Setdb2 controls convergence and extension movements during zebrafish gastrulation by transcriptional regulation of dvr1

Developmental Biology ∎ (∎∎∎∎) ∎∎∎–∎∎∎

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Q1 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66

Contents lists available at ScienceDirect

Developmental Biology journal homepage: www.elsevier.com/locate/developmentalbiology

Setdb2 controls convergence and extension movements during zebrafish gastrulation by transcriptional regulation of dvr1 Ting-Ting Du a,1, Peng-Fei Xu c,1, Zhi-Wei Dong b, Hong-Bo Fan b, Yi Jin a, Mei Dong b, Yi Chen a, Wei-Jun Pan b, Rui-Bao Ren a, Ting-Xi Liu a,b, Min Deng b,n, Qiu-Hua Huang a,n a State Key Laboratory for Medical Genomics, Shanghai Institute of Hematology, RuiJin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China b Key Laboratory of Stem Cell Biology, Institute of Health Sciences, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, China c Department of Cell Biology, University of Virginia, Charlottesville, VA, USA

art ic l e i nf o

a b s t r a c t

Article history: Received 21 March 2014 Received in revised form 15 May 2014 Accepted 24 May 2014

As the primary driving forces of gastrulation, convergence and extension (C&E) movements lead to a medio-lateral narrowing and an anterior–posterior elongation of the embryonic body axis. Histone methylation as a post-translational modification plays a critical role in early embryonic development, but its functions in C&E movements remain largely unknown. Here, we show that the setdb2-dvr1 transcriptional cascade plays a critical role in C&E movements during zebrafish gastrulation. Knockdown of Setdb2, a SET domain-containing protein possessing a potential histone H3K9 methyltransferase activity, induced abnormal C&E movements, resulting in anterior–posterior shortening and mediolateral expansion of the embryonic axis, as well as abnormal notochord cell polarity. Furthermore, we found that Setdb2 functions through fine-tuning the expression of dvr1, a ligand of the TGF-β superfamily, to an appropriate level to ensure proper C&E movements in a non-cell-autonomous manner. In addition, both overexpression and knockdown of Dvr1 at the one-cell stage resulted in defects at epiboly and C&E. These data demonstrate that Setdb2 is a novel regulator for C&E movements and acts by modulating the expression level of dvr1, suggesting that Dvr1 acts as a direct and essential mediator for C&E cell movements. & 2014 Published by Elsevier Inc.

Keywords: Zebrafish Convergence and extension Histone methylation TGF-β superfamily setdb2 dvr1

Introduction Gastrulation is a pivotal morphogenetic process in vertebrate development to establish the three germ layers: endoderm, mesoderm and ectoderm. Vertebrate gastrulation consists of three evolutionarily conserved morphogenetic cell movements: epiboly, involution and C&E movements (Keller, 2002). In zebrafish gastrulation, C&E movements narrow the germ layers mediolaterally and elongate them anteroposteriorly to sculpt the body plan (Solnica-Krezel, 2006; Warga and Kimmel, 1990). Previous studies indicate that non-canonical Wnt/PCP signaling is a crucial regulator of C&E movements during vertebrate gastrulation. The core members of the non-canonical Wnt/PCP pathway have been shown to regulate C&E movements (Jessen et al., 2002; Topczewski et al., 2001). Zebrafish mutants silberblick (slb;wnt11) and pipetail (ppt; wnt5) have been identified as required for normal C&E movements and show typical phenotypes of impaired C&E movements, such as n

Corresponding authors. E-mail addresses: [email protected] (M. Deng), [email protected] (Q.-H. Huang). 1 These authors contributed equally to this work.

anterior–posterior shortening and medio-lateral expansion of the embryonic axis (Heisenberg et al., 2000; Rauch et al., 1997). In addition to Wnt/PCP signaling, several other signaling pathways are also known to participate in proper C&E movements, including the BMP (Myers et al., 2002a; von der Hardt et al., 2007), Jak/Stat (Miyagi et al., 2004; Yamashita et al., 2002; Yamashita et al., 2004), Eph-Ephrins (Chan et al., 2001; Jones et al., 1998; Oates et al., 1999) and PDGF-PI3K signaling pathways (Ataliotis et al., 1995; Ghil and Chung, 1999). Meanwhile, genes involved in cell adhension and the extracellular matrix have also been shown to participate in the C&E movements (Bakkers et al., 2004; Coyle et al., 2008; Kim et al., 1998; Williams et al., 2012; Yamamoto et al., 1998). It has been seen that histone methylation plays a critical role in regulating gene expression during embryonic development (Kouzarides, 2007; Li, 2002), but it remains largely unknown whether histone methylation has effects on gastrulation, especially on C&E cell movements (Tsai et al., 2011). Our previous study identified the zebrafish setdb2 (SET domain bifurcated 2) gene in our large-scale sequencing database and genomewide-survey and developmental expression mapping of zebrafish SET-domain-containing genes. Further investigation indicated that Setdb2 possesses potential transcriptional repression activity through catalyzing trimethylation at

http://dx.doi.org/10.1016/j.ydbio.2014.05.022 0012-1606/& 2014 Published by Elsevier Inc.

Please cite this article as: Du, T.-T., et al., Setdb2 controls convergence and extension movements during zebrafish gastrulation by transcriptional regulation of dvr1. Dev. Biol. (2014), http://dx.doi.org/10.1016/j.ydbio.2014.05.022i

67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92

T.-T. Du et al. / Developmental Biology ∎ (∎∎∎∎) ∎∎∎–∎∎∎

2

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66

histone H3 lysine 9 (H3K9me3) and that it restricts dorsal organizer formation and regulates left-right asymmetry by suppressing fgf8a activity (Xu et al., 2010). Interestingly, in Setdb2-deficient embryos, we also observed shortened and laterally expanded notochord and somites stained for ntla and myod1, respectively. Furthermore, undulating midlines were frequently observed between 18 and 28 h post fertilization (hpf) in setdb2 morphant embryos (Xu et al., 2010), resembling defects in C&E movements (Hammerschmidt et al., 1996; Oishi et al., 2006; Sumanas et al., 2001). These abnormal patterns of midline genes raise the possibility that Setdb2 is required for appropriate C&E movements during zebrafish gastrulation. Members of the transforming growth factor beta (TGF-β) superfamily regulate fundamental cell processes, including cell growth, cell differentiation, apoptosis, cellular homeostasis, adhesion and migration, particularly in the germ layer formation and body axis patterning during early embryonic development (Wu and Hill, 2009). Members of the TGF-β superfamily, such as BMPs, have been shown to contribute to proper gastrulation movements during embryonic development (Myers et al., 2002a; von der Hardt et al., 2007). The decapentaplegic and Vg-related 1 (dvr1) gene, the zebrafish ortholog of Xenopus Vg1 and mammalian Gdf1, also belongs to the TGF-β superfamily of ligands (Andersson et al., 2007; Birsoy et al., 2006; Dohrmann et al., 1996). In Xenopus, Vg1 has been reported to be involved in Smad2 phosphorylation and mesoderm induction. The Vg1-depleted embryos develop with abnormal gastrulation and reductions in anterior and dorsal structures (Birsoy et al., 2006). Additionally, Xenopus Vg1, mouse Gdf1 and human GDF1 have been related to the establishment of left-right asymmetry. In Xenopus, altered expression of Vg1 on the right side of 16-cell embryos or disruption of Vg1 signaling on the left side randomizes cardiac and visceral left-right orientation (Hyatt et al., 1996). Gdf1-/- mice exhibit abnormalities in left-right axis formation, such as visceral situs inversus and right pulmonary isomerism (Rankin et al., 2000). Further, a study in a Finnish family with five siblings affected with right atrial isomerism (RAI, a subtype of heterotaxy syndrome), asplenia and situs anomalies shows that the affected children are compound heterozygotes for truncating mutations in the GDF1 gene (Kaasinen et al., 2010). Meanwhile, more functions of zebrafish Dvr1 have gradually been identified. Ectopic expression of dvr1 in wild-type embryos results in severe dorsalization, and Dvr1 signaling in zebrafish acts similarly to Nodal and depends on EGF-CFC coreceptors for the interaction with and activation of Activin receptors (Cheng et al., 2003). Furthermore, embryonic depletion of Dvr1 via morpholino (mo) (morpholino phosphorodiamidate antisense oligonucleotides) induces abnormal development, resulting in absent cardiac looping, pericardial edema and impaired trunk and neural development (Li et al., 2012). A recent study has also indicated that Dvr1 is responsible for left-right asymmetry by enabling the transfer of a left-right signal from KV to the LPM (Peterson et al., 2013). However, the function of Dvr1 relevant to zebrafish gastrulation has not previously been explored. In this study, we show that the setdb2 gene is required for the establishment of proper C&E movements in a non-cell-autonomous manner during zebrafish gastrulation. Furthermore, Setdb2 regulates the expression of dvr1, and the characteristic defects of C&E movements observed in setdb2 morphant embryos can be largely restored by Dvr1, a member of TGF-β superfamily, suggesting a novel signaling pathway for regulating the C&E movements.

Results Setdb2 is required for the C&E movements during zebrafish gastrulation Our previous study illustrated that the SET-domain-containing protein Setdb2 possesses potential transcriptional repression

activity and can negatively regulate the dorsal organizer formation by suppressing the expression of fgf8a. During that study, we also observed inappropriate C&E movements, including shortened and laterally expanded mesodermal structures and undulations of the midline (Xu et al., 2010). These features prompted us to hypothesize that Setdb2 is involved in C&E movements during gastrulation. To test this hypothesis, we examined the role of Setdb2 with antisense morpholino-mediated knockdown. Western blot analysis showed dose-dependent reductions in the endogenous Setdb2 protein level in setdb2 morphant embryos (Fig. S1A). Embryos injected with setdb2 mo displayed a mildly shortened anteriorposterior axis at the end of gastrulation (Fig. S2B,E) and 3-somite stage (3s) (Fig. S2D,E) compared with the control. At 22 hpf, the knockdown of Setdb2 caused a dramatically shortened and curved body axis (data not shown). To further characterize these defects at the tissue level, we examined the expression levels of several specific genes. At the tail-bud and 3-somite stages, setdb2 morphants showed a wider and shorter notochord (ntla) (Fig. 1A,B,G), broader neural plate (dlx3b) (Fig. 1C,D,G), wider rhombomeres (egr2b) (Fig. 1E–G) and a slightly posteriorly located prechordal plate (ctsl1b) compared to the control embryos (Fig. 1C,D,G, insets). These observations, combined with compressed and laterally expanded somites (myod1) (Xu et al., 2010), suggest that the C&E movements were impaired in Setdb2-knockdown embryos. At the cellular level, we observed an abnormality in cell polarity in Setdb2-depleted embryos at the 3-somite stage. We analyzed the shape and orientation of notochord cells. In control embryos, the notochord cells exhibited an average of length-towidth ratio (LWR) of 3.89 70.951, and 65% of notochord cells oriented their long axes within a 201 arc relative to a line perpendicular to the notochord (Figs. 1H,J,K, n ¼4 embryos, 80 cells). By contrast, in setdb2 morphants, the notochord cells showed reduced LWR of 1.81 7 0.501 and more random mediolateral alignment. Only 20% of notochord cells oriented their long axes within a 201 arc relative to a line perpendicular to the notochord (Figs. 1I–K, n ¼4 embryos, 80 cells). In addition, the impaired morphological cell polarity in setdb2 morphants were first observed at the end of gastrulation (Fig. S3B,E,F). All of these results suggest that Setdb2 is required for C&E movements during zebrafish gastrulation.

Setdb2 regulates C&E movements by modulating the expression of dvr1 To further investigate the molecular mechanisms of Setdb2 in C&E movements, we collected control and Setdb2-depleted embryos at the onset of gastrulation to perform microarray analysis. The expression levels of 56 genes were changed by more than 2.0-fold in setdb2 morphants in comparison to the controls. Those deregulated genes were categorized by the biological function based on published studies and the online software DAVID for functional annotation of bioinformatics microarray analysis. Six of these genes were associated with embryonic morphogenesis. The expression of these genes was validated with quantitative PCR (Table S1). Among the six candidate genes, the zebrafish dvr1 gene, a homolog of Xenopus Vg1 and mammalian Gdf1, has been implicated in mesoderm induction and is required for establishing leftright asymmetry (Birsoy et al., 2006; Cheng et al., 2003; Peterson et al., 2013). The previous study showed that Dvr1-depleted embryos in zebrafish had a shortened body axis and curved neural tube (Li et al., 2012). These specific phenotypes raise the possibility that Dvr1 may be a downstream target of Setdb2 during the gastrulation process and C&E.

Please cite this article as: Du, T.-T., et al., Setdb2 controls convergence and extension movements during zebrafish gastrulation by transcriptional regulation of dvr1. Dev. Biol. (2014), http://dx.doi.org/10.1016/j.ydbio.2014.05.022i

67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132

T.-T. Du et al. / Developmental Biology ∎ (∎∎∎∎) ∎∎∎–∎∎∎

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66

3

Fig. 1. Setdb2 knockdown induces defects in C&E movements during gastrulation. (A–G) Whole-mount in situ hybridization and statistical analyses of the expression patterns of ntla, ctsl1b, dlx3b and egr2b at the indicated developmental stages in embryos injected with setdb2 control mo or setdb2 mo. Insets show magnified views of the relative position of ctsl1b and dlx3b. The dorsal views on embryos in A and B were flat-mounted. Embryos were animal views in C and D with head to the left and dorsal views in E and F with head to the top. L1 and L2, the length and width of the notochord at 10 hpf; L3, the length of the posterior shift of prechordal plate relative to the overlying neural plate; Angle, between the two sides of neural plate; L4, the width of rhombomeres 3 and 5. Asterisks indicate statistically significant differences (n, Po 0.05; nn , P o 0.01; nnn, Po 0.001). (H–K) Representative confocal images of notochord cells in setdb2 control embryos (n ¼4 embryos, 80 cells) and setdb2 morphants (n¼ 4 embryos, 80 cells) co-injected at the one-cell stage with 100 pg mRNA coding for membrane-targeted RFP. Anterior is upwards. The boundaries of the notochord are marked with dotted white lines. Length-width ratio (LWR) of notochord cells and mediolateral alignment (ML alignment) at the 3-somite stage were measured with ImageJ2X. The rose diagrams were drawn using Rose.NET. Scale bar: 20 μm. The statistical significance of difference is indicated by asterisks (nnnn, Po 0.0001).

The upregulation of dvr1 was the first detected at 5.5 hpf in setdb2 morphants and was maintained during gastrulation (Fig. 2A). Meanwhile, we checked the transcriptional expression levels of both setdb2 and dvr1 in wild-type embryos from 2.5 hpf to 10 hpf (Fig. 2B). There was a highly negative correlation between setdb2 and dvr1 gene transcription, especially during gastrulation. Based on these observations, we hypothesized that the upregulation of dvr1 might be responsible for the impaired C&E movements caused by the inhibition of Setdb2. One prediction based on the above hypothesis is that reducing dvr1 expression should rescue the defects caused by Setdb2 knockdown, whereas overexpression of dvr1 would mimic the defects induced by the inhibition of Setdb2. To test this hypothesis and determine whether Dvr1 acts downstream of Setdb2 to regulate C&E movements, we first designed two non-overlapping

morpholinos, dvr1 moATG (blocking translation) and dvr1 moSB (blocking splicing), to specifically silence its functions in zebrafish. To test the knockdown efficiency and specificity of the two morpholinos, we co-injected 16 ng of either dvr1 control mo or dvr1 moATG with 200 pg of an EGFP mRNA reporter containing morpholino-targeting sequences into one-cell-stage embryos. The results indicated that dvr1 moATG was able to efficiently and specifically block the translation of the EGFP reporter (Fig. S1B). Next, we checked the dvr1 mRNA level at various developmental stages in embryos injected with either control mo or 16 ng dvr1 moSB. The dvr1 mRNA level was gradually reduced in dvr1 moSB embryos from the zygotic transcription stage to the end of gastrulation (Fig. S1C). We then co-injected 16 ng setdb2 mo with 4 ng dvr1 moATG at the one-cell stage. We found that the broader presumptive

Please cite this article as: Du, T.-T., et al., Setdb2 controls convergence and extension movements during zebrafish gastrulation by transcriptional regulation of dvr1. Dev. Biol. (2014), http://dx.doi.org/10.1016/j.ydbio.2014.05.022i

67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132

4

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66

T.-T. Du et al. / Developmental Biology ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Fig. 2. Dvr1 acts downstream of Setdb2 to affect C&E movements. (A) The relative expression levels of dvr1 in control and setdb2 morphant embryos from 2.5 hpf to 10 hpf. The upregulation of dvr1 was firstly detected at 5.5 hpf in the setdb2 morphants. (B) The relative expression levels of setdb2 and dvr1 from 2.5 hpf to 10 hpf in wild-type embryos. (C) The expression of the setdb2 gene was unaffected in two types of dvr1 morphants compared to the control. Gapdh expression was analyzed as a control in quantitative real-time PCR. (D-L) dvr1 moATG partially rescued the C&E movement defects of setdb2 morphants. The embryos were injected at the one-cell stage with 16 ng setdb2 control mo, 16 ng setdb2 mo or 16 ng setdb2 mo plus 4 ng dvr1 moATG. Whole-mount in situ hybridization of the expression patterns of fgf8a and ntla were performed at the indicated developmental stages. L1, the width of the presumptive hindbrain at 10 hpf; L2 and L3, the width and length of the notochord. Dorsal views of embryos in G-I were flat-mounted. Embryos were dorsal views in left side of D–F with head to the top and lateral views in right side of D-F. Embryos were lateral views in upper of J–L and dorsal views in nether of J–L with head to the left. (M) Statistical analyses of the expression patterns of fgf8a and ntla were performed. Asterisks indicate statistically significant differences (nn, Po 0.01; nnnn, P o0.0001).

Please cite this article as: Du, T.-T., et al., Setdb2 controls convergence and extension movements during zebrafish gastrulation by transcriptional regulation of dvr1. Dev. Biol. (2014), http://dx.doi.org/10.1016/j.ydbio.2014.05.022i

67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132

T.-T. Du et al. / Developmental Biology ∎ (∎∎∎∎) ∎∎∎–∎∎∎

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66

hindbrain (fgf8a) (Fig. 2D–F, M), the wider and shorter notochord (ntla) (Fig. 2G–I, M), and even the curved notochord at 22 hpf induced by Setdb2 deficiency were rescued to nearly normal (19/ 21, n ¼ 21) (Fig. 2J–M). Consistent with this result, knockdown of Dvr1 with moSB in setdb2 morphant embryos also partially restored the defects of C&E to normal (Fig. S4). As published previously, misexpression of dvr1 can induce ectopic expression of the mesodermal markers ntla and gsc in zebrafish or Xenopus (Birsoy et al., 2006; Cheng et al., 2003). After injecting 80 pg dvr1 mRNA, we also observed the upregulation of the dorsal-mesodermal genes flh and fgf8a at the shield stage (data not shown). Meanwhile, this high level of dvr1 mRNA in embryos resulted in severe defects during epiboly (Fig. 3B,F). We next proceeded to inject dvr1 mRNA with varying dosage into embryos. Among embryos injected with 40 pg dvr1 mRNA, most of them showed obvious delays in epiboly and developed to approximately 70% epiboly stage at 10 hpf compared to the control (Fig. 3C,F). Embryos injected with lower amounts of dvr1 mRNA (20 and 15 pg) showed mildly delayed epiboly (Fig. 3C,D,F). Finally, the epiboly process could be completed when we injected 10 pg dvr1 mRNA into the embryos, but 40% of them had impaired C&E movements, such as mild defects in extension (Fig. 3E–G). Meanwhile, whole-mount in situ hybridization also showed a wider and slightly shorter notochord (ntla) (13/40, n ¼ 40) (Fig. 3H,I), broader neural plate (dlx3b) (16/32, n ¼32) (Fig. 3J,K, green dotted line), wider and slightly shorter somites (myod1) (15/ 31, n¼ 31) (Fig. 3L,M) and a slightly posteriorly located prechordal plate (ctsl1b) (13/32, n ¼32) (Fig. 3J,K) in the embryos injected with 10 pg dvr1 mRNA. After injection of dvr1 mRNA, we also observed the undulated notochord stained with RNA probes for ntla at 22 hpf (Fig. S5). Additionally, the impaired morphological cell polarity in embryos injected with 10 pg dvr1 mRNA was also observed at the end of gastrulation (Fig. S3D–F). These C&E defects partially resembled what had been observed in setdb2 morphants. Additionally, there was no significant change in the setdb2 mRNA level in dvr1 moATG and dvr1 moSB morphant embryos compared to the control (Fig. 2C). Taken together, these results suggest that aberrant upregulation of dvr1 is responsible for the impaired C&E movements caused by the inhibition of Setdb2 and that the setdb2-dvr1 transcriptional cascade is essential for regulating proper C&E movements. In our previous studies, knockdown of Setdb2 resulted in randomization of left-right asymmetry (Xu et al., 2010). Other published studies suggest a close relationship among C&E movements, ciliogenesis and left-right asymmetry. Most core members of non-canonical Wnt signaling pathway that participate in C&E movements also play important roles during cilia formation and left-right asymmetry patterning (Matsui and Bessho, 2012). Furthermore, Vg1/Dvr1/Gdf1 have all been reported to be involved in regulating the left-right axis formation (Hyatt et al., 1996; Peterson et al., 2013; Rankin et al., 2000). These reports led us to test whether the upregulation of dvr1 is also responsible for the abnormal left-right asymmetry induced by the inhibition of Setdb2. We found that the randomization of the asymmetric gene spaw at 20 hpf and the heart laterality marker myl7 at 30 hpf in setdb2 morphants could not be rescued by co-injection with 4 ng dvr1 moATG (Fig. S6). These results indicate that Dvr1 acts as a downstream target of Setdb2 to specifically regulate C&E movements but not left-right asymmetry patterning during zebrafish development. Knockdown of Dvr1 showed defects in C&E cell movements Previous studies have reported that both the gain of function and loss of function of components of the non-canonical Wnt signaling pathway can cause abnormal C&E movements (Kilian et

5

al., 2003; Matsui et al., 2005; Ungar et al., 1995). To further investigate the effect of Dvr1 on C&E movements during zebrafish gastrulation, we examined the effects of Dvr1 knockdown on the C&E process. Whole-mount in situ hybridization showed that injection of dvr1 moATG resulted in a wider presumptive hindbrain (fgf8a) (31/33, n ¼33) (Fig. 4A,B), broader neural plate (dlx3b) (41/ 63, n ¼ 63) (Fig. 4C,D, green dotted line), compressed and laterally expanded somites (myod1) (30/33, n¼ 33) (Fig. 4E,F) and undulated notochord (ntla) (23/33, n ¼33) (Fig. 4G,H). Similar to setdb2 morphants, compared to the control embryos, dvr1 moATG also caused a slightly posteriorly located prechordal plate (ctsl1b) with respect to the anterior edge of the neural plate (25/63, n ¼63) (Fig. 4C,D). Moreover, we also found that dvr1 moSB could induce phenotypes of C&E similar to those observed in dvr1 moATG and setdb2 morphants (Fig. 4I–P). Meanwhile, knockdown of Dvr1 could also lead to defects in cell polarity at the end of gastrulation (Fig. S3C,E,F). Further, we verified the specificity of C&E defects caused by Dvr1 inhibition by co-injecting dvr1 mRNA with dvr1 moATG. The phenotype of compressed and laterally expanded somites (myod1) in dvr1 moATG morphant was nearly rescued in a dose-dependent manner (Fig. 4Q). To assess how Dvr1 affects cell movements during zebrafish gastrulation, we performed a cell-tracing experiment with UVmediated photoactivation (Hatta et al., 2006). We used UV radiation to label a small cluster of cells on the dorsal side of the margin or the middle of the lateral margin at 6 hpf and analyzed the cell migration at the tail-bud stage. In comparison with the control embryos, the anterior migration of labeled cells was perturbed in dvr1 moATG and dvr1 moSB morphants (Fig. 5F,J,Q). Meanwhile, in dvr1 moATG and dvr1 moSB morphants, the convergence of the laterally marked cells toward the dorsal side was also impaired (Fig. 5H,L,Q). In addition to dvr1 morphants, we also observed phenotypes of mildly reduced extension of the axial mesoderm and severely reduced movement of labeled lateral cells towards the dorsal side in setdb2 morphants (Fig. 5N,P,Q). In conclusion, both loss of Dvr1 and gain of Dvr1 resulted in aberrant C&E cell migration, suggesting that an appropriate expression level of dvr1 is crucial for normal C&E cell movements. The nature of the relationship between C&E cell movements and cell fate specification has recently attracted attention (Chen et al., 2012; Myers et al., 2002b; van Eekelen et al., 2010; Xu et al., 2013). To rule out the possibility that the impaired C&E movements resulted from aberrant cell fate specification in dvr1 morphants, we analyzed the expression of dorsal (chd, gsc, flh, cyc), ventral (bmp4, bmp2b, tbx6) and mesodermal (ntla) patterning genes in dvr1 morphants using whole-mount in situ hybridization. At the shield stage, the dvr1 moATG and dvr1 moSB morphants showed morphologically normal development (Fig. S5A, bright field). The expression of the genes mentioned above showed no obvious alterations in morphants with either dvr1 moATG or moSB, except for the decreased expression level of bmp4 in dvr1 moATG embryos (Fig. S7A,B). Consistent with this observation, there is no obvious early patterning defect in bmp4 mutants (Stickney et al., 2007). Thus, these results raise the possibility that the zygotic but not maternal expression of dvr1 causes impaired C&E movements without interference in cell fate decisions. It has been reported that a fraction of morpholinos used in zebrafish show off-target effects due to p53-induced apoptosis (Robu et al., 2007). To conform that the phenotypes of C&E movement defects observed in setdb2 and dvr1 morphants were gene-specific effects, we co-injected setdb2 mo, dvr1 moATG and dvr1 moSB, respectively, with p53 mo. The embryos injected with setdb2 mo, dvr1 moATG or dvr1 moSB showed a shortened body axis, curved somites and neural apoptosis of the head at 22 hpf. After co-injection with p53 mo, the phenotypes of head apoptosis were largely eliminated, but the shortened body axis and curved

Please cite this article as: Du, T.-T., et al., Setdb2 controls convergence and extension movements during zebrafish gastrulation by transcriptional regulation of dvr1. Dev. Biol. (2014), http://dx.doi.org/10.1016/j.ydbio.2014.05.022i

67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132

6

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66

T.-T. Du et al. / Developmental Biology ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Fig. 3. Forced expression of dvr1 impairs epiboly and C&E movements. (A–F) Embryos were either uninjected or injected with 80, 40, 20, 15 or 10 pg dvr1 mRNA at the onecell stage and were imaged at 10 hpf. Base on the severity and stage of morphological phenotypes, the embryos were grouped into normal bud stage (blue), 50% epiboly (red), 70% epiboly (orange), 90% epiboly (green) and C&E defects (gray). Statistical analyses of the phenotypes are shown in F. The dotted lines in A and E indicate the angle between the most anterior and posterior ends of the body axis. The arrows in B-D indicate the front edge of epiboly process. All embryos were lateral view. (G) The average extension angle of uninjected embryos and embryos injected with 10 pg dvr1 mRNA which had impaired extension (approximately 40% of total 10 pg dvr1 mRNA injected embryos) was plotted. The statistical significance of this difference is indicated by asterisks (nnnn, P o0.0001). (H-M) Embryos at one-cell stage were injected with control or dvr1 mRNA (50 pg EGFP mRNA or 10 pg dvr1 mRNA) and stained with RNA probes for ntla, ctsl1b, dlx3b and myod1 at 10 hpf. White and black double-headed arrows denote the width and length of the notochord and somites, respectively. The green dotted lines indicate the angle between the two sides of the neural plate. Embryos were dorsal view in H-I and L-M with head to the top and animal view in J and K with head to the left.

Please cite this article as: Du, T.-T., et al., Setdb2 controls convergence and extension movements during zebrafish gastrulation by transcriptional regulation of dvr1. Dev. Biol. (2014), http://dx.doi.org/10.1016/j.ydbio.2014.05.022i

67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132

T.-T. Du et al. / Developmental Biology ∎ (∎∎∎∎) ∎∎∎–∎∎∎

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66

7

Fig. 4. Knockdown of the dvr1 gene leads to abnormal C&E movements. (A–H) Whole-mount in situ hybridization for fgf8a, dlx3b, ctsl1b, myod1 and ntla were performed in control and dvr1 moATG morphants at the indicated developmental stages. (I–P) Whole-mount in situ hybridization for ntla, dlx3b, ctsl1b, egr2b and myod1 were performed in control and dvr1 moSB morphant embryos at the indicated developmental stages. The percentage of phenotypes were as follows: (J) 17/19 (ntla), (L) 42/44 (dlx3b), 13/44 (ctsl1b), (N) 23/25 (egr2b), (P) 34/34 (myod1). Red double-headed arrows denote the width of the presumptive hindbrain in the lateral view or the width of rhombomeres 3 and 5 in the dorsal view. Green dotted lines indicate the angle between the two sides of the neural plate. Black and blue double-headed arrows show the width and length of somites in E-F and O-P, respectively. Black double-headed arrows in I and J indicate the length of the notochord. Embryos were dorsal view in left side of A-B, E-F, I-J and M-P with head to the top, animal view in C-D and K-L with head to the left, lateral view in right side of A and B with head to the top, lateral view in left side of G and H and dorsal view in right side of G and H with head to the left. (Q) The C&E defects were specifically caused by Dvr1 inhibition. Whole-mount mRNA in situ hybridization of the expression pattern of myod1 in control, dvr1 moATG-injected and dvr1 moATG plus dvr1 mRNA-injected embryos at 9-somite stage.

somites were still present (Fig. S8, left column). Consistent with these results, the co-injection of p53 mo in setdb2 and dvr1 morphants could not restore the phenotype of wider and shorter notochord (ntla) compared to the control (Fig. S8, right column), indicating that the defects of C&E movements in setdb2 or dvr1 morphants were induced by inhibition of Setdb2 or Dvr1, and not by p53-dependent off-target effects. The setdb2-dvr1 transcriptional cascade regulates C&E cell movements in a non-cell-autonomous manner To further examine how the setdb2-dvr1 transcriptional cascade affects C&E movements during gastrulation, in a cell-autonomous or non-cell-autonomous manner, one-cell-stage donor embryos were injected with Rhodamine-dextran or with fluoresceindextran plus setdb2 mo or dvr1 moSB. At the shield stage, we cotransplanted dorsal mesendodermal cells of both wild-type embryos (Rhodamine-dextran) and morphants (fluorescein-dextran plus setdb2 mo) into the dorsal side of wild-type host embryos and tested the extension. At the end of gastrulation, cells from both wild-type and setdb2-mo donor embryos properly migrated to the animal pole to a similar extent (Fig. 6C).

Furthermore, when cells from the middle of the lateral blastoderm margin of wild-type embryos or morphants were co-transplanted into the same location of wild-type host embryos and imaged at the bud stage, we also observed that these donor cells converged to the dorsal side in a similar way (Fig. 6D). When we cotransplanted dorsal mesendodermal cells of donor embryos into the dorsal side of setdb2-morphant host embryos, we found that donor cells from both wild-type embryos and morphants migrated to the animal pole posteriorly without any obvious differences (Fig. 6E). Similarly, when cells from the middle of the lateral blastoderm margin of these donor embryos were co-transplanted into setdb2-morphant host embryos, they failed to converge to the dorsal side (Fig. 6F). In addition, based on the function of Dvr1 as a ligand of the TGF-β superfamily, we predict that knockdown of Dvr1 also impaired C&E movements in non-cell-autonomous manner. As expected, when we co-transplanted dorsal mesendodermal cells or lateral mesendodermal cells derived from either wild-type or dvr1 moSB donor embryos into dvr1 moSB host embryos at dorsal or lateral sites, respectively, the C&E movements of the transplanted cells were impaired (Fig. 6G–J). These results demonstrate that a setdb2-dvr1 transcriptional cascade regulates zebrafish C&E movements in a non-cell-autonomous way.

Please cite this article as: Du, T.-T., et al., Setdb2 controls convergence and extension movements during zebrafish gastrulation by transcriptional regulation of dvr1. Dev. Biol. (2014), http://dx.doi.org/10.1016/j.ydbio.2014.05.022i

67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132

8

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66

T.-T. Du et al. / Developmental Biology ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Fig. 5. Identification of the essential role of Dvr1 in C&E cell movements using a cell-tracing strategy. (A–P) Tracking cell movements in C&E using UV-mediated photoactivation. At the one-cell stage, the embryos were injected with either 100 pg mRNA coding for Kaede protein plus dvr1 control morpholino or kaede mRNA plus dvr1/ setdb2 morpholinos. Embryos were labeled by UV radiation at the shield stage and analyzed at the tail-bud stage. Dorsal mesendodermal cells in control (A,B, n¼13), dvr1 moATG (E,F, n¼21), dvr1 moSB (I,J, n ¼10) and setdb2 mo (M,N, n ¼14). Lateral mesendodermal cells in control (C,D, n¼ 13), dvr1 moATG (G,H, n¼ 22), dvr1 moSB (K,L, n ¼10) and setdb2 mo (O,P, n¼18). (Q) Statistical analysis of anterior migration of labeled dorsal mesendodermal cells and dorsal migration of lateral mesendodermal cells. All images were lateral views. Asterisks indicate statistically significant differences (nnn, Po 0.001; nnnn, Po 0.0001). V, ventral side. D, dorsal side.

Please cite this article as: Du, T.-T., et al., Setdb2 controls convergence and extension movements during zebrafish gastrulation by transcriptional regulation of dvr1. Dev. Biol. (2014), http://dx.doi.org/10.1016/j.ydbio.2014.05.022i

67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132

T.-T. Du et al. / Developmental Biology ∎ (∎∎∎∎) ∎∎∎–∎∎∎

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66

9

Fig. 6. The setdb2-dvr1 transcriptional cascade regulates C&E cell movements in a non-cell-autonomous manner. (A-B) Representative images of transplantation for analysis of C&E cell movements at the shield stage. Red and green cells are those labeled with Rhodamine-dextran or fluorescein-dextran in donor embryos, respectively. (C-D) Dorsal or lateral mesendodermal cells derived from wild-type and setdb2-mo donor embryos, respectively, were transplanted into the corresponding regions of wild-type host embryos at the shield stage and imaged at 11 hpf. (E-F) Dorsal or lateral mesendodermal cells derived from wild-type and setdb2-mo donor embryos, respectively, were transplanted into the corresponding regions of setdb2 mo-injected host embryos at the shield stage and imaged at 11 hpf. (G-H) Dorsal or lateral mesendodermal cells derived from wild-type and dvr1 moSB donor embryos, respectively, were transplanted into the corresponding regions of wild-type host embryos at the shield stage and imaged at 11 hpf. (I-J) Dorsal or lateral mesendodermal cells derived from wild-type and dvr1 moSB donor embryos, respectively, were transplanted into the corresponding regions of dvr1 moSB-injected host embryos at the shield stage and imaged at 11 hpf. The percentages of phenotypes were as follows: (C) 7/9, (D) 6/6, (E) 7/8, (F) 9/9, (G) 8/10, (H) 6/6, (I) 8/8, (J) 9/10. The angles in C, E, G and I indicate the anterior migration of dorsal mesendodermal cells after transplantation. The white arrows in D, F, H and J indicate the dorsal migration of lateral mesendodermal cells after transplantation. All embryos were lateral views with head to the top.

Discussion As an essential and fascinating aspect of embryonic development, the complex C&E cell movements have been intensively studied. Many genes involved in the regulation of this process have been identified, but how these genes coordinate the process of C&E deserves further study. Here, we uncovered a novel requirement for Setdb2, a histone H3K9 methyltransferase, in regulating C&E movements. More importantly, we found that Setdb2 functions by fine-tuning the expression of dvr1 to an appropriate level to ensure proper C&E movements in a noncell-autonomous manner. Specifically, we first revealed that Dvr1, a ligand of the TGF-β superfamily, which has been shown to specify cell fates and regulate the formation of left-right asymmetry, also participates in cell movements during epiboly and C&E. This study provides novel insights into the mechanism that controls C&E cell movements during gastrulation. During vertebrate gastrulation, C&E movements play a pivotal role in shaping the embryonic body. The signaling pathways underlying the morphogenetic cell movements of the C&E process have been well established (Roszko et al., 2009). Although more and more reports show that epigenetic modifications of histones have essential roles in early embryonic development, whether epigenetic regulators participate in the C&E process remains largely unclear. Recently, the results of chromatin immunoprecipitation (ChIP) in zebrafish have shown the complexity and importance of the transcriptional regulatory function of histone modifications before, during and after the time of zygotic gene activation (ZGA), suggesting an instructive role of histone modifications for the control of developmental gene activation and repression (Andersen et al., 2012; Lindeman et al.; Lindeman et al., 2010). In this study, we provide a unique example of the importance of histone modifications in developmental gene repression during early zebrafish development: Setdb2, an epigenetic regulator responsible for H3K9me3, contributes to control of the C&E cell movements through negative regulation of dvr1 gene expression. Specifically, when we assessed the transcriptional expression

levels of setdb2 and dvr1 individually in wild-type embryos (Fig. 2B), we found a highly negative correlation between their transcriptional patterns. These data suggest that the chromatin state of dvr1 may be directly modified by Setdb2. Deciphering the H3K9me3 modification enrichment profile on the dvr1 promoter is a major goal in the future. Afterward, we plan to clarify whether the Setdb2 is recruited to the promoter of the dvr1 gene to directly exert its transcriptional repression activity and to identify any other transcriptional factors that may be involved in the proper positioning of Setdb2 at this promoter. Meanwhile, based on the finding that the knockdown with dvr1 moATG or moSB in setdb2 morphants only partially restored the C&E defects to normal (Fig. 2D–M, Fig. S4), we propose the possibility that other deregulated genes identified from microarray data may also participate in the regulation of C&E movements as candidate downstream targets of Setdb2. The process of C&E involves massive and complex cell movements to establish and shape the embryonic body. The behaviors of different cell groups have been well identified. Depending on the localization of the cells along the dorsal-ventral axis, cell behaviors can be divided into several types, as follows: I, cells of the most ventral domain cells migrate towards the vegetal pole; II, cells of the lateral domain cells primarily undergo directed cell migration towards the dorsal side; III, cells of the lateral-dorsal domain show medial planar and polarized radial intercalations; IV, cells of the dorsal domain have anterior-directed migration in the most anterior cells or mediolateral intercalations in the posterior cells (Glickman et al., 2003; Jessen et al., 2002; Myers et al., 2002a; Roszko et al., 2009; Sepich et al., 2000; Solnica-Krezel, 2006; Topczewski et al., 2001; Ulrich et al., 2003; Yin et al., 2008). The specific cell behaviors are precisely coordinated in time and space by various signaling pathways. Recently, the molecular mechanisms underlying the individual cell behaviors have been elucidated partially (Solnica-Krezel, 2006). Based on the present study, the impairment of anterior-directed migration in the dorsal domain was mild in both setdb2 and dvr1 morphants. In contrast, the convergence of the lateral or dorsal domain and the extension of

Please cite this article as: Du, T.-T., et al., Setdb2 controls convergence and extension movements during zebrafish gastrulation by transcriptional regulation of dvr1. Dev. Biol. (2014), http://dx.doi.org/10.1016/j.ydbio.2014.05.022i

67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132

10

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66

T.-T. Du et al. / Developmental Biology ∎ (∎∎∎∎) ∎∎∎–∎∎∎

posterior cells in the dorsal domain were seriously affected (Figs. 1, 4). The differences in the behaviors of cell groups may reflect variations in the underlying molecular mechanism. Therefore, these results thus prompt us to further investigate precisely which cell groups and behaviors are driven by the setdb2-dvr1 transcriptional cascade. How does the setdb2-dvr1 transcriptional cascade control the C&E movements? The key question is which downstream targets of Dvr1 mediate the specific cell movements in setdb2 morphants. Our transplantation experiment showed that the setdb2-dvr1 transcriptional cascade and inhibition of Dvr1 both effected C&E movements in a non-cell-autonomous manner (Fig. 6). These results, along with previous studies of the dvr1 gene, strongly suggest that the upregulated dvr1 gene in setdb2 morphants may act as the ligand for TGF-β signals to activate downstream signals. Although the critical roles of Gdf1/Vg1/Dvr1 during embryonic development have been well defined, the molecular mechanism and the signaling pathway activated by these ligands have not been well verified. Previous reports suggest that there are similarities in the functions and signaling properties of Vg1/Gdf1 and other subgroups of TGF-β ligands, such as Nodal/Activin. In Xenopus, maternal Vg1 is required for Smad2 phosphorylation (Birsoy et al., 2006). Mesoderm induction by Gdf1 is blocked by truncation of the Activin type II receptor (inhibiting several TGF-β signaling pathways including Activin, Bmp4 and Nodal) and Smad2-interaction domain of FAST1 (blocking formation of the Smad2-Smad4 complex) (Wall et al., 2000). An additional study indicates that aVg1 acts similarly to Nodal and depends on EGFCFC coreceptors for the interaction with and activation of Activin receptors (Cheng et al., 2003). In our study, we also detected the specifically increased expression level of lefty1, which inhibits the Nodal signaling pathway (Schier and Talbot, 2005), in setdb2 morphants at the onset of gastrulation, but not lefty2, cyclops or squint (data not shown). Determining whether Nodal signaling is activated requires further testing. All of these results led us to speculate that the setdb2-dvr1 transcriptional cascade may resemble that of Nodal in the regulation of C&E movements or may instead mediate C&E movements through a unique pathway that has not been elucidated. In our study, we also observed the defects of cell polarity in setdb2 morphants, dvr1 morphants and dvr1 overexpression embryos (Fig. 1, Fig. S3). As generally known, the Wnt/PCP pathway is a crucial regulator of C&E movements and essential for polarized cell behaviors. In order to address the possibility that the Wnt/PCP pathway was compromised in the impaired C&E movements caused by aberrant setdb2-dvr1 transcriptional cascade, we checked the expression level of the Wnt/PCP core components (wnt5b, wnt11, vangl2, gpc4, prickle1a, dvl2, dvl3, fzd2, fzd7a, fzd7b, cdc42-2, rac1 and rhoab) in setdb2 mo microarray data and did not observe obvious alterations in these genes (data not shown). In addition, we also analyzed these Wnt/PCP core genes in dvr1 moATG morphants by real-time PCR. The expression of the genes mentioned above showed no obvious alterations in dvr1 moATG morphants, except for the mild up-regulation of wnt11 (data not shown). Based on the results above, we have not found obvious aberrent transcriptional level of primary Wnt/PCP core genes in setdb2 or dvr1 morphant embryos. The involvement of the Wnt/ PCP pathway and the genetic interaction between Wnt/PCP pathway and setdb2-dvr1 transcriptional cascade need to be further investigated. During gastrulation, the embryonic cells undergo massive migration to acquire proper localization and to sculpt the body axis. Meanwhile, cells start to make cell fate decisions by expressing specific genes (Myers et al., 2002b). The key point is the relationship between cell movement behaviors and cell fate decisions during gastrulation. Previous studies have shown that

maternal Wnt/β-catenin activates squint and dharma/bozozok to participate in organizer formation and cell specification (SolnicaKrezel and Driever, 2001). Moreover, this signal also causes the activation of stat3, which is not essential for cell fate specification but is required for C&E movements (Miyagi et al., 2004; Yamashita et al., 2002). These observations indicate that maternal Wnt/βcatenin signaling may mediate cell fate specification and cell movements via parallel pathways. In our previous study, we found that the setdb2 mo induced abnormal cell fate specification at the onset of gastrulation (Xu et al., 2010) (Fig. S7). Further study indicated that the expanded dorsal organizer but not the C&E defects induced by Setdb2 knockdown could be well restored to normal by co-injecting fgf8a morpholino (data not shown). These observations suggest that setdb2-dvr1 transcriptional cascade regulates C&E movements in parallel with, rather than downstream of, cell fate specification. Similarly, in our study, we observed the downregulation of bmp4 in dvr1 moATG embryos. It suggests that maternal expression of dvr1 gene may modulate cell fate specification and raises the question of whether aberrant cell fate determination induced by maternal Dvr1 inhibition is also responsible for abnormal cell movement.

Materials and methods Fish care Zebrafish maintenance, breeding, and staging were performed under standard conditions as described previously (Kimmel et al., 1995). Plasmid construction The fragments of gsc, egr2b, myod1 and myl7 were cloned into pCS2þ, pBSK or pGEM-Teasy vector. The plasmids dlx3b, ctsl1b, bmp2b, bmp4, spaw, cyclops, flh, chordin, fgf8a, ntla, tbx6 and the chimeric constructs of dvr1 for dvr1 mRNA synthesis were generous gifts from other laboratories (Acknowledgments). Whole-mount in situ hybridization Whole-mount in situ hybridization with mRNA antisense probes were performed as described previously (Thisse and Thisse, 2008). The stained embryos were mounted in 4% methylcellulose and photographed using the Nikon SMZ1500 stereomicroscope. Morpholinos and mRNA injection Setdb2 control mo: 50 –CACGGTTGAGGACATTTAAATCACT-30 , Setdb2 mo: 50 –CAGGGTTCAGGAGATTTTAATGACT-30 , dvr1 control mo: 50 GCTGTCAGCAGCACCAACAACATTA-30 , dvr1 moATG: 50 –GCTCTGAGGAGGACCAAGAACATTA-30 , dvr1 moSB: 50 –GTGGTCCGTAAGATCCTCACCTTGA-30 , p53 mo: 50 –TCTTGGCTGTCGTTTTGCGCCATT G-30 . All morpholino antisense oligonucleotides were purchased from GENE TOOLS. Capped mRNAs were synthesized with the Message machine kit according to manufacturer's instructions (Ambion). All morpholinos and mRNAs were injected at the one-cell stage. Generation of anti-Setdb2 antibody and western blot A rabbit polyclonal antiserum against zebrafish Setdb2 protein was generated using a C-terminal peptide of zebrafish Setdb2 as an antigen source, and used to immune a rabbit (willget biotech). Western blot was performed as described previously (Fu et al., 2009). Signals were detected with Mouse anti-histone H3

Please cite this article as: Du, T.-T., et al., Setdb2 controls convergence and extension movements during zebrafish gastrulation by transcriptional regulation of dvr1. Dev. Biol. (2014), http://dx.doi.org/10.1016/j.ydbio.2014.05.022i

67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132

T.-T. Du et al. / Developmental Biology ∎ (∎∎∎∎) ∎∎∎–∎∎∎

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66

antibody (1:10,000) and rabbit anti-zebrafish Setdb2 antibody (1:1000). Immunofluorescence Whole-mount immunofluorescence with primary antibody against rabbit DsRed (1:500 dilution, clontech) and Alexa Fluor 546-labeled secondary antibody (donkey anti-rabbit, 1:500 dilution, Invitrogen) was performed as described previously (Xu et al., 2010). Cell polarity analyses The embryos were injected with 16 ng setdb2 control mo plus 100 pg membrane-RFP mRNA (membrane labeling), or setdb2 mo plus 100 pg membrane-RFP mRNA. The stained embryos were mounted in 2% low-melting point agarose for visualization of notochord and the images were collected using a OLYMPUS-FV1000 confocal microscope equipped with 60  water-immersion objectives. LWRs and ML alignment were measured with ImageJ2X. The rose diagrams were drawn using Rose.NET. Microarray analysis The embryos were injected with 16 ng setdb2 control mo or 16 ng setdb2 mo at one-cell stage and collected at 6 hpf. About 5– 10 μg total RNA was extracted for cDNA generation, amplified and hybridized to Affymetrix Zebrafish Genome Arrays. Two independent experiments of biological replicates (setdb2 control mo and setdb2 mo) were performed. The microarray assay and data analysis were finished by Gene Tech (Shanghai) Company Limited. The microarray data has been loaded into ArrayExpress (www.ebi. ac.uk/arrayexpress). The ArrayExpress accession account are as follows: Username: Reviewer_E-MTAB-2207, Password: UMFCB80d.

11

Statistical analysis Data analyses were performed in excel and GraphPad Prism6. All statistics were performed using a two-tailed Student's t-test. Results were presented as means 7 SD.

Funding This work was supported by the National Natural Science Foundation of China (31000636).

Author Contributions T.-T.D. and P.-F.X. performed experiments and analyzed data; Z.-W.D., H.-B.F., Y.J., M.D. and Y.C. assisted with experiments; T.-T. D., P.-F.X., W.-J.P., R.-B.R., T.-X.L., M.D. and Q.-H.H. designed the research plan; T.-T.D., P.-F.X. and Q.-H.H. wrote the paper.

Acknowledgments We thank Dr. Bernard Thisse and Christine Thisse at University of Virginia for providing dlx3b, ctsl1b, bmp2b and bmp4 plasmids, Dr. Joseph Yost at University of Utah for providing spaw and Cyclops plasmids, Dr. Steve Wilson at University College London for providing flh, chordin and fgf8a plasmids, Dr. Michael Rebagliati at University of Iowa for providing ntla plasmid, Dr. Sharon L. Amacher at University of California, Berkeley for providing tbx6 plasmid and Dr. Alexander F. Schier at Harvard University for providing the chimeric constructs of dvr1. We thank Hao Yuan for providing p53 morpholino. We thank Dr. Bernard Thisse and Dr. Christine Thisse for careful reading of the manuscript and helpful suggestions. We thank Shan-He Yu, Lei Wang, Xiao-E Jia, Lei Gao, Yuan-Liang Zhang, Yin-Yin Xie for excellent technical assistance. We thank Dr. Xiao-Jian Sun and all members of the laboratory for helpful discussions.

Quantitative real-time PCR Real-time quantitative RT-PCR was performed using SYBR Green (Toyobo Engineering). Zebrafish housekeeping gene gapdh was used as an internal control. The QPCR primer sequences used in this study were described in Table S2.

Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.ydbio.2014.05.022. References

Cell-tracking by UV-mediated photoactivation Embryos were injected with 400 pg mRNA coding for Kaede protein plus dvr1 control mo, or kaede mRNA plus dvr1/setdb2 morpholinos at one-cell stage. The embryos were radiated by UV at shield stage. The cell movements images were performed on OLYMPUS-FV-1000 confocal microscope under 20  objectives at the shield and the tail-bud stages. Transplantation experiments At one-cell stage, the donor embryos were injected with 2 nl of 0.5% rhodamine-dextran alone, or 2 nl of 0.5% fluorescein-dextran (MW 10,000, Molecular Probes) together with 16 ng setdb2 mo/ dvr1 moSB. The host embryos were uninjected or injected with setdb2 mo/dvr1 moSB. At 6 hpf, the dorsal or lateral mesendodermal cells of two donor embryos were transplanted to the corresponding regions of the host embryos. The images were collected at 6 hpf and 10.5 hpf under OLYMPUS-FV-1000 confocal microscope with 20  objectives.

Andersen, I.S., Ostrup, O., Lindeman, L.C., Aanes, H., Reiner, A.H., Mathavan, S., Alestrom, P., Collas, P., 2012. Epigenetic complexity during the zebrafish midblastula transition. Biochem. Biophys. Res. Commun. 417, 1139–1144. Andersson, O., Bertolino, P., Ibanez, C.F., 2007. Distinct and cooperative roles of mammalian Vg1 homologs GDF1 and GDF3 during early embryonic development. Dev. Biol. 311, 500–511. Ataliotis, P., Symes, K., Chou, M.M., Ho, L., Mercola, M., 1995. PDGF signalling is required for gastrulation of Xenopus laevis. Development 121, 3099–3110. Bakkers, J., Kramer, C., Pothof, J., Quaedvlieg, N.E., Spaink, H.P., Hammerschmidt, M., 2004. Has2 is required upstream of Rac1 to govern dorsal migration of lateral cells during zebrafish gastrulation. Development 131, 525–537. Birsoy, B., Kofron, M., Schaible, K., Wylie, C., Heasman, J., 2006. Vg 1 is an essential signaling molecule in Xenopus development. Development 133, 15–20. Chan, J., Mably, J.D., Serluca, F.C., Chen, J.N., Goldstein, N.B., Thomas, M.C., Cleary, J. A., Brennan, C., Fishman, M.C., Roberts, T.M., 2001. Morphogenesis of prechordal plate and notochord requires intact Eph/ephrin B signaling. Dev. Biol. 234, 470–482. Chen, Q., Takada, R., Takada, S., 2012. Loss of Porcupine impairs convergent extension during gastrulation in zebrafish. J. Cell. Sci. 125, 2224–2234. Cheng, S.K., Olale, F., Bennett, J.T., Brivanlou, A.H., Schier, A.F., 2003. EGF-CFC proteins are essential coreceptors for the TGF-beta signals Vg1 and GDF1. Genes Dev. 17, 31–36. Coyle, R.C., Latimer, A., Jessen, J.R., 2008. Membrane-type 1 matrix metalloproteinase regulates cell migration during zebrafish gastrulation: evidence for an interaction with non-canonical Wnt signaling. Exp. Cell Res. 314, 2150–2162.

Please cite this article as: Du, T.-T., et al., Setdb2 controls convergence and extension movements during zebrafish gastrulation by transcriptional regulation of dvr1. Dev. Biol. (2014), http://dx.doi.org/10.1016/j.ydbio.2014.05.022i

67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132

12

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 Q2 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63

T.-T. Du et al. / Developmental Biology ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Dohrmann, C.E., Kessler, D.S., Melton, D.A., 1996. Induction of axial mesoderm by zDVR-1, the zebrafish orthologue of Xenopus Vg1. Dev. Biol. 175, 108–117. Fu, Y.F., Du, T.T., Dong, M., Zhu, K.Y., Jing, C.B., Zhang, Y., Wang, L., Fan, H.B., Chen, Y., Jin, Y., Yue, G.P., Chen, S.J., Chen, Z., Huang, Q.H., Jing, Q., Deng, M., Liu, T.X., 2009. Mir-144 selectively regulates embryonic alpha-hemoglobin synthesis during primitive erythropoiesis. Blood 113, 1340–1349. Ghil, J.S., Chung, H.M., 1999. Evidence that platelet derived growth factor (PDGF) action is required for mesoderm patterning in early amphibian (Xenopus laevis) embryogenesis. Int. J. Dev. Biol. 43, 329–334. Glickman, N.S., Kimmel, C.B., Jones, M.A., Adams, R.J., 2003. Shaping the zebrafish notochord. Development 130, 873–887. Hammerschmidt, M., Pelegri, F., Mullins, M.C., Kane, D.A., Brand, M., van Eeden, F.J., Furutani-Seiki, M., Granato, M., Haffter, P., Heisenberg, C.P., Jiang, Y.J., Kelsh, R. N., Odenthal, J., Warga, R.M., Nusslein-Volhard, C., 1996. Mutations affecting morphogenesis during gastrulation and tail formation in the zebrafish, Danio rerio. Development 123, 143–151. Hatta, K., Tsujii, H., Omura, T., 2006. Cell tracking using a photoconvertible fluorescent protein. Nat. Protoc. 1, 960–967. Heisenberg, C.P., Tada, M., Rauch, G.J., Saude, L., Concha, M.L., Geisler, R., Stemple, D. L., Smith, J.C., Wilson, S.W., 2000. Silberblick/Wnt11 mediates convergent extension movements during zebrafish gastrulation. Nature 405, 76–81. Hyatt, B.A., Lohr, J.L., Yost, H.J., 1996. Initiation of vertebrate left-right axis formation by maternal Vg1. Nature 384, 62–65. Jessen, J.R., Topczewski, J., Bingham, S., Sepich, D.S., Marlow, F., Chandrasekhar, A., Solnica-Krezel, L., 2002. Zebrafish trilobite identifies new roles for Strabismus in gastrulation and neuronal movements. Nat. Cell Biol. 4, 610–615. Jones, T.L., Chong, L.D., Kim, J., Xu, R.H., Kung, H.F., Daar, I.O., 1998. Loss of cell adhesion in Xenopus laevis embryos mediated by the cytoplasmic domain of XLerk, an erythropoietin-producing hepatocellular ligand. Proc. Natl. Acad. Sci. USA 95, 576–581. Kaasinen, E., Aittomaki, K., Eronen, M., Vahteristo, P., Karhu, A., Mecklin, J.P., Kajantie, E., Aaltonen, L.A., Lehtonen, R., 2010. Recessively inherited right atrial isomerism caused by mutations in growth/differentiation factor 1 (GDF1). Hum. Mol. Genet. 19, 2747–2753. Keller, R., 2002. Shaping the vertebrate body plan by polarized embryonic cell movements. Science 298, 1950–1954. Kilian, B., Mansukoski, H., Barbosa, F.C., Ulrich, F., Tada, M., Heisenberg, C.P., 2003. The role of Ppt/Wnt5 in regulating cell shape and movement during zebrafish gastrulation. Mech. Dev. 120, 467–476. Kim, S.H., Yamamoto, A., Bouwmeester, T., Agius, E., Robertis, E.M., 1998. The role of paraxial protocadherin in selective adhesion and cell movements of the mesoderm during Xenopus gastrulation. Development 125, 4681–4690. Kimmel, C.B., Ballard, W.W., Kimmel, S.R., Ullmann, B., Schilling, T.F., 1995. Stages of embryonic development of the zebrafish. Dev. Dyn. 203, 253–310. Kouzarides, T., 2007. Chromatin modifications and their function. Cell 128, 693–705. Li, E., 2002. Chromatin modification and epigenetic reprogramming in mammalian development. Nat. Rev. Genet. 3, 662–673. Li, X., Ma, Y., Li, D., Gao, X., Li, P., Bai, N., Luo, M., Tan, X., Lu, C., Ma, X., 2012. Arsenic impairs embryo development via down-regulating Dvr1 expression in zebrafish. Toxicol. Lett. 212, 161–168. Lindeman, L.C., Andersen, I.S., Reiner, A.H., Li, N., Aanes, H., Ostrup, O., Winata, C., Mathavan, S., Muller, F., Alestrom, P., Collas, P., Prepatterning of developmental gene expression by modified histones before zygotic genome activation. Dev Cell., 21, 993–1004. Lindeman, L.C., Winata, C.L., Aanes, H., Mathavan, S., Alestrom, P., Collas, P., 2010. Chromatin states of developmentally-regulated genes revealed by DNA and histone methylation patterns in zebrafish embryos. Int. J. Dev. Biol. 54, 803–813. Matsui, T., Bessho, Y., 2012. Left-right asymmetry in zebrafish. Cell. Mol. Life Sci. 69, 3069–3077. Matsui, T., Raya, A., Kawakami, Y., Callol-Massot, C., Capdevila, J., RodriguezEsteban, C., Izpisua Belmonte, J.C., 2005. Noncanonical Wnt signaling regulates midline convergence of organ primordia during zebrafish development. Genes Dev. 19, 164–175. Miyagi, C., Yamashita, S., Ohba, Y., Yoshizaki, H., Matsuda, M., Hirano, T., 2004. STAT3 noncell-autonomously controls planar cell polarity during zebrafish convergence and extension. J. Cell Biol. 166, 975–981. Myers, D.C., Sepich, D.S., Solnica-Krezel, L., 2002a. Bmp activity gradient regulates convergent extension during zebrafish gastrulation. Dev. Biol. 243, 81–98. Myers, D.C., Sepich, D.S., Solnica-Krezel, L., 2002b. Convergence and extension in vertebrate gastrulae: cell movements according to or in search of identity? Trends Genet. 18, 447–455. Oates, A.C., Lackmann, M., Power, M.A., Brennan, C., Down, L.M., Do, C., Evans, B., Holder, N., Boyd, A.W., 1999. An early developmental role for eph-ephrin interaction during vertebrate gastrulation. Mech. Dev. 83, 77–94. Oishi, I., Kawakami, Y., Raya, A., Callol-Massot, C., Izpisua Belmonte, J.C., 2006. Regulation of primary cilia formation and left-right patterning in zebrafish by a noncanonical Wnt signaling mediator, duboraya. Nat. Genet. 38, 1316–1322. Peterson, A.G., Wang, X., Joseph Yost, H., 2013. Dvr1 transfers left-right asymmetric signals from Kupffer's vesicle to lateral plate mesoderm in zebrafish. Dev. Biol. 382, 198–208.

Rankin, C.T., Bunton, T., Lawler, A.M., Lee, S.J., 2000. Regulation of left-right patterning in mice by growth/differentiation factor-1. Nat. Genet. 24, 262–265. Rauch, G.J., Hammerschmidt, M., Blader, P., Schauerte, H.E., Strahle, U., Ingham, P. W., McMahon, A.P., Haffter, P., 1997. Wnt5 is required for tail formation in the zebrafish embryo. Cold Spring Harb. Symp. Quant. Biol. 62, 227–234. Robu, M.E., Larson, J.D., Nasevicius, A., Beiraghi, S., Brenner, C., Farber, S.A., Ekker, S. C., 2007. p53 activation by knockdown technologies. PLoS Genet. 3, e78. Roszko, I., Sawada, A., Solnica-Krezel, L., 2009. Regulation of convergence and extension movements during vertebrate gastrulation by the Wnt/PCP pathway. Semin. Cell Dev. Biol. 20, 986–997. Schier, A.F., Talbot, W.S., 2005. Molecular genetics of axis formation in zebrafish. Annu. Rev. Genet. 39, 561–613. Sepich, D.S., Myers, D.C., Short, R., Topczewski, J., Marlow, F., Solnica-Krezel, L., 2000. Role of the zebrafish trilobite locus in gastrulation movements of convergence and extension. Genesis 27, 159–173. Solnica-Krezel, L., 2006. Gastrulation in zebrafish – all just about adhesion? Curr. Opin. Genet. Dev. 16, 433–441. Solnica-Krezel, L., Driever, W., 2001. The role of the homeodomain protein Bozozok in zebrafish axis formation. Int. J. Dev. Biol. 45, 299–310. Stickney, H.L., Imai, Y., Draper, B., Moens, C., Talbot, W.S., 2007. Zebrafish bmp4 functions during late gastrulation to specify ventroposterior cell fates. Dev. Biol. 310, 71–84. Sumanas, S., Kim, H.J., Hermanson, S., Ekker, S.C., 2001. Zebrafish frizzled-2 morphant displays defects in body axis elongation. Genesis 30, 114–118. Thisse, C., Thisse, B., 2008. High-resolution in situ hybridization to whole-mount zebrafish embryos. Nat. Protoc. 3, 59–69. Topczewski, J., Sepich, D.S., Myers, D.C., Walker, C., Amores, A., Lele, Z., Hammerschmidt, M., Postlethwait, J., Solnica-Krezel, L., 2001. The zebrafish glypican knypek controls cell polarity during gastrulation movements of convergent extension. Dev. Cell. 1, 251–264. Tsai, Y.J., Pan, H., Hung, C.M., Hou, P.T., Li, Y.C., Lee, Y.J., Shen, Y.T., Wu, T.T., Li, C., 2011. The predominant protein arginine methyltransferase PRMT1 is critical for zebrafish convergence and extension during gastrulation. FEBS J. 278, 905–917. Ulrich, F., Concha, M.L., Heid, P.J., Voss, E., Witzel, S., Roehl, H., Tada, M., Wilson, S. W., Adams, R.J., Soll, D.R., Heisenberg, C.P., 2003. Slb/Wnt11 controls hypoblast cell migration and morphogenesis at the onset of zebrafish gastrulation. Development 130, 5375–5384. Ungar, A.R., Kelly, G.M., Moon, R.T., 1995. Wnt4 affects morphogenesis when misexpressed in the zebrafish embryo. Mech Dev. 52, 153–164. van Eekelen, M., Runtuwene, V., Overvoorde, J., den Hertog, J., 2010. RPTPalpha and PTPepsilon signaling via Fyn/Yes and RhoA is essential for zebrafish convergence and extension cell movements during gastrulation. Dev. Biol. 340, 626–639. von der Hardt, S., Bakkers, J., Inbal, A., Carvalho, L., Solnica-Krezel, L., Heisenberg, C. P., Hammerschmidt, M., 2007. The Bmp gradient of the zebrafish gastrula guides migrating lateral cells by regulating cell-cell adhesion. Curr. Biol. 17, 475–487. Wall, N.A., Craig, E.J., Labosky, P.A., Kessler, D.S., 2000. Mesendoderm induction and reversal of left-right pattern by mouse Gdf1, a Vg1-related gene. Dev. Biol. 227, 495–509. Warga, R.M., Kimmel, C.B., 1990. Cell movements during epiboly and gastrulation in zebrafish. Development 108, 569–580. Williams, B.B., Cantrell, V.A., Mundell, N.A., Bennett, A.C., Quick, R.E., Jessen, J.R., 2012. VANGL2 regulates membrane trafficking of MMP14 to control cell polarity and migration. J. Cell Sci. 125, 2141–2147. Wu, M.Y., Hill, C.S., 2009. Tgf-beta superfamily signaling in embryonic development and homeostasis. Dev. Cell. 16, 329–343. Xu, P.F., Zhu, K.Y., Jin, Y., Chen, Y., Sun, X.J., Deng, M., Chen, S.J., Chen, Z., Liu, T.X., 2010. Setdb2 restricts dorsal organizer territory and regulates left-right asymmetry through suppressing fgf8 activity. Proc. Natl. Acad. Sci. USA 107, 2521–2526. Xu, X., Shuen, W.H., Chen, C., Goudevenou, K., Jones, P., Sablitzky, F., 2013. Swap70b is required for convergent and extension cell movement during zebrafish gastrulation linking Wnt11 signalling and RhoA effector function. Dev. Biol. 386, 191–203. Yamamoto, A., Amacher, S.L., Kim, S.H., Geissert, D., Kimmel, C.B., De Robertis, E.M., 1998. Zebrafish paraxial protocadherin is a downstream target of spadetail involved in morphogenesis of gastrula mesoderm. Development 125, 3389–3397. Yamashita, S., Miyagi, C., Carmany-Rampey, A., Shimizu, T., Fujii, R., Schier, A.F., Hirano, T., 2002. Stat3 controls cell movements during zebrafish gastrulation. Dev. Cell. 2, 363–375. Yamashita, S., Miyagi, C., Fukada, T., Kagara, N., Che, Y.S., Hirano, T., 2004. Zinc transporter LIVI controls epithelial-mesenchymal transition in zebrafish gastrula organizer. Nature 429, 298–302. Yin, C., Kiskowski, M., Pouille, P.A., Farge, E., Solnica-Krezel, L., 2008. Cooperation of polarized cell intercalations drives convergence and extension of presomitic mesoderm during zebrafish gastrulation. J. Cell Biol. 180, 221–232.

Please cite this article as: Du, T.-T., et al., Setdb2 controls convergence and extension movements during zebrafish gastrulation by transcriptional regulation of dvr1. Dev. Biol. (2014), http://dx.doi.org/10.1016/j.ydbio.2014.05.022i

64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125