Developmental Cell

Short Article S1P-Yap1 Signaling Regulates Endoderm Formation Required for Cardiac Precursor Cell Migration in Zebrafish Hajime Fukui,1 Kenta Terai,2 Hiroyuki Nakajima,1 Ayano Chiba,1 Shigetomo Fukuhara,1 and Naoki Mochizuki1,3,* 1Department of Cell Biology, National Cerebral and Cardiovascular Center Research Institute, Fujishirodai 5-7-1, Suita, Osaka 565-8565, Japan 2Laboratory of Function and Morphology, Institute of Molecular and Cellular Biosciences, The University of Tokyo, 1-1-1 Yayoi, Bunkyo-ku, Tokyo 113-0032, Japan 3JST-CREST, National Cerebral and Cardiovascular Center Research Institute, Fujishirodai 5-7-1, Suita, Osaka 565-8565, Japan *Correspondence: [email protected] http://dx.doi.org/10.1016/j.devcel.2014.08.014

SUMMARY

To form the primary heart tube in zebrafish, bilateral cardiac precursor cells (CPCs) migrate toward the midline beneath the endoderm. Mutants lacking endoderm and fish with defective sphingosine 1-phosphate (S1P) signaling exhibit cardia bifida. Endoderm defects lead to the lack of foothold for the CPCs, whereas the cause of cardia bifida in S1P signaling mutants remains unclear. Here we show that S1P signaling regulates CPC migration through Yes-associated protein 1 (Yap1)-dependent endoderm survival. Cardia bifida seen in spns2 (S1P transporter) morphants and s1pr2 (S1P receptor-2) morphants could be rescued by endodermal expression of nuclear localized form of yap1. yap1 morphants had decreased expression of the Yap1/Tead target connective tissue growth factor a (Ctgfa) and consequently increased endodermal cell apoptosis. Consistently, ctgfa morphants showed defects of the endodermal sheet and cardia bifida. Collectively, we show that S1pr2/Yap1-regulated ctgfa expression is essential for the proper endoderm formation required for CPC migration.

INTRODUCTION During embryogenesis, zebrafish cardiac precursor cells (CPCs) originating from the anterior lateral plate mesoderm migrate toward the midline between the endoderm and the yolk syncytial layer (YSL) to form the cardiac tube (Staudt and Stainier, 2012). The function of the endoderm as a foothold for CPCs is evidenced by two laterally positioned hearts (cardia bifida) in the zebrafish endodermal loss-of function mutants (cas/sox32, sox17, oep, fau/gata5, and bon; Stainier et al., 1996). Similarly, endodermdeficient embryos, including Foxp4-deficient and Gata4-deficient mice, exhibit cardia bifida (Li et al., 2004; Kuo et al., 1997), indicating that the proper transcription factor-regulated endoderm formation and maintenance are essential for CPC migration.

We and others have demonstrated that mutants of sphingosine 1-phosphate (S1P) transporter, Spns2, which is expressed in the YSL during early embryogenesis, exhibit cardia bifida (Kawahara et al., 2009; Osborne et al., 2008). Furthermore, a mutant of S1pr2 (mil), one of five S1P receptors (S1pr1-5) shows cardia bifida (Kupperman et al., 2000), suggesting that S1P/S1pr2 signaling contributes to CPC migration. However, it is unclear where and how S1P/S1pr2 signaling regulates CPC migration. Yes-associated protein 1 (YAP1) is a final effector molecule downstream of Hippo signaling and functions as a transcription cofactor for TEA domain (TEAD)-binding family (TEAD1-4) transcription factors in the nucleus (Zhao et al., 2008). The hippo pathway is an evolutionally conserved kinase cascade that controls organ size and function by regulating cell proliferation, survival, and differentiation (Harvey et al., 2003; Zhao et al., 2007). LATS1/2 kinases activated by MST1/2 phosphorylate Ser of HXRXXS motifs of YAP1, thereby exporting YAP1 from the nucleus (Dong et al., 2007; Zhao et al., 2007). Therefore, YAP1/ TEAD-dependent transcription is inhibited by LATS1/2 activation. In mammals, Hippo signaling consisting of MST1/2 (mammalian orthologs of fruit fly, Hippo), SAV1, MOB1, and LATS1/2 form a canonical kinase-mediated signaling cascade that determines the localization of YAP1 (Zhao et al., 2008). Knocking down Yap1 in zebrafish leads to general developmental delay including cardiogenesis due to increased apoptosis and decreased proliferation (Hu et al., 2013; Jiang et al., 2009); however, it remains unclear where Yap1 specifically functions during embryogenesis. There are several YAP1/TEAD-promoted genes: connective tissue growth factor (CTGF), cysteine-rich angiogenic inducer 61 (CYR61), neural apoptosis inhibitory protein, cyclin-dependent kinase 6, BCL2, and BIRC5 (Hong and Guan, 2012; Landin Malt et al., 2013; Xie et al., 2013). CTGF consists of four domains: an insulin-like growth factor binding protein domain, a von Willebrand factor C (VWC) domain, a thrombospondin type 1 repeat domain, and a carboxy-terminal cysteine knot (CT) domain. Among these domains, the VWC domain and CT domain are reported to bind to bone morphogenetic proteins/tumor growth factor-b1 and to Wnt coreceptor lowdensity lipoprotein receptor-related protein 6, respectively, to induce biological function (Abreu et al., 2002; Mercurio et al., 2004).

128 Developmental Cell 31, 128–136, October 13, 2014 ª2014 Elsevier Inc.

Developmental Cell S1P-Yap1 Signaling for Endoderm Formation

Figure 1. Yap1/Tead-Dependent Transcription Is Required for Cardiac Precursor Cell Migration toward Midline (A) 3D-rendered confocal stack merged images (bright field and fluorescence) of Tg(myosin light polypeptide 7 [myl7]:nuclear localization signal-tagged tandem EosFP [Nls-tdEosFP]) embryos (48 hr postfertilization [hpf]) uninjected (left) and injected with morpholino (MO) indicated (center and right). tdEosFP fluorescence is shown as green. Images displayed in the following figures are the 3D-rendered images of a stack. Images are ventral views unless otherwise described (anterior to the top). The confocal images and WISH images in the following figures are a set of representative images of at least four independent experiments. (B) WISH analyses of the embryos (24 hpf) of the wild-type (left) and those injected with the MO indicated (center and right) using antisense probe for myl7. A set of representative images of four independent experiments is shown. (C) A schematic illustration of zebrafish (z) Yap1 and EGFP-zYtip. The numbers, amino acids (aa); EGFP-zYtip, EGFP fused with a Tead binding domain (aa 7-103 of zYap1). (D) Luciferase activity of the 293T cells transfected with the plasmids coding upstream activation sequence (UAS) followed by luciferase cDNA together with the plasmids indicated at the bottom. zYap1 binding to human TEAD2 fused with a DNA-binding domain of GAL4 (GAL4 db-hTEAD2) could promote UAS-dependent transcription of luciferase. The graph shows the average with SEM of the relative luciferase unit normalized by the luciferase activity of untransfected cells (n = 6). (E) Images of Tg(myl7:Nls-mCherry) embryos at 48 hpf uninjected (left) and injected with mRNAs indicated at one-cell stage with 50 pg (center) and 200 pg (right) EGFP-zytip mRNA. mCherry image (pseudocolor magenta), EGFP image, and bright field image are merged. (F) Quantitative analyses of the incidence of cardia bifida of (E). The graph shows the percentage calculated by the number of the embryos exhibiting cardia bifida among the embryos injected with mRNAs indicated at the bottom. Total number of the embryos injected with mRNA is indicated on the top of column. Hereafter, total number of the embryos investigated is indicated on the top of the column. xp < 0.0001. See also Figure S1.

The nuclear localization of YAP1/TAZ is induced by S1P and lysophosphatidic acids (LPA) that activate the group of G protein-coupled receptors (S1PR1-5 and LPAR1-6). Both S1P and LPA inactivate the Hippo signaling by inhibiting the phosphorylation of LATS1/2 kinases through the activation of G12/13-RhoGEF-Rho signaling, thereby accelerating the nuclear accumulation of YAP1/TAZ (Yu et al., 2012). Moreover, S1pr2G13 signaling is reported to regulate CPC migration (Ye and Lin, 2013). These previous data prompted us to test whether S1P regulates CPC migration by regulating Yap1. Here, we demonstrate that S1P/S1pr2 signaling controls endoderm maintenance via Yap1-dependent ctgfa expression and that the proper formation of the endoderm is essential for CPC migration.

RESULTS Yap1/Tead-Dependent Transcription Regulates CPC Migration To investigate whether Yap1 is essential for CPC migration, we examined the effect of Yap1 depletion on the CPC migration by injecting morpholino antisense oligonucleotides (MO) for yap1 into transgenic (Tg) zebrafish embryo Tg(myl7:NlstdEosFP) in which nuclear localization signal-tagged tandem Eos fluorescent protein (Nls-tdEosFP) was expressed under the control of myosin light polypeptide 7 (myl7) promoter. At 48 hr postfertilization (hpf), yap1 morpahnts exhibited cardia bifida as did s1pr2 morphants (Figure 1A). Consistently, wholemount in situ hybridization (WISH) analyses revealed that the

Developmental Cell 31, 128–136, October 13, 2014 ª2014 Elsevier Inc. 129

Developmental Cell S1P-Yap1 Signaling for Endoderm Formation

endogenous myl7-positive CPCs were positioned bilaterally in the yap1 morphants as well as the s1pr2 morphants (Figure 1B). The effect of yap1 MO on splicing-block was confirmed by RT-PCR. Moreover, cardia bifida found in the yap1 morphants was rescued by injection of zebrafish (z) yap1 mRNA (Figures S1A and S1B available online), indicating that cardia bifida is ascribed to the depletion of Yap1. YAP1 directly associates with TEAD in the nucleus to promote the expression of target genes that exert biological functions. To test whether Yap1/Tead-dependent transcription is essential for CPC migration, we examined the effect of interference of the association of Yap1 with Tead on CPC migration. For that, we constructed the plasmid expressing zTead-binding domain fused with enhanced green fluorescent protein (EGFP)-zYtip (von Gise et al., 2012) as a dominant-negative form of zYap1 (Figure 1C). zYtip inhibited Yap1-dependent transcription as demonstrated by a significant decrease of Yap1-dependent luciferase expression (Figure 1D). Then, we examined the effect of EGFP-zYtip on CPC migration by injecting EGFP-zytip mRNA into Tg(myl7:Nls-mCherry) embryos. CPC migration was dose-dependently inhibited (Figures 1E and 1F), indicating that Yap1/Tead-dependent transcription is essential for CPC migration. Yap1 Functions Downstream of S1P-S1pr2 Signaling in the Endoderm S1pr2 and Spns2 are expressed in the endoderm and in the YSL, respectively (Kawahara et al., 2009; Kupperman et al., 2000). S1P-S1PR2 signaling suppresses LATS1/2 phosphorylation, thereby inducing nuclear accumulation of YAP1 (Yu et al., 2012). Therefore, we assumed that Yap1 functions downstream of S1P signaling for CPC migration. To localize Yap1 during CPC migration, we first needed to exclude the possibility that Yap1 functions downstream of S1pr2 in the YSL. Depletion of neither S1pr2 nor Yap1 in the YSL affected CPC migration, whereas that of Spns2 resulted in cardia bifida (Figures 2A and 2B), indicating that Yap1 does not function in the YSL. We next examined whether Yap1 functions downstream of S1P signaling in the endoderm by the rescue experiments. Cardia bifida shown in the s1pr2 morphants, the spns2 morphants, and the yap1 morphants was reversed by the overexpression of Lats1/2-kinaseresistant Yap1 (Yap1-5SA) in the endoderm by sox17 promoter using a transient Tol2 transposon-based expression system (Figure 2C). These data indicate that Yap1 functions downstream of S1P-S1pr2 signaling in the endoderm to regulate CPC migration. S1pr2 mutant (mil) embryos show defects of the anterior endoderm (Kupperman et al., 2000). We, therefore, examined the morphogenesis of the anterior endoderm in the yap1 morphants. In the control zebrafish embryos at 26 hpf, one of the endoderm marker genes, foxa2 mRNA was clearly expressed in the pharyngeal endoderm and notochord (Figure 2D). Of note, an anterior endodermal space between bilateral pharyngeal endoderm was widened in both the yap1 morphants and the s1pr2 morphants compared to the control, although the pharyngeal endoderm and the notochord were comparable among the morphants and the control. In these widened regions, the expression of foxa2 mRNA appeared to be reduced. We assumed that the failure of CPC migration toward midline is the consequence of the defect of anterior endoderm. To address

this question, we depicted both endoderm and CPCs with double Tg (sox17:GFP);(myl7:Nls-mCherry) embryos. At 26 hpf, bilateral CPCs merged at the midline in the control embryos that showed no defects in the endoderm, whereas CPCs failed to coalesce at the midline in both the yap1 morphants and the s1pr2 morphants that showed defects of the anterior endoderm (Figure 2E, asterisk). The defects of anterior endoderm found in the yap1 morphants were rescued by injection of zyap1 mRNA (Figure S1C). Recently, G13 (encoded by gna13: gna13a and gna13b) is reported to function downstream of S1pr2 in the zebrafish endoderm for cardiac precursor cell migration (Ye and Lin, 2013). Moreover, the S1P-S1PR2-G12/13-RhoGEF-Rho signal inhibits LATS1/2 and in turn induces nuclear translocation of Yap1 (Yu et al., 2012). To delineate the S1P-mediated signal that controls the endoderm formation, we examined the effect of depletion of G13 (both G13a and G13b) and depletion of Lats1/2 in the s1pr2 morphants and the gna13 morphants. Depletion of G13 resulted in the defect of the anterior endoderm that was observed in the yap1 morphants and in the s1pr2 morphants (Figure 2E, asterisk). Depletion of Lats1/2 rescued the anterior endodermal defects found in the s1pr2 morphants and in the gna13 morphants, suggesting that Lats1/2 function downstream of G13 in the endoderm (Figures 2E and 2F). We confirmed that the injection of lats1/2 MOs at the dose we used for examining genetic interaction did not affect any phenotype (Figure S1E). In addition, a synergistic increase in the endoderm defect with cardia bifida was found in the embryos treated with a low dose of both yap1 MO and gna13 MOs, suggesting the genetic interaction of gna13 and yap1 genes (Figure 2G). To examine whether Yap1-dependent transcription is activated in the endoderm, we developed a Tg line that enables us to monitor the Yap1 nuclear translocation-dependent transcriptional activation that is reflected by GFP expression via Gal4/UAS system. Yap1-dependent transcription was confirmed by the disappearance of GFP in the yap1 morphants (Figure S2A). During the convergence of endoderm (14-somite stage), Yap1-dependent transcription was detected in the endoderm marked by sox17:DsRed (Figures S2B and S2C) and by foxa2 expression at 20 hpf (Figure S2D). In addition, it was detected in the CPCs marked by myl7 expression and in the heart (Figures S2D and S2E), suggesting that Yap1 functions in both endoderm and CPCs. Furthermore, forced expression of EGFP-zYtip in the endoderm resulted in cardia bifida associated with the defect of the anterior endoderm (Figures 2H and 2J), while that in the CPCs did not affect endoderm formation and their migration (Figures 2I and 2J). These results suggest that the S1pr2-G13-mediated inhibition of Lats1/2 signal and the Yap1/Tead-dependent transcription in the endoderm are essential for proper endoderm formation and CPC migration. In addition, these results imply that Yap1-dependent transcription in the CPCs is dispensable for CPC migration. Yap1 Regulates Cell Survival of the Endodermal Sheet To investigate what causes the defects of the anterior endoderm in the yap1 morphants, we time-lapse imaged the endodermal cells marked by sox17 promoter-driven GFP (Figure 3A). During the eight to 16-somite stage, an endodermal sheet was converged toward the midline (Movie S1). The endodermal sheet was correctly formed between bilateral pharyngeal endoderm by

130 Developmental Cell 31, 128–136, October 13, 2014 ª2014 Elsevier Inc.

Developmental Cell S1P-Yap1 Signaling for Endoderm Formation

Figure 2. Yap1 Functions Downstream of S1P-S1pr2 Signaling in the Endoderm (A) Confocal images of Tg(myl7:Nls-tdEosFP) embryos at 48 hpf injected with MO indicated at the bottom into the yolk syncytial layer (YSL). (B) Quantitative analyses of the incidence of cardia bifida of (A). Graph shows the percentage of the embryos with cardia bifida among the total embryos injected with MO indicated at the bottom. (C) The incidence of cardia bifida of Tg(myl7:Nls-tdEosFP) embryos injected with MO and the plasmids (30 pg/embryo) transiently expressing either EGFP (G) or EGFP-tagged Lats1/2 kinase-insensitive Yap1 (Y) under the control of sox17 promoter using Tol2 transposon system. (D) WISH analyses of foxa2 expression of the embryos (26 hpf) uninjected (left) and injected with MO indicated at the bottom (center and right). Note that foxa2 mRNA was detected in the endoderm and in the notochord and that the space between bilateral pharyngeal endoderm was widened in the yap1 morphants and s1pr2 morphants. A set of representative images of four independent experiments is shown as dorsal views. (E) 3D-rendered two-photon laser-scanned z-stack images of Tg(sox17:GFP);(myl7:Nls-mCherry) embryos uninjected (left) and injected with MO indicated at the bottom. gna13 MO stands for both gna13a and gna13b MOs. Asterisks indicate the defect of the endoderm. (F and G) Quantitative analyses of incidence of the defects in the endoderm with cardia bifida in the embryos treated with the MOs indicated at the bottom. (H and I) Images of those injected with the plasmid indicated at the top (25 pg/embryo) into the embryo indicated at the top. Note that forced expression of zYtip in the endoderm resulted in the defect of the endoderm with cardia bifida while that in the CPCs did not. Asterisk indicates the endodermal defect. (J) Quantitative analyses of the results obtained in (H) and (I). x p < 0.0001, yp < 0.01, zp < 0.05. See also Figure S2.

the 12-somite stage and was maintained during the postdevelopmental stage in the control morphants (Figure 3A and Movie S1). In clear contrast, some endodermal cells underwent fragmentation in the anterior endoderm and failed to form endodermal sheet properly in the yap1 morphants (Figure 3A, arrows; and Movie S2). Similarly, fragmentation was found in the anterior

endoderm of the s1pr2 morphants (Figure S3) as well as the gna13 morphants (Figure S3 and Movie S3). We further ascertained whether the defect of the anterior endoderm with fragmentation of the cells is associated with cardia bifida. The increase in fragmented endodermal cells paralleled the increased incidence of cardia bifida in the yap1 morphants and

Developmental Cell 31, 128–136, October 13, 2014 ª2014 Elsevier Inc. 131

Developmental Cell S1P-Yap1 Signaling for Endoderm Formation

Figure 3. Yap1 Regulates Endodermal Cell Survival (A) Time-sequential (from eight-somite stage [8s] to 16-somite stage [16s]) images of Tg(sox17:GFP) embryos injected with MO indicated at the left. Arrows denote the fragmentation of sox17 promoter-driven GFP-positive cells. A set of representative images of eight independent experiments is shown. Dorsal view. (B) Incidence of cardia bifia (black) and normal heart (white) in the morphants exhibiting less than 5 fragmented endodermal cells or more than 5 fragmented endodermal cells in the embryos treated with the MO indicated at the bottom. (C) Bar graph shows the number of the TUNELpositive cells among the GFP-positive endodermal cells of Tg(sox17:GFP) embryos at 14-somite stage (14s) injected with MO indicated at the bottom. ctrl MO, control morpholino. yp < 0.01 (versus ctrl morphants). (D) Bar graph shows the number of the TUNELpositive cells in a S1pr2 mutant (mil), heterozygous embryo (s1pr2m93/+) and homozygous embryos (s1pr2m93/m93). yp < 0.01. (E) Gene expression profile analyzed by RNA-seq using RNAs obtained from the Tg(sox17:GFP) embryos uninjected (control) and injected with yap1 MO by sorting of GFP-positive endodermal cells. The graph shows the ratio of the reads per kilobase per million reads (RPKM) of the yap1 morphants-derived RNAs divided by that of the control. Among the genes we analyzed, those regulating cell survival are listed. See also Figure S3 and Movies S1, S2, and S3.

the s1pr2 morphants (Figure 3B). These data suggest that the apoptosis of endodermal cells in these morphants might cause the defect of the anterior endoderm followed by the failure of CPC migration. Therefore, we examined whether the apoptosis occurs in the endoderm of the yap1 morphants and the s1pr2 morphants at the 14-somite stage. There was a significant increase of the TUNEL-positive cells in the anterior endoderm of both morphants (Figure 3C) and the s1pr2 mutant (mil) (Figure 3D). Apoptosis in the endoderm with the defect in the yap1 morphants was significantly rescued by injection of zyap1 mRNA (Figure S1D), suggesting that the apoptosis is the cause of the endodermal defects found in the yap1 morphant.

Next, we aimed at identifying the Yap1/Tead-promoted genes required for endodermal cell survival with RNA-seq. We analyzed the reads per kilobase per million reads as an indicator of mRNAs in Tg(sox17:GFP)-expressing cells of the control embryos and the yap1 morphants at 12 hpf. Among the potential genes that might regulate cell survival, the expression of connective tissue growth factor a (ctgfa) and cysteine-rich angiogenic inducer 61 (cyr61) was remarkably reduced in the yap1 morphants compared with the control embryos (Figure 3E). Because the expression of CTGF and CYR61 is reported to be transcriptionally promoted by mammalian YAP1/TEAD (Hong and Guan, 2012), we focused on both genes. Ctgfa Cell-Autonomously Regulates Endoderm Formation that Is Necessary for CPC Migration To unravel the role for Ctgfa and Cyr61 in endodermal cell survival, we examined the effect of depletion of either Ctgfa or Cyr61 on the endoderm formation by imaging both endoderm and CPCs using double Tg (sox17:GFP);(myl7:Nls-mCherry) embryos (Figure 4A). At 28 hpf, the ctgfa morphants but not the cyr61 morphants exhibited the defect of anterior endoderm

132 Developmental Cell 31, 128–136, October 13, 2014 ª2014 Elsevier Inc.

Developmental Cell S1P-Yap1 Signaling for Endoderm Formation

Figure 4. Endodermal Cells Cell-Autonomously Regulate Their Survival in a Manner Dependent on the Expression of Yap1-Promoted ctgfa (A) 3D-rendered two-photon laser-scanned z-stack images of Tg(sox17:GFP);(myl7:Nls-mCherry) embryos (28 hpf) injected with MO indicated at the bottom. An asterisk shows the defect of the endoderm. Dorsal view. (B and C) Incidence of endoderm defects with cardia bifida of Tg(sox17:GFP);(myl7:Nls-mCherry) embryos injected with MO indicated at the bottom. Total number of the embryos analyzed in each morphant group is indicated at the top of the column. (D and E) Quantitative-PCR analyses of expression of sox17 mRNA (D) and ctgfa mRNA (E) in the endoderm (GFP+ cells) and extra-endoderm (GFP- cells) of the Tg(sox17:GFP) embryos injected with MO indicated at the bottom. GFP+ and GFP- cells from the embryo at 12 hpf were sorted by GFP fluorescence (n = 4). (F) WISH analyses of ctgfa mRNA expression in the wild-type embryo (left) and in the yap1 morphant (right; 22 hpf). A set of representative images of four independent experiments is shown. Dorsal view. (G) Schematic illustration of injection of MO into the endoderm-fated cells (simultaneous injection of both acvr1bT206D and sox32 mRNAs) and into the CPC-fated cells (simultaneous injection of both gata5 and smarcd3b mRNAs) with rhodamine dextran to visualize the cells injected with MO and mRNAs. MO, mRNA, and rhodamine dextran were coinjected into one blastomere at the 64-cell stage. (H) Representative 3D-rendered confocal images of the Tg(sox17:GFP) embryo injected with mRNAs indicated at the top with the MOs indicated above the image as explained in (G). (legend continued on next page)

Developmental Cell 31, 128–136, October 13, 2014 ª2014 Elsevier Inc. 133

Developmental Cell S1P-Yap1 Signaling for Endoderm Formation

associated with cardia bifida (Figures 4A and 4B). The WISH analyses of foxa2 mRNA revealed the widened endodermal space, in which less expression of foxa2 mRNA was observed in the ctgfa morphants, but not in the cyr61 morphants (Figure S4A). Consistently, fragmented endodermal cells and TUNEL-positive apoptotic cells were increased in the anterior endoderm of the ctgfa morphants but not in the cyr61 morphants (Figures 3B and 3C). We next tested the genetic interaction between yap1 and ctgfa genes. Although the embryos injected with a low-dose morpholino for either yap1 or ctgfa showed low incidence of cardia bifida, those injected with a low dose of both yap1 and ctgfa MOs showed a synergistic increase of cardia bifida (Figure 4C). sox17 mRNA of the endodermal cells and the cells outside of the endoderm of the yap1 morphants were comparable to those of the control morphants (Figure 4D); however, ctgfa mRNA in the endoderm and outside of the endoderm was remarkably decreased in the yap1 morphants (Figure 4E). The WISH analyses supported the data that the expression of ctgfa mRNA in the yap1 morphants was reduced compared to that in the wildtype embryos (Figure 4F). These data indicate that ctgfa gene expression is also regulated by Yap1 in the endoderm. Indeed, in the embryos at 16 hpf, ctgfa mRNA expression overlapped foxa2 mRNA expression but not nkx2.5 mRNA expression that is expressed in the mesoderm, especially in CPCs (Figure S4B). Unexpectedly, transient expression of EGFP-tagged zCtgfa in the endoderm did not reverse but worsened the defects of the anterior endoderm (Figure S4C). This is probably due to the inappropriate amount of Ctgfa expression or inadequate balance of domain expression of Ctgfa for the endoderm. Finally, we investigated the cell-autonomous function of Yap1 and Ctgfa for the endodermal cells and for the CPCs. We tested the cell-autonomous function in vivo by knocking down either Yap1 or Ctgfa in either endoderm-fated cells or CPC-fated cells (Figure 4G; Aoki et al., 2002; David and Rosa, 2001; Lou et al., 2011). Depletion of either Yap1 or Ctgfa in the endoderm-fated cells resulted in the endodermal defects (Figures 4H and 4I), whereas that in the CPC-fated cells resulted in proper CPC migration and endoderm formation (Figures 4J and 4K), indicating that Yap1-dependent Ctgfa expression in the endoderm is cell-autonomously essential in the endoderm formation. Secreted CTGF is known to bind to fibronectin (Hoshijima et al., 2006), which supports the endoderm as an extracellular matrix and regulates the survival of the cells attaching to itself. Hence, we examined the attachment of the sox17 promoterdriven GFP-positive endodermal cells obtained from the control morphants, the yap1 morphants, and the ctgfa morphants to the fibronectin-coated dishes. The endodermal cells from the yap1 morphants and the ctgfa morphants exhibited less adhesion and spreading on the fibronectin-coated dishes than those

from the control (Figures 4L and 4M). These data suggest that Ctgfa secreted from the endoderm might regulate the endodermal cell attachment to the extracellular matrix in a cellautonomous manner. DISCUSSION Our present study supports that the proper formation of endoderm is essential for CPC migration toward the midline (Varner and Taber, 2012). Impairment of midline CPC migration has been observed in the endodermal mutants and in the S1P signaling mutants. Here, we demonstrate that the cardia bifida found in the S1P signaling morphants is ascribed to the endodermal defects caused by impaired Yap1-mediated signaling. There have been no reports accounting for the relevance of the cardia bifida found in the S1P signaling mutants to those found in the mutants showing endoderm defects. Endoderm not only serves as a foothold for CPC, but also secretes molecules that directly or indirectly regulate CPC migration and cardiogenesis. The YSL-derived fibronectin is required for both endoderm formation and CPC migration (Sakaguchi et al., 2006). Fibroblast growth factors and bone morphogenetic proteins expressed in the endoderm cooperatively regulate early cardiogenesis (Alsan and Schultheiss, 2002; Tirosh-Finkel et al., 2010). Therefore, structurally and functionally proper endoderm formation is essential for cardiogenesis. Yap1 functions in the endoderm to promote the expression of Ctgfa that regulates endodermal cell survival. We revealed that S1P-S1pr2 signaling induced Yap1/Tead-dependent ctgfa expression via Lats1/2 inactivation in the endoderm. Moreover, this signal in the endoderm is cell-autonomously essential for the endodermal formation. S1pr2-G13 signal controls cardiac migration in zebrafish (Ye and Lin, 2013). Recently, S1P-triggered signaling was reported to induce nuclear translocation of YAP1 by inhibiting LATS1/2 in cultured cells (Yu et al., 2012), although the significance of this signaling has not been demonstrated in vivo. YAP1 functions as the final effector molecule of the Hippo signaling to regulate cell survival and to determine the size of the organs. Our data add an essential role for Yap1 in vivo during embryogenesis that needs S1P signaling for the endoderm formation. We provide evidence that Ctgfa maintains the endodermal sheet that is required for CPC migration. CTGF functions in various tissue and organ development processes, such as angiogenesis, chondrogenesis, and palatogenesis (Chen and Lau, 2009; Parada et al., 2013), because it has four domains presumably to exhibit pleiotropic biological functions (Holbourn et al., 2008). Indeed, CTGF interacts with fibronectin through its CT domain (Hoshijima et al., 2006). The fibronectin released from the YSL is essential for endoderm formation (Nair and

(I) Quantitative analyses of incidence of endodermal defects in the embryos examined in (H). (J) Representative 3D-rendered confocal images of the Tg(sox17:GFP);(myl7:Nls-tdEosFP) embryo injected with mRNAs indicated at the top with the MOs indicated above the image as explained in (G). (K) Quantitative analyses of incidence of heart defects in the embryos examined in (J). (L and M) Attachment and spread of the endodermal cells obtained at 12 hpf from Tg(sox17:GFP) embryos injected with MO indicated at the top to the fibronectincoated dish. Arrowheads indicate the cells spreading after attachment to the dish. A panel of representative result of three independent experiments is shown in (L). Scale bars represent 50 mm. (M) The graph shows the number of the spreading GFP-positive cells after attachment. yp < 0.01. See also Figure S4.

134 Developmental Cell 31, 128–136, October 13, 2014 ª2014 Elsevier Inc.

Developmental Cell S1P-Yap1 Signaling for Endoderm Formation

Schilling, 2008). We found that ctgfa was downregulated in the yap1 morphants and that the endodermal cells of the yap1 morphants failed to attach to fibronectin-coated dishes. Collectively, these data indicate the biological function of fibronectin-bound Ctgfa for cell survival and suggest the domain-specific biological function of Ctgfa. In summary, we delineate the signaling pathway that is required for the precise endodermal formation associated with CPC midline migration: S1P/S1pr2 signaling inhibits Lats1/2, presumably resulting in nuclear translocation of Yap1 that functions as a cotranscription factor of Tead to promote Ctgfa expression. EXPERIMENTAL PROCEDURES Zebrafish—Danio rerio—Strains and Transgenic Lines The experiments using zebrafish were approved by the institutional animal committee of National Cerebral and Cardiovascular Center and performed according to the guidelines of the institute. We used the AB strain as wild-type and the Tg(sox17:GFP) fish as monitoring the endoderm development (Mizoguchi et al., 2008). mil mutant fish and UAS:GFP fish were kindly provided by G. Crump (Balczerski et al., 2012; Kupperman et al., 2000) and K. Kawakami (Asakawa and Kawakami, 2008), respectively. Microinjection of Oligonucleotide and mRNA We designed the start codon-targeted ctgfa MO and cyr61 MO. The efficiency of both MOs was confirmed by EGFP expression in the embryos injected of EGFP-tagged ctgfa or cyr61 mRNA together with corresponding MO (Figures S4D and S4E). The details of these MOs and other MOs are described in the Supplemental Experimental Procedures. To analyze the effect of MO in the endoderm-fated cells or in the CPC-fated cells, both acvr1bT206D and sox32 mRNAs and both gata5 and smarcd3b mRNAs were injected with either yap1 MO or ctgfa MO into the one blastomere cell at the 64-cell stage. Data Analysis and Statistics All columns were indicated as mean ± SEM. Statistical significance of multiple groups was determined by one-way ANOVA with Bonferroni’s post hoc test. The rescue experiments of MO injection with Tol2 DNAs, lats1/2 MOs on cardiac migration, and TUNEL assay of s1pr2m93 mutant were analyzed with Student’s t test. SUPPLEMENTAL INFORMATION Supplemental Information includes Supplemental Experimental Procedures, four figures, and three movies and can be found with this article online at http://dx.doi.org/10.1016/j.devcel.2014.08.014. ACKNOWLEDGMENTS We thank Y. Kikuchi for Tg(sox17:GFP) fish; D.Y. Stainier for pEGFP-sox17 promoter plasmid; K.D. Poss for eef1a1l1 plasmid; G. Crump for S1pr2 mutant (mil) fish; K. Kawakami for UAS:GFP reporter fish; the Zebrafish International Resource Center (ZIRC, NIH-NCRR #RR12546) for Tg(sox17:DsRed) fish; H. Takeda and A. Shimada for teaching the manipulation of embryos; and M. Sone, W. Koeda, and A. Okamoto for their technical assistance. This work was supported in part by grants-in-aid for scientific research on innovative ‘‘Neuro-Vascular Wiring’’ (no. 22122003 to N.M.) of The Ministry of Education, Culture, Sports, Science, and Technology, Japan; Scientific Research (B) (no. 24370084 to N.M.) from the Japan Society for the Promotion of Science; the Core Research for Evolutional Science and Technology (CREST) program of Japan Science and Technology Agency (to N.M.); and by grants-in-aid for young scientists (B) (to H.F.). Received: January 8, 2014 Revised: May 26, 2014 Accepted: August 11, 2014 Published: October 13, 2014

REFERENCES Abreu, J.G., Ketpura, N.I., Reversade, B., and De Robertis, E.M. (2002). Connective-tissue growth factor (CTGF) modulates cell signalling by BMP and TGF-beta. Nat. Cell Biol. 4, 599–604. Alsan, B.H., and Schultheiss, T.M. (2002). Regulation of avian cardiogenesis by Fgf8 signaling. Development 129, 1935–1943. Aoki, T.O., David, N.B., Minchiotti, G., Saint-Etienne, L., Dickmeis, T., Persico, G.M., Stra¨hle, U., Mourrain, P., and Rosa, F.M. (2002). Molecular integration of casanova in the Nodal signalling pathway controlling endoderm formation. Development 129, 275–286. Asakawa, K., and Kawakami, K. (2008). Targeted gene expression by the Gal4UAS system in zebrafish. Dev. Growth Differ. 50, 391–399. Balczerski, B., Matsutani, M., Castillo, P., Osborne, N., Stainier, D.Y.R., and Crump, J.G. (2012). Analysis of sphingosine-1-phosphate signaling mutants reveals endodermal requirements for the growth but not dorsoventral patterning of jaw skeletal precursors. Dev. Biol. 362, 230–241. Chen, C.C., and Lau, L.F. (2009). Functions and mechanisms of action of CCN matricellular proteins. Int. J. Biochem. Cell Biol. 41, 771–783. David, N.B., and Rosa, F.M. (2001). Cell autonomous commitment to an endodermal fate and behaviour by activation of Nodal signalling. Development 128, 3937–3947. Dong, J., Feldmann, G., Huang, J., Wu, S., Zhang, N., Comerford, S.A., Gayyed, M.F., Anders, R.A., Maitra, A., and Pan, D. (2007). Elucidation of a universal size-control mechanism in Drosophila and mammals. Cell 130, 1120–1133. Harvey, K.F., Pfleger, C.M., and Hariharan, I.K. (2003). The Drosophila Mst ortholog, hippo, restricts growth and cell proliferation and promotes apoptosis. Cell 114, 457–467. Holbourn, K.P., Acharya, K.R., and Perbal, B. (2008). The CCN family of proteins: structure-function relationships. Trends Biochem. Sci. 33, 461–473. Hong, W., and Guan, K.L. (2012). The YAP and TAZ transcription co-activators: key downstream effectors of the mammalian Hippo pathway. Semin. Cell Dev. Biol. 23, 785–793. Hoshijima, M., Hattori, T., Inoue, M., Araki, D., Hanagata, H., Miyauchi, A., and Takigawa, M. (2006). CT domain of CCN2/CTGF directly interacts with fibronectin and enhances cell adhesion of chondrocytes through integrin alpha5beta1. FEBS Lett. 580, 1376–1382. Hu, J., Sun, S., Jiang, Q., Sun, S., Wang, W., Gui, Y., and Song, H. (2013). Yes-associated protein (yap) is required for early embryonic development in zebrafish (danio rerio). Int. J. Biol. Sci. 9, 267–278. Jiang, Q., Liu, D., Gong, Y., Wang, Y., Sun, S., Gui, Y., and Song, H. (2009). yap is required for the development of brain, eyes, and neural crest in zebrafish. Biochem. Biophys. Res. Commun. 384, 114–119. Kawahara, A., Nishi, T., Hisano, Y., Fukui, H., Yamaguchi, A., and Mochizuki, N. (2009). The sphingolipid transporter spns2 functions in migration of zebrafish myocardial precursors. Science 323, 524–527. Kuo, C.T., Morrisey, E.E., Anandappa, R., Sigrist, K., Lu, M.M., Parmacek, M.S., Soudais, C., and Leiden, J.M. (1997). GATA4 transcription factor is required for ventral morphogenesis and heart tube formation. Genes Dev. 11, 1048–1060. Kupperman, E., An, S., Osborne, N., Waldron, S., and Stainier, D.Y.R. (2000). A sphingosine-1-phosphate receptor regulates cell migration during vertebrate heart development. Nature 406, 192–195. Landin Malt, A., Georges, A., Silber, J., Zider, A., and Flagiello, D. (2013). Interaction with the Yes-associated protein (YAP) allows TEAD1 to positively regulate NAIP expression. FEBS Lett. 587, 3216–3223. Li, S., Zhou, D., Lu, M.M., and Morrisey, E.E. (2004). Advanced cardiac morphogenesis does not require heart tube fusion. Science 305, 1619–1622. Lou, X., Deshwar, A.R., Crump, J.G., and Scott, I.C. (2011). Smarcd3b and Gata5 promote a cardiac progenitor fate in the zebrafish embryo. Development 138, 3113–3123.

Developmental Cell 31, 128–136, October 13, 2014 ª2014 Elsevier Inc. 135

Developmental Cell S1P-Yap1 Signaling for Endoderm Formation

Mercurio, S., Latinkic, B., Itasaki, N., Krumlauf, R., and Smith, J.C. (2004). Connective-tissue growth factor modulates WNT signalling and interacts with the WNT receptor complex. Development 131, 2137–2147. Mizoguchi, T., Verkade, H., Heath, J.K., Kuroiwa, A., and Kikuchi, Y. (2008). Sdf1/Cxcr4 signaling controls the dorsal migration of endodermal cells during zebrafish gastrulation. Development 135, 2521–2529. Nair, S., and Schilling, T.F. (2008). Chemokine signaling controls endodermal migration during zebrafish gastrulation. Science 322, 89–92. Osborne, N., Brand-Arzamendi, K., Ober, E.A., Jin, S.W., Verkade, H., Holtzman, N.G., Yelon, D., and Stainier, D.Y. (2008). The spinster homolog, two of hearts, is required for sphingosine 1-phosphate signaling in zebrafish. Curr. Biol. 18, 1882–1888. Parada, C., Li, J., Iwata, J., Suzuki, A., and Chai, Y. (2013). CTGF mediates Smad-dependent transforming growth factor b signaling to regulate mesenchymal cell proliferation during palate development. Mol. Cell. Biol. 33, 3482–3493. Sakaguchi, T., Kikuchi, Y., Kuroiwa, A., Takeda, H., and Stainier, D.Y.R. (2006). The yolk syncytial layer regulates myocardial migration by influencing extracellular matrix assembly in zebrafish. Development 133, 4063–4072.

Tirosh-Finkel, L., Zeisel, A., Brodt-Ivenshitz, M., Shamai, A., Yao, Z., Seger, R., Domany, E., and Tzahor, E. (2010). BMP-mediated inhibition of FGF signaling promotes cardiomyocyte differentiation of anterior heart field progenitors. Development 137, 2989–3000. Varner, V.D., and Taber, L.A. (2012). Not just inductive: a crucial mechanical role for the endoderm during heart tube assembly. Development 139, 1680– 1690. von Gise, A., Lin, Z., Schlegelmilch, K., Honor, L.B., Pan, G.M., Buck, J.N., Ma, Q., Ishiwata, T., Zhou, B., Camargo, F.D., and Pu, W.T. (2012). YAP1, the nuclear target of Hippo signaling, stimulates heart growth through cardiomyocyte proliferation but not hypertrophy. Proc. Natl. Acad. Sci. USA 109, 2394– 2399. Xie, Q., Chen, J., Feng, H., Peng, S., Adams, U., Bai, Y., Huang, L., Li, J., Huang, J., Meng, S., and Yuan, Z. (2013). YAP/TEAD-mediated transcription controls cellular senescence. Cancer Res. 73, 3615–3624. Ye, D., and Lin, F. (2013). S1pr2/Ga13 signaling controls myocardial migration by regulating endoderm convergence. Development 140, 789–799. Yu, F.X., Zhao, B., Panupinthu, N., Jewell, J.L., Lian, I., Wang, L.H., Zhao, J., Yuan, H., Tumaneng, K., Li, H., et al. (2012). Regulation of the Hippo-YAP pathway by G-protein-coupled receptor signaling. Cell 150, 780–791.

Stainier, D.Y.R., Fouquet, B., Chen, J.N., Warren, K.S., Weinstein, B.M., Meiler, S.E., Mohideen, M.A.P.K., Neuhauss, S.C.F., Solnica-Krezel, L., Schier, A.F., et al. (1996). Mutations affecting the formation and function of the cardiovascular system in the zebrafish embryo. Development 123, 285–292.

Zhao, B., Wei, X., Li, W., Udan, R.S., Yang, Q., Kim, J., Xie, J., Ikenoue, T., Yu, J., Li, L., et al. (2007). Inactivation of YAP oncoprotein by the Hippo pathway is involved in cell contact inhibition and tissue growth control. Genes Dev. 21, 2747–2761.

Staudt, D., and Stainier, D. (2012). Uncovering the molecular and cellular mechanisms of heart development using the zebrafish. Annu Rev Genet 46, 397–418.

Zhao, B., Ye, X., Yu, J., Li, L., Li, W., Li, S., Yu, J., Lin, J.D., Wang, C.Y., Chinnaiyan, A.M., et al. (2008). TEAD mediates YAP-dependent gene induction and growth control. Genes Dev. 22, 1962–1971.

136 Developmental Cell 31, 128–136, October 13, 2014 ª2014 Elsevier Inc.