Integration of Canonical and Noncanonical Wnt Signaling Pathways Patterns the Neuroectoderm Along the Anterior–Posterior Axis of Sea Urchin Embryos

Three different Wnt signaling pathways function to restrict the anterior neuroectoderm state to the anterior end of the sea urchin embryo, a mechanism of anterior fate restriction that could be conserved among deuterostomes.


Introduction
Wnt signaling pathways play fundamental roles in many developmental processes. One of the earliest and most crucial of these roles is the activation of gene regulatory programs that specify different cell fates along the embryo's primary anteriorposterior (AP) axis. Recent comparative analyses suggest that Wnt/b-catenin signaling is an ancient AP patterning mechanism that establishes posterior identity in most metazoan embryos [1][2][3][4][5][6][7][8][9]. In invertebrate deuterostome embryos, which include cephalochordates, urochordates, hemichordates, and echinoderms, localized determinants cause stabilization of b-catenin in posterior blastomeres. This stabilized b-catenin enters nuclei in which it activates genes that specify endomesoderm, marking the site of gastrulation at what corresponded to the vegetal pole of the egg and forming the posterior end of the developing embryo [2,4,10,11]. In the sea urchin embryo, in which the molecular mechanisms of endomesoderm specification are best understood [12], the first evidence of Wnt signaling after fertilization is the presence of b-catenin in the nuclei (nb-catenin) of posterior cells, beginning at the 16-to 32-cell stage. During the next few cleavages, a detectable gradient of nb-catenin forms in the posterior half of the embryo, with the highest concentration at the posterior pole [4]. This gradient of Wnt/b-catenin signaling is both necessary and sufficient to activate the gene regulatory networks that establish mesoderm and endoderm cell fates in a posterior-to-anterior wave during late cleavage stages [13,14].
Wnt/b-catenin signaling also transforms the initial regulatory state that specifies anterior neuroectoderm (ANE) development in those deuterostome embryos in which it has been examined [15][16][17][18][19][20][21][22]. In the sea urchin embryo, we refer to this neuroectoderm as ANE because it becomes restricted to a region derived from the animal pole of the egg, which is located opposite to the posterior end of the embryo (see [23]). The initial regulatory state of early sea urchin embryos activates ANE specification by the 32-cell stage, when genes encoding the earliest ANE regulatory proteins are expressed broadly throughout the anterior half of the embryo [22,24]. These early factors include Six3, which is expressed at the anterior end of bilaterian embryos [25] and has been shown by functional studies to be critical for the specification of anteriormost neuroectoderm in diverse embryos including Tribolium castaneum [26], sea urchins [24], zebrafish, and mouse [27].
Beginning around the 60-cell stage, a progressive posterior-toanterior down-regulation of ANE factor gene expression in most of the anterior half of the embryo occurs by an unknown mechanism that requires posterior Wnt/b-catenin signaling [22]. This process continues during blastula stages until the ANE regulatory state is confined to a disk of cells around the anterior pole of the mesenchyme blastula [24]. Interestingly, an unknown signal from posterior Wnt/b-catenin signaling also appears to be necessary to pattern the anterior ectoderm along the AP axis in Saccoglossus kowalevskii [2], which belongs to the hemichordates, a sister clade to echinoderms.
Remarkably, Six3 activates a large cohort of genes in the sea urchin ANE that are orthologs of genes expressed in the vertebrate ANE (forebrain/eye field) ( Figure 1A), raising the possibility that the common ancestor of sea urchins and vertebrates may have shared this ANE regulatory program [24]. Similar to the sea urchin embryo, an initial widespread regulatory state in late blastula/early gastrula stages of vertebrate embryos supports expression of genes encoding early anterior forebrain/eye field factors throughout the presumptive neuroectoderm, including six3 [28,29]. Simultaneously, secreted antagonists from the organizer block bone morphogenetic protein (BMP) signaling on the dorsal side and a high-to-low, posterior-to-anterior gradient of nb-catenin forms in the presumptive neuroectoderm. This Wnt/b-catenin signaling gradient is part of a mechanism that activates genes encoding posterior neuroectoderm factors while down-regulating anterior (forebrain/eye field) factors in the posterior neuroectoderm [16,18,21]. By these mechanisms, the neural plate is formed and expression of the presumptive forebrain/eye field factors is restricted to cells at its anterior end, where Wnt antagonists protect them from posteriorization [30][31][32][33]. Multiple Wnts, Fzl receptors, and Wnt antagonists (i.e., Wnt8, Wnt3a, Wnt1, Fzl8, and Dkk1) have been implicated in posteriorization of the neuroectoderm in vertebrate embryos, as well as members of the fibroblast growth factor (FGF), retinoic acid (RA), and transforming growth factorbeta (TGF-b) signaling pathways [28,30,34,35]. However, the exact functions of these pathways in AP neuroectoderm patterning have been difficult to determine because of their earlier functions as well as the complex cell movements of gastrulation during this process [28,30,36]. Moreover, the interactions among these various pathways in AP neuroectoderm patterning are not well understood.
Recent studies suggest that early AP neuroectoderm patterning in vertebrate embryos is independent of information from the dorsal organizer. In both Xenopus and zebrafish, the initial widespread regulatory state promotes neuroectoderm specification throughout most of the embryo in the absence of b-catenin, which blocks dorsal organizer formation, as well as BMP2, BMP4, and BMP7. Expression of neuroectoderm markers is radialized around the AP axis in these embryos, but remarkably they retain normal AP neuroectoderm patterning, and the ANE expands posteriorly when both maternal and zygotic Wnt/b-catenin function is blocked [20,21]. Interestingly, in the sea urchin embryo, the action of Wnt/b-catenin signaling in early patterning of the neuroectoderm along the AP axis [22] is also separate and distinct from the dorsal-ventral patterning mechanism because it occurs before Nodal and BMP signaling are activated and is required for their expression [37][38][39][40]. The inhibition of Wnt/b-catenin signaling, and consequently the loss of expression of Nodal and BMP, causes a large majority of cells to differentiate into ANE [22,24,41]. Thus, the developmental regulatory mechanisms used by vertebrate embryos for ANE development have striking similarities to those used by sea urchin embryos and may therefore represent an ancestral deuterostome mechanism.
Here we show that the Wnt-dependent restriction of neuroectoderm to the anterior pole involves not only Wnt/b-catenin but also a series of linked steps mediated by Wnt/JNK signaling through Wnt1, Wnt8, and Fzl5/8, the homolog of vertebrate Fzl8. Coordinated progression of signaling through these Wnt pathways and activation of the secreted Wnt antagonist Dkk1 in anteriormost blastomeres establish the definitive ANE around the anterior pole. Signaling through a second Wnt receptor, Fzl1/2/7, and its activation of PKC suppress Wnt/b-catenin and Wnt/JNK ANE restriction activities to coordinate the correct temporal progression of ANE restriction. Collectively, signaling through three different Wnt signaling pathways provides precise spatiotemporal control of neuroectoderm AP patterning along the AP axis.

Wnt/b-Catenin Signaling Prevents ANE Specification in Posterior Blastomeres
FoxQ2 and Six3 are essential for the specification of the ANE and are the earliest ANE regulatory genes to be expressed. Their transcripts accumulate in the anterior half of the 32-cell embryo but are never detectable in the posterior half ( Figure 1B,C) [22,24]. We reasoned that posterior repression might depend on Wnt/b-catenin signaling because this pathway is activated in posterior blastomeres by the 16-cell stage [4,14,42,43]. To test this possibility, we blocked nb-catenin by injecting embryos with mRNA encoding either Tcf-Engrailed (Tcf-Eng) [42] or Axin [44] and examined foxq2 expression at the 32-cell stage (Figure 1B,F; Figure S1). In both cases, foxq2 and six3 were expressed in every blastomere during early cleavage stages (32-cell, Figure 1F; 120cell, Figure 1G) and ubiquitous expression persisted until late mesenchyme blastula stage (24 hpf) (Figures 1G,H and S1Ae-g). As expected, each perturbation resulted in formation of dauer blastulae with a thickened neuroepithelium covering most of the embryo that produced greatly increased numbers of serotonergic neurons throughout ( Figure 1I versus E; Figure S1Ah versus Ad). These 4-d embryos phenocopied DCadherin mRNA-injected embryos, which previously were shown to lack nb-catenin in all

Author Summary
The initial regulatory state of most cells in many deuterostome embryos, including those of vertebrates and sea urchins, supports anterior neural fate specification. It is important to restrict this neurogenic potential to the anterior end of the embryo during early embryogenesis, but the molecular mechanisms by which this re-specification of posterior fate occurs are incompletely understood in any embryo. The sea urchin embryo is ideally suited to study this process because, in contrast to vertebrates, anterior-posterior neuroectoderm patterning occurs independently of dorsal-ventral axis patterning and takes place before the complex cell movements of gastrulation. In this study, we show that a linked, three-step process involving at least three different Wnt signaling pathways provides precise spatiotemporal restriction of the anterior neuroectoderm regulatory state to the anterior end of the sea urchin embryo. Because these three pathways impinge on the same developmental process, they could be functioning as an integrated Wnt signaling network. Moreover, striking parallels among gene expression patterns and functional studies suggest that this mechanism of anterior fate restriction could be highly conserved among deuterostomes. but the four vegetal-most blastomeres, the small micromeres during cleavage stages [24]. Together, these data indicate that the factors that activate ANE specification operate in all early blastomeres in these Wnt/b-catenin-deficient embryos and likely are part of a ubiquitous maternal regulatory state. Moreover, these observations indicate that the first step in suppressing the ANE in the posterior half of the embryo depends on the repression or rapid down-regulation of ANE regulatory gene transcription by Wnt/bcatenin signaling.

Fzl5/8 Signaling and JNK Activity Are Required to Down-Regulate the ANE Regulatory State in Posterior Ectoderm Cells
Previous studies have shown that restriction of foxq2 expression to the anterior pole depends on posterior Wnt/b-catenin signaling. However, Wnt/b-catenin signaling has never been detected in the anterior half of the embryo (the presumptive ectoderm, blue in Figure 1Ja), suggesting that an intermediate signal(s) downstream of posterior Wnt/b-catenin signaling must mediate this second phase of ANE restriction (Figure 1Jb versus Jc; the gray region in this and subsequent figures represents the posterior ectoderm and the orange arrows indicate the second phase of restriction). We hypothesized that this intermediate signal ( Figure 2C, signal X) might also involve Wnt signaling, and we tested this idea by exploring the functions of the Wnt [Frizzled (Fzl)] receptors in ANE restriction. Two of the four sea urchin receptors, Fzl5/8 and Fzl1/2/7, were expressed during ANE restriction ( Figure S2A) and also in the appropriate cells to mediate this process ( Figure  S2Ba-h), making them excellent candidates for transducing Wnt signals that eliminate the ANE regulatory state from the posterior ectoderm ( Figure S2Bi-Bl).
To determine whether Fzl5/8 signaling has a role in neuroectoderm AP patterning, we injected embryos either with morpholinos targeting Fzl5/8 or with mRNA encoding a Cterminal truncated form of the receptor (DFzl5/8) that acts as a dominant negative by competing for Wnt ligands [45]. In contrast to embryos injected with Axin or Tcf-Eng mRNA, those expressing DFzl5/8 mRNA had normal foxq2 transcript levels and distributions at the 32-cell stage (cf., Figure 2Aa,Af), suggesting that Fzl5/8 signaling is not required for the initial Wnt/b-catenin-dependent down-regulation of foxq2 mRNA in the posterior half of the embryo. Further evidence that Fzl5/8 is not required for early Wnt/b-catenin activity is provided below. However, at mesenchyme blastula stage (24 hpf), DFzl5/8-injected embryos expressed foxq2 ectopically throughout the anterior half of the embryo, indicating that the second phase of its restriction to the anterior pole requires Fzl5/8 function (cf., Figure 2Ab,Ag). Expression of foxq2 also was not correctly restricted in two different Fzl5/8 morphants, although the phenotype was less pronounced (cf., Figure 2Ab,Ag versus Figure S3D,E). We used DFzl5/8 for further studies because it gave the more penetrant phenotype, likely because it blocked signaling through both maternal and zygotic Fzl5/8. Importantly, eliminating expression of six3, the critical upstream ANE regulator, from the posterior ectoderm also required functional Fzl5/8 signaling (Figure 2Ad,Ai). Furthermore, the transcript levels per embryo for genes in the 24 hpf Six3-dependent ANE regulatory network ( Figure 2B) were significantly elevated in DFzl5/8-containing mesenchyme blastula embryos. Interestingly, one of these was zygotic fzl5/8 mRNA itself (Figure 2Ac,Ah), indicating that Fzl5/8 function is required to down-regulate fzl5/8 mRNA levels in the posterior ectoderm. Finally, 3-d pluteus larvae injected with DFzl5/8 had an expanded thick neuroepithelium with a greatly increased number of serotonergic neurons (Figure 2Aj). In contrast, the thickened neuroepithelium in normal pluteus-stage embryos was restricted to a small region that produced only 4-6 serotonergic neurons ( Figure 2Ae). These observations indicate that a Fzl5/8 signalingdependent process eliminates the ANE regulatory state required for serotonergic neural development from the posterior ectoderm ( Figure 2C).
In addition to the ubiquitous maternal and anterior zygotic expression of fzl5/8 at mesenchyme blastula stage ( Figure S2e [46] showed that Fzl5/8 signaling in these posterior cells works through the c-Jun N-terminal kinase (JNK) pathway to initiate primary invagination movements later during gastrulation. This observation raised the possibility that the earlier ANE restriction process mediated by Fzl5/8 in posterior ectoderm may also depend on the JNK pathway. jnk mRNA was present ubiquitously during ANE restriction ( Figure S2C), and indeed, foxq2 failed to restrict to the anterior pole in embryos injected with a spliceblocking JNK morpholino ( Figure S3A). This JNK morphant phenotype was weaker than the DFzl5/8 phenotype (cf., Figure 2G and Figure S3A), probably because some normal JNK transcripts persisted in the embryo ( Figure S3J). It is also possible that maternally synthesized JNK protein persisted in these embryos. As an additional test, we treated embryos with the specific JNK inhibitor, (L)-JNKI1 [46,47], beginning at fertilization, which produced embryos expressing foxq2 throughout the anterior half of the embryo, mimicking exactly the DFzl5/8 phenotype (Figure 2, cf. Ag,Al). Moreover, fzl5/8 and six3 expression was not restricted to the anterior pole ( Figure 2Am,An), and these embryos also had an expanded, thickened neuroepithelium and an increased number of serotonergic neurons, as seen in DFzl5/8-injected embryos ( Figure 2Aj). These results indicate that the second phase of ANE restriction that down-regulates the ANE regulatory state in the anterior half (i.e., the posterior ectoderm) depends on Fzl5/ 8 function. Moreover, they suggest that JNK activity transduces a Wnt signal X through this Wnt receptor, the production of which depends on Wnt/b-catenin activity in the posterior half of the embryo (signal X, Figure 2C).

Wnt1 and Wnt8 Signals Restrict the ANE to the Anterior Pole
To identify the link between Wnt/b-catenin signaling and Fzl5/ 8, we first searched for genes encoding Wnt ligands that are expressed by the 60-cell stage (i.e., the beginning of the second phase of ANE restriction) in posterior blastomeres and that also depend on Wnt/b-catenin activity. We confirmed the previously reported expression profile of wnt8, which is activated by Wnt/bcatenin [48,49]: At the 60-cell stage, wnt8 was expressed in both the micromeres and the adjacent blastomere tier (veg 2 ) ( Figure 3Aa). Similarly, wnt1 expression was first detected midway through the 60-cell stage (9 hpf) in the micromeres, and it also depended on nb-catenin (Figures 3Aa and S4A,C). As 3.5-d pluteus stage embryo. (I) A 3.5-d embryo misexpressing Axin mRNA. White box outlines the ANE. Serotonergic neurons (green), DAPI (nuclei, blue). (J) Schematic showing the stages of ANE restriction in 32-cell stage (6 hpf) (a), early blastula stage (7-15 hpf) (b), and mesenchyme blastula stage (24 hpf) embryos (c). ANE (blue), nb-catenin (orange nuclei at 32-cell stage), endomesoderm territory (orange at blastula stages), posterior ectoderm (gray at blastula stages), and unknown restriction mechanism activated by posterior nb-catenin (orange arrows). doi:10.1371/journal.pbio.1001467.g001 development progressed, wnt8 expression first moved into the next most anterior tier of blastomeres (veg 1 ) and then, during late blastula stages (18 and 24 hpf), into both veg 1    posterior regions of the anterior hemisphere. As restriction proceeded, wnt8 continued to be expressed near cells expressing ANE marker genes, whereas wnt1 expression was more posterior. In order to evaluate whether these secreted ligands were required for ANE restriction in posterior ectoderm, we performed knockdown experiments by injecting either of two different morpholinos designed against each. As shown in Figure 3B, embryos injected with either Wnt1 or Wnt8 morpholinos failed to down-regulate foxq2 expression in posterior ectoderm. ANE restriction was more strongly perturbed in Wnt1 morphants, even though the cells producing it were more distant from the site of action than those producing Wnt8. This raised the possibility that Wnt1 is necessary for later Wnt8 expression. However, this was not the case because, at blastula stage (16 hpf), Wnt8 expression was normal in Wnt1 morphants ( Figure S4D). The converse was also true: wnt1 expression did not depend on Wnt8 ( Figure S4E). We conclude that production of each of these ligands depends on Wnt/b-catenin signaling, but they do not depend on each other but act in parallel in ANE restriction.
These results suggest that Wnt1 and Wnt8 spatiotemporally link posterior Wnt/b-catenin signaling to Fzl5/8-mediated downregulation of ANE factors in the posterior ectoderm. To test this hypothesis, we first showed that overexpressed Wnt1 or Wnt8 completely eliminated foxq2 expression in the ANE (Figure 3Cb,Cd). We then tested whether this foxq2 downregulation required active Fzl5/8. Strikingly, DFzl5/8 strongly blocked the suppression of foxq2 expression mediated by either Wnt1 or Wnt8 (100% rescue of Wnt1 or Wnt8 misexpression phenotype; n = 63 and 67, respectively) (Figures 3Cc,Ce and S5Bb,Bd). These results strongly support the conclusion from the Wnt1 and Wnt8 loss-of-function analyses that Fzl5/8-mediated ANE restriction in the posterior ectoderm requires these ligands. Furthermore, suppression of foxq2 expression by both Wnt 1 and Wnt8 also required JNK activity, since the JNK inhibitor rescued the loss-of-ANE phenotypes produced by misexpression of Wnt8 (81% of embryos rescued; n = 126) (Figure 3Ca versus Cf) and, to a lesser extent, Wnt1 (55% of embryos had low to normal foxq2 expression; n = 83) ( Figure S5A). These data indicate that Wnt1, Wnt8, and Fzl5/8 function in a Wnt/JNK signaling pathway to effect the second phase of ANE restriction.

Fzl1/2/7 Signaling and PKC Activity Are Necessary for ANE Specification
The expression pattern of the gene encoding the other early Wnt receptor, fzl1/2/7, suggests that Fzl1/2/7 signaling also could affect neuroectoderm restriction ( Figure S2Ba-Bd). We tested this possibility by morpholino knockdown. We were surprised to find that neither six3 nor foxq2 was activated at the 32-to 60-cell stage (Figure 4Aa versus Ah and Figure 4C) and neither mRNA was detectable throughout the normal time of ANE restriction (Figure 4Ah-k, Af-versus Am). As expected, zygotic fzl5/8 expression, which depends on Six3 [24], also required Fzl1/2/7 (Figure 4Ae versus Al). As well, the expression of all other known regulatory factors that depend on Six3 at mesenchyme blastula stage (24 hpf) also required Fz1/2/7 function ( Figure 4D). Moreover, the ectoderm in 3-to 4-d Fzl1/ 2/7 morphants lacked a thickened columnar epithelium corresponding to the ANE in normal embryos ( Figure S3F). In 4-d pluteus larvae, which normally have well-established neurons in the ANE, the large majority of Fzl1/2/7 morphants had none (37/41 embryos) (Figure 4Ag versus An, green). They also had a severely reduced number of ciliary band neurons, as assayed by the pan-neural marker SynaptotagminB (Figure 4An, 1e11 antibody, magenta). These results indicate that Fzl1/2/7-mediated signaling is essential for establishment and maintenance of the early neuroectoderm regulatory state, which in turn subsequently is required for the specification and differentiation of all neurons (Figure 4Ag).
The Fzl1/2/7 morphant phenotype is opposite to the Axin or Tcf-Eng misexpression phenotypes as well as those produced by DFzl5/8 misexpression or treatment with the JNK inhibitor or JNK morpholino. These observations raise the possibility that Fzl1/2/7 transduces a different Wnt signal, possibly through the Ca 2+ pathway. Although the architecture of the Ca 2+ pathway downstream of Fzl receptors is not yet well established, one important player in other systems is conventional Protein Kinase C (PKC) [50,51]. In the sea urchin embryo, genes encoding conventional PKC isoforms are expressed maternally and throughout development and at least one is activated by the 60cell stage ( Figure 4B) [52]. To test the hypothesis that pPKC, like Fzl1/2/7, is necessary for maintaining ANE specification, we treated embryos with the specific PKC inhibitor, Bisindolylmaleimide 1, which blocks activation through phosphorylation of most Ca 2+ -dependent PKC isoforms by competing for the ATP binding site [53]. Treatment with this inhibitor at 1-3 mM strongly reduced the level of pPKC ( Figure 4B), but had no detectable deleterious effects on the morphology of embryos during ANE restriction. Importantly, the level of pPKC in Fzl1/2/7 morphants was as low as that produced by the PKC-specific inhibitor ( Figure 4B), indicating that Fzl1/2/7 function is required for activation of this kinase. Similar to Fzl1/2/7 morphants, foxq2 expression was never initiated in embryos treated with the inhibitor continuously from fertilization to mesenchyme blastula stage (24 hpf) (Figure 4Ao-r). Moreover, six3 and fzl5/8 were not expressed (Figure 4As,At), and in a large majority of embryos (36/ 39) serotonergic neurons did not develop (Figure 4Au and Figure  S3Ia versus Ic), showing that neural differentiation was severely compromised in treated embryos. While these experiments demonstrate that activation of PKC is required for the ANE regulatory state and that Fzl1/2/7 is required for that activation, they do not conclusively prove that Fzl1/2/7 signals through the Ca 2+ pathway because PKC activation can occur by other mechanisms. We conclude that Fzl1/2/7 signaling and PKC activity are each essential for early neuroectoderm specification.

Fzl1/2/7 Signaling and PKC Activity Antagonize the ANE Restriction Mechanism
Our findings that a Wnt signaling branch utilizing Fzl1/2/7 and PKC activity is necessary for initiating expression of upstream ANE regulatory factors was entirely unexpected because at early stages, Wnt signaling is thought to antagonize this process. We hypothesized that Wnt signaling through this receptor is necessary either for the expression of regulatory genes that specify the ANE or for antagonizing the ANE restriction mechanism from the very earliest stages. To distinguish between these alternatives, we first asked whether Fzl1/2/7 signaling is part of the maternal mechanism that can drive ubiquitous expression of ANE regulatory genes in the absence of Wnt/b-catenin signaling. Within each of three batches of embryos, we injected one set of fertilized eggs with Axin mRNA, a second set with Fzl1/2/7 morpholino, and a third with both Fzl1/2/7 morpholino and Axin mRNA ( Figure 5A). As shown above, foxq2 was expressed throughout the embryo in the absence of nb-catenin, whereas it was completely undetectable in more than 90% (52/57) of embryos lacking Fzl1/2/7. However, it was expressed at high levels throughout all Fzl1/2/7-deficient embryos (47/47) when Wnt/b-catenin signaling was also blocked. These results indicate that maternal factors are still capable of activating foxq2 in embryos lacking Fzl1/2/7 and that the loss of foxq2/ANE fate in Fzl1/2/7 morphants requires a functional Wnt/b-catenin pathway. Thus, Fzl1/2/7 signaling is not a positive regulator of the initial maternal regulatory state that supports ANE specification, but rather it inhibits the Wnt/b-catenin-dependent ANE restriction mechanism.
To test if Fzl1/2/7 also antagonizes the Fzl5/8-JNK-dependent second phase of ANE restriction, we asked whether blocking Fzl5/ 8 or JNK function could rescue ANE specification in embryos lacking Fzl1/2/7 signaling ( Figure 5B,C). Similar to the above experiments, in three different batches of embryos, we found that blocking the function of either Fzl5/8 or the JNK pathway rescued the expanded expression of foxq2 in 99% (n = 72) or 93% (n = 70), respectively, of embryos also lacking Fzl1/2/7. These results suggest that Fzl1/2/7 antagonizes Fzl5/8-JNK-mediated ANE restriction. In the final set of experiments, we tested whether PKC signaling also antagonizes Fzl5/8-JNK-mediated ANE restriction ( Figure 5D). Using the same approach, we injected one set of fertilized eggs with DFzl5/8, treated a second with the PKC inhibitor, and a third was treated with PKC inhibitor and injected with DFzl5/8. Blocking the function of Fzl5/8 in these embryos rescued the expression of foxq2 in a large majority of embryos (77% rescue; n = 83), demonstrating that, like Fzl1/2/7, PKC antagonizes the ANE restriction mechanism by antagonizing Fzl5/8 signaling. Collectively, these results support the idea that the Fzl1/2/7-dependent suppression of Fzl5/8-mediated ANE restriction works through PKC ( Figure 5G).
The data suggest that Fzl1/2/7 signaling antagonizes Fzl5/8-JNK-mediated down regulation of genes necessary for ANE specification. Because Fzl1/2/7 functions as early as the 32-cell stage to maintain expression of ANE markers, it might also antagonize Fz5/8 indirectly by down-regulating Wnt/b-catenin activity. To test this possibility, we measured the level of Wnt/bcatenin signaling in 120-cell embryos (12 hpf) during the early stages of ANE restriction using the TCF-luciferase reporter plasmid, TopFlash [54]. Three different batches of embryos that had been injected with DFzl5/8 or treated with PKC inhibitor showed no significant difference in TopFlash activity when compared to controls ( Figure 5E), suggesting that neither of these proteins affects early Wnt/b-catenin signaling. In contrast, TopFlash activity increased ,2.5-fold on average in embryos lacking Fzl1/2/7 compared to controls ( Figure 5E), indicating that signaling through Fzl1/2/7 negatively regulates Wnt/bcatenin activity in cleavage-stage embryos. Recently published experiments showed that introduction of mRNA encoding a dominant negative form of Fzl1/2/7 caused a reduction in TopFlash activity in cleavage-stage embryos and a loss of endoderm specification [55]. While this appears to conflict with our results, it is important to realize that interference with Fzl1/ 2/7 activity by misexpression of DFzl1/2/7 can interfere with the function of maternal Fzl1/2/7, whereas Fzl1/2/7 morpholino cannot. In keeping with this, embryos in which zygotic Fzl1/2/7 synthesis was blocked with a morpholino still expressed Wnt1 and Wnt8 ( Figure 5F), whereas these are not expressed in embryos injected with DFzl1/2/7 [55]. Thus, in Fzl1/2/7 morphants, these Wnt ligands up-regulate Wnt/b-catenin-and Wnt/Fzl5/8mediated ANE restriction, whereas the absence of these ligands in DFzl5/8-containing embryos leads to a reduction in Wnt/bcatenin activity. Collectively, these data suggest that Fzl1/2/7 signaling and PKC activity provide a buffer that limits the rate of ANE down-regulation by both of these Wnt signaling pathways ( Figure 5G). A possible concern in the DFz5/8 and DFz1/2/7 experiments is that elevating the levels of these proteins might influence the balance of signaling between the Wnt signaling pathways, for example, by competing for common components. To test this possibility, we overexpressed either wild-type Fzl5/8 or Fzl1/2/7 mRNA. In both cases, embryos developed normally and had normal foxq2 expression patterns ( Figure S6A). Next we showed that elevating the levels of Fzl1/2/7 mRNA did not change DFzl5/8's ability to prevent ANE restriction ( Figure S6B) or prevent elimination of the ANE by excess Wnt1 mRNA ( Figure  S6C). Taken together these data indicate that the levels of endogenous Fzl receptors are not limiting. These data contrast with the Wnt1 and Wnt8 misexpression results, which showed that excess ligand can dramatically up-regulate ANE restriction (Figure 3Cb,d), suggesting that it is the levels of Wnt ligand in time and space and not those of the Wnt receptors that control the ANE restriction mechanism

Dkk1 Antagonism of Wnt Signaling Protects the Final ANE Territory
Around the mesenchyme blastula stage (24 hpf), restriction of the ANE is complete and it constitutes a separate regulatory domain at the anterior end of the embryo with well-defined borders. Expression of fzl5/8 is also restricted to this domain (Figure 2), raising the question of why Fzl5/8-mediated signaling does not continue to down-regulate the ANE regulatory state there. We hypothesized that the secreted Wnt antagonist, Dkk1, might play a role because, in most of the major clades, competition between anterior Wnt antagonism by Dkk1 and posterior Wnt signaling has been shown to regulate cell fates along the primary (AP) axis [8]. Very low-level dkk1 expression was detectable as early as the 120-cell stage by qPCR ( Figure 6A), and increased during the time of ANE restriction, reaching maximal levels by the mesenchyme blastula stage (24 hpf). At this time dkk1 expression could be detected by in situ hybridization at the anterior end of the embryo as well as in a ring of cells surrounding the future site of gastrulation ( Figure 6A, inset). Thus, dkk1 was expressed at the right time and place to prevent anterior Wnt-mediated ANE down-regulation. Interestingly, expression of dkk1 depended on Fzl5/8 signaling ( Figure 6B), raising the possibility that Fzl5/8 signaling limits its own activity in anterior cells by promoting a negative feedback mechanism through this Wnt antagonist.
To test whether Dkk1 protects the ANE regulatory state from Wnt-mediated down-regulation, we monitored the expression of a set of genes encoding ANE regulatory factors in Dkk1 morphants at mesenchyme blastula stage by in situ hybridization. Each of the genes tested was severely down-regulated in these embryos (Figures 6Cg-k and S3H), and no serotonergic neurons developed in 4-d plutei (Figure 6Cl). Furthermore, overexpression of Dkk1 mRNA prevented restriction of foxq2 expression (Figure 6Db) and rescued foxq2 expression in embryos also overexpressing wnt1 mRNA (88% rescue; n = 65) (Figure 6Dc versus d). Together these results indicate that Dkk1 can block the Wnt1/Fzl5/8-JNKdependent ANE restriction mechanism. Interestingly, overexpression of Dkk1 also rescued foxq2 expression in Fzl1/2/7 morphants (98% rescue; n = 64) (Figure 6Ec,Ed), suggesting that it may interfere with either Wntb-catenin or Fzl5/8 signaling or both. There is some support for both possibilities. First, Dkk1 likely inhibits Fzl5/8 activity because the morphological phenotype (unpublished data) and foxq2 expression pattern (cf., Figures 1F,  2Ag, and 6Db) of Dkk1 mRNA-injected embryos were more similar to those of embryos lacking functional Fzl5/8 than to those lacking Wnt/b-catenin signaling (cf., Figure 6Db and Figures 1H  and 2Ag). Second, misexpressed Dkk1 can also interfere with endomesodermal gene expression, which depends on the Wnt/bcatenin pathway ( Figure S4B).

Discussion
The data presented here show that patterning the neuroectoderm along the AP axis of the early sea urchin embryo depends on an elegant spatiotemporal coordination and integration of the activities of three different Wnt signaling pathways. Throughout this process, a balance is achieved between the initial regulatory mechanisms that can specify the ANE ubiquitously, those that subsequently suppress it in posterior regions, and those that limit ANE suppression. The consequence is that ANE tissue is stably positioned only at the anterior pole of the embryo by the mesenchyme blastula stage. To summarize our current model (Figure 7), the first phase of ANE restriction requires Wnt/bcatenin and occurs very rapidly in posterior blastomeres by the 32to 60-cell stage. Wnt/b-catenin signaling simultaneously activates expression of Wnt1 and Wnt8; these cells and these ligands initiate the second phase of ANE down-regulation in the posterior ectoderm (non-ANE ectoderm in the anterior hemisphere) by activating the Fzl5/8-JNK pathway, beginning around the 60-cell stage. As development progresses, Wnt1 and Wnt8 mRNAs accumulate in more anterior blastomeres, behind the front of ANE down-regulation. Whether these secreted ligands diffuse to the overlying ectoderm to directly activate Fzl5/8 or whether they act indirectly to stimulate production of other Wnt ligands that signal through this receptor is not known. Regardless, it is clear that Wnt1, Wnt8, Fzl5/8, and JNK are all required for full ANE downregulation in the posterior ectoderm and suppression of transcription of fzl5/8 itself. Clearly, Fzl5/8 plays a pivotal role in the ANE restriction process because it is necessary not only for the second phase of ANE restriction but also to stop that process in the third phase of ANE patterning when Fzl5/8 signaling leads to the expression of the Wnt receptor antagonist Dkk1 at the anterior pole. Thus, the coordination between the timing of auto-repression of fzl5/8 transcription and activation of Dkk1 by Fzl5/8 ensures that this negative feedback loop reproducibly defines the ANE at the anterior pole of the embryo by mesenchyme blastula stage (24 hpf). The relative timing of Dkk1 production in the anterior ectoderm and ANE restriction in the rest of the embryo is critical and carefully controlled by a third Wnt pathway working through Fzl1/2/7 and PKC activities that limit Wnt/b-catenin and Wnt/ JNK functions during the first two phases of ANE clearance.
Because all of these Wnt pathways affect the same developmental process (i.e., the specification of ANE versus non-ANE fates along the primary axis), they may function as components of an interactive Wnt signaling network rather than as separate pathways with different roles. Yet it appears that posterior Wnt/bcatenin and anterior Wnt/JNK signaling define two adjacent early regulatory domains in the sea urchin embryo. While our data suggest that these two signaling pathways activate different downstream regulatory programs in order to down-regulate ANE factors, both pathways are linked spatially and temporally by the activities of at least two common signaling components, Wnt1 and Wnt8 [49,56]. These results are in keeping with recent evidence that individual Wnt ligands are able to activate distinct Wnt signaling branches, often in the same or adjacent territories [57][58][59]. However, it remains to be determined whether Wnt1 and/or Wnt8 act directly on cells in the anterior hemisphere in the ANE restriction process, although it is interesting that wnt8 expression moves into posterior ectoderm cells as ANE factors move out. Alternatively, Wnt1 and Wnt8 may act indirectly by reinforcing the nb-catenin gradient in the anterior-most cells of the  posterior half of the embryo (i.e., near the equator), activating production of an unidentified intermediate Wnt ligand that is secreted from even more anterior cells and that activates Fzl5/8 signaling.
We found that the cardinal ANE regulatory genes, six3 and foxq2, are not expressed in Fzl1/2/7 knockdowns. This unexpected phenotype is the exact opposite of the ANE expansion produced by interference with Wnt/b-catenin, Fzl5/8, and JNK signaling. The function of Fzl1/2/7 begins as early as the 32-cell stage, around the time that nb-catenin is first detectable in posterior blastomeres [4] and at least 2 h before the ANE restriction process mediated by Fzl5/8 and JNK is observed. Since Fzl1/2/7 signaling significantly suppresses Wnt/b-catenin signaling during early cleavage stages, we propose that it reduces Wnt/b-catenindependent Fzl5/8 and JNK activities. This model suggests that Fzl1/2/7 signaling is essential for controlling the rate of progression of the ANE restriction mechanism along the AP axis, providing a ''timing buffer'' that prevents premature elimination of the ANE regulatory state during the early cleavage and blastula stages. We propose that one function of this Fzl1/2/7 ''timing buffer'' is to allow sufficient Fzl5/8-dependent accumulation of Dkk1 in the ANE by later blastula stages to protect it from Wnt signals and define its borders.
Since Fzl1/2/7 does not appear to signal through either the Wnt/JNK or the Wnt/b-catenin pathways during ANE restriction, we propose that it transduces signals through the Wnt/Ca 2+ pathway. This mechanism may be similar to the situation in several other systems where Wnt/Ca 2+ signaling affects early development either through the intracellular messengers CamKI, Calcineurin, and the transcription factor, NF-AT, or through PKC [60]. Similar to what we report here, the Wnt/Ca 2+ pathway has been shown to antagonize Wnt/b-catenin signaling during vertebrate D/V axis specification [61,62]. Interestingly, we found that blocking PKC activity with either an inhibitor or with Fzl1/ 2/7 morpholino had exactly the same effect on phosphorylated PKC levels and on the Fzl5/8-JNK-dependent re-specification of ANE to ectoderm fate. However, inhibiting the function of Fzl1/ 2/7 elevated Wnt/b-catenin activity, whereas loss of PKC activity did not. This result suggests that Fzl1/2/7 signaling activates two branches that affect ANE restriction, one that antagonizes early Wnt/b-catenin activity and another, mediated by pPKC, that blocks Fzl5/8-mediated ANE restriction in the anterior hemisphere. Thus, if Fzl1/2/7 mediates Wnt/Ca 2+ signaling in the sea urchin embryo, it could affect several different downstream parallel pathways, any or all of which are necessary to prevent premature and complete elimination of the ANE regulatory state. Moreover, the involvement of Wnt/Ca 2+ signaling in AP neuroectoderm patterning would be a first.
These considerations suggest that the function of Fzl1/2/7 in the early embryo is context-dependent, and we propose that the balance of information sent by this receptor through different Wnt signaling pathways is essential for correct specification and patterning. Recent data from several laboratories suggest that the same Fzl receptors can activate different Wnt signaling pathways, even in the same cells [50,60,63]. For example, the sea urchin Fzl1/2/7 homologue, Fz7, activates Wnt/b-catenin signaling and D/V axis specification in the early Xenopus embryo [63], but it also later activates Wnt/JNK and possibly the PKC signaling pathways that are required in the same general territory for convergent extension movements during gastrulation [57,64,65]. In the sea urchin embryo, our results and those of Lhomond et al. (2012) [55] are consistent with two early roles for Fzl1/2/7 -one stimulating endoderm specification via nb-catenin through maternal Fzl1/2/7 in posterior blastomeres and another produced by zygotic Fzl1/2/7 that antagonizes early Wnt/bcatenin and subsequent Wnt/JNK signaling through an alternative Wnt pathway (Ca +2 ) that operates throughout the embryo. The balance between these pathways may favor Wnt/b-catenin signaling in the posterior half of the cleavage stage embryo because of localized Wnt/b-catenin pathway-specific co-factors in that part of the embryo [66,67].
Striking parallels are emerging in the regulatory mechanisms that sea urchin and vertebrate embryos use to establish neural regulatory states at the anterior pole. Both embryos require Six3 for anterior neural development and share many homologous factors [24,27]. Moreover, as shown here, in the absence of Wnt/ b-catenin, and consequently of Nodal, BMP, and all other known signaling pathways, the regulatory state of all of the cells in the sea urchin embryo supports development of ANE from the very beginning of its specification. These data indicate that an initial ubiquitous maternal regulatory state activates ANE specification and that one of the most important roles of posterior Wnt/bcatenin signaling is to break the symmetry of this neuralpromoting state. Similarly, in vertebrate embryos, an initial regulatory state is capable of activating ANE markers throughout the embryo in the absence of Wnt, Nodal, and BMP signaling [15,[20][21][22]68]. Thus, this initial, broad activation of ANE specification, and its subsequent down-regulation, could be a widely shared property of deuterostome embryos.
The Wnt-dependent mechanism used for AP neuroectoderm patterning is still incompletely understood in vertebrates, in part because complex cell movements during patterning and the involvement of Wnt signaling in earlier specification events obscure the spatial and temporal relationships among the individual players [28,30,36]. In vertebrates, the only known Wnt pathway involved in the early restriction of ANE factors to the anterior pole is Wnt/b-catenin signaling [16][17][18]. Here, we show for the first time, to our knowledge, that the anterior Dkk1posterior Wnt/b-catenin neuroectoderm patterning mechanism observed in vertebrates exists in a nonchordate deuterostome. These data strongly suggest the general Dkk1-Wnt/b-catenin AP patterning mechanism present in extant pre-bilaterian embryos was likely co-opted to pattern the neuroectoderm along the AP axis in the deuterostome ancestor. In addition to a posterior-toanterior gradient of Wnt/b-catenin signaling, AP neuroectoderm patterning in the sea urchin embryo also requires Wnt/JNK signaling and an additional pathway mediated by Fzl1/2/7 that may function in Wnt/Ca 2+ signaling. At present these are completely novel findings, but the fact that orthologs of several Wnt signaling components that function in these additional pathways in sea urchins (Fzl8, Wnt1, Wnt8, Dkk1) ( Figure 8A,C) also are involved in posteriorizing the neural plate of vertebrate embryos ( Figure 8A) [17,31,69] raises the possibility that this entire multistep mechanism was present in the common echinoderm/ vertebrate ancestor and still operates to specify anterior neural identity in deuterostome embryos. Supporting this view, recent studies in hemichordates indicate that expression of homologues of sea urchin foxq2, sfrp1/5, and six3 demarcate an anterior-most region of the embryo that is homologous to the vertebrate anterior neural ridge secondary patterning center [19]. Interestingly, these factors are initially broadly expressed and restricted to this region by an unknown mechanism that depends on posterior Wnt/bcatenin signaling and appears to require Fzl5/8 function in the anterior part of the embryo ( Figure 8D). Moreover, there are similarities in the expression patterns of ANE genes (dkk1, dkk3, six3, foxq2) and those specifying endomesoderm (wnt1 and wnt8) between the invertebrate chordate amphioxus and the sea urchin embryo ( Figure 8B). For example, foxq2 is initially expressed in a  [17,21,29,[73][74][75][76][77]. (B) In amphioxus embryos, foxq2 is expressed throughout the anterior half, and wnt8 throughout the vegetal plate, of early gastrula embryos (left-hand diagram). By late gastrula, foxq2 and the putative ANE factors dkk1, six3, and dkk3 are expressed in the anterior-most ectoderm. wnt8 and wnt1 are expressed posterior to these putative ANE factors, consistent with a role in the restriction of foxq2, six3, and dkk3 expression to the anterior pole (middle diagram). Data taken from [78][79][80][81][82]. (C) Sea urchin embryo ANE factors are initially expressed throughout the presumptive ectoderm, and wnt1 and wnt8 are both expressed in the posterior half of blastula stage embryos (left-hand diagram). Then, ANE factors are progressively down-regulated from posterior ectoderm by a Wnt1, Wnt8, Fzl5/8, and Dkk1-dependent mechanism (middle diagram). In the absence of Wnt/b-catenin and TGF-b signaling, the entire sea urchin embryo expresses ANE factors (right-hand diagram). Data taken from this study and [22,24]. (D) In blastula stage hemichordate embryos, foxq2 is expressed broadly in the anterior half of the embryo (data show that six3 and fzl5/8 also are expressed broadly by early gastrula stages) (left-hand diagram). By late blastula stages, putative ANE factors foxq2, six3, and sfrp1/5 are restricted to the anterior-most ectoderm, and functional data show that sfrp1/5 restriction involves Fzl5/8 (middle diagram). In the absence of Wnt/b-catenin signaling the entire hemichordate embryo expresses putative ANE factors six3 and sfrp1/5 (right-hand diagram). Data taken from [2,19,83,84]. doi:10.1371/journal.pbio.1001467.g008 broad region and subsequently restricted to the anterior-most region. It can be completely cleared from this region of the embryo by LiCl treatment, which can elevate Wnt/b-catenin signaling [11], raising the possibility that amphioxus also uses the same ANE patterning mechanism. Thus, there is accumulating evidence that the ANE clearance mechanism described here may be used in a wide variety of deuterostomes. However, to date, only the work reported here reveals the intricate, interdependent Wnt signaling mechanisms that are required to confine the ANE regulatory state to the anterior end of the embryo.

Animals, Embryos, and Treatments
Strongylocentrotus purpuratus sea urchins were obtained from Point Loma Marine Invertebrate Lab, Lakeside, CA; The Cultured Abalone, Goleta, CA; or Marinus, Garden Grove, CA. Embryos were cultured in artificial seawater at 15uC. For drug treatments, eggs attached to a protamine sulfate-coated plate were fertilized in the presence of 2 mM 4-Aminobenzoic acid (PABA), and fertilization envelopes were removed by shear force. Treatments with the cell-permeable JNK Inhibitor 1, (L)-form, (EMD/ Calbiochem) and the PKC inhibitor, Bisindolylmaleimide 1 (EMD/Calbiochem), were performed by diluting the stock solution to 50 mM or 3 mM, respectively. JNK Inhibitor 1, (L)form is a specific inhibitor that blocks interactions between JNK and its transcriptional substrates, such as c-Jun and c-Fos, resulting in a knockout phenotype [46,47]. Bisindolylmaleimide 1 is a selective inhibitor that specifically competes with the ATP binding site of most PKC isoforms [53]. As controls for the PKC inhibitor experiments, DMSO was added alone. These experiments were repeated with at least three different embryo batches, and each produced the same results.

Preparation of cDNA Clones
The 24-h blastula total cDNA was used to obtain full-length clones for dkk1, frizzled5/8, frizzled1/2/7, wnt1, and a partial clone of jnk by PCR. The following primers were based on the sea urchin genome sequence: Sp-dkk1

mRNA and Morpholino Injections
Full-length dkk1 and wnt1 cDNA sequences were inserted into pCS2+ vector for mis-expression studies. DFzl5/8-pCS2 and Wnt8-pCS2 were obtained from Jenifer Croce (CNRS/Villefranche sur Mer, France) and Christine Byrum (College of Charleston, Charleston, SC), respectively. pCS2 constructs were linearized with Not1 and mRNA was synthesized with mMessage Machine kit (Ambion), purified by LiCl precipitation and ,20 pl injected at the following concentrations: Fzl1/2/7 mRNA = 1.0-1.5 mg/mL; Fzl5/8 mRNA = 2.0 mg/mL; DFzl5/8 mRNA = 2.0 mg/mL; Wnt1 mRNA = 0.01-0.1 mg/mL; Wnt8 mRNA = 0.5-1.0 mg/mL; Dkk1 mRNA = 3.0 mg/mL; Axin mRNA = 1.0 mg/mL; Tcf-Eng mRNA = 0.5-1.0 mg/mL. S. purpuratus EST sequences for wnt1, fzl1/2/7, and fzl5/8 as well as sequence information from 59 RACE on dkk1 were used to generate translation-blocking morpholino oligonucleotides. A splice-blocking morpholino oligonucleotide was designed for the second exon-intron boundary of wnt8, which produces transcripts encoding a protein lacking sequence from the second exon, which was verified by PCR ( Figure S2A). The morpholinos were produced by Gene-Tools LLC (Eugene, OR). The sequences and injection concentrations were: Embryos were injected immediately after fertilization with solutions containing FITC, 20% glycerol, and mRNA and/or morpholino oligonucleotides. All injected embryos were cultured at 15uC. Microinjection experiments were performed using at least three different batches of embryos, and each experiment consisted of 50-150 embryos unless otherwise stated. Experiments were scored only if a change in phenotype or marker expression was seen in at least 85%-90% of the manipulated embryos.
Quantitative PCR (qPCR) qPCR was performed as described previously [24]. Each experiment was repeated with embryos from at least three different mating pairs, and each PCR reaction was carried out in triplicate. The primer set information can be found in Table S1. For developmental expression analysis, the number of transcripts per embryo was estimated based on the Ct value of the z12 transcript [70].

Whole-Mount in Situ Hybridization
The probes for each gene analyzed correspond to the full-length cDNA sequence. Alkaline phosphatase and three-color fluorescent in situ hybridization were carried out as previously described [24,71]. For the three-color in situ hybridization, foxq2 was labeled with fluorescein and detected with Cy5-TSA, wnt1 was labeled with DNP and detected with Cy3-TSA, and wnt8 was labeled with DIG and detected with fluorescein-TSA.

Immunohistochemistry
Embryos were fixed in 2%-4% paraformaldehyde in artificial seawater at RT for 20 min and washed 5 times in phosphatebuffered saline containing 0.1% Tween-20. Embryos were incubated with primary antibodies at 4uC overnight at a dilution of 1:1,000 for serotonin (Sigma, St. Louis, MO) and synaptotag-minB/1e11 [72]. Primary antibodies were detected by incubating embryos with Alexa-coupled secondary antibodies for 1 h at RT. Nuclei were stained with DAPI.

Western Analysis
Protein extracts were prepared by adding 30 mL of lysis buffer (25 mM Tris-HCL pH 7.4; 150 mM NaCl; 5 mM EDTA; with PhosSTOP phosphatase and Complete Mini protease inhibitor cocktails; Roche, Indianapolis, IN) to a pellet of 300 injected embryos. Embryos were crushed 4-5 times with a pestle, immediately spun at 16,000 RCF for 15 min at 4uC, and the supernatant was stored at 280uC until use. Samples were thawed on ice and 46 NuPage Running Buffer containing 4% SDS and 10% 2-ME was added. Samples were heated at 80uC for 3-5 min and 20 mL of each sample was run on 4%-12% NuPage Bis-Tris gradient gel (Invitrogen, Grand Island, NY), transferred to nitrocellulose. Membranes were probed overnight at 4uC in Phosphate Buffered Saline+0.1% Tween-20 (PBST)+3% BSA with a poly-clonal Phospho-PKC(pan) (b11 Ser660) antibody (Cell Signaling Technology, Danvers, MA) (1:250) that recognizes a region that includes serine 660 and detects endogenous levels of phosphorylated PKC a, b1, b11, d, e, g, and h. The recognition sequence is conserved in S. purpuratus PKC isoforms. Membranes were washed 3-5 times in room temperature PBST and probed for 1 h at room temperature in PBST+3% Bovine Serum Albumin with an enhanced chemiluminescent anti-rabbit IgG horseradish peroxidase secondary antibody (GE Healthcare, Piscataway, NJ). After 3-5 more washes in PBST, the membranes were developed and imaged.

Luciferase Assays
Promega Dual Luciferase Reporter System (Promega) was used to perform dual luciferase assays. Embryos (350-400) were injected with linearized TopFlash-Firefly Luciferase (REF) and Endo16-Renilla Luciferase plasmids at concentrations of 20 ng/ mL and 10 ng/mL, respectively, along with 10 ng/mL of linearized genomic DNA carrier. The Firefly and Renilla luciferase signals were recorded with a plate style luminometer using Promega's suggested protocol. The level of luciferase activity was normalized to the level of Renilla activity for each experiment. All experiments were repeated three times using separate batches of embryos.