Involvement of Wnt Signaling Pathways in the Metamorphosis of the Bryozoan Bugula neritina

In this study, we analyzed the metamorphosis of the marine bryozoan Bugula neritina. We observed the morphogenesis of the ancestrula. We defined three distinct pre-ancestrula stages based on the anatomy of the developing polypide and the overall morphology of pre-ancestrula. We then used an annotation based enrichment analysis tool to analyze the B. neritina transcriptome and identified over-representation of genes related to Wnt signaling pathways, suggesting its involvement in metamorphosis. Finally, we studied the temporal-spatial gene expression studies of several Wnt pathway genes. We found that one of the Wnt ligand, BnWnt10, was expressed spatially opposite to the Wnt antagonist BnsFRP within the blastemas, which is the presumptive polypide. Down-stream components of the canonical Wnt signaling pathway were exclusively expressed in the blastemas. Bnβcatenin and BnFz5/8 were exclusively expressed in the blastemas throughout the metamorphosis. Based on the genes expression patterns, we propose that BnWnt10 and BnsFRP may relate to the patterning of the polypide, in which the two genes served as positional signals and contributed to the polarization of the blastemas. Another Wnt ligand, BnWnt6, was expressed in the apical part of the pre-ancestrula epidermis. Overall, our findings suggest that the Wnt signaling pathway may be important to the pattern formation of polypide and the development of epidermis.


Introduction
Life began in oceans and marine invertebrates were the first to evolve [1]. Marine larval metamorphosis is therefore more ancient than terrestrial metamorphosis. Understanding marine larval metamorphosis is the key to unmask the evolutionary history of the animal kingdom [1,2,3,4]. To understand how metamorphosis is evolved in different taxa and ultimately the evolution origin of animals, we need extensive understanding on the molecular mechanisms of development in different marine invertebrate taxa. Yet, comparing to what we know about the terrestrial vertebrates and invertebrates, the development of marine invertebrates is obviously understudied.
Bryozoans, also known as ectoprocts [5], belong to the protostomal lophotrochrozoa clade, which is the third major branch of bilaterian animals [6]. The molecular mechanisms of metamorphosis in Bryozoans are largely unknown. Among the phylum, anatomical changes during the initial metamorphosis in Bugula neritina were previously observed in great detail [7,8,9,10]. B. neritina larvae can be obtained in large numbers and their synchronous metamorphosis can be easily triggered [11,12], making it a good species for study. In addition, during metamorphosis, the polypide, consisting of the lophophore, digestive tract, nerve ganglia and most of the musculature, and the cystid, consisting of the epidermis and a lightly calcified chitinous housing, are built de novo [8,13,14]. These dramatic transformations make B. neritina a good model for the study of morphogenesis in bryozoans.
Recently, our lab generated a transcriptome dataset from various metamorphic stages of B. neritina [15]. Based on the results from GO annotation and KEGG mapping, we suggested that Wnt signaling pathways should play a major role during the metamorphosis of B. neritina. The canonical Wnt pathway is activated by the binding of Wnt ligand to the receptor Frizzled [16]. The Wnt/ Frizzled binding inhibits degradation of the key protein b-catenin and leads to the cytoplasmic accumulation ofb-catenin, which is translocated into the nucleus [17]. Nucleatedb-catenin binds with Tcf/Lef transcription factors and activates target genes that regulate cell proliferation [18,19,20]. In non-canonical signaling pathways, activation of down-stream activities is independent of bcatenin and relies on different signal transduction mechanisms [21,22]. While the non-canonical Wnt pathways were implicated in planar cell polarization [23] and convergent extension in tissue growth [24], the canonical Wnt pathway is broadly used by animals, ranging from vertebrates to planarians, to pattern the primary body axis. In pre-bilaterians such as sponges, hydras and cnidarians, which have an oral-aboral axis with overt radial symmetry about it, the canonical Wnt pathway controls animalvegetal axial patterning during embryogenesis as well as oralaboral axial patterning during metamorphosis [24,25]. In bilaterians, the canonical Wnt signaling has been implicated in dorsal-ventral (D-V) axis patterning as well as anterior-posterior (A-P) axis specification during embryonic as well as postembryonic development in nematodes, planarians and various vertebrate models [26,27,28,29]. In nearly all examined animals, Wnts were posteriorly expressed whereas Wnt inhibitors were expressed in the anterior pole. Such a highly conserved expression pattern together with the results from gene perturbation experiments suggested that Wnts may be important universal posteriorizing factors [30,31]. We wondered whether or not and how the Wnt pathway regulates axial patterning in bryozoans. Specifically, we would like to know if Wnts expressions also bias toward the posterior end in bryozoans.
In this study, we firstly studied the anatomy of pre-ancestrula at different time points by Hematoxylin Eosin (HE) staining and Toluidine blue staining. We staged the metamorphosis of B. neritina into different pre-ancestrula stages (the intermediate metamorphic stages). We then preformed DAVID, an annotation based enrichment analytical tool, to identify over-represented KEGG pathways in B. neritina transcriptome. Finally, we profiled the spatio-temporal expression patterns of two Wnts, the antagonist sFRP, three Frizzleds, b-catenin and GSK3b in different preancestrula stages.

Histology of pre-ancestrula stages
All the time points discussed below refer to Fig. 1A and Fig. 1B. A set of portraits (Fig. 1C) modified from [8] and based on the results from histological staining shows the anatomy of B. neritina at various pre-ancestrula stages. The detailed histology of B. neritina larvae was reported in [8] and [9]. In this paper, we will refer to the primary axis of swimming larva and pre-ancestrula as anteriorposterior (A-P) axis and apical-basal axis respectively. The larval A-P axis is defined based on larval swimming direction and is corresponding to aboral-oral axis used in earlier histological studies on bryozoans larvae [7,8,9,10]. In marine benthos biology, the apical-basal axis is generally used to represent the primary axis of sessile invertebrates such as hydras and sponges [32,33]. The basal end is referred as the end where organism attached to the substrate and the apical end is referred as the end furthest from the attachment. The apical-basal axis of pre-ancestrula should not be confused with the cellular axis of epithelial cells.
Previous studies described two phases in the metamorphosis of B. neritina. The first phase is short, characterized by drastic morphological changes, while the second phase is characterized by the gradual development of polypide from the blastemas and the development of cystid from the internal sac. Here, we divided the second phase into three stages: the early pre-ancestrula, the mid pre-ancestrula, and the late pre-ancestrula stages.
The early pre-ancestrula stage (0-4 h post-attachment). In the early pre-ancestrula stage, continuous proliferation of the apex of the epidermis results in the elongation of the body of the preancestrula [9], transforming the pie-shaped pre-ancestrula at 1 h post-attachment into a tubular body at 4 h post-attachment. At 4 h post-attachment, a basal adhesion disc, consisting of fibrous longitudinal cells as shown in both the HE stain and the Toluidine blue stain, can be identified (highlighted in red boxes in Fig. 1A and 1B). A layer of shell material is also visible outside the epidermis (indicated by arrows in Fig. 1A and 1B). The blastemas are internalized during the first phase of metamorphosis, with both the epidermal and mesodermal blastemas fold inward and the epidermal blastema interior to the mesodermal blastema. The blastemas eventually become a U-shaped tissue with an apical-to-basal orientation (Fig. 1). We divided the blastemas into the apical half (apical blastemas), which develops into the lophophore, and the basal half (basal blastemas), which develops into the pharynx and the stomach. During the early pre-ancestrula stage, there is no obvious differentiation of the blastemas, except that the whole tissue increases in length.
The mid pre-ancestrula stage (8-16 h post-attachment). The beginning and duration of this stage are variable. In this stage, the tubular body continues to elongate until it reached around 800 mm. The mesodermal blastema and the internalized larval pallial epithelium differentiate into the tentacle sheath, a layer of cells that separates the epidermal blastema from the parenchyme cells. The basal epidermal blastema first differentiates into a hollow sphere and subsequently into a hollow tube. Concomitantly, the apical epidermal blastema undergoes substantial proliferation and differentiation until, at 12 h post-attachment, a structure with finger-like protrusions is visible.
The late pre-ancestrula stage (24-36 h post-attachment). This stage is characterized by the presence of coleomic cavity. Parenchyme cells extended from the point where the polypide attached to the basal adhesion disc. At this stage, the length of the body no longer increases. The pharynx is emerging from the region above the basal epidermal blastema below the apical epidermal blastema and the basal epidermal blastema are developing into the stomach. The apical epidermal blastema develops into a palm-like structure, with a ring of growing tentacles radiating from the base of the lophophore.

Database for Annotation, Visualization and Integrated Discovery (DAVID) analysis on B. neritina transcriptome
A list of over-represented KEGG pathways is shown in Table 1. Majority of the enriched KEGG pathways are related to fatty acid or amino acid metabolisms. For instance, TCA cycle is found to be 5.99 folds over-represented and was the most enriched KEGG pathway. Several enriched KEGG pathways, such as RNA polymerase (5.07 folds enrichment), Aminoacyl-tRNA biosynthesis (3.11 folds enrichment) and Ribosome (2.83 folds enrichment) are related to translation and transcription. In term of signal transduction pathways, Wnt signaling pathways are found to be over-represented. More than two-folds enrichment is detected. 22 genes in the transcriptome are associated with Wnt signaling pathways.

Temporal gene expression patterns
The results of qRT-PCR assays are shown in Fig. 2. The gene expression levels in swimming larvae were taken as the base point for comparison in each analysis. Among the eight genes, the temporal gene expression patterns of BnWnt6, BnWnt10, BnFz1/2/ 7, BnsFRP, and Bnbcatenin were similar. Their expression levels peaked in the early pre-ancestrula stage, decreased in late the preancestrula stage and then increased in the ancestrula stage to the same level as in the early pre-ancestrula stage. The gene expression levels of BnFz5/8 and BnFz4/9/10 were more stable, with less than a single-fold fluctuation, during the full metamorphosis. BnGSK3b was substantially down-regulated during the two pre-ancestrula stages but was up-regulated during the ancestrula stage.

Spatial gene expression patterns
BnWnt6. The results of BnWnt6 Whole mount in situ Hybridization (WISH) and a section of WISH are showed in Fig. 3. BnWnt6 was expressed in the dorsal side of the wall region of the internal sac. During the attachment process, BnWnt6expressing cells become part of the apical epidermis peripheral to the central axis. At 2 h post-attachment, BnWnt6 expression was  [7] and according to the histology of each stage. A time line for metamorphosis was defined by the overall morphology of the pre-ancestrula stage and the anatomy of the blastemas (the presumptive polypide). Note that the anterior-posterior axis and the apical-basal axis were used to represent the primary body axis of swimming larvae and pre-ancestralae respectively. AB, apical blastemas; AD, adhesion disc; An: anterior; BB, basal blastema; Ep, epidermis; Co, coelom; CS, calcified shell; Lo, (presumptive) lophophore; MB, mesodermal blastema; PC, parenchyme cells; Ph, (presumptive) pharynx; Po: posterior; St, (presumptive) stomach; TS, (presumptive) tentacle sheath. doi:10.1371/journal.pone.0033323.g001 restricted to a small patch of cells on the apical epidermis just below the apex.    In the early pre-ancestrula stage (2 h post-attachment), BnWnt10 was expressed in the apical blastemas while BnsFRP was expressed in the basal tip of the blastemas. Expression of BnWnt10 decreased following an apical-to-basal gradient. The three Frizzled receptors, Bnbcatenin and BnGSK3b continued to be expressed exclusively in the blastemas.
Starting from 4 h post-attachment, visualization of gene expression in the interior became more difficult. We therefore semi-thin sectioned samples at 4 h (early pre-ancestrula stage), 8 h, 12 h, 16 h (mid pre-ancestrula stage) and 24 h post-attachment (late pre-ancestrula stage) before preforming in situ hybridization (refer as Section In Situ Hybridization SISH). Serial SISH sections of BnsFRP, BnFz5/8 and Bnbcatenin were shown in Fig. 5 and Fig. 6. BnFz5/8 and Bnbcatenin were expressed exclusively in the blastemas in all stages (Fig. 6). BnsFRP was expressed at distinct region within the developing polypide at different pre-ancestrula stages (Fig. 5). At 4 h post-attachment, BnsFRP was still expressing at the basal epidermal blastemas. At 8 h post-attachment BnsFRP was expressed at the intersection between the apical and basal epidermal blastema. At 16 h post-attachment, when the digestive tract is emerging from the basal epidermal blastema in the mid and late pre-ancestrula stage, BnsFRP was expressed in the junction between the apical and basal blastemas and the exterior of the emerging digestive tract and but not in the basal pole of the polypide (Fig. 5). At 24 h post-attachment, BnsFRP expression was detectable only at the base of the developing lophophore. The overall gene expression patterns are depicted in Fig. 7.

Discussion
In this study, we aimed at uncovering the molecular mechanisms underlying metamorphosis of B. neritina by analyzing genes expression patterns. A basic histological reference map is needed to correctly interpret the spatial gene expression patterns. Reed et al. [10] studied morphogenetic movements in the initiation phase of metamorphosis in detail. Woollacott et al. [7] divided the metamorphosis into the initial phase and second phase and reported the anatomy of early and advanced (late) pre-ancestrula. In this study, we divided the second phase of the metamorphosis into three pre-ancestrula stages based on the anatomy of the developing polypide and the overall morphology of pre-ancestrula. The mid-preancestrula stage, when the pharynx and stomach are developing from the basal blastemas, was not described previously. This set up a basic histological reference map for subsequent gene expression studies. In addition, we described a novel structure, the basal adhesion disc which is characterized by elongated and interlocked fibrous cells. We speculate that the adhesion disc may provide mechanical reinforcement to pre-ancestrula. Earlier histological studies on B. neritina did not describe this organ.
In our previous study, we sequenced the cDNA generated from various metamorphic stages of B. neritina using high-throughput 454 sequencing technology. A large amount of relatively low abundant but very important genes were detected and mapped to different signal transduction pathways such as Wnt signaling pathways, the Mitogen-activated protein kinase (MAPK) pathway and the cell apoptosis pathway. This led us to hypothesize that these signal transduction pathways have major roles during larval metamorphosis in B. neritina [15]. Yet, we did not perform any enrichment analysis to support our prediction. In this study, we analyzed the B. neritina transcriptome dataset using the Database for Annotation, Visualization and Integrated Discovery (DAVID) tool. The underlying principle of DAVID enrichment analysis is that, in any biological sample, if a biological process is abnormal in a given study, co-functioning genes or genes in relevant groups should have higher chance to be detected by high throughput screening technologies and hence over-represented in the corresponding gene list [34]. A gene list containing 7183 annotated contigs from the B. neritina transcriptome was submitted to DAVID enrichment analysis. We found significant enrichment of KEGG pathways related to energy metabolism and translation and transcription, which suggested intense energy consumption and active cellular proliferation during metamorphosis. This findings are in concert with the results from our previous proteomic analysis, which found substantial up-regulation of  proteins directly involved in or indirectly related to energy metabolism and de novo protein synthesis [35]. More importantly, consistent with our prediction, we found that genes related to Wnt signaling pathways were over-represented in the B. neritina transcriptome, suggesting that Wnt pathways may have important functions in the metamorphosis.
The results from the qRT-PCR assay further supported the argument that Wnt pathways are important signal transduction pathways in the metamorphosis of B. neritina. The relative gene expression levels of BnWnt6, BnWnt10, and downstream component BnFz1/2/7, and Bnbcatenin substantially increased during metamorphosis while an opposite trend in the expression level of the negative regulator, BnGSK3b, was observed. Co-expression of BnWnt10 and Bnbcatenin in the blastemas suggested that the canonical Wnt signaling pathway is involved in the development of polypide from the blastemas. In the larval stage and the early preancestrula stage (2 h post-attachment), BnWnt10 and BnsFRP expressions were spatially opposite to each other in the blastemas. Throughout metamorphosis, BnsFRP expression was detected in the bottom part of the apical blastema, the upper part of the basal blastema and the exterior of the developing digestive tract, but never at the top of the apical blastemas.
Wnt ligands are known morphogens [36,37] while sFRP is one of the antagonists that counteract Wnt signaling [38,39]. The Wnt/B-catenin activity gradient as well as Wnt inhibitors such as sFRP expression act as graded positional cues to establish the primary body axis and latter direct cell specification in embryogenesis, post-embryonic development and development of multiple tissues [36,37,40,41]. For instance, Wnt genes participate in epithelial-mesenchymal signaling and may specify region identity in the anterior foregut in mouse embryo [42]. Based on the spatial gene expression patterns and knowledge on the function of Wnts and sFRP in tissues patterning, we hypothesize that, as early as the larval stage, BnWnt10 and BnsFRP set up a local positional signal in the blastemas. During metamorphosis, local expressions of BnWnt10 and BnsFRP patterned the blastemas. BnWnt10 expression triggers the canonical Wnt pathway and results in the development of lophophore from the apical blastemas. On the other hand, BnsFRP inhibits the canonical Wnt pathway activity, leading to the development of digestive tract from the basal blastemas.
It was proposed that the central role of Wnt signaling is to promote posterior rather than anterior aspects of animal tissues [30,31]. In almost all examined bilaterians, Wnts were expressed in the vegetal pole of embryos or the posterior of larval or postembryonic forms [30]. The A-P axis, which is the primary axis of bilaterians, defines the mouth as the anterior and the anus as the posterior [43]. Bryozoans have a U-shape digestive tract. The mouth and anus lie adjacent to each other and point at the same direction, the apical pole. Based on the development of polypide, the apical epidermal blastema, which is the precursor of the feeding apparatus lophophore, should be regarded as the anterior whereas the basal epidermal blastema, which is the precursor of the whole digestive tract, should be regarded as the posterior. The expressions of BnWnt10 in the apical epidermal blastema and BnsFRP in the basal epidermal blastema are therefore right opposite to that of other bilaterians. We suggest that the role of Wnt in A-P axial patterning in the polypide of B. neritina may be different from that in the organogenesis in other bilaterians, with the pre-cautions that there may be additional Wnts involve in the development of polypide and we have only examined the expression pattern of one of these Wnts.
The expression pattern of BnsFRP in the mid and late preancestrula stage suggested that BnsFRP may serve as an posterior signal in the emerging lophophore and, at the same time, relate to the differentiation of the pharynx, which is the anterior portion of the digestive tract. Numerous spatial expression studies have pointed out the requirement of sFRP in gut development in both vertebrates and invertebrates [44,45,46]. For instance, sFRP1 and sFRP2 might mediate mesenchymal induction of stomach epithelium in mouse [42]. Inactivation of sFRP1 and sFRP2 leads to reduction in fore-stomach length in mouse embryos [47]. Negative regulation of Wnt5a signaling by sFRP was suggested to control oriented cell division and apicobasal polarity in the epithelium of developing gut [47].
The exact functions of BnWnt10 and BnsFRP have to be confirmed by gene perturbation experiments. If BnWnt10 and BnsFRP were indeed patterning the lophophore or the polypide as a whole, perturbation of their expression should give rise to distinct phenotypes. We predict that, for example, knockdown of BnWnt10 will produce ancestrulae with defective lophophores or ectopic digestive tracts, and vice versa for knockdown of BnsFRP. Unfortunately, gene manipulation techniques have not been established in any bryozoan species.
Another Wnt ligand, BnWnt6, was expressed in the apical part of the developing epidermis. The pre-anccestrula epidermis will become the living part of the ancestrula cystid and is responsible for synthesis of the chitinous calcified shell during metamorphosis. In chick embryo, Wnt6 acts through the bcatenin-independent non-canonical pathway. And Wnt6 expression, derived from the ectoderm, is necessary for chick neural crest induction [48]. In this study, we did not detect co-localization of BnWnt6 and Bnbcatenin expression, suggesting that BnWnt6 also acts through thebcateninindependent non-canonical pathway. So far, there is no study reporting the function of that particular apical field of epidermal cells. Whether BnWnt6 expression is related to the axial development of ancestrula or associated with stem cells [49] in the epidermis remains to be elucidated. Additional spatial expression analysis of BnWnt6 on the latter pre-ancestrula and ancestrula stages may give us more hints on the possible role of BnWnt6 in the morphogenesis of ancestrula. Interestingly, Fuch et al. [50] reported that another Wnt ligand, BnWnt1, was expressed in a different region of the internal sac -the neck-wall region junction. Our result could be supportive of their claim that the internal sac is regionalized and its cells might be differentially involved in adult body wall patterning.
To conclude, we have examined the gene expression patterns of various components of Wnt signaling pathways. The results suggested that the canonical Wnt pathway may be involved in the development of polypide. In particular, apical BnWnt10 and basal BnsFRP expression within the blastemas suggested a possible role for the canonical Wnt pathway in patterning the polypide as a whole or only the lophophore. However, BnWnt10 and BnsFRP expression patterns are different from a posterior Wnts and anterior Wnt inhibitors expression patterns observed in other bilaterians. On the other hand, BnWnt6 was expressed in the apical epidermis. We speculate that BnWnt6 acts through non-canonical Wnt pathway and may be associated with the development of the epidermis. This study suggested, for the first time, a possible role of Wnt signaling in pattern formation in bryozoan lophophores/ polypides.

Larval sample preparation
Adult B. neritina colonies were collected from the floating rafts of a fish farm in Trio Beach, Hong Kong (22u21919 N, 114u16915E) between February and April, 2010. Permission to collect B. neritina colonies from the fish farm was given by the fish farm owner. Adult colonies were maintained in a 21uC flow-through seawater system at the Coastal Marine Laboratory (Hong Kong University of Science and Technology) for no more than 7 days before use.
Free-swimming larvae were collected according to the procedures described in our previous study [15]. B. neritina larvae were induced to metamorphose by placing them in the dark for one hour. Larvae that did not attach after one hour were discarded. At different time points after their initial attachment, samples of the various pre-ancestrula stages (intermediate stage of metamorphosis) specified by the number of hour(s) post-attachment and the ancestrula stage (the first zooid or juvenile, 52 h post-attachment) were collected for sectioning or whole mount in situ hybridization. They were first fixed with fixative and then scraped off from the petri-dish, or, for molecular studies, directly scraped off and stored under cryogenic conditions until further use.

Specimen embedding and histological staining
Swimming larvae and pre-ancestrulae at various time points (1 h, 2 h, 4 h, 8 h, 16 h and 24 h post-attachment) were fixed in 4% paraformaldehyde in autoclaved filtered seawater for 1 h at room temperature. Samples were then dehydrated in ethanol and stored at 220uC until further use. Paraffin (McCormick Scientific, Richmond, IL, United States) embedding was performed according to the standard protocol. Semi-thin sections (7 mm) were prepared by a Leica 820II microtome (Leica instruments GmBH, Nussloch, Germany). Around 10 sections were placed on one polylysine coated slide and allowed to dry under a frame hood for 2 h. The sections were then de-waxed in xylene for 5 min and then rehydrated in descending concentrations of ethanol (100%, 95%, 90%, 75%, 50%, MilliQ water). Hematoxylin and Eosin staining and Toluidine blue staining were preformed according to [51]. Histological stained sections were then mounted on a Dpx mountant (Fluka, St. Louis, MO, United States) and visualized with a microscope under a bright-field illumination setting.

DAVID analysis of the Bugula neritina transcriptome
The B. neritina transcriptome was blasted against the newest version of the uniprot database (released in July 2011) in the blastx mode using in-house scripts. With the E value cutoff at 1e 28 , a total of 7234 genes had significant matches in the database. This gene list (uniport accession numbers provided in the supplementary information) was submitted to the DAVID webpage (DAVID v6.7) [52] for analysis. The detailed procedures and underlying statistics of DAVID can be found in [34] and [53]. The human (Homo sapiens), mouse (Mus musculus), zebrafish (Danio rerio), fruitfly (Drosophila melanogaster), and nematode (Caenorhabditis elegans) genomes were selected as the background, because these genomes have the most detailed annotations available. The results of the enrichment analysis by DAVID were filtered with the enrichment P value cutoff at 0.01 which is more stringent than recommended (#0.05) [53].

Rapid amplification of cDNA ends (RACE)
Total RNA was isolated from swimming larvae and samples from the early pre-ancestrula (4 h post-attachment), the late preancestrula (24 h post-attachment), and the ancestrula stages using TRIzol Reagent (Invitrogen, Carlsbad, CA, United States). Total RNA extraction and cDNA synthesis procedures for the RACE template are given in [15].
Gene specific primers (GSPs) of Wnt pathway-related genes were designed based on the annotated sequence of the reads from the B. neritina transcriptome (NCBI accession number: SRA010777.2). The sequence of the reads and the corresponding gene-specific primers are listed in Table S1. The cDNA template for the 59RACE was prepared by adding an oligo dC tail to the 39cDNA ends by a terminal deoxynucleotidyl transferase (TdT) (USB, Cleveland, Ohio, United States) reaction. Nested PCR was performed such that the gene-specific 59RACE1 primer and an oligodG(10)-adaptor primer were used for the first round of PCR. Gene-specific 59RACE2 primer and adaptor primer were used for the second round of PCR. The 39 RACE was the same as the 59 RACE except that the cDNA template was not modified by the TdT reaction. An oligodT adaptor primer was used instead of an oligodG adaptor primer and gene-specific 39RACE primers were used. PCR products of both 39 and 59 RACE were gel purified and then ligated to the pMD18-simple T vector (TaKaRa Bio Inc., Dalian, China). Transformations, insert screening, and sequencing reactions were performed as described in [15].
Gene orthology assignment and signal peptide prediction Gene orthology assignments for all genes were determined using the bioinformatics software Mega 5.0 [54]. Reference gene sequences from different organisms were downloaded from the NCBI protein sequence database. Prior to phylogenetic analysis, reference sequences together with the corresponding B. neritina genes were aligned using clusterW according to their amino acid sequences. In Wnt genes orthology assignment, mismatches in the alignment was manually corrected according to the conserved cysteine residues and two Asn-linked glycosylation sites. We used the Maximum likelihood method to conduct phylogenetic analysis. One thousand bootstraps were calculated. Signal peptide prediction was carried out by the SignalP 3.0 Server [55] Quantitative real-time polymerase chain reaction (qRT-PCR) To capture the temporal expression patterns during metamorphosis, four stages -swimming larvae, early pre-ancestrulae (4 h post-attachment), late pre-ancestrulae (24 h post-attachment), and ancestrulae were chosen for qRT-PCR analysis. Total RNA extraction and cDNA synthesis procedures were the same as above except that a random hexamer primer was used in cDNA synthesis. The primer sequences are listed in Table S2. qRT-PCR assays for each gene were performed in triplicate. All of the qRT-PCR assays were carried out using iTaqSYBR Green Supermix with ROX (BioRad Life Science, Hercules, CA, United States) and were run on a Stratagene mx3000p PCR machine (Agilent Technologies, Santa Clara, CA, United States). The results were normalized using the housekeeping gene 18S of B. neritina. The relative changes were calculated using the 22DDCT method [56].
In situ hybridization (whole-mount ISH (WISH) and sections ISH (SISH)) The DNA templates for ISH RNA probe synthesis of each gene were prepared by PCR amplification of the 39 RACE clone using the gene-specific 39RACE2 primer and the T7-adaptor primer. Digoxigenin-labeled probes were synthesized from the PCRproduct templates according to the protocol supplied with the DIG RNA labeling kit (Roche Diagnostics, Nutley, NJ, USA).
For WISH, swimming larvae and 1 h post-attachment and 2 h post-attachment pre-ancestrulae were fixed in 3.7% formalin in AFSW overnight at 4uC. To collect the pre-ancestrulae samples, swimming larvae were allowed to attach and then fixed on a 90 mm petri dish. Briefly, the fixed samples were dehydrated in 100% methanol at 220uC until further processing. The samples were rehydrated by descending concentrations of methanol in PBS-0.1%TritonX100 (PBST). The larvae were washed and penetrated by proteinase K (New England Biolabs, Ipswich, MA, USA) treatment. Swimming larvae were incubated for 7 min while samples from the two pre-ancestrula time points were incubated for 15 min. Samples were post-fixed by 4% PFA and washed before pre-hybridization at 56uC for 1 h in a hybridization mix. RNA probe hybridization was conducted at 56uC overnight. In each hybridization, no more than a 30 ng antisense RNA probe was used. The hybridized samples were washed with hybridization mix (without sperm DNA or heparin). The samples were blocked for 1 h and then incubated in 1:5000 antiDIG-AP antibody (Roche Diagnostics, Mannheim, Germany) at 4uC overnight with orbital rotation. They were washed and then treated with alkaline Tris buffer (100 mM Tris-HCl pH 9.5, 100 mM NaCl, 50 mM MgCl2, 0.1% Tween20) before further incubation in AP stain development solution (Roche Diagnostics, Mannheim, Germany) at room temperature in the dark. Staining was stopped by TE buffer when the samples developed a blue or purple color band. The samples were visualized under a microscope (Olympus, Tokyo, Japan) with a dark-field illumination setting. B. neritina samples were not fully transparent even after the WISH protocol, sectioning was necessary to visualize the precise expression pattern, especially the genes expressed in the interior of the specimen. WISH specimens were embedded in paraffin, semi-thin sectioned, de-waxed, mounted, and visualized as described above.
For SISH, pre-ancestrulae at 4 h, 8 h, 16 h, 24 h postattachment were fixed in 4% PFA for 1 hour. Embedding and paraffin sectioning by microtome were completed as described above. Sections were de-waxed in freshly prepared xylene for 5 min and rehydrated in a descending concentration of ethanol (100%, 90%, 75%, DEPC treated MilliQ water, freshly prepared). Penetration by Proteinase K, post-fixation, pre-hybridization, antisense probe hybridization, probe washing, non-specific antigen blocking, antibody incubation, and stain development were the same as for WISH except all the procedures were carried out on glass slides.