Evolutionary conservation of the placodal transcriptional network during sexual and asexual development in chordates



Background: An important question behind vertebrate evolution is whether the cranial placodes originated de novo, or if their precursors were present in the ancestor of chordates. In this respect, tunicates are of particular interest as they are considered the closest relatives to vertebrates. They are also the only chordate group possessing species that reproduce both sexually and asexually, allowing both types of development to be studied to address whether embryonic pathways have been co-opted during budding to build the same structures. Results: We studied the expression of members of the transcriptional network associated with vertebrate placodal formation (Six, Eya, and FoxI) in the colonial tunicate Botryllus schlosseri. During both sexual and asexual development, each transcript is expressed in branchial fissures and in two discrete regions proposed to be homologues to groups of vertebrate placodes. Discussion: Results reinforce the idea that placode origin predates the origin of vertebrates and that the molecular network involving these genes was co-opted in the evolution of asexual reproduction. Considering that gill slit formation in deuterostomes is based on similar expression patterns, we discuss possible alternative evolutionary scenarios depicting gene co-option as critical step in placode and pharynx evolution. Developmental Dynamics 242:752–766, 2013. © 2013 Wiley Periodicals, Inc.


The origin of vertebrates and the appearance of their novelties is still a debated issue. According to the “New Head” hypothesis, key innovations for vertebrate success were the neural crest and the neurogenic placodes (Northcutt and Gans, 1983). Thus, it is of interest to analyze if these crucial components originated de novo in the vertebrate lineage, or if they had precursors in the common ancestor of all chordates. In addition to vertebrates, the chordates include the cephalochordates and the tunicates, with the latter now considered most closely related to vertebrates (Delsuc et al., 2008). Sequenced tunicate genomes are very compact compared with those of vertebrates (Satou et al., 2008), and their study helps reconstruct the genetic situation of the common ancestor (Lemaire, 2011). Therefore, the tunicates can be useful living models for the study of the evolution of structures which have evolved high complexity in vertebrates (Shimeld and Holland, 2000).

In vertebrates, the cranial placodes are patches of thickened embryonic ectoderm that give rise to many sense organs and ganglia of the vertebrate head. The network of regulatory links between transcription factors involved in vertebrate placode development is still incompletely characterized, but a set of genes marks a contiguous region of non-neural ectoderm early in embryogenesis, the so-called preplacodal domain that gives rise to anterior (adenohypophyseal, olfactory, and lens) and posterior (otic, epibranchial, and lateral line) placodes. Of interest, a few of these genes, such as those of the Eya, Six1/2, and Six4/5 gene families, continue to be expressed in all placodes and their derivatives (reviewed in Schlosser, 2010). Orthologues of these transcription factors were found to have an expression domain adjacent to the anterior neural plate border and/or in sensory cells differentiating from non-neural ectoderm in both tunicates and in the cephalochordate amphioxus (Bassham and Postlethwait, 2005; Mazet et al., 2005; Kozmik et al., 2007), and more recently also at the anterior-most rim neuroectoderm of an insect (Posnien et al., 2011). However, the entire molecular pathway characterizing the cranial placodes can be found only in vertebrates (Holland and Holland, 2001; Holland, 2005; Schlosser, 2005, 2007). Intriguingly, Six1/2, Eya, and FoxI are similarly expressed during gill slit formation in all investigated deuterostomes developing these structures (Sahly et al., 1999; Solomon et al., 2003; Bessarab et al., 2004; Kozmik et al., 2007; Schlosser, 2007; Gillis et al., 2012).

In addition to molecular evidence, developmental and structural data support the idea that tunicates possess embryonic areas comparable to some cranial placodes of vertebrates (reviewed in Graham and Shimeld, 2013). In Ciona intestinalis embryos there are two thickened ectodermal domains, an anterior stomodeal and a posterior atrial, from which sensory organs derive (Manni et al., 2004). Each domain expresses a subset of placodal genes, such as Pax, Six, Eya, and FoxI and is proposed to have homology to vertebrate olfactory-adenohypophyseal and otic-lateral line placodes, respectively (Mazet et al., 2005).

Tunicates are a very diverse group and include both solitary and colonial species. Colonial species can form similar zooids through the different developmental pathways found in sexual and asexual reproduction, allowing comparison of these processes in the same organism. The colonial tunicate Botryllus schlosseri is a model species for this kind of study (Manni and Burighel, 2006); moreover, it is phylogenetically distant from C. intestinalis (Tatian et al., 2011), so comparison between the two species helps to reconstruct the hypothetical basal tunicate state. Three asexual (blastogenic) generations coexist in the colony: the adults, their buds (primary buds), which in turn produce budlets (secondary buds) (Fig. 1A). The bud rudiment appears as a thickened disc on the zooid wall. It initially organizes in a double vesicle, with the inner epithelium derived from the parental peribranchial leaflet and the outer one from the parental epidermis. The inner vesicle gives rise to the main organs, such as the branchial and peribranchial chambers, the gut, and the nervous system. Other organs, such as the heart and the gonads, derive from mesenchymal cells that arrive in the bud by means of blood circulation (Fig. 1B) (Manni et al., 2007).

Figure 1.

Botryllus schlosseri life cycle (A) and secondary bud development (B). A: Ozooid and blastozooids are shown in dorsal view; modified from (Manni and Burighel, 2006). B: Budlet development is depicted from its appearance until a new generation occurs at its peribranchial epithelium. Ventral view, modified from (Manni et al., 2007).

Bud and embryo develop following different mechanisms with a completely different starting point; the zygote in embryogenesis and specific pluripotent somatic territories of the parental zooid in blastogenesis. However, for some structures, such as the nervous system and the branchial stigmata, developmental similarities have been found (Manni et al., 1999, 2002). For example, the nervous system of the larva differentiates from the neural plate, whereas in blastozooids, it derives from a territory of the inner bud vesicle, however in both it shows a placodal-like derivation with several ultrastructural similarities (Burighel et al., 1998; Manni et al., 1999). In addition, a molecular study has shown expression of the Pitx gene, which is normally involved in vertebrate adenohypophysial placode development, in nervous system formation during both developmental pathways (Tiozzo et al., 2005).

We here characterize B. schlosseri orthologues of Six1/2, Six3/6, Eya, and FoxI, and report their spatiotemporal expression patterns during both embryogenesis and blastogenesis. Our results show that these genes are expressed both in larva and bud during branchial fissure formation, and in two domains along the anterior–posterior axis. We hypothesize that the latter are placodal homologue territories that can be recognized during not only sexual but also asexual development of tunicates, and discuss this in the context of both placode ancestry and the co-option of gene networks.


Molecular and Phylogenetic Analyses of cDNA Clones of Six1/2, Six3/6, Eya, and FoxI from Botryllus schlosseri

From initial reverse transcriptase polymerase chain reaction (RT-PCR) and additional 5′- and 3′-rapid amplification of cDNA ends (RACE) experiments, we obtained cDNA clones coding for two members of the Six gene subclass, one of the Eya gene family and one of the FoxI gene family (from EMBL ID: HE681849 to EMBL ID: HE681860). The predicted amino acid sequence of each transcript (data not shown) was used for molecular phylogenetic analyses (Figs. 2-4). In accordance with the phylogenetic analyses, we name the B. schlosseri gene sequences as BsSix1/2 (B. schlosseri Sine oculis homeobox 1/2), BsSix3/6 (B. schlosseri Sine oculis homeobox 3/6), BsEya (B. schlosseri Eyes absent) and BsFoxI (B. schlosseri Forkhead box I), respectively.

We note that cDNA clones of Six1/2 that have been recovered differ from each other in few positions, as do clones of Eya (and resulting in one case in a non synonymous change; data not shown). Probably these are allelic variants. Moreover, none of our Six1/2 3′-RACE clones possessed a poly A tail at 3′ end, suggesting that the identified 3′ untranslated region (UTR) is incomplete. The presence of the spliced leader sequence (Gasparini and Shimeld, 2011) at the 5′ end of the Six and Eya cDNA clones (data not shown) confirms that the full 5′ UTRs have been characterized. Finally, we also note that our FoxI transcript is probably incomplete in its open reading frame (ORF) at the 5′ end, as the putative protein is shorter at this terminal than FoxI proteins in C. intestinalis (data not shown). We cannot exclude, however, that B. schlosseri may possess different transcripts derived from alternative splicing, among which our short transcript could have a complete ORF starting from the first Methionine codon. We also cannot exclude the possibility that B. schlosseri may have additional FoxI paralogs, considering that three FoxI genes have been identified in the C. intestinalis genome (Yagi et al., 2003), and that our phylogenetic analysis (Fig. 3) has insufficient resolution to robustly resolve relationships between ascidian FoxI genes.

Figure 2.

Molecular phylogenetic analysis of Six protein family. Unrooted phylogram constructed with Bayesian methodology and based upon the Six homeodomain and Six-type domains. B. schlosseri sequences (dark highlighted) fall into groups with their respective orthologs. Six1/2, Six3/6, and Six4/5 clades are encircled in gray; in the Six1/2 and Six3/6 clades, the vertebrate sequences are grouped together. Numbers adjacent to nodes indicate posterior probabilities. The scale bar represents 0.1 amino acid substitutions per site. Each sequence is named to reflect the respective species: Bs, Botryllus schlosseri; Bf, Branchiostoma floridae; Ci, Ciona intestinalis; Dm, Drosophila melanogaster; Gg, Gallus gallus; Mm, Mus musculus; Od, Oikopleura dioica; Pd, Platynereis dumerilii; Tc, Tribolium castaneum; Xt, Xenopus tropicalis.

Figure 3.

Molecular phylogenetic analysis of Eya protein family. Bayesian tree constructed using the predicted amino acid sequences coding for Eya from various organisms. The analysis is based upon a trimmed alignment that includes the EYA domain. B. schlosseri sequence (dark highlighted) falls into a clade that includes sequences from other tunicates (gray box). Numbers adjacent to nodes indicate posterior probabilities. The scale bar represents 0.1 amino acid substitutions per site. Sequences are named as in Figure 2, apart from: Sp, Strongylocentrotus purpuratus; Dj, Dugesia japonica.

Figure 4.

Bayesian analyses of Fox protein family. A: Simplified unrooted phylogenetic tree based upon a trimmed alignment that includes the fork head domain of Fox proteins from different organisms. The B. schlosseri sequence (dark highlighted) falls into the FoxI clade (gray box). B: Unrooted Bayesian analysis using a trimmed alignment of FoxI sequences, in which B. schlosseri FoxI is clustered with C. intestinalis FoxIb. The vertebrate sequences are grouped together. Numbers adjacent to nodes indicate posterior probabilities. Scale bars represent 0.1 amino acid substitutions per site in both trees. Sequences are named as in Figure 2.

Expression Analysis During Blastogenesis Indicates That Placodal Homologue Territories Can Be Recognized in Asexual Development of Tunicates

In situ hybridizations (ISHs) for each characterized transcript were performed during blastogenesis (Figs. 5, 6). No expression was detected in secondary bud primordia in the prospective parental peribranchial wall before thickening occurs, although we cannot exclude an undetectably low level of gene expression. In the initial stages, when the bud appears as an elongated double vesicle, the BsSix1/2 probe marked two defined areas recognizable as the presumptive regions of the peribranchial epithelium (Fig. 5A). These regions are subject to invagination movements and subsequently elongate as folds to divide the inner vesicle into the branchial and peribranchial chambers. In the following developmental stages, BsSix1/2 expression was maintained during the differentiation of the peribranchial chambers extending to the atrial primordium (Fig. 5B), which derives from posterodorsal fusion of peribranchial chambers. The transcript was also detected in the areas involved in siphon formation. Each siphon rudiment is constituted of two thickened epithelial discs: an outer epidermal disc, and an inner one belonging to the branchial or atrial epithelium (in the case of the oral siphon or atrial siphon, respectively). BsSix1/2 marked the inner discs in both siphon rudiments. At the level of the differentiating oral siphon, the signal was not uniformly distributed but mainly localized in areas facing the lumen on opposite sides; these areas represent regions from which tentacles will develop (Fig. 5C). In the adult, the eight tentacles form a ring that extends from the small ectodermic fold of the velum.

Figure 5.

ISH of BsSix1/2 and BsEya-C during blastogenesis. A–C: BsSix1/2 expression in secondary bud. Arrows on (A) indicate stigmata of the primary bud. Arrowheads point to presumptive peribranchial epithelium on (A), atrial primordium on (B) and labeled velum and tentacle rudiments of the oral siphon on (C). Insets show this diagrammatically with blue marking expression, and with plane of section shown on (B) and (C). Arrows here indicate dorsoventral axis (arrow tip is dorsal), with the perlateral axis indicated as a bar perpendicular to this. D,E: BsSix1/2 expression in primary bud. D: The arrow marks the oral siphon. Arrowheads mark the stigmata. F: BsSix1/2 expression in adult blood cells (arrowheads). G: BsEya-C expression in ventral region of a secondary bud (arrowheads). The inset shows this diagrammatically. H,I: BsEya-C expression in primary buds in siphons (arrows) and stigmata (arrowheads). The early secondary bud lacks expression. J: Diagram of a primary bud showing BsSix and BsEya-C co-expression in blue. Each dotted arrow gives the position of the respective section (D,E,H,I) and these are also shown schematically alongside. bc, branchial chamber; be, branchial epithelium; ep, epidermis; en, endostyle; g, gut; pb, primary bud; pc, peribranchial chamber; pe, peribranchial epithelium; sb, secondary bud; os, oral siphon; tn, tentacles. Scale bars = 20 μm.

Figure 6.

In situ hybridization (ISH) of BsSix3/6 and BsFoxI during blastogenesis. A–C: BsSix3/6 expression in the primary bud. A,B: Black arrowheads indicate expression in the anterior cerebral ganglion, black arrows expression in the ciliated duct and white arrows the anterior nerves. C: Arrowheads show expression in the oral siphon. Underneath each image is a diagram showing the plane of section (dotted arrows) and expression domains in blue. D,E: BsFoxI expression in the secondary bud (arrow) and peribranchial epithelium (arrowheads). The diagrams above show this schematically, with blue indicating expression and plane of section indicated for (E). F: Summary expression patterns of the four genes during blastogenesis. Budlet development is depicted on the left from its appearance until formation of atrial folds. On the right is a primary bud. The dotted arrow gives the position of the section in D. Structures that expressed various genes are color coded as indicated. bc, branchial chamber; g, gut; cg, cerebral ganglion; nc, neural complex primordium; n.d., not determined; os, oral siphon; pc, peribranchial chamber; tn, tentacle. Scale bars = 20 μm.

In the primary bud, BsSix1/2 expression was maintained but uniformly distributed along the siphon rudiments and in the stigmata primordia (Fig. 5A,D–E). The stigmata form from discrete thickened areas of the peribranchial and branchial epithelia. At this level, the peribranchial epithelium evaginates, then contacts and fuses with the branchial epithelium, giving rise to the branchial fissures (Manni et al., 2002). BsSix1/2 expression was maintained in the stigmata primordia all through their differentiation. The adult blastozooid completely lacked BsSix1/2 expression, except for some labeled blood cells scattered in the circulatory system (Fig. 5F).

For the ISH of BsEya, we constructed two different probes, because in C. intestinalis alternate splicing of the Eya gene is known to generate different transcripts with different spatial expression (Mazet et al., 2005). The first probe (BsEya-N) is designed in the 5′ region of the transcript, which in C. intestinalis Eya is differently spliced (Mazet et al., 2005). The second probe (BsEya-C) contains the Eya domain encoded by the 3′ region, and known to be present in all the identified C. intestinalis splice variants.

BsEya-N showed expression limited to some circulating blood cells (data not shown). However, the BsEya-C probe showed an expression pattern (Fig. 5G–I) that mainly corresponds to that obtained for BsSix1/2 (Fig. 5J), supporting the hypothesis that Eya protein regulates Six protein function as a cofactor (Jemc and Rebay, 2007). As well as in some circulating blood cells, BsEya-C first appeared in the secondary bud, when the vesicle begins to elongate according to its anteroposterior axis, although at this stage its expression differed from BsSix1/2, being confined to a single ventral domain (Fig. 5G). In the primary bud, BsEya-C expression followed precisely the pattern seen for BsSix1/2 in the developing siphons and stigmata (Fig. 5H,J).

BsSix3/6 expression becomes recognizable in the primary bud, at the level of the neural gland (including its ciliated duct) and the cerebral ganglion (Fig. 6A,B); in the latter, it was confined to a few neurons enclosed between the two anterior nerves. During oral siphon development, expression was identified only in the branchial epithelial disc, which will form the siphon inner wall, an area covered with tunic that extends until the tentacle/velum region (Fig. 6C). BsSix3/6 was maintained during the development of all these structures until the adult stage.

BsSix3/6 expression in the developing neural complex partially overlaps the expression domain of the Bs-Pitx, which includes the ciliated duct and the same region of cerebral ganglion (Tiozzo et al., 2005). In vertebrates, Six3/6 genes express in some nervous structures (i.e., forebrain, retina, and rostral placodes), promoting and regulating their development (Schlosser, 2006; Kumar, 2009), and Pitx genes are necessary for the differentiation of adenohypophyseal endocrine cells, and mesencephalic GABAergic and dopaminergic neurons (Abeliovich and Hammond, 2007; Zhu et al., 2007; Waite et al., 2011). Both are anteriorly confined preplacodal genes, and it has been suggested they may define an extended anterior (adenohypophyseal, olfactory, lens) placodal area (Schlosser, 2010). From our results, we suggest that, in B. schlosseri, Six3/6, and Pitx are together required for the differentiation of the ciliated duct and anterior part of cerebral ganglion, confirming that the anterior part of neural complex is a putative placodal homologue as previously proposed (for review, see Mackie and Burighel, 2005; Manni and Burighel, 2006).

ISHs for each gene were performed at early blastogenetic stages, with only BsFoxI becoming recognizable very early when signal covered the entire primordium of the secondary bud when it appeared as a thickened disc on the parental peribranchial wall (Fig. 6D). As differentiation proceeded, the signal lost this ubiquitous pattern, and when the inner vesicle began to subdivide by folding into the branchial and peribranchial chambers, BsFoxI localized to the thickened prospective peribranchial epithelium (Fig. 6E) from which also differentiates atrium and the respective atrial siphon epithelium (Manni et al., 2007). In the primary bud, expression was restricted to the peribranchial epithelium facing the branchial sheet (Fig. 6D). Here, it was maintained until the stage at which the stigmata primordia were represented by discrete thickened areas.

Our data reveal that Eya, Six1/2, Six3/6, and FoxI genes have specific spatiotemporal expression patterns in blastogenesis (Fig. 6F). In particular, during differentiation of siphons we identify expression of: (i) BsFoxI in secondary bud at the level of the prospective region of atrial epithelium, (ii) BsSix1/2 and BsEya in secondary bud in branchial/atrial epithelium when these two layers come into contact with epidermis to form respectively the oral and atrial siphons of the adult, (iii) BsSix3/6 in primary bud at the level of the branchial epithelium of siphon promordia, then maintained until it evaginates to contact the overlying oral epidermis in the adult. Previous data showed that the oral siphon rudiment also expresses Bs-Pitx (Tiozzo et al., 2005). In particular, Bs-Pitx expression is seen on the branchial evagination as a ring that surrounds the future siphon aperture and persists during velum and tentacle formation. Both siphons are rich in primary sensory cells, sensitive to vibrations, whereas the oral tentacles and velum also possess the secondary sensory cells of the coronal organ, sensitive to contact (Mackie et al., 2006; Burighel et al., 2011). The above data show that these genes are localized to territories which undergo thickening and evagination, and from which sensory cells differentiate. These morphogenetic events also recall the formation of placodal-derived structures.

Comparison of Gene Expression in Sexual and Asexual Development Indicates a Co-option of the Related Transcriptional Network in Blastogenesis

We also studied the expression profile of each gene in embryos, larvae and oozooids at different stages (Fig. 7). BsSix1/2 expression was not detected during embryonic and swimming larval phases. Signal first appeared when the larva settled and began metamorphosis. At this stage, the probe marked the siphon rudiments in the inner epithelial components (branchial and atrial, respectively) and the oozooid stigmata rudiments (protostigmata) (Fig. 7A,B). In both cases, BsSix1/2 transcript was recognizable during differentiation of these structures, but was absent in the adult oozooid. BsSix3/6 was expressed in oral siphon, neural gland and cerebral ganglion from metamorphosed larva (Fig. 7A,C) to adult oozooid. In contrast to blastogenesis, the differentiating ganglion was entirely labeled and did not show regionalized expression.

Figure 7.

Expression of BsSix1/2, BsSix3/6, BsEya-C, and BsFoxI in sexual development. A: Sketched oozooid summarizing the expression patterns of the four genes. B–D: BsSix1/2, BsSix3/6, and BsEya-C expression in metamorphosed larva. The oral siphon (white arrows) expressed all three genes, the atrial siphon (black arrows) and the prostigmata express BsSix1/2 and BsEya-C. E,F: BsFoxI expression in the late tail bud (E) and swimming larva (F). Expression marks peribranchial chamber epithelium (arrowheads) and two opposite regions (gray arrows) of the bud. Each inset diagram shows the plane of section. E,F: Underneath each image is a diagram showing the expression domains in blue. Dotted lines give the position of the sections shown above. Black arrows indicate anteroposterior axis (arrow tip is anterior), with the perlateral axis indicated as a bar perpendicular to this. ac, atrial chamber; as, atrial siphon; ba, blood ampulla; bc, branchial chamber; cd, ciliated duct of the neural gland; cg, cerebral ganglion; en, endostyle; g, gut; ng, neual gland; ns, larval nervous system; os, oral siphon; pc, peribranchial chamber; ph, photolith; pt, protostigmata; ta, tail. Scale bars = 20 μm. Lower diagrams after (Manni and Burighel, 2006).

The BsEya-C probe gave results similar to those obtained with Six1/2, as also seen during the blastogenesis (Fig. 7A,D). BsFoxI was expressed starting from late tail bud stage (just before hatching) in the prospective peribranchial chambers (Fig. 7E). As for the bud, the expression was initially extended to the whole peribranchial epithelium, then became first limited to the regions facing the branchial component and, eventually, restricted to the prospective protostigma areas in larvae which had completed metamorphosis. In the swimming larva, the expression of BsFoxI was also recognizable in the early bud, labeling two opposite regions of the inner vesicle (Fig. 7F). This bud originates from somatic cells of the embryo and will give rise to the first blastozooid of the colony.

Our data offer the possibility to compare in the same species the expression of developmental genes in larval stages/metamorphosis and blastogenesis. In B. schlosseri, Six1/2 and Eya are expressed in developing siphon cells and in peribranchial cells of stigmata during both metamorphosis and blastogenesis; Six3/6 is expressed in the oral siphon, in the neural gland and cerebral ganglion in oozooid and blastozooid; FoxI is expressed in bud rudiment and peribranchial chambers both in larval development and in blastogenesis. Therefore, despite blastogenesis and embryogenesis displaying completely different starting points, the two adult forms show a strong anatomical resemblance, and our results indicate that the molecular network involving these genes takes part in organogenesis of correspondent structures in both these developmental pathways. We note that, even if the cytological organization and innervation pattern of branchial fissures are the same, the major morphological difference between adult oozooids and blastozooids regards their quantity, size and orientation (few big protostigmata elongated dorsoventrally in oozooid, definitive row of stigmata running and elongated anteroposteriorly in blastozooid). With this respect published (Gasparini et al., 2011) and present data agree with equivalent gene expression patterns during both developmental processes. Because the interaction between Six-Eya genes appeared early during animal evolution (Vopalensky and Kozmik, 2009), and FoxI is also an ancient gene (at least as old as deuterostomes), we suggest that in tunicates, where coloniality probably evolved many times secondarily (Perez-Portela et al., 2009), the gene network was co-opted from embryogenesis.

Expression in Branchial Fissure Formation Depicts Alternative Evolutionary Scenarios of Gene Co-option in Placode and Pharynx Evolution

In B. schlosseri late embryogenesis and in the primary bud, the branchial fissures or stigmata form from discrete thickened areas of the peribranchial and branchial epithelia (Manni et al., 2002) which are not considered neurogenic placodes (Manni et al., 2004). Our data show that Six1/2, Eya and FoxI are initially expressed in the thickened stigmata primordia; later, when the two epithelia contact and fuse, only Six1/2 and Eya expression is maintained (Fig. 8). This expression profile represents an example of the presence of this gene network in a process unrelated to placodal differentiation. It is known that these genes play roles in organogenesis and cell-type specification of other structures including vertebrate pharyngeal arches/pouches (Sahly et al., 1999; Solomon et al., 2003; Bessarab et al., 2004; Schlosser, 2007) and, at least for Six1/2 and Eya (no data are available for FoxI), amphioxus and hemichordate gill slits (Kozmik et al., 2007; Gillis et al., 2012). These data suggest a common expression of the transcriptional network during deuterostome pharyngeal fissure formation. Moreover, they render possible an evolutionary scenario depicting co-option of a pharyngeal network into placodes (Schlosser, 2005), assuming placodes evolved in the chordate lineage. The opposite scenario could also be considered if the presence of a placode precursor in insects is accepted such that placode evolution is as old as the bilaterian common ancestor (Posnien et al., 2011).

Figure 8.

Schematic drawing representing expression of Six1/2, Eya, and FoxI during differentiation of a nonplacodal structure, such as a branchial fissure (stigma). A–C: Peribranchial and branchial epithelia contact each other (A), fuse (B), and perforate (C) to create branchial fissures. The exact pattern of gene expression during the process is indicated (adapted from Manni et al., 2002).

Comparison of Expression Profiles Between Species Shows Divergence in Timing Within Tunicates

While most studies on tunicate development have focused on ascidians such as B. schlosseri and C. intestinalis, the planktonic larvaceans have also been investigated. Bassham and Postlethwait (2005) identified expression of Six1/2, Six3/6, Eya, and Pitx in the larvacean Oikopleura dioica in two regions of ectoderm (one rostral and one more posterior) from which develop stomodeum, related sensory organs, ciliated duct, and posterior sensory structures. These expression domains in O. dioica are in regions and structures similar to the expression domains of their C. intestinalis orthologs.

Our results in B. schlosseri sexual development show that the Six1/2, Six3/6, Eya, and FoxI genes are expressed in ectodermal derivatives only during postembryonic development. This contrasts with the reported situation both in C. intestinalis, where the expression of these genes starts from the neurula stage (Mazet et al., 2005), and O. dioica, where the related gene expression has been detected from the early tail bud stage (Bassham and Postlethwait, 2005). However, although from a temporal point of view our results differ from those of other tunicates, from a spatial point of view there are strong similarities (Fig. 9). Indeed, in evaluating the gene expression patterns, we have also to consider the heterochronic shift in the timing of adult organ differentiation during sexual development in the species. Larvaceans differ from ascidians in that they do not undergo a radical metamorphosis, maintaining the tadpole body plan throughout their life. On the other hand, in solitary ascidians such as C. intestinalis, adult organ differentiation typically begins after the body axis reorganization and resorption of larval structures that accompanies attachment during metamorphosis. Colonial ascidian larvae like B. schlosseri, conversely, differentiate adult organs precociously in the brooded larva (Davidson et al., 2004; Jacobs et al., 2008).

Figure 9.

Relationship between vertebrate placodes and placodal areas both in Ciona intestinalis and Botryllus schlosseri.Upper boxes, diagrams showing schematic vertebrate and C. intestinalis head regions (viewed dorsally in embryos, neural tube indicated in dark gray). Vertebrate placodes are color coded. Their fate in an adult vertebrate is indicated on the left, in a schematic lateral view of a vertebrate head based on a lamprey. The C. intestinalis ectoderm domains that express members of the placode markers are color coded according to proposed homology to vertebrate placodes. The fate of these areas in the adult is indicated on the right. The box at the bottom shows that both the larva/metamorphosing oozooid (the neural tube is marked by dotted line, being present in the larva but absent in the oozooid) and the bud of B. schlosseri (left and right drawings, respectively) possess two domains marked by placodal genes, color coded according to their proposed homology to vertebrate placodes. The fates of the two domains in the adults are shown in the central drawing. Upper boxes after (Mazet et al., 2005). For data on Pitx in B. schlosseri, see (Tiozzo et al., 2005).


Considering that the main genes involved in placode formation are widely distributed through the animal kingdom, where they are involved in the specification of neuronal/neurosecretory cells, current views suggest that the molecular machinery that permitted placode evolution in vertebrates was present in the ancestral chordate (Schlosser, 2008; Graham and Shimeld, 2013), if not bilaterian (Posnien et al., 2011). In tunicate embryos, these genes show expression recalling that of placode differentiation in vertebrates. In the latter, a panplacodal domain of non-neural ectoderm is evidenced early in development by Six1/2 and Eya, and later, two sub-domains, corresponding to the anterior and the posterior placodal regions are marked by Six3/6 and Pitx, and by FoxI, respectively. Similar sub-domains have been described in C. intestinalis embryos: the anterior one, from which the palps, the oral siphon, the ciliated duct of the neural gland and anterior neurons derive, is marked by Six1/2, Six3/6, Eya, and Pitx and has been considered similar to the vertebrate hypophyseal and olfactory placodes; the posterior one, giving rise to the atrial primordia, from which atrium, peribranchial chambers, and atrial siphon derive, is evidenced by Six1/2, Six4/5, Eya, and FoxI and has been compared with the otic/lateral line placodes (Mazet et al., 2005). Our results in B. schlosseri show that, in the larva and postmetamorphic development, the same genes are expressed in territories attributable to these two sub-domains or in their derivatives (Fig. 9). Moreover, they show that a corresponding expression profile can be recognized in the bud. These data, and the evidence from expression patterns during gill slit formation in deuterostome sexual development and tunicate budding, indicate evolutionary plasticity of this network to be recruited in different developmental programs for formation of the same structures within tunicates. Therefore, these results indicate for the first time that the molecular network involving these genes is present in distantly related solitary and colonial tunicates, and that it was co-opted when the developmental pathway of asexual reproduction evolved in tunicates, for formation of structures similar to those derived through embryogenesis.


Animals and Embryos

Colonies of Botryllus schlosseri (Styelidae, Stolidobranchiata) were collected in the Lagoon of Venice, cultured according to Sabbadin's technique (Sabbadin, 1955) and fed with Liquifry Marine (Liquifry Co., Dorking, England). Fertilization is internal and embryos develop in the parent. Embryogenesis is synchronized with the colonial cycle and, before zooid regression, the tadpole larvae hatch and escape through the atrial siphon. After a short period of swimming life, the larva metamorphoses into a sessile oozooid, which initiates asexual reproduction.

The transparency of the colonies allowed us to follow the daily development in vivo of buds under the stereomicroscope, thereby permitting the selection of appropriate stages. Embryos were extracted from the colony using a tungsten needle. Metamorphosing larvae were followed in their development in oozooids and collected 1 day and 3 days postmetamorphosis. Embryos, larvae, oozooids, and colonies were anesthetized with MS222 (Sigma) before fixation for subsequent in situ hybridization experiments.

Identification of Transcripts

Total RNA was extracted from mixed developmental stages of fresh B. schlosseri colonies by means of Qiagen RNeasy Mini Kit. Pools of cDNA were obtained using SuperScript First-Strand Synthesis System for RT-PCR (Invitrogen). cDNA pools were used for the amplification of the Six1/2, Six3/6, Eya, and FoxI gene families (see Table 1 for PCR details and primer sequences). Final products were purified with QIAquick Gel Extraction Kit (Qiagen), cloned into pCRII (Invitrogen) then transformed in NovaBlue T1 Singles Competent Cells (Novagen). Plasmid DNA was isolated with QIAprep Spin Miniprep Kit (Qiagen) and each clone sequenced on both strands.

Table 1. Details of the PCR for the amplification of the Six1/2, Six3/6, Eya and FoxI gene families
 Step (and PCR type)sampleCyclingPrimer namePrimer Sequenceb
  1. a

    Adapted from Arendt et al. (2002).

  2. b

    Primer degeneracy uses standard IUB codes.

BsSix1/2aI (degenerated)cDNA1 min 30 sec 94°C; (1 min 94°C; 2 min 42°C; 4 min 72°C) × 5; 5 min 72°CFwBsSix1/2-15′-CCNWSITTYGGNTTYACNCARGA-3′
II (nested degenerated)Step I solution1min 30 sec 94°C; (1 min 94°C; 2 min 47°C; 4 min 72°C) × 35; 15 min 72°CFwBsSix1/2-1n5′-ARGTNGCNTGYGTITGYGARGT-3′
BsSix3/6I (degenerated)cDNA1 min 30 sec 94°C; (1 min 94°C; 2 min 42°C; 4 min 72°C) × 35; 15 min 72°CFwBsSix3/6-15′-AGATTAGCACGNTTYYTNTGG-3′
II (seminested degenerated)Step I solutionSame of aboveFwBsSix3/6-1n5′-AAGATTCCTATGGWSNYTNCC-3′
BsEyaI (semidegenerated)cDNA2 min 30 sec 94°C; (30 sec 94°C; 2 min 50°C; 1 min 72°C) × 35; 15 min 72°CFwBsEya-15′-GARGARTGYGAYCARGTNCA-3′
II (nested semidegenerated)Step I solution2 min 30 sec 94°C; (30 sec 94°C; 2 min 53.2°C; 1 min 72°C) × 35; 15 min 72°CFwBsEya-15′-GARGARTGYGAYCARGTNCA-3
BsFoxI(degenerated)cDNA1 min 30 sec 94°C; (1 min 94°C; 2 min 59.6°C; 4 min 72°C) × 35; 15 min 72°CFwBsFoxI-15′-GGAACTGATCAARHTNGTNMG-3′

Using the SMART RACE cDNA Amplification Kit (Clontech), pools of cDNAs, obtained from a pool of poly-A enriched RNA by means of a Dynabeads mRNA Purification Kit (Invitrogen), were tested with specific primers for transcripts coding for Six1/2, Six3/6, Eya, and FoxI (see Table 2 for primer sequences). Touchdown PCRs were done following the subsequent cycling conditions: (94°C for 30 sec, 72°C for 3 min) for 5 cycles, (94°C for 30 sec, 70°C for 30 sec, 72°C for 3 min) for 5 cycles and (94°C for 30 sec, 68°C for 30 sec, 72°C for 3 min) for 30 cycles. 5′- and 3′-RACE fragments were gel purified (QIAquick Gel Extraction Kit, Qiagen), cloned into pCRII vector (Invitrogen) and then sequenced on both strands.

Table 2. Primers Used in RACE Reactions

Molecular Phylogeny

Alignments were constructed with ClustalX 2.0 (Larkin et al., 2007) using default parameters on datasets respectively formed by 34 sequences for the Six analysis, 16 sequences for Eya, 121 sequences for Fox and 9 sequences for FoxI. (See Supp. File S1, which is available online.) Alignments were then trimmed to conserved domains, so that most of the sequence regions uninformative for the phylogenetic analyses have been removed. The regions used correspond to the following residues of the predicted amino acid sequences that have been obtained: (i) from 46 to 250 of BsSix1/2 and from 69 to 272 of BsSix1/3, respectively, for the Six analyses; (ii) from 128 to 396 of BsEya, for the Eya analysis; (iii) from 142 to 238 for Fox analysis; (iv) from 70 to 491 for FoxI analysis (data not shown). Bayesian phylogenetic analyses were conducted in MrBayes3.1 using default settings (Ronquist and Huelsenbeck, 2003). The analyses were continued for 1 million generations, examined for convergence and the first 25% discarded when compiling summary statistics and consensus trees. Phylogenetic trees were viewed in Treeview (Page, 1996) and then imported into Corel Draw 14 for labeling.

Below are the accession numbers of the sequences (in square brackets) downloaded from NCBI, apart from: (§) that are from JGI database, (°) that are from JGI archive accessible by means of ANISEED database (Tassy et al., 2010) and (#) that are the EMBL-bank IDs for consensus sequences of BsEya, BsSix1/2, BsSix3/6, and BsFoxI, respectively (data not shown).

Figure 2 sequences' accession numbers

BsSix1–2 [HE681850, HE681849, HE681851](#), BsSix3–6 [HE681853, HE681852, HE681854](#), DmSine_oculis [NP_476733.1], DmSix3 [AAD39863.1], DmSix4A [NP_649256.1], PdSix2 [CAC86663.1], PdSix3 [CAR66435.1], XtSix1 [NP_001093693.1], XtSix2 [NP_001093745.1], XtSix3 [fgenesh1_pg.C_scaffold_25000028](§), XtSix4 [e_gw1.68.116.1](§), XtSix6 [NP_001093696.1], MmSix1 [NP_033215.2], MmSix2 [BAA11825.1], MmSix3 [NP_035511.2], MmSix4 [NP_035512.1], MmSix5 [NP_035513.1], MmSix6 [NP_035514.1], BfSix1–2b [YOS_2605529:2](§), BfSix3–6b [YOS_2610351:1](§), BfSix4–5 [YOS_2599917:2](§), CiSix1–2 [fgenesh3_pg.C_chr_03p000638](§), CiSix3–6 [ci0100153410](°), CiSix4–5 [ci0100131152](°), OdSix3–6b [AAZ23145.1], OdSix3–6a [AAZ23143.1], OdSix1–2 [AAZ23141.1], TcSine_oculis [XP_972167.2], TcOptix [NP_001106938.1], GgSix1 [NP_001038150.1], GgSix2 [NP_001038160.1], GgSix3 [NP_989695.1], GgSix4 [CAB41947.1], GgSix6 [NP_990325.1].

Figure 3 sequences' accession numbers

BsEya [HE681856, HE681855, HE681857](#), DmEya [NP_723188.1], MmEya1 [NP_034294.2], MmEya2 [NP_034295.1], MmEya3 [NP_997592.1], MmEya4 [NP_034297.2], BfEya [fgenesh2_pg.scaffold_123000072](§), SpEya [XP_001182519.1], TcEya [XP_974387.1], OdEya [AAZ23131.1], CiEya [ci0100137801](°), DjEya [CAD89531.1], GgEya1 [XP_418290.2], GgEya2 [NP_990246.1], GgEya3 [XP_417715.2], GgEya4 [XP_419735.2].

Figure 4 sequences' accession numbers

BsFoxI [HE681859, HE681858, HE681860](#), DmSpp1 [NP_476730.1], DmSpp2 [NP_476834.1], DmFoxF [AAK85150.1], Dm59A [NP_523814.1], DmFoxN14 [NP_524302.1], DmFoxA [NP_524542.1], Dm64A [NP_523912.1], DmFoxC [NP_524202.1], Dm96Ca [NP_524495.1], Dm96Cb [NP_524496.1], DmCG32006 [NP_726538.1], DmCG11152 [NP_651951.1], DmCG9571 [NP_608369.1], DmCG16899 [NP_649932.1], DmMnf [NP_729672.1], Dm3F [NP_570076.2], DmFoxO [NP_996204.1], DmFoxN23 [NP_511071.3], BfFoxAa [fgenesh2_pm.scaffold_42000007](§), BfFoxAb [estExt_fgenesh2_pg.C_420051](§), BfFoxB [CAD44627.1], BfFoxAB [ACE79151.1], BfFoxC [CAH69694.1], BfFoxD [estExt_fgenesh2_kg.C_2440002](§), BfFoxEa [estExt_GenewiseH_1.C_2440072](§), BfFoxEb [estExt_fgenesh2_pg.C_11730001](§), BfFoxEc [estExt_fgenesh2_pg.C_4550022](§), BfFoxEd [e_gw.559.26.1](§), BfFoxEe [e_gw.559.27.1](§), BfFoxEf [e_gw.455.40.1](§), BfFoxEg [e_gw.455.41.1](§), BfFoxEh [e_gw.251.61.1](§), BfFoxEi [e_gw.251.62.1](§), BfFoxF [fgenesh2_pg.scaffold_6000134](§), BfFoxG [estExt_gwp.C_4610023](§), BfFoxH [ACE79158.1], BfFoxI [fgenesh2_pg.scaffold_167000105](§), BfFoxJ1 [estExt_fgenesh2_pg.C_170078](§), BfFoxJ23 [fgenesh2_pg.scaffold_96000044](§), BfFoxK [ACE79146.1], BfFoxL1 [fgenesh2_pg.scaffold_6000114](§), BfFoxL2 [fgenesh2_pg.scaffold_20000015](§), BfFoxM [ACE79148.1], BfFoxN14b [estExt_fgenesh2_pg.C_1860025](§), BfFoxN14a [ACE79137.1], BfFoxN23 [ACE79140.1], BfFoxO [ACE79159.1], BfFoxQ1 [estExt_gwp.C_60230](§), BfFox2 [estExt_fgenesh2_pg.C_1200108](§), BfFox3 [e_gw.590.38.1](§), BfFox5 [fgenesh2_pg.scaffold_370000019](§), BfFoxQ2c [ACE79147.1], CiFox5 [NP_001071790.1], CiFox4 [NP_001071789.1], CiFox2 [NP_001071788.1], CiFox1 [NP_001071787.1], CiFoxQ [NP_001071718.1], CiFoxP [NP_001071939.1], CiFoxO [NP_001071717.1], CiFoxN23 [NP_001071716.1], CiFoxM [NP_001071993.1], CiFoxL2 [BAE06446.1], CiFoxK [NP_001071715.1], CiFoxJ2 [NP_001071714.1], CiFoxIc [NP_001071713.1], CiFoxIb [NP_001071712.1], CiFoxIa [NP_001071711.1], CiFoxHb [BAE06440.1], CiFoxHa [NP_001071992.1], CiFoxG [BAE06438.1], CiFoxF [NP_001071710.1], CiFoxE [BAE06436.1], CiFoxD [NP_001071709.1], CiFoxC [NP_001071708.1], CiFoxB [NP_001122339.1], CiFoxAb [BAE06432.1], CiFoxAa1 [NP_001072032.1], CiFoxAa2 [NP_001071991.1], MmFoxA1 [AAH96524], MmFoxA2 [AAX90601], MmFoxA3 [NP032286], MmFoxB1 [NP071773], MmFoxB2 [EDL41545], MmFoxC1 [Q61572], MmFoxC2 [NP038547], MmFoxD1 [NP032268], MmFoxD2 [O35392], MmFoxD3 [Q61060], MmFoxD4 [NP032048], MmFoxE1 [CAD29716], MmFoxE3 [Q9QY14], MmFoxF1 [NP034556], MmFoxF2 [NP034355], MmFoxG1 [AAH46958], MmFoxH1 [O88621], MmFoxI1 [CAI25209], MmFoxI2 [NP899016], MmFoxJ1 [NP032266], MmFoxJ2 [NP068699], MmFoxJ3 [AAH58231], MmFoxK2 [CAM21743], MmFoxK1 [NP951031], MmFoxL1 [NP032050], MmFoxL2 [NP036150], MmFoxM1 [O08696], MmFoxN1 [NP032264], MmFoxN4 [AAL06288], MmFoxN2 [NP851305], MmFoxO1 [Q9R1E0], MmFoxO3 [NP062714], MmFoxO4 [NP061259], MmFoxO6 [CAE00176], MmFoxP1 [P58462], MmFoxP2 [NP444472], MmFoxP4 [NP083043], MmFoxP3 [Q99JB6], MmFoxQ1 [CAC60398], MmFoxR1 [Q3UTB7], MmFoxR2 [Q3UM89], MmFoxS1 [NP_034356], XtFoxI1e [NP_988949.1], XtFoxI1c [NP_988949.1], XtFoxI2 [NP_001016544.1].

In Situ Hybridization (ISH)

RNA probes for Six1/2 (724 bp, EMBL ID: HE681849), Six3/6 (1106 bp, EMBL ID: HE681854), FoxI (1167 bp, EMBL ID: HE681860) and Eya (two probes of 905 bp, EMBL ID: HE681855, and 517 bp, EMBL ID: HE681857, respectively) were obtained from 5′- or 3′-RACE fragments in pCRII vector (Invitrogen). B. schlosseri larvae and colonies for ISH were fixed overnight in freshly prepared MOPS buffered (0.1 M MOPS (Sigma), 1 mM MgSO4 (Sigma), 2 mM EGTA (Fluka), 0.5 M NaCl, 4% paraformaldehyde (TAAB). ISH experiments were performed as previously described (Degasperi et al., 2009). Sections were photographed on a Leica 5000B light microscope with a Leica DFC 480 digital photo camera. Images and sketches were then organized with Corel Draw 14.


This study was founded by “Progetto Di Eccellenza 2008–2009” from “Fondazione Cassa di Risparmio di Padova e Rovigo” and grants from Università degli Studi di Padova to F.G., P.B., and L.M. S.M.S. thanks the BBSRC for support.