Blocking Dishevelled signaling in the noncanonical Wnt pathway in sea urchins disrupts endoderm formation and spiculogenesis, but not secondary mesoderm formation


  • Christine A. Byrum,

    Corresponding author
    1. Department of Zoology, University of Hawaii at Manoa, Honolulu, Hawaii
    2. Developmental, Cell and Molecular Biology Group, Duke University, Durham, North Carolina
    3. Department of Biology, College of Charleston, Charleston, South Carolina
    • Department of Biology, College of Charleston, Charleston, SC 29401
    Search for more papers by this author
  • Ronghui Xu,

    1. Department of Zoology, University of Hawaii at Manoa, Honolulu, Hawaii
    Search for more papers by this author
  • Joanna M. Bince,

    1. Department of Zoology, University of Hawaii at Manoa, Honolulu, Hawaii
    Search for more papers by this author
  • David R. McClay,

    1. Developmental, Cell and Molecular Biology Group, Duke University, Durham, North Carolina
    Search for more papers by this author
  • Athula H. Wikramanayake

    Corresponding author
    1. Department of Zoology, University of Hawaii at Manoa, Honolulu, Hawaii
    2. Department of Biology, The University of Miami, Coral Gables, Florida
    • Department of Biology, The University of Miami, 1301 Memorial Drive, Coral Gables, FL 33146
    Search for more papers by this author


Dishevelled (Dsh) is a phosphoprotein key to beta-catenin dependent (canonical) and beta-catenin independent (noncanonical) Wnt signaling. Whereas canonical Wnt signaling has been intensively studied in sea urchin development, little is known about other Wnt pathways. To examine roles of these beta-catenin independent pathways in embryogenesis, we used Dsh-DEP, a deletion construct blocking planar cell polarity (PCP) and Wnt/Ca2+ signaling. Embryos overexpressing Dsh-DEP failed to gastrulate or undergo skeletogenesis, but produced pigment cells. Although early mesodermal gene expression was largely unperturbed, embryos exhibited reduced expression of genes regulating endoderm specification and differentiation. Overexpressing activated beta-catenin failed to rescue Dsh-DEP embryos, indicating that Dsh-DEP blocks endoderm formation downstream of initial canonical Wnt signaling. Because Dsh-DEP-like constructs block PCP signaling in other metazoans, and disrupting RhoA or Fz 5/8 in echinoids blocks subsets of the Dsh-DEP phenotypes, our data suggest that noncanonical Wnt signaling is crucial for sea urchin endoderm formation and skeletogenesis. Developmental Dynamics 238:1649–1665, 2009. © 2009 Wiley-Liss, Inc.


In sea urchin embryos, both the canonical and noncanonical Wnt pathways are known to act in cell fate specification (Croce et al.,2006a; Beane,2007; Kumburegama and Wikramanayake,2007). The canonical Wnt pathway is critical in specification of the endomesoderm, a cell subpopulation that later subdivides to form endoderm and mesoderm (Emily-Fenouil et al.,1998; Wikramanayake et al.,1998; Logan et al.,1999), but much remains to be learned about the roles of noncanonical Wnt signaling in echinoids.

In general, both the canonical and noncanonical Wnt signaling pathways are initiated in a similar manner. A Wnt ligand binds to a Frizzled (Fz) receptor thereby activating the cytoplasmic protein Dishevelled (Dsh). Downstream of Dsh, however, the transducing pathways diverge. In the so-called canonical Wnt pathway, Dsh cooperates with other molecules to block phosphorylation of cytoplasmic β-catenin by GSK-3β (Wharton,2003; Wallingford and Habas,2005). This prevents degradation of β-catenin and allows the protein to accumulate in the nucleus where, in conjunction with TCF/LEF, it activates transcription of downstream target genes. In the noncanonical Wnt pathways, binding of a Wnt ligand to Fz activates effectors other than β-catenin, but signal transduction by means of these pathways also requires Dsh. Noncanonical Wnt signaling pathways include (A) the planar cell polarity (PCP) pathway, which signals through the Rho/ROCK and Rac/JNK signaling cascades (Weber et al.,2000; Park et al.,2006), and (B) the Wnt/Ca2+ pathway, which activates release of intracellular calcium and calcium sensitive kinases like PKC and CamKII (Sheldahl et al.,2003). Like the canonical Wnt pathway, the noncanonical Wnt pathways are important in development. In vertebrates, PCP signaling regulates cell movements during gastrulation and neurulation (Wallingford et al.,2000; Heisenberg et al.,2000; Wallingford and Harland,2002), is crucial for cardiogenesis (Eisenberg and Eisenberg,1999; Pandur et al.,2002), and acts in formation of the organ of corti (Montcouquiol et al.,2003; Dabdoub et al.,2003). Developmental processes affected by Wnt/Ca2+ signaling include avian myoblast differentiation (Anakwe et al.,2003; Kohn and Moon,2005), zebrafish dorsal axis specification (Kohn and Moon,2005), and cell–cell communication in the mature mouse retina (Yu et al.,2004).

The ability of Dsh to transduce canonical and noncanonical Wnt signals primarily depends on three domains within the protein: the DIX, PDZ, and DEP domains (Penton et al.,2002). The DIX and PDZ domains function together in β-catenin dependent Wnt signaling (Tada and Smith,2000; Rothbacher et al.,2000), whereas the PDZ and DEP domains cooperate in both the PCP (Tada and Smith,2000) and Wnt/Ca2+ (Sheldahl et al.,2003) pathways. Because specific Dsh domains participate in diverting Wnt signals to either canonical or noncanonical pathways, Dsh deletion constructs can be used to selectively block these Wnt signaling pathways (e.g., Axelrod et al.,1998; Tada and Smith,2000; Rothbacher et al.,2000; Weitzel et al.,2004).

Two recent studies suggest that noncanonical Wnt signaling acts in sea urchin development. In an investigation of the Wnt receptor Frizzled 5/8 (Fz 5/8), Croce et al. (2006a) showed that overexpressing a dominant negative form (FzTM1), blocked formation of endoderm and spicules without impairing production of mesenchyme-like cells or pigment cells. It was noted that the dominant negative phenotype was rescued by coexpression of a constitutively active form of RhoA, a protein that acts in PCP signal transduction in other bilaterians. These results, and supporting studies, led Croce et al. (2006a) to hypothesize that Fz 5/8 signals through the PCP pathway. A second relevant study focused on RhoA (Beane et al.,2006; Beane,2007). Like Croce et al. (2006a), Beane et al. found that blocking RhoA activity either with pharmacological inhibitors or a dominant negative form of RhoA also prevented invagination and spicule formation, but allowed production of mesenchyme-like cells. A similar phenotype resulted when activity of the RhoA effector ROCK was blocked using a chemical inhibitor. Based on their observations, Beane et al. concluded that the observed defects were also caused by inhibition of the PCP pathway.

In a previous investigation, we studied roles of Dsh in regulating patterning along the sea urchin animal–vegetal (A–V) axis. This work examined the early distribution of Dsh during embryogenesis and documented roles of the protein in canonical Wnt signaling (Weitzel et al.,2004). Using GFP-tagged Dsh (in Lytechinus variegatus), we showed that GFP-Dsh accumulates selectively in the vegetal cortex of fertilized eggs and that it continues to accumulate in vegetal regions through the 32-cell stage. We also demonstrated that Dsh plays a critical role in nuclear localization of β-catenin during canonical Wnt signaling. By stabilizing β-catenin in the vegetal cells, Dsh regulates entry of β-catenin into the nucleus where β-catenin cooperates with TCF/LEF to activate gene transcription (Weitzel et al.,2004).

In this study, we have used the Dsh-DEP construct to clarify the role of noncanonical Wnt signaling in sea urchin development and to test the hypothesis that blocking PCP signaling interferes with formation of the archenteron and spicules. Dsh-DEP lacks both the DIX and PDZ domains and has previously been shown to block PCP signaling in other bilaterians (Axelrod et al.,1998; Tada and Smith,2000). In flies, for example, Dsh-DEP overexpression disrupts planar cell polarity as characterized by random orientation of wing hairs and body bristles, mispatterning of the ommatidiae, and impaired dorsal closure (Axelrod et al.,1998). In Xenopus, Dsh-DEP overexpression in animal pole explants blocks convergent extension of this tissue in response to activin treatment (Tada and Smith,2000). Studies suggest that this Dsh-DEP construct competitively prevents localization of endogenous Dsh to the cell membrane, thereby inhibiting noncanonical Wnt signaling (Axelrod et al.,1998; Pan et al.,2004). Finally, there is additional evidence that overexpressing the DEP domain perturbs Wnt/Ca2+ signaling because Dsh-DEP overexpression blocks Fz-mediated activation of PKC (a Wnt/Ca2+ component) in Xenopus embryos (Sheldahl et al.,2003).

Inhibiting PCP signaling by overexpressing either the Fz 5/8 construct, FzTM1 (Croce et al.,2006a), or a dnRhoA (Beane et al.,2006) construct likely disrupts only a subset of the processes regulated by noncanonical Wnt pathways in the sea urchin. Because there are four different Fz homologues present in the sea urchin genome (Croce et al.,2006b), it is possible that more than one signals through the PCP pathway or that a homologue other than Fz 5/8 transduces Wnt/Ca2+ signals. In addition, dnRhoA only blocks the Rho/ROCK portion of the PCP pathway and is unlikely to affect either Rac/JNK or Wnt/Ca2+ signaling. Because Dsh-DEP should block Dsh activities in both branches of the PCP and Wnt/Ca2+ pathways, we predicted that phenotypic characteristics of the FzTM1 and dnRhoA embryos would occur in Dsh-DEP embryos, but that Dsh-DEP embryos would exhibit a broader range of phenotypic anomalies because Dsh-DEP likely blocks pathways unaffected by FzTM1 or dnRhoA.

Our findings show that overexpressing Dsh-DEP by mRNA injection blocks archenteron invagination, early endodermal specification, and skeletogenesis in sea urchins, but has little effect on secondary mesenchyme cells. Thus, the hypothesis that PCP signaling directly or indirectly participates in the endoderm pathway and in spiculogenesis is supported. The activities of Dsh-DEP logically fit with the other known activities of a noncanonical Wnt pathway in the macromere progeny (Croce et al.,2006a; Beane et al.,2006; Beane,2007) and it is clear that Dsh-DEP overexpression does not influence signaling of the micromeres to the macromeres. Targets of early endomesoderm specification impacted by noncanonical Wnt signaling have also been identified. Cumulatively, these findings further highlight the importance of Dsh during early development of sea urchin embryos as well as the multiple roles of Dsh in directly or indirectly regulating gene expression through noncanonical Wnt pathways.


Dsh-DEP Blocks Archenteron Invagination and Spiculogenesis, but Not Formation of Primary Mesenchyme and Pigment Cells

Dsh-DEP, a deletion construct comprised of the portions of Dsh C-terminal to the PDZ domain, was generated to study noncanonical Wnt signaling in sea urchin embryos (Fig. 1). As hypothesized, overexpressing this construct by injecting Dsh-DEP mRNA into the fertilized egg produced embryos that exhibited characteristics typical of both FzTM1 and dnRhoA embryos. Consistent with expectations that Dsh-DEP would have more comprehensive effects than FzTM1 or dnRhoA, we found that the Dsh-DEP construct caused a broader range of phenotypic effects than either FzTM1 or dnRhoA (details are discussed later in this study). Cumulatively, our results support the hypothesis that Fz 5/8 and RhoA signal through a noncanonical Wnt pathway, most likely the PCP pathway.

Figure 1.

The Dsh construct LvDsh-GFP and the deletion construct LvDsh-DEP are shown.

The phenotypic effects of overexpressing Dsh-DEP by mRNA injection into fertilized eggs of L. variegatus, L. pictus, or S. purpuratus were similar for each species (Figs. 2A–D and 4A–C for S. purpuratus; Figs. 3A, 6, and 8 for L. variegatus; Figs. 6C and 8C for L. pictus). Early development was comparable to that of controls before the swimming blastula and early mesenchyme blastula stages. By the time controls reached the swimming blastula stage, both control and Dsh-DEP embryos clearly produced cilia, however most Dsh-DEP embryos failed to swim. Instead, they tended to spin on the bottom of the dish, possibly due to defects in ciliary coordination. Of interest, a recent study showed that Dsh was required for the apical positioning of basal bodies during vertebrate mucociliary epithelial development (Park et al.,2008). These authors further showed that Dsh was required for the planar polarization mediating the coordinated beating of the cilia in this epithelium. It is currently unknown if similar ciliary defects are seen in Dsh-DEP sea urchin embryos, but it is an intriguing possibility that will be investigated in future studies. Also, in mesenchyme blastulae injected with Dsh-DEP mRNA, the blastula wall was thicker than in controls and embryos were smaller. Embryos failed to elongate and, in many, mesenchyme-like cells moved into the blastocoel. During the stage when normal embryos gastrulate, embryos injected with high concentrations of Dsh-DEP mRNA failed to invaginate or form a blastopore. Instead, these embryos remained rounded and the thickened epithelium became more squamous. Occasionally the embryonic epithelium remained thick at one pole, but in most embryos the blastula wall was uniformly thin. When controls reached the pluteus stage, some Dsh-DEP embryos produced small spicules, but spicule formation was greatly reduced. These embryos continued to be rounded and the spicules formed failed to extend or did not contribute to arm formation. Pigment cells were generally present at this stage (Fig. 2D). As expected, overexpression of Dsh-DEP mRNA produced a phenotype distinct from that seen when canonical Wnt signaling is blocked (for example, by overexpressing the Dsh-DIX deletion construct; Fig. 3). This further supports the hypothesis that overexpressing the Dsh-DEP construct blocks noncanonical Wnt pathways.

Figure 2.

A–H: Phenotypes produced by Dsh-DEP overexpression in Strongylocentrotus purpuratus at 21 (A), 31 (B), 43 (C), and 72 (D) hours postfertilization (hpf) compared with normal development (E–H). Note that Dsh-DEP mRNA-injected embryos form mesenchyme-like cells, but fail to form a gut. Spicule formation is also reduced or absent in most. Also, during early development, the blastula wall is uniformly thickened in Dsh-DEP embryos. Similar phenotypes were also observed in Lytechinus variegatus and Lytechinus pictus. Scale bars = 100 μm.

Figure 4.

Range of phenotypes observed in Dsh-DEP mRNA-injected embryos. A: Some embryos continue to form a gut (S. purpuratus, 43 hours postfertilization [hpf]), but this percentage drops as Dsh-DEP mRNA concentration increases. These embryos often fail to form arms and produce fewer spicules. B: At high concentrations, embryos lack a gut although they continue to produce mesenchyme-like cells (S. purpuratus, 48 hpf). Spicule formation is reduced or absent. C: At mid to high concentrations, many embryos form a severe Dsh-DEP phenotype in which the embryo is hollow but contains numerous mesenchyme-like cells (S. purpuratus, 48 hpf). This could be due, in part, to RNA toxicity, but embryos injected with comparable picomolar concentrations of other mRNAs showed lower levels of this phenotype. D: Graphical representation of a single injection session showing a typical distribution of these phenotypes. Dsh-DEP mRNA concentrations are given in pg/pl. Numbers in parentheses represent the total number of embryos counted for each treatment.

Figure 3.

Comparison of phenotypes in Lytechinus variegatusDsh-DEP, and Dsh-DIX mRNA-injected embryos. A:Dsh-DEP mRNA-injected embryo after 2 days. B:Dsh-DIX mRNA-injected embryo after 3 days. The Dsh-DIX construct is the deletion construct that was previously shown to block nuclear localization of β-catenin, thereby blocking canonical Wnt signaling (Weitzel et al.,2004). The Dsh-DIX mRNA-injected embryo is severely animalized and has lost all endomesodermal structures including pigment, skeletogenic, and mesenchymal cells. C: Uninjected control, 3 days.

Figure 6.

Spatial expression of ectodermal and mesodermal markers in control (Cont) and Dsh-DEP (DEP) mRNA-injected embryos. A,C: Brightfield images. B,D: Fluorescent images. A–D: Expression of the oral ectodermal marker EctoV was similar in both control (A,B) and Dsh-DEP mRNA-injected (C,D) embryos (L. pictus, 53 hpf). E–H:Alx1, a marker of primary mesenchyme cells (PMCs). I–L:Delta, responsible for specification of secondary mesenchyme, were also similarly expressed in Dsh-DEP mRNA-injected and control embryos both at 5 hpf (120–256 cell stage; E,F,I,J) and the mesenchyme blastula stages (G,H,K,L). M–R:GCM, a marker of pigment cells, was expressed in both Dsh-DEP mRNA-injected and control prism embryos (Q,R) as well as early gastrula embryos (O,P), but expression of GCM appears to be delayed in Dsh-DEP expressing embryos (N) at the mesenchyme blastula stage as compared to controls (M). Also, ingression of pigment cells is clearly delayed in early gastrulae overexpressing Dsh-DEP (P) compared with controls (O). Arrows in figures point to cells expressing the mesodermal marker. 5 hpf, 120–256 cell stage; EG, early gastrula; MB, mesenchyme blastula; Pr, prism stage. Alx1, Delta, and GCM markers were detected in L. variegatus by whole-mount in situ hybridization (WMISH) of L. variegatus embryos. EctoV was detected by immunofluorescence.

Figure 8.

Spatial expression of endodermal markers in control (Cont) and Dsh-DEP (DEP) mRNA-injected embryos. A–D: Brightfield images (A,C) and fluorescence images (B,D) of embryos immunostained for the Endo1 protein. Endo1 was present in control prism stage larvae (B), but not in embryos overexpressing Dsh-DEP (D). E–Z:Blimp1 (E–J), GataE (K–N), Brachyury (O–T), and Endo16 (U–Z) mRNAs were present in control embryos, but absent or reduced in Dsh-DEP embryos. MB, mesenchyme blastula; 5 hpf, 120- to 256-cell stage; EG, early gastrula; LG, late gastrula; Pr, prism stage. Blimp1, GataE, Brachyury, and Endo16 markers were detected by whole-mount in situ hybridization (WMISH) of L. variegatus embryos. Endo1 was detected by immunofluorescence in L. pictus.

Embryos overexpressing lower Dsh-DEP mRNA concentrations (less than 1.8 pg/pl) often produced a gut but remained spherical when controls had reached the prism or early pluteus stage (Fig. 4A). As concentrations of injected Dsh-DEP mRNA increased, embryos were less likely to form a gut (Fig. 4B) and at highest concentrations a “severe phenotype” was observed in which the blastocoel became occluded with mesenchyme-like cells (Fig. 4C). This phenotype was also observed at mid-level concentrations, but much less frequently. Although the severe phenotype is similar to that caused by mRNA toxicity, injecting comparable picomolar concentrations of a control mRNA (Green Lantern, GL) was much less likely to cause this result. In S. purpuratus embryos injected with a 5.7–6.5 μM concentration of either Dsh-DEP or GL mRNA (a concentration equivalent to a 1.8–2.1 pg/pl concentration of Dsh-DEP), on average, 16% of the Dsh-DEP mRNA-injected embryos exhibited a severe phenotype, whereas only 6% of the GL mRNA-injected embryos, on average, exhibited this phenotype. This indicates that the severe Dsh-DEP phenotype is specific to Dsh-DEP overexpression and that it is unlikely to be caused by mRNA toxicity. Because frequencies of the severe phenotype rose at highest Dsh-DEP mRNA concentrations (e.g., 2.6 pg/pl), we subsequently used Dsh-DEP mRNA concentrations of 2.1 pg/pl or less to minimize the possibility of artifacts due to RNA toxicity.

Although the phenotypes described were most common, considerable variation existed within a single batch of embryos (Fig, 4D). To minimize this variation, embryos were co-injected with rhodamine dye, and embryos with similar fluorescence intensity compared. Although fluorescent intensities were similar, Dsh-DEP embryos co-injected with rhodamine continued to exhibit the different phenotypes, indicating a variable response to very small differences in Dsh-DEP concentrations.

Dsh-DEP Acts in Descendents of Macromeres and/or Mesomeres but Not Micromeres

To test which cell populations are affected by Dsh-DEP overexpression, cell transplantation experiments were performed on 16-cell Lytechinus variegatus embryos. At this stage, a signal originating in the micromeres induces endomesoderm formation in the macromeres. Also the micromeres are specified to become primary mesenchyme cells (PMCs), cells that later produce the skeletal structures. To determine whether Dsh-DEP blocks signals from the micromeres, micromeres were removed from a Dsh-DEP mRNA-injected embryo and transplanted to an uninjected micromereless embryo (Fig. 5A). If Dsh-DEP blocked signal transduction in the micromeres and their descendents, then these chimeras should have impaired endomesoderm specification. As a result, they would produce embryos with the Dsh-DEP skeletal phenotype, that is, a reduced skeleton or no skeleton. In addition, these chimeras would be expected to show defects in endoderm formation if Dsh-DEP interfered with the endomesoderm-inducing signals that originate in the micromeres. Instead, in 94% of the chimeric recombinants containing micromeres overexpressing Dsh-DEP mRNA, skeletogenesis was normal (Fig. 5A; Table 1). Observed rates of skeletogenesis in these chimeras were similar to those in uninjected controls whereas only 3% of Dsh-DEP mRNA-injected controls formed a normal skeleton. Therefore, skeletogenic defects in Dsh-DEP embryos are not due to the Dsh-DEP construct altering Dsh-dependent skeletogenic functions in the micromeres. Furthermore, because 94% of the chimeras produced endoderm (vs. 91% of uninjected controls and 38% of the Dsh-DEP mRNA-injected controls), the endoderm-inducing signals from the micromeres were also unaffected by Dsh-DEP overexpression.

Figure 5.

Dsh-DEP blocks cell fate specification in the macromeres and/or mesomeres, but not in the micromeres. A: Dsh-DEP micromeres recombined with control micromereless embryos developed normally (1 day postfertilization [dpf]). B: Control micromeres recombined with Dsh-DEP micromereless embryos (micromeres removed at 16-cell stage) produced the Dsh-DEP phenotype (1 day pf).

Table 1. Summary of Cell Transplantation Experiments
Phenotypes:Skeletogenesis NormalEndoderm Present
Chimeric Embryos  
Dsh-DEP micromeres/Uninjected Micromereless (n = 16)15 (94%)15 (94%)
Uninjected micromeres/Dsh-DEP Micromereless (n = 16)0 (0%)5 (31%)
Whole Embryos  
Uninjected Controls (n = 42)37 (88%)38 (91%)
Dsh-DEP mRNA-Injected Controls (n = 37)1 (3%)14 (38%)

In the reciprocal experiment, normal micromeres were transplanted onto a Dsh-DEP mRNA-injected micromereless embryo (Fig. 5B; Table 1). In this case, 94% of the chimeras produced an intermediate to strong Dsh-DEP phenotype (compared with 5% of uninjected controls and 95% of the Dsh-DEP mRNA-injected controls). None of the chimeras with uninjected micromeres exhibited normal skeleton formation and endoderm (a complete gut or a partial gut) was present at levels similar to those observed in Dsh-DEP mRNA-injected whole embryos. These results indicate that the signal blocked by Dsh-DEP occurs not in the micromeres, but in the macromeres and/or mesomeres, and that overexpressing Dsh-DEP blocks endoderm formation and skeletogenesis by modifying signal transduction responses in the macromeres and/or mesomeres.

Ectoderm and Early Mesoderm Development Proceeds in Embryos With Compromised Noncanonical Wnt Signaling

To characterize the impact of overexpressing Dsh-DEP on cell fate specification and differentiation, injected embryos were analyzed using antibody staining, whole-mount in situ hybridization (WMISH), and quantitative polymerase chain reaction (QPCR). In all cases, WMISH was performed in L. variegatus and QPCR data were collected from S. purpuratus.

Ectodermal patterning was evaluated using the monoclonal antibody EctoV, a marker of the oral ectoderm and foregut in plutei (Coffman and McClay,1990). We found that EctoV expression was similar in both control plutei and Dsh-DEP embryos at the same stage of development (L. pictus; Fig. 6B,D). Comparable distributions of EctoV were obtained in L. variegatus and S. purpuratus (not shown). In each case, the EctoV marker was present throughout the oral surface in control embryos, and in Dsh-DEP embryos the marker covered approximately half of the cells. This suggests that oral/aboral polarity was not disrupted in the Dsh-DEP embryos.

In the sea urchin, two main cell types form mesoderm: the skeletogenic primary mesenchyme cells (PMCs) and the nonskeletogenic secondary mesenchyme cells (SMCs). A key gene acting in specification of the skeletogenic cells is the transcription factor pmar1. By repressing HesC, a second repressor, pmar1 activates specification of the micromeres to a) form skeletogenic cells and b) produce a signal (the early signal) that induces macromeres to form endomesoderm (Olivieri et al.,2002; Revilla-i-Domingo et al.,2007). Using QPCR, levels of pmar1 expression were evaluated in 120–256 cell Dsh-DEP and control embryos (Fig. 7). This comparison revealed that levels of pmar1 were much higher in Dsh-DEP embryos than in controls.

Figure 7.

Quantitative polymerase chain reaction (QPCR) analysis of endomesoderm gene expression in Dsh-DEP mRNA-injected S. purpuratus embryos at the 120- to 256-cell, mesenchyme blastula, and prism stages. In all cases, samples were normalized to Ubiquitin and compared with GFP mRNA-injected controls. Light gray indicates genes associated with endoderm formation and dark gray indicates genes associated with mesoderm formation. Numbers in parentheses correspond to number of QPCR trials. For each trial, the sample was run in triplicate. Horizontal lines on the graph indicate the ΔΔCT levels considered significant.

Although levels of pmar1 are unusually high in Dsh-DEP embryos, gene expression downstream of pmar1 is unperturbed. For example, spatial expression of Alx1, a gene that specifies large micromeres to become PMCs and a known marker of large micromeres and PMCs (Ettensohn et al.,2003), was detected using WMISH (Fig. 6E–H). When expression of Alx1 was examined in 120–256 cell embryos and mesenchyme blastulae, each Dsh-DEP mRNA-injected stage had levels of expression similar to the corresponding control. Levels of Alx1 expression, detected using QPCR, were also similar in control and Dsh-DEP mRNA-injected mesenchyme blastulae (Fig. 7). Thus, although pmar1 levels were elevated in Dsh-DEP embryos, this did not interfere with specification of micromeres by Alx1. Both control and Dsh-DEP mRNA-injected embryos produced similar levels of Alx1, confirming that early micromere specification can proceed without noncanonical Wnt signaling. The reduced formation of spicules observed later in Dsh-DEP embryos is likely not due to disruption of PMC specification, but is hypothesized to result due to perturbation of later differentiation events. Signals from the ectoderm and endoderm are also necessary for skeletogenesis and it is hypothesized that these later signals are disrupted by the overexpression of Dsh-DEP, not early signaling in the micromeres (Armstrong et al.,1993; Benink et al.,1997; Duloquin et al.,2007; Röttinger et al.,2008).

Another gene indirectly regulated by pmar1 is Delta, which codes for a ligand of the Notch receptor. Notch-Delta signaling is known to initiate secondary mesenchyme specification in the sea urchin (Sweet et al.,2002). The expression of Delta, evaluated by WMISH in 120–256 cell embryos and mesenchyme blastulae (Fig. 6I–L), was comparable for corresponding stages of Dsh-DEP mRNA-injected embryos and controls. Because ectopic expression of pmar1 in S. purpuratus causes ectopic expression of Delta (Oliveri and Davidson,2004), we expected that Delta would be misexpressed if the observed increased pmar1 levels were due to ectopic expression. Because spatial expression of Delta, was similar in Dsh-DEP embryos and controls, it is likely that the observed increases in pmar1 expression are not due to ectopic expression of this molecule, but to higher levels of pmar1 in the micromeres.

Finally, expression of Gcm another regulator of secondary mesenchyme specification (pigment cell specification; Ransick et al.,2002; Ransick and Davidson,2006), was assessed by WMISH. Expression of Gcm was compared in Dsh-DEP and control embryos at the mesenchyme blastula, early gastrula, and prism stages (Fig. 6M–R). Although Gcm expression was delayed in Dsh-DEP embryos and the timing of secondary mesenchyme cell morphogenesis also lagged in Dsh-DEP embryos (few migrating blastocoelar cells or pigment cells were detected in early gastrulae), Gcm expression was obvious in both the early gastrula and prism stages. These results were also confirmed by QPCR (Fig. 7). Mesenchyme blastulae containing injected Dsh-DEP mRNA (24 hours postfertilization [hpf]) expressed lower levels of Gcm than controls, but later, at the prism stage (48 hpf), Gcm expression did not differ noticeably in Dsh-DEP embryos and controls. Based on these results, overexpressing Dsh-DEP delays formation of some Gcm-dependent forms of secondary mesenchyme, but does not completely inhibit it.

We conclude that early specification of many ectodermal and mesodermal cells proceeds almost normally in Dsh-DEP embryos. Based on distribution of the EctoV marker, Dsh-DEP embryos produce a subpopulation of cells expressing oral ectoderm as well as a subpopulation lacking oral ectoderm (presumably the aboral ectodermal cells). Also, despite elevated pmar1 levels, early specification of the PMCs occurs in the typical micromere territory and, based on expression of genes downstream of pmar1, early specification of the micromeres is unperturbed. Although expression of Gcm is delayed, embryos recover by early gastrulation to express normal levels of Gcm and produce pigment cells.

Endodermal Gene Expression Is Reduced in Embryos With Compromised Noncanonical Wnt Signaling

Because Dsh-DEP RNA injected embryos failed to form guts, expression of genes acting in endoderm specification was evaluated. Expression of Wnt8, Blimp1 (formerly named Krox), GataE, Brachyury (Bra), and Endo16 was determined using WMISH (L. variegatus; Fig. 8) and/or QPCR (S. purpuratus; Fig. 7) and expression of Endo1 was assayed using immunohistochemistry (Fig. 8A–D). Wnt8, a key molecule that regulates formation of the endomesoderm downstream of β-catenin, is initially expressed in the 16-cell stage micromeres and subsequently in the micromere descendents and macromeres beginning at the 32- to 60-cell stage (Wikramanayake et al.,2004; Minokawa et al.,2005). At the 60-cell stage, the veg1 and veg2 tiers are produced when the macromeres undergo equatorial cleavage. The veg1 layer lies closer to the animal pole, whereas the veg2 layer is more vegetal. By the mesenchyme blastula stage, Wnt8 expression is localized to the veg2 endoderm and the veg1 cells. When Dsh-DEP mRNA-injected embryos were compared with controls at the mesenchyme blastula stage, levels of Wnt8 mRNA were similar as assayed by QPCR (Fig. 7). This suggests that Wnt8, one of the earliest endomesodermal genes in the sea urchin endomesodermal network, is not affected by overexpression of Dsh-DEP and that Dsh-DEP may act on a pathway downstream or parallel to Wnt8.

Blimp1 is an early transcriptional regulator required for endomesodermal specification, and later for endoderm specification (Livi and Davidson,2007). WMISH showed that 120–256 cell and mesenchyme blastula controls expressed Blimp1 in a ring of cells (Fig. 8E,G) that later become the midgut and hindgut of the gastrula (not shown). Blimp1 expression was reduced or absent in the Dsh-DEP mRNA-injected embryos at equivalent stages (Fig. 8F,H). In prism stage controls, Blimp1 expression was high in the midgut and hindgut (Fig. 8I), but in the Dsh-DEP mRNA-injected embryos Blimp1 expression was often restricted to a patch of cells in the blastula wall (Fig. 8J) or Blimp1 expression was lost. The distribution of Blimp1 in Dsh-DEP mRNA-injected embryos is somewhat surprising given the findings of Smith et al. (2007). Smith concluded that Wnt8, signaling through nuclear β-catenin, drives Blimp1 expression. If canonical Wnt signaling was driving expression of Blimp1, one would not expect overexpression of Dsh-DEP, a construct that blocks noncanonical Wnt signaling, to reduce transcription of Blimp1. Thus, an important noncanonical Wnt signal may be present early in endoderm specification because Blimp1 is an early gene in that network. Another possibility is that noncanonical Wnt signaling is required for maintenance of Blimp1 transcription.

A second transcription factor important in endoderm specification is GataE, a zinc finger transcriptional regulator of several endodermal and mesodermal genes (Hinman and Davidson,2003a; Lee and Davidson,2004). GataE is expressed early in endomesoderm specification and later becomes confined to endodermal cells. During gastrulation, GataE is concentrated at the archenteron tip and in endoderm precursors surrounding the blastopore (Fig. 8K). By the late prism stage, it becomes restricted to the coelomic pouches and midgut (Fig. 8M). Unlike the controls, Dsh-DEP overexpressing embryos at the mesenchyme blastula (not shown) and late gastrula stages (Fig. 8L) either had reduced levels of GataE expression or lacked GataE expression altogether. By the prism stage, Dsh-DEP embryos only expressed a few patches of GataE (Fig. 8N) and again expression was greatly reduced compared with that observed in controls. GataE expression was not examined in embryos before the mesenchyme blastula stage.

We also examined the expression of Brachyury (Bra), another important transcription factor in endoderm specification. Gross and McClay (2001) hypothesized that this transcription factor regulates endoderm morphogenesis, acting downstream of GataE and several other endomesodermal genes. WMISH showed that in control mesenchyme blastulae, Bra is expressed in cells fated to be endoderm (Fig. 8O). By the late gastrula stage, it is expressed around the blastopore and in cells that will become the stomodeum (Fig. 8Q). At both stages, Dsh-DEP overexpressing embryos often expressed Bra, but levels of expression were reduced compared with the controls (Fig. 8P,R). Also, Bra was restricted to the vegetal region of Dsh-DEP overexpressing late gastrulae while stomodeal staining was less obvious. In prism stage controls, Bra is expressed in the stomodeum and vegetally near the anus (Fig. 8S), but in embryos overexpressing Dsh-DEP mRNA, Bra expression was either absent or restricted to a small portion of the blastula wall (Fig. 8T). QPCR studies showed that levels of Bra expression in mesenchyme blastulae and prism stage embryos misexpressing Dsh-DEP were reduced compared with those in GFP mRNA-injected controls (Fig. 7). This difference was much greater at the mesenchyme blastula stage than the prism stage.

Two other genes were also used to evaluate endodermal specification and differentiation, Endo1 (immunostaining) and Endo16 (WMISH and QPCR). Endo1 is a glycoprotein that localizes to the apical and basolateral surfaces of endodermal cells. It is first expressed in the archenteron during invagination, and as the gut differentiates it becomes restricted to midgut and hindgut cells (Wessel and McClay,1985). Dsh-DEP mRNA-injected embryos and controls were stained at the prism stage with the Endo1 antibody (Fig. 8A–D). Although Endo1 was clearly present in controls, it was absent in Dsh-DEP mRNA-injected embryos. Expression of the other marker, Endo16, was also determined in L. variegatus mesenchyme blastulae, early gastrulae, and late gastrulae (Fig. 8U–Z) using WMISH. Endo16 accumulates in the veg2 tier at the mesenchyme blastula stage, is concentrated in the archenteron as gastrulation proceeds, and spreads into the veg1-derived endoderm (the hindgut and midgut) by the late gastrula stage (Nocente-McGrath et al.,1989; Ransick et al.,1993; Ransick and Davidson,1998). When controls had reached the mesenchyme blastula stage (Fig. 8U), Dsh-DEP mRNA-injected embryos at the same stage (Fig. 8V) exhibited little or no expression of Endo16. In early gastrulae and late gastrulae injected with Dsh-DEP mRNA (Fig. 8X,Z), Endo16 was present in a ring at the presumed vegetal pole, but expression was lower in the Dsh-DEP mRNA-injected embryos than in controls (Fig. 8W,Y). Levels of expression (determined by QPCR) were somewhat reduced in prism stage Dsh-DEP mRNA-injected embryos when compared with GFP mRNA-injected controls (Fig. 7).

Cumulatively, these results provide strong evidence that Dsh-DEP mRNA-injected embryos fail to express key endodermal genes and suggest that blocking noncanonical Wnt signaling in the sea urchin embryo interferes with early endoderm specification.

Dsh-DEP Embryos Are Not Rescued by Overexpression of Actβ-Catenin

At least three different pathways use Dsh in cell signaling: the canonical Wnt pathway (Klingensmith et al.,1994), the PCP pathway (Axelrod et al.,1998), and the Wnt/Ca2+ pathway (Sheldahl et al.,2003). Based on the phenotype of the Dsh-DEP embryo, it seemed unlikely that Dsh-DEP merely blocks the canonical Wnt pathway. Embryos in which canonical Wnt signaling has been blocked by overexpression of (A) delta cadherin (Wikramanayake et al.,1998; Logan et al.,1999), (B) GSK-3β (Emily-Fenouil et al.,1998), (C) dominant negative TCF/LEF (Huang et al.,2000; Vonica et al.,2000), or (D) Dsh-DIX (Weitzel et al.,2004; Fig. 3B) are phenotypically distinct from Dsh-DEP embryos.

To further demonstrate that Dsh-DEP does not block Dsh upstream of β-catenin in the canonical Wnt pathway, Dsh-DEP mRNA was coinjected with a stable form of Xenopus β-catenin (actβ-cat) mRNA (Yost et al.,1996; Fig. 9C,G). If Dsh-DEP only blocked the canonical Wnt pathway, then activating β-catenin downstream of Dsh should either rescue endomesodermal expression or produce the excessive endomesoderm levels observed in embryos overexpressing either actβ-cat (Fig. 9B; Wikramanayake et al.,1998) or constitutively active TCF/LEF (Vonica et al.,2000). It is also noteworthy that embryos lacking canonical Wnt signaling due to the overexpression of cadherin can be rescued by coexpressing actβ-cat (Wikramanayake et al.,1998). When Dsh-DEP and actβ-cat mRNA were coexpressed (Fig. 9C), the actβ-cat construct failed to rescue endomesoderm formation, and phenotypes of the coinjected embryos were identical to those injected with Dsh-DEP mRNA alone. Thus, the Dsh-DEP phenotype cannot be rescued by activating the canonical Wnt pathway downstream of Dsh. This supports the conclusion that Dsh-DEP is not preferentially blocking the canonical Wnt signal upstream of β-catenin.

Figure 9.

Phenotypes produced in L. variegatus following Dsh-DEP/activated β-catenin mRNA coinjections. A–C: Live embryos, pluteus stage; Buffer-injected controls develop normally (A), embryos injected with activated β-catenin (actβcat) mRNA are severely vegetalized and primarily form endoderm and mesoderm (B), and embryos coinjected with actβcat and Dsh-DEP mRNA exhibit the Dsh-DEP phenotype (C). D–F:GataE expression assayed by whole-mount in situ hybridization (WMISH), 20 hours postfertilization (hpf); control (D), embryos injected with actβcat mRNA (E), and embryos coinjected with actβcat and Dsh-DEP mRNA (F; note loss of GataE). G: Sea urchin embryos coinjected with Dsh-DEP and activated β-catenin mRNA exhibit the Dsh-DEP phenotype at rates comparable to Dsh-DEP mRNA-injected embryos. The pg/pl concentrations injected are indicated next to treatments, followed by the number of embryos observed (in parentheses). Embryos exhibited the Dsh-DEP phenotype (hollow with mesenchyme-like cells), were vegetalized (formed excessive amounts of endoderm and mesoderm at the expense of ectoderm), or developed in a manner similar to controls, forming either a partial gut (blind gut) or a whole gut.

To test whether endoderm formation was blocked in the actβ-cat/Dsh-DEP mRNA coinjected embryos, expression of two endodermal markers, GataE (Hinman and Davidson,2003a; Lee and Davidson,2004) and Endo16 (Romano and Wray,2003), was compared in four situations: (1) uninjected embryos, (2) actβ-cat mRNA-injected embryos, (3) Dsh-DEP mRNA-injected embryos, and (4) actβ-cat/Dsh-DEP mRNA coinjected embryos. Although GataE is clearly expressed in control embryos (Fig. 9D) and actβ-cat mRNA-injected embryos (Fig. 9E), expression of this gene was reduced or absent in both the Dsh-DEP embryos (not shown) and the actβ-cat/Dsh-DEP embryos (Fig. 9F). Similar results were obtained when the expression of Endo16 was examined in these embryos (not shown). Thus, endoderm formation is reduced in Dsh-DEP embryos and cannot be rescued by activating the canonical Wnt pathway downstream of Dsh. These results indicate that GataE and Endo16 are either (A) activated downstream of a noncanonical Wnt pathway or (B) maintenance of GataE and Endo16 gene expression depends on inputs from a noncanonical Wnt pathway that is disrupted by Dsh-DEP.


This study used Dsh-DEP, a truncated form of Dsh that has been shown to block noncanonical Wnt signaling through a dominant-negative mechanism, to disrupt noncanonical Wnt signaling in sea urchins. Overexpression of Dsh-DEP had striking effects on sea urchin development and produced embryos that lacked endoderm and failed to undergo skeletogenesis. Although skeletogenesis was inhibited in Dsh-DEP mRNA-injected embryos, early mesoderm specification occurred and other mesodermal derivatives, including pigment cells and mesenchyme-like cells developed relatively normally. Our results indicate that defects in embryo development induced by overexpressing Dsh-DEP, including the skeletal defects, were due in large to the inhibition of noncanonical Wnt signals in the macromere derivatives. These noncanonical Wnt signals appear to regulate the segregation of endoderm from nonskeletogenic mesoderm downstream of endomesoderm specification. Additionally, these signals non–cell-autonomously regulate the differentiation of the PMCs. These observations reveal important new insights into the role of noncanonical Wnt signaling in the development of the sea urchin embryo.

Dsh-DEP Blocks Noncanonical Wnt Signaling, but Not Canonical Wnt Signaling

Dsh is known to participate in both canonical and noncanonical Wnt signaling. Results from the current study clearly demonstrate that Dsh-DEP overexpression in sea urchins did not block the early canonical Wnt signal specifying endomesoderm. This conclusion is supported by four observations:

  • 1The Dsh-DEP phenotype differs from that of embryos lacking canonical Wnt signaling (Fig. 3). Blocking the canonical Wnt signal produces completely animalized embryos that fail to undergo both endomesoderm specification and ectoderm patterning (Emily-Fenouil et al.,1998; Wikramanayake et al.,1998; Logan et al.,1999; Huang et al.,2000; Vonica et al.,2000; Weitzel et al.,2004), whereas overexpressing Dsh-DEP does not appear to affect polarization of the embryo along the oral–aboral axis and permits formation of additional cell types including the pigment cells and mesenchyme cells.
  • 2The animalized phenotype in embryos with blocked canonical Wnt signaling can be rescued by overexpressing actβ-cat. Coexpressing Dsh-DEP with actβ-cat, however, fails to rescue endoderm formation, indicating that Dsh-DEP does not block Dsh in a β-catenin dependent Wnt pathway and that Dsh-DEP acts in a pathway downstream or parallel to the canonical Wnt pathway.
  • 3Early zygotic genes transcribed in response to canonical Wnt signaling such as pmar1 and Wnt8 are expressed in Dsh-DEP embryos. One would not expect this result if Dsh-DEP blocked canonical Wnt signaling.
  • 4In 16-cell stage embryos, overexpressing Dsh-DEP in the micromeres did not affect endomesoderm formation in uninjected macromeres and the resulting embryos developed normally. Because the micromere inductive signals depend on a canonical Wnt signal, this experiment supports the conclusion that Dsh-DEP does not interfere with early canonical Wnt signaling.

Clearly, Dsh-DEP has no significant effect on canonical Wnt signaling. Instead, as in other organisms, this construct likely interferes with noncanonical Wnt pathways. As mentioned earlier, Croce et al. (2006a) hypothesized that a dominant-negative form of Frizzled 5/8 (FzTM1) inhibited PCP signaling in L. variegatus and Paracentrotus lividus. If this hypothesis is correct, then Dsh-DEP embryos should exhibit characteristics typical of the FzTM1 phenotype. Both FzTM1 and Dsh-DEP embryos lacked endoderm and an organized skeleton, but produced primary mesenchyme-like cells and pigment cells. In both FzTM1 and Dsh-DEP embryos, expression of ectodermal and mesodermal markers was similar to that in control embryos (with the exception of early GCM expression in the Dsh-DEP embryos). Also both FzTM1 and Dsh-DEP mRNA-injected embryos, exhibited reduced expression of endodermal markers relative to levels observed in the control embryos. Despite these similarities, the FzTM1 and Dsh-DEP phenotypes differed. Dsh-DEP embryos failed to express genes acting in early endoderm specification, whereas FzTM1 embryos had normal endodermal specification until the mesenchyme blastula stage. This difference is not surprising. Blocking Dsh activity with Dsh-DEP should have more comprehensive effects than blocking Fz 5/8 alone because Dsh should be able to interact with other Fz gene products (Croce et al.,2006b). Thus, our findings support the hypothesis that Dsh-DEP affects a noncanonical Wnt pathway acting earlier in endoderm specification than Fz 5/8. Because overexpressing Dsh-DEP caused defects specific to endoderm formation we hypothesize that this noncanonical Wnt signal regulates endoderm development downstream of the initial endomesoderm specification by β-catenin dependent Wnt signaling.

Phenotypically, Dsh-DEP mRNA-injected sea urchin embryos resemble those lacking RhoA signaling (dnRhoA). Because RhoA is a key effector of PCP signaling downstream of Dsh in other bilaterians, Beane et al. (2006) hypothesized that blocking RhoA activity during early sea urchin development would cause the same effects as inhibiting PCP pathway activity. We found that overexpressing Dsh-DEP caused several of the same responses observed when RhoA activity was blocked. In both cases, embryos lacked an archenteron, but produced mesenchyme-like cells. Also, expression of the endodermal genes GataE and Endo16 (Beane,2007) decreased in both, but Bra levels, although decreased in Dsh-DEP and FzTM1 (Croce et al.,2006a) embryos, were normal in dnRhoA embryos (Beane et al.,2006). Thus, regulation of Bra does not depend on the RhoA link of the noncanonical Wnt pathway. Another difference between Dsh-DEP and dnRhoA embryos is that GataE was strongly down-regulated in Dsh-DEP overexpressing embryos at the mesenchyme blastula stage and later, but less so, in dnRhoA overexpressing embryos. Again, Dsh-DEP blocks numerous activities including the subset of targets perturbed by dnRhoA. These observations further support our conclusion that signaling blocked by Dsh-DEP includes the RhoA-regulated PCP pathways as well as other noncanonical Wnt pathways.

Although overexpressing activated RhoA (actRhoA) rescued FzTM1 embryos, actRhoA did not rescue Dsh-DEP embryos (not shown). This is further evidence that Dsh-DEP blocks pathways in addition to the RhoA portion of the PCP pathway. Perhaps Fz 5/8 selectively signals through the Rho/ROCK portion of the PCP pathway, but Dsh-DEP blocks an additional noncanonical Wnt pathway such as the Wnt/Ca2+ pathway. Sheldahl et al. (2003) showed that overexpressing Dsh-DEP in Xenopus blocked activities of the Wnt/Ca2+ pathway. Also, Croce et al. (2006a) found that blocking Wnt/Ca2+ signaling with a PKC inhibitor impaired skeletogenesis but allowed formation of the archenteron and PMC ingression. When we chemically blocked Wnt/Ca2+ pathway components (unpublished data), we found that, depending on the inhibitor, endoderm formation or skeletogenesis were impaired. Because blocking the Wnt/Ca2+ signal produces a subset of the effects observed in Dsh-DEP embryos, these results indicate that Dsh-DEP overexpression may block both noncanonical Wnt pathways in the sea urchin.

Portions of the Sea Urchin Endomesoderm Network Affected by Dsh-DEP

The Sea Urchin Endomesoderm Gene Regulatory Network (GRN) ( (Davidson et al.,2002a,b; Oliveri and Davidson,2004; Levine and Davidson,2005) is an invaluable tool for understanding how molecular interactions are modified in Dsh-DEP overexpressing embryos. The earliest molecule perturbed in this study is pmar1. Within the PMC network, pmar1 is a key factor in micromere specification, acting as an upstream regulator of the micromere signal that induces endomesoderm in the macromeres (Oliveri et al.,2002,2003). As we noted in the results section, although pmar1 expression is elevated in Dsh-DEP embryos, spatial expression of pmar1 is expected to be similar to that in control embryos. One might suspect that increasing pmar1 expression in these cells would affect downstream targets, but mesenchyme blastulae overexpressing Dsh-DEP displayed normal expression of both Delta and Alx1, two molecules downstream of pmar1. Examination of the PMC network explains why this occurs. pmar1 affects micromere specification by repressing HesC, which in turn represses Delta, Alx1, and other PMC genes (Revilla-i-Domingo et al.,2007). If control embryos normally produce sufficient amounts of pmar1 to repress HesC activity, one would not expect increased expression of pmar1 in the Dsh-DEP embryos to affect Delta and Alx1 because pmar1 cannot further suppress HesC. Thus, the increased levels of pmar1 expression have little to no effect on early specification of PMCs.

In addition, analysis of endodermal gene expression in Dsh-DEP embryos by QPCR and WMISH revealed decreased expression of Blimp1, Bra, GataE, and Endo16, but normal expression of Wnt8. Within the GRN, Otx is upstream of all of these genes. This is supported by the QPCR data appearing on the Sea Urchin Endomesodermal website which indicate that Blimp1, Bra, GataE, and Endo16 expression all drop when transcriptional activity of Otx is decreased by overexpressing a form of Otx fused to the Engrailed repressor domain (Davidson et al.,2002a,b; We hypothesize that overexpressing Dsh-DEP perturbs the Endoderm network in mesenchyme blastulae and gastrulae by disrupting portions of the network downstream of Otx. One possibility is that Dsh-DEP blocks a pathway that would normally influence production of a zygotic form of Otx (Otxα or Otxβ). The disrupted Otx will then be unable to act in transcription of Blimp1, Bra, GataE, or Endo16. It is interesting that embryos overexpressing Engrailed-SpOtx (Li et al.,1999) are superficially similar to the Dsh-DEP overexpressing embryos. These embryos produce mesenchyme-like cells and pigment cells, but lack endoderm. They differ from the Dsh-DEP overexpressing embryos in that expression of the oral marker EctoV is distributed throughout the embryo, but expression of aboral markers (CyIIIa, Spec1, Spec2a) is reduced. Also, as in the Dsh-DEP overexpressing embryos, Engrailed-SpOtx embryos cannot be rescued by coinjection of activated β-catenin. If Dsh-DEP affects only one of the two zygotic forms of Otx (Otxα or Otxβ), one might expect Dsh-DEP overexpressing embryos to be different from embryos produced by overexpressing Engrailed-SpOtx, in which both Otxα and Otxβ dependent transcription is affected. These inhibitory effects of Dsh-DEP on Otx targets suggest that a noncanonical Wnt signaling pathway may block Otx early in development. This was not predicted based on the Fz5/8 and RhoA work. In those studies, blocking either RhoA or Fz5/8 did not disrupt early expression of the Otx target Bra in the mesenchyme blastula (although it does disrupt Bra expression later in the gastrula stage of Fz5/8). Overexpression of the Dsh-DEP construct, however, decreased expression of Bra in the mesenchyme blastula. This suggests that Dsh-DEP overexpression interferes with signaling of Otx at the mesenchyme blastula stage and later and that the pathway inhibited by Dsh-DEP may have functions independent of the RhoA and Fz5/8-dependent PCP pathways. Because GataE is known to act in formation of pigment cells earlier in development, it may be possible that this pathway is not disrupted before the mesenchyme blastula stage. Further testing will be needed to determine whether this is the case.

In conclusion, we have shown that Dsh-DEP reduces or prevents expression of key genes necessary for endoderm formation including Blimp1, GataE, Bra, and Endo16. Comparison of the Dsh-DEP phenotype and gene expression patterns to those of sea urchin embryos lacking either Fz 5/8 or RhoA activity supports the hypothesis that Fz 5/8 and RhoA both signal through the PCP pathway. In addition, characteristics unique to the Dsh-DEP embryos suggest that this construct also blocks signaling through an additional earlier noncanonical Wnt pathway. Future use of this construct in conjunction with other methods should be valuable to those wishing to learn more about the range of Dsh-dependent pathways influencing both skeletogenesis and endoderm formation during sea urchin development.


Care of Animals/Embryos

Strongylocentrotus purpuratus, Lytechinus variegatus, and Lytechinus pictus were used in the experiments. S. purpuratus were obtained from Charles Hollahan (Santa Barbara, CA) or Marinus Scientific (Long Beach, CA); L. pictus came from Marinus Scientific; and L. variegatus were imported from Duke Marine Lab (Beaufort, NC) or Sea Life, Inc. (Tavernier, FL). S. purpuratus and L. pictus were kept in refrigerated seawater aquaria systems at 12–15°C, whereas L. variegatus was maintained at 22°C.

To induce spawning, 1 ml of 0.5 M potassium chloride was injected into the adult urchin up to 3 times. Eggs were collected in seawater and stored at room temperature or mildly chilled on ice (depending on the species). Sperm were collected dry and kept on ice or at 4°C. Sperm were diluted 1:1,000 in seawater and 50 μl was added to approximately 10 ml of artificial seawater containing eggs. Embryos of S. purpuratus and L. pictus were raised at 15°C, and L. variegatus embryos were raised at 18–23°C.

Subcloning and Plasmid Construction

L. variegatus Dsh cDNA (Genbank AY624074) was used as a template to PCR-amplify the LvDsh-DEP (subsequently referred to as Dsh-DEP) coding region. The oligonucleotide primers used to amplify Dsh-DEP were 5′-cgcggatccatgGCTAAATGTTGGGA- CCCT TC-3′ and 5′-ATCACCTTCACCCTCTCCAC-3′. In the first primer, the underlined bases indicate restriction enzyme sites that were used for subcloning of the gene fragments into the pCS2+ vector. The amplified Dsh-DEP fragment was digested with BamH1 and subcloned into the corresponding sites in the pCS2+ vector. All PCR reactions were performed using Vnt Polymerase (NEB, Beverly, MA). Restriction digests were performed to determine whether size and orientation of constructs was correct and constructs were sequenced to ensure fidelity of the PCR reactions.


Plasmid constructs were linearized and RNA was synthesized using the mMessage mMachine kit (Ambion, Austin, TX). RNA was isolated by phenol-chloroform extraction, followed by quick spin column purification (Roche, Indianapolis, IN) and isopropanol precipitation. RNA was diluted in 25–40% glycerol for injection.

Spawned eggs were washed several times in artificial seawater and were then treated with acidic seawater (pH 5.0) or put through Nytex mesh to remove the jelly coat. Using a mouth pipette, eggs were transferred to a 1% protamine sulfate coated dish and fertilized in artificial seawater containing 3-amino 1,2,4-triazole to prevent hardening of the fertilization envelope. Equivalent picomolar concentrations of RNA were injected into experimental and control embryos when comparisons were done. Concentrations used are as follows: Dsh-DEP (1.0–2.6 pg/pl), Dsh-DIX (1.5–2.0 pg/pl; Weitzel et al.,2004), actβ-cat (0.01 pg/pl; Yost et al.,1996), actRhoA (0.05 pg/pl; Beane et al.,2006), and GFP (0.5–1.5 pg/pl). Injected embryos were transferred from the protamine sulfate coated dish to an agar-coated dish sometime between the 60-cell and unhatched blastula stages.


Uninjected and Dsh-DEP injected L. variegatus embryos were allowed to develop until the 8-cell stage at which time they were pretreated with calcium-free seawater (McClay,1986). At the 16-cell stage, using glass needles, micromeres or mesomeres from the Dsh-DEP injected embryos were transplanted to uninjected host embryos, replacing the same cell type. After embryos were returned to filtered artificial seawater and cells had firmly adhered, embryos were placed in a 23°C incubator and allowed to develop until the prism stage.


Embryos were fixed in 2% paraformaldehyde in seawater for 20 min at room temperature and post-fixed in cold 100% MeOH for 1–10 min on ice. They were then rinsed 4 times in 1× PBS or seawater (5 min rinses) and placed into a blocking solution (1× PBS containing 3% bovine serum albumin) for 15–20 min. Embryos were incubated in 1° antibody in block (Endo1 1:25 or EctoV 1:5) for 1 hr at room temperature or overnight at 4°C. After three 1× PBS rinses (5 min each), embryos were blocked for 15–20 min and then transferred to 2° antibody in block (anti-mouse TRITC Texas Red, 1:150). After four 1× PBS rinses, embryos were transferred to 1× PBS containing 4% sodium azide and viewed using fluorescence microscopy. Endo1 (Wessel and McClay,1985) is an endodermal marker of the mid- and hindgut and EctoV (Coffman and McClay,1990) is a marker of oral ectoderm.


Mesodermal and endodermal genes to be evaluated by WMISH or QPCR were selected using the Sea Urchin Endomesodermal Gene Regulatory Network (GRN) ( (Davidson et al.,2002a,b; Oliveri and Davidson,2004; Levine and Davidson,2005) as a predictive guide. This network is a logic-based representation of the gene regulatory interactions occurring during endomesoderm formation in S. purpuratus (Davidson et al.,2002a,b; Oliveri and Davidson,2004; Levine and Davidson,2005).

WMISH was performed on L. variegatus embryos as described in Croce et al. (2003). WMISH probes were produced from L. variegatus sequences based, in most cases, on information from S. purpuratus. The probes used were synthesized from LvAlx1 (Ettensohn et al.,2003; Wu and McClay,2007; gift from S. Wu), LvBrachyury (Harada et al., 1995; Peterson et al.,1999; gift, J. Croce), LvDelta (Zhu et al.,2001; Sweet et al.,2002; gift, J. Croce), LvEndo16 (Nocente-McGrath et al.,1989; Yuh and Davidson,1996; Romano and Wray, 2006; gift, J. Croce), LvGataE (Hinman and Davidson,2003a; Lee and Davidson,2004; gift, J. Croce), LvGCM (Ransick et al.,2002; gift, S-Y. Wu/J. Croce), and LvBlimp1 (Krox; Wang et al.,1996; Hinman and Davidson,2003b; gift, K. Walton/J. Croce).

To prepare samples for QPCR analysis, control (embryos microinjected with 4.1–4.4 pM GFP mRNA) and experimental samples (embryos microinjected with 4.1–4.4 pM Dsh-DEP mRNA) of S. purpuratus collected at the 120–256 cell (5 hpf), mesenchyme blastula (24 hpf), and prism (48 hpf) stages were placed into TriReagent (Molecular Research Center, Cincinnati, OH) at a concentration of 1 embryo/μl. For each stage 125–150 embryos were collected. Phenol chloroform extractions were performed to isolate RNA and contaminating DNA was removed using RNase-free DNase (Ambion). Extracted RNA was reverse transcribed into cDNA using a SuperScript II RT H-Kit (Invitrogen, Carlsbad, CA) and a Bio-Rad SYBR Green kit was used to set up QPCR reactions. For each set of embryos, a sample was processed without reverse transcriptase to ensure that no DNA remained in the extracted RNA. From a 20-μl cDNA sample, 1 μl was used for each QPCR reaction.

Several of the QPCR primers were kindly provided by Dr. Takuya Minokawa (ubiquitin, Gcm, Eve, Pmar1, Krl, and Delta). Others were designed using SeqWeb software. These primer sets were designed to amplify sequences that were 150–250 bp long. All QPCR reactions were done in triplicate. SYBR Green Chemistry was used in the reactions and data was collected using a MJ Opticon2 QPCR machine. The cycle threshold (CT) data for 3 cycles of each gene analyzed was averaged and this data was normalized against the average CT values for ubiquitin mRNA. Ubiquitin is uniformly expressed in the sea urchin during the first 24 hr of development (Wessel et al.,1989; Gong et al.,1991). The cycle threshold difference (ΔΔCT) was then calculated as the normalized CT value for the control (the ΔCT value for GFP) minus the normalized experimental CT value (the ΔCT value for Dsh-DEP).


We thank Jenifer Croce, Simon Wu, Hyla Sweet, and Kate Walton for generously providing plasmids to synthesize whole-mount in situ hybridization probes or for providing synthesized probes. The LvDsh:GFP cDNA was kindly provided by Charles Ettensohn and Michelle Illies. We thank Takuya Minokawa for the QPCR primers, Donna Au for providing preliminary data concerning the Ca++/Wnt pathway, and Cyndi Bradham, Wendy Beane, Esther Miranda, Jeff Peng, and Shalika Kumburegama for their input. Finally, thanks to Jenifer Croce for reviewing earlier versions of this manuscript. C.A.B. was funded by a American Heart Association Postdoctoral Fellowship, D.R.M. was funded by the NIH, and A.H.W. was funded by the NSF, a Hawaii Community Foundation Grant, and a Hawaii State Brin Grant.