induced severe defects. The noto- also produced similar defects with changes in gene expression. chord was short, broader, and bent. Pictures shown in F–K are shown at the These results indicate that the abnormal CE movement results same magnification. (L) Injection of mRNA encoding zNDaam1 or the soluble from the disruption of Daam1 and EphB functions (additional form of EphB1 induced the abnormalities of CE movement in 19.2% or 11.6% supporting data are shown in SI Figs. 9 and 10). of injected embryos, respectively, as judged by expression of myoD at 14 hpf. To determine whether a functional relationship existed be- Coinjection of these mRNAs induced the same abnormality in 49.4% of tween Daam1 and EphB1, we coexpressed NDaam1 and the embryos. (M) Injection of a soluble mEphB1-FC mRNA induced the CE defect soluble form of EphB1 in embryos. Coexpression of these two in 29.8% of injected embryos. Coinjection with zCDaam1a mRNA rescued this ϭ defect, and the CE abnormality was observed in only 11.4% of embryos. proteins produced defects in 49.4% of injected embryos (n 97), Injection of zCDaam1 mRNA alone induced the defect in 9.0% of embryos. whereas the defects were observed at lower rates when either NDaam1 (19.2%, n ϭ 52) or the soluble EphB1 (11.6%, n ϭ 42) were expressed individually (Fig. 3L). The relationship was also (n ϭ 59). Coexpression with zCDaam1 rescued the abnormalities explored with a rescue experiment using a constitutively active in a significant number (11.4%, n ϭ 44). Injection of CDaam1 form of Daam1 (zCDaam1) (Fig. 3M). Misexpression of the mRNA alone induced slight defects in only 9.0% of embryos soluble form of EphB1 resulted in defects in 28.8% of embryos (n ϭ 67).

6710 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0608946104 Kida et al. Downloaded by guest on September 29, 2021 A

0 min 20 min 40 min 60 min 80 min 100 min

B C

15 min 19 min 23 min 52 min 82 min 102 min

D E F 12 hpf Wnt11 MO

GFP- Phalloi GFP- Daam1 din Daam1

Fig. 4. Subcellular localization of Daam1 in notochord cells. (A) Serial notochord pictures of a notochord cell taken at 20-min intervals. EGFP-Daam1-positive morphogenesis vesicles (red arrows) moved from the caudal cell surface to the center (green arrowhead). Daam1-positive fibers were formed on the caudal side of the cell (blue line). (B) Serial pictures of another notochord cell taken at 4-min inter- cellular dynamics vals. EGFP-Daam1-positive vesicles (red arrows) moved from the caudal cell of notochord cells surface to the center (green arrowhead). (C) Serial pictures of the cell shown in (A). GFP-tagged Daam1 accumulated in the fibers near the caudal cell surface (blue bars). (D and E) Subcellular localization of GFP-tagged Daam1 is Fig. 5. Noncanonical Wnt signaling and Daam1-mediated endocytosis of shown in green. At this stage, Daam1 colocalized with F-actin in the cell cortex EphB. On binding to EphrinBs, EphB receptor molecules are phosphorylated and the fibers formed near the caudal surface (yellow arrowheads). Note that and recruited to make a ternary complex with Dvl2 and Daam1. This complex the colocalization is more evident at the lateral ends that make tight junctions is incorporated into the endocytic vesicles in a dynamin-dependent manner. with surrounding tissues (blue arrowheads in E). (F) When wnt11 MO was Active noncanonical Wnt signaling is necessary for the formation of this injected, the notochord cells did not elongate but, rather, formed small spiky ternary complex. After the endocytic removal of EphB, cells dissociate to enter protrusions where Daam1 colocalized. Note that Daam1 did not localize in the the next steps of CE cell movement, migration and elongation. In the absence endocytic vesicles or the fibers in the cell cortex and, instead, stayed at the cell of the noncanonical Wnt signal or in the presence of the canonical Wnt signal, membrane. EphB receptors associate with Dvl2 alone to form a nonfunctional complex that remains on the cell surface. This likely takes place after dephosphoryla- tion by an unknown phophatase activity. Consequently, cells do not dissoci- Dynamic Subcellular Localization of Daam1 in the Zebrafish Noto- ate. In CE movement, dissociation and elongation of cells, mediated through chord. To further confirm the roles of Daam1 in the CE move- the EphB/Dvl2/Daam1 complex and the Dvl2/Daam1/ROCK cascade, respec- ment, we observed the dynamic subcellular localization of tively, are tightly coupled to achieve orchestrated morphogenetic movement. EGFP-fused Daam1 proteins, which were expressed from an EGFP-Daam1 fusion gene driven by the Flh promoter (22) in the was evident (Fig. 4D). This dynamic change of localization zebrafish notochord. EGFP-Daam1 proteins were detected in implies that Daam1-mediated Eph endocytosis releases the some vesicles in the cortex of cells at the early intercalation adhesive properties of the caudal cell surface of the notochord A stages (Fig. 4 ). In addition to this, a bright accumulation of cells. In the next step, Daam1 is involved in the elongation of cell signals was observed in the center of the cells. Although we failed shape, associating with F-actin that aligns along the direction of to stain this organelle with an antibody, we suspected it was the cell extension and contributes to tight adhesive connections with Golgi apparatus, because colocalization of Daam1 with the Golgi surrounding tissues. apparatus was confirmed in HEK293 cells in vitro (data not When MO against wnt11, which plays important roles during shown). Ten hours after injection of this construct, GFP-positive the CE cell movement (23), was injected, the notochord cells vesicle-like dots actively internalized in the caudal half of the changed their cell shape, forming small protrusions where cells (Fig. 4B). As the CE movement proceeded, the cells became Daam1 colocalized (Fig. 4E). In these cells, Daam1 did not compressed, and GFP signals accumulated in fibers elongating localize in the endocytic vesicles or the fibers in the cell cortex. laterally near the caudal surface (Fig. 4C). These findings This indicates that the subcellular localization depends on the indicate that the Daam1 protein changes its subcellular local- noncanonical Wnt cascade. ization dynamically. We also found that the subcellular localization of Daam1 was asymmetrical, with the caudal half active Discussion BIOLOGY in endocytosis (Fig. 4A and SI Movies 1 and 2). Based on our data, we propose the following model to describe the role of EphB and Dvl2 in noncanonical Wnt signaling (Fig. DEVELOPMENTAL Cytoskeleton and Daam1. Next, we explored the nature of the 5). EphB receptor tyrosine kinases are phosphorylated after Daam1-positive fibers formed in the later stages. As the CE cell clustering triggered by binding to their cognate ligands Eph- movement proceeded, cells became compressed. In these elon- rinBs. In the absence of noncanonical Wnt signaling or in the gating notochord cells, phalloidin-positive F-actin structures presence of canonical Wnt signaling, the EphB receptors do not were evident at 12 hpf. In these stages, Daam1 colocalized with make a complex with Daam1. In the presence of noncanonical the cytoskeleton formed at the cell cortex (Fig. 4D) and the Wnt signaling, phosphorylated EphB molecules are recruited to lateral ends (Fig. 4E, blue arrowheads). When the two images make a ternary complex with Dvl2 and Daam1. This complex is were merged, colocalization of these two distinct components transported to the endocytic vesicles, whose formation depends

Kida et al. PNAS ͉ April 17, 2007 ͉ vol. 104 ͉ no. 16 ͉ 6711 Downloaded by guest on September 29, 2021 on dynamin. The subsequent endocytic removal of EphB mol- RhoB(N19) plasmids were kindly provided by C. Rondanino and ecules induces cell repulsion, followed by the initiation of CE cell G. Apodaca (University of Pittsburgh, Pittsburgh, PA) (20). movement (Fig. 5). In CE cell movement, dissociation and Mouse EphBs and EphrinBs were kindly provided by M. Tanaka elongation of cells are mediated through the EphB/Dvl2/Daam1 (National Cancer Center Research Institute, Tokyo, Japan) (30). ternary complex and the Dvl2/Daam1/ROCK cascade (8), re- Mouse EphB1 and EphB2 cDNAs were obtained by RT-PCR spectively. The two cascades are coordinated to achieve the from mouse cDNAs. Mutant forms of EphB1 were constructed orchestrated morphogenetic movement of cells (Fig. 5). with the PCR techniques and appropriate primers. Mutants of Wnt11 regulates the E-cadherin-mediated cohesion of the me- mouse Dvl2 (FL-Dvl2, MT-Dvl2-DIX, -PDZ, -DEP) and Xdd1 soendoderm progenitors (9). In this process, Rab5, one of the key were kindly gifted by R. Habas (University of Medicine and regulators of early endocytosis and intracellular trafficking, medi- Dentistry of New Jersey, Newark, NJ) (8) and S. Y. Sokol ates Wnt11 activity. The dynamic modulation of cell cohesion (Mount Sinai School of Medicine, New York, NY) (4), respec- through the interaction of Wnt11 and Rab5 as well as the trafficking tively. Zebrafish dynamin1 cDNA (IMAGp998N1014600Q) was of E-cadherin is pivotal for the migration of many cell types. The obtained from the RZPD, and their mutants were constructed notochord, however, is less sensitive to changes in Rac5 activity (9). by PCR-based techniques. Maintenance of the fish colony and Our data reveal that the noncanonical Wnt cascade regulates this whole-mount in situ hybridization were performed as described process through endocytosis in a distinctive way, namely the switch- (31). pFlh-EGFP was a kind gift of M. E. Halpern (Carnegie ing of EphB-mediated repulsion and adhesion of cells. Further- Institute, Baltimore, MD) (22). more, our data identifies the direct molecular links, highlighting Daam1, Dvl2, and dynamin as central players that commence the Immunocytochemistry, Immunoprecipitation, and Western Blotting. dynamic behaviors of the notochord cells. Transfection, immunocytochemistry, and Western blotting of In axonal pathfinding, EphBs and Wnts have been identified HEK293 cells were performed as described (32). For phalloidin as guidance molecules for spinal cord commissural axons and staining, embryos were fixed overnight in 4% paraformaldehyde– corticospinal axons (24, 25). Recently, Wnt3, a classical mor- PBS solution. After fixation, embryos were lysed for2hwith phogen, has been identified as a guidance factor, counterbal- PBDT (PBS/1%BSA/0.1% Triton X-100/1% DMSO). Embryos ancing EphB/EphrinB1 activity (26). Although the precise mo- were incubated with primary antibodies overnight and then lecular nature of their interaction has not been elucidated, it is washed and incubated with a second antibody conjugated with evident that a functional link between EphB/EphrinB signaling Alexa Fluor 594 or Alexa Fluor 594-conjugated Phalloidin for and the Wnt cascades exists. Our results indicate that Daam1, actin staining. For nuclear staining, cells were incubated with which is also expressed in the nervous tissues (27), might be DAPI for 10 min. involved in this phenomenon, acting at the intersection of two MOs or mRNA Injection into Zebrafish Eggs. MOs were designed and different cascades to organize the correct patterning and path- synthesized by Gene Tools (Philomath, OR). Sequences were as finding. It is likely that Daam1 and the noncanonical Wnt follows: zDaam1a:5Ј-GGCTCAAGGGATAATGGGAAC- pathway may also have a similar involvement in the progression GAGG; zDaam1a:5Ј-GGCTCAAGGGATAATGGGAACG- of tumors such as colorectal cancer (28). AGG; zDaam1b:5Ј-AGCTATGACCCCCTCTCAAAATGGC; Ј Materials and Methods zDaam2:5-AGCTGGCAATGCGAACATGGCTTCC; zEphb2: 5Ј-TGCAGTCGCCGTCGTGGAGTCCATC; zEphb4:5Ј- Antibody and Reagents. Antibodies against Myc-, Flag-, and AGCTCCATCGCGGAATCACGAGTGT; zDnm1:5Ј-GTAAT- HA-tags were purchased from Santa Cruz Biotechnology (Santa CAAAATTGTCCTACCGTCAG; and zWnt11: GAAAGTTC- Cruz, CA), Sigma (St. Louis, MO), and Rockland (Gilbertsville, CTGTATTCTGTCATGTC. Control MOs had four base PA), respectively. Anti-phospho-tyrosine antibody was pur- mismatches. Oligonucleotides were solubilized in Danieau solution. chased from Zymed (San Francisco, CA). FM1–43FX and Alexa ␮ For in vitro synthesis of mRNAs, linearized plasmids were used as Fluor 594-conjugated Transferrin (50 g/ml; Invitrogen, Carls- templates in the RiboMAX Large Scale RNA Production system bad, CA) was used for the incorporation assay as described (29). (Promega, Madison, WI). Synthesized mRNAs were purified after FM1–43FZ is a derivative of FM4–64 suitable for immunohis- treatment with RNase-free DNase and dissolved in nuclease-free tochemistry. The anti-Daam1 antibody was prepared by immu- water. Injection of MOs or mRNAs and in situ hybridization were nizing rabbits with a synthesized peptide corresponding to the performed as described (31). 1,051- to 1,078-aa region of the human Daam1 protein. We also used another anti-Daam1 antibody (M05), which was obtained We thank David G. Wilkinson (National Institute of Medical Research, from Abnova (Taipei, Taiwan). London, U.K.), Raymond Habas, Terry P. Yamaguchi, and Wataru Shoji (Institute of Development, Aging and Cancer) for their helpful discus- cDNA Probes and in Situ Hybridization. Zebrafish Daam1a cDNA sions and Stephan W. Wilson (University College London, London, (IMAGp964D2115Q) was purchased from the RZPD (Berlin, U.K.), Marnie E. Halpern, Sergei Sokol, Gerard Apodaca, Christine Germany) clone bank (www.rzpd.de). HA-tagged zebrafish Rondanino, Masamitsu Tanaka, and Masa-aki Nakaya (Kanazawa Uni- NDaam1a (1–417-aa region) and Myc-tagged zebrafish versity, Kanazawa, Japan) for supplying materials. This work was sup- ported by a Grant-in-Aid for Scientific Research on Priority Areas from CDaam1a (439–1,079-aa region) were constructed from this the Ministry of Education, Science, Sports, and Culture of Japan (to cDNA by using RT-PCR techniques. Myc-tagged mouse full- T.O.), a Creative Basic Research Grant from the Ministry of Education, length Daam1 was the kind gift of T. Yamaguchi (National Science, Sports, and Culture of Japan (to T.O.), and an Exploratory Cancer Institute, Frederick, MD). GFP-RhoB and GFP- Research Program for Young Scientists (to Y.S.K.).

1. Adler PN (1992) BioEssays 14:735–741. 9. Ulrich F, Krieg M, Schotz EM, Link V, Castanon I, Schnabel V, Taubenberger 2. Veeman MT, Axelrod JD, Moon RT (2003) Dev Cell 5:367–377. A, Mueller D, Puech PH, Heisenberg CP (2005) Dev Cell 9:555–564. 3. Wehrli M, Tomlinson A (1995) Development (Cambridge, UK) 121:2451– 10. Myers DC, Sepich DS, Solnica-Krezel L (2002) Dev Biol 243:81–98. 2459. 11. Yamashita S, Miyagi C, Carmany-Rampey A, Shimizu T, Fujii R, Schier AF, 4. Sokol SY (1996) Curr Biol 6:1456–1467. Hirano T (2002) Dev Cell 2:363–375. 5. Strutt D (2003) Development (Cambridge, UK) 130:4501–4513. 12. Chan J, Mably JD, Serluca FC, Chen JN, Goldstein NB, Thomas MC, Cleary 6. Wodarz A, Nusse R (1998) Annu Rev Cell Dev Biol 14:59–88. JA, Brennan C, Fishman MC, Roberts TM (2001) Dev Biol 234:470–482. 7. Strutt D (2001) Curr Biol 11:R506–R509. 13. Chen H, Fre S, Slepnev VI, Capua MR, Takei K, Butler MH, Di Fiore PP, De 8. Habas R, Kato Y, He X (2001) Cell 107:843–854. Camilli P (1998) Nature 394:793–797.

6712 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0608946104 Kida et al. Downloaded by guest on September 29, 2021 14. Xu Q, Holder N, Patient R, Wilson SW (1994) Development (Cambridge, UK) 24. Imondi R, Kaprielian Z (2001) Development (Cambridge, UK) 128:4859–4871. 120:287–299. 25. Liu Y, Shi J, Lu CC, Wang ZB, Lyuksyutova AI, Song X, Zou Y (2005) Nat 15. Wilkinson DG (2003) Nat Cell Biol 5:851–853. Neurosci 8:1151–1159. 16. Poliakov A, Cotrina M, Wilkinson DG (2004) Dev Cell 7:465–480. 26. Lee HS, Bong YS, Moore KB, Soria K, Moody SA, Daar IO (2006) Nat Cell 17. Marston DJ, Dickinson S, Nobes CD (2003) Nat Cell Biol 5:879–888. Biol 8:55–63. 18. Zimmer M, Palmer A, Kohler J, Klein R (2003) Nat Cell Biol 5:869–878. 27. Kida Y, Shiraishi T, Ogura T (2004) Brain Res Dev Brain Res 153:143–150. 19. Fernandez-Borja M, Janssen L, Verwoerd D, Hordijk P, Neefjes J (2005) J Cell 28. Batlle E, Bacani J, Begthel H, Jonkheer S, Gregorieff A, van de Born M, Malats Sci 118:2661–2670. N, Sancho E, Boon E, Pawson T, et al. (2005) Nature 435:1126–1130. 20. Rojas R, Ruiz WG, Wang E, Kinlough CL, Poland PA, Hughey RP, Dunn KW, 29. Irie F, Okuno M, Pasquale EB, Yamaguchi Y (2005) Nat Cell Biol 7:501– Apodaca G (2004) J Biol Chem. 509. 21. Damke H, Baba T, Warnock DE, Schmid SL (1994) J Cell Biol 127:915–934. 30. Tanaka M, Kamo T, Ota S, Sugimura H (2003) EMBO J 22:847–858. 22. Gamse JT, Thisse C, Thisse B, Halpern ME (2003) Development (Cambridge, 31. Shoji W, Isogai S, Sato-Maeda M, Obinata M, Kuwada JY (2003) Development UK) 130:1059–1068. (Cambridge, UK) 130:3227–3236. 23. Heisenberg CP, Tada M, Rauch GJ, Saude L, Concha ML, Geisler R, Stemple 32. Kida Y, Maeda Y, Shiraishi T, Suzuki T, Ogura T (2004) Development DL, Smith JC, Wilson SW (2000) Nature 405:76–81. (Cambridge, UK) 131:4179–4187. BIOLOGY DEVELOPMENTAL

Kida et al. PNAS ͉ April 17, 2007 ͉ vol. 104 ͉ no. 16 ͉ 6713 Downloaded by guest on September 29, 2021