Gene expression profiles in Ciona intestinalis tailbud embryos

Ascidians belong to the subphylum Urochordata, which is one of the three chordate groups. Adult ascidians are sessile and specialized for filter feeders, but in most species their fertilized eggs develop rather quickly into tadpole-type larvae. The ascidian tadpole consists of about 2600 cells that form several distinct types of tissues (Satoh, 1994). The tadpole is organized into a trunk and tail. The trunk contains a dorsal central nervous system (CNS) with two sensory organs (otolith and ocellus), endoderm, mesenchyme, trunk lateral cells (TLCs) and trunk ventral cells (TVCs). The tail contains a notochord flanked dorsally by the nerve cord (non-neuronal ependymal cells), ventrally by endodermal strand, and bilaterally by three rows of muscle cells. The entire surface of the larva is covered by an epidermis. This configuration of the ascidian tadpole is thought to represent one of the most simplified and primitive chordate body plans (reviewed by Satoh and Jeffery, 1995; Di Gregorio and Levine, 1998; Satou and Satoh, 1999; Wada and Satoh, 2001; Nishino and Satoh, 2001).

The ascidian embryogenesis is well documented. The cleavage pattern is invariant, and cleavage is bilaterally symmetrical. Gastrulation is initiated around the 118-cell stage, and it involves epibolic movements of ectodermal cells and migration of endodermal and mesodermal cells inside the embryo. Neurulation is accomplished by folding of the presumptive neural plate, as in vertebrate embryos. Then, the tailbud embryo is formed, which eventually develops into a tadpole larva. In addition, the lineage of embryonic cells is characterized by detailed descriptions of the epidermis, CNS, endoderm, mesenchyme, TLCs, muscle and notochord (Conklin, 1905; Ortolani, 1955; Nishida and Satoh, 1983; Nishida, 1987; Nicol and Meinertzhagen, 1988). These research findings show that the ascidian embryo is an appropriate experimental system with which to explore cellular and molecular mechanisms that underlie the embryonic cell specification and pattern formation of the embryo (reviewed by Satoh, 1994; Satoh, 1999; Nishida, 1997; Jeffery, 2001). Regulatory mechanisms of specific developmental gene expressions as well as cell-cell interactions are able to be investigated at single cell level (e.g. Yasuo and Satoh, 1993; Nakatani and Nishida, 1994; Corbo et al., 1997).

The genome size of Ciona intestinalis is estimated to be about 160 Mb, and the number of genes approximately 15,500 (Simmen et al., 1998). This genome size and gene number are comparable with those of Drosophila. C. intestinalis is a cosmopolitan species used by researchers worldwide, and they spawn all year round. Their eggs are self-fertile and the generation time is about 2-3 months. These conditions have allowed us to screen mutants that affect developmental processes (Nakatani et al., 1999 in Ciona savignyi; Sordino et al., 2000 in Ciona intestinalis).

Given these research results, we decided to attempt cDNA projects of C. intestinalis, in collaboration with Ciona genome project consortium members. We have focused on the gene expression profiles of embryos at the tailbud stage because it is expected that, at this stage, most of tissues and organs that constitute the future larva begin to be differentiated; therefore, the genes responsible for the formation of these tissues are expressed. We had several specific aims. One was to obtain molecular markers for each of the tissues and/or organs. Although molecular markers for differentiation of embryonic cells have been well characterized in another species (Halocynthia roretzi), to date, in Ciona embryos, molecular markers have been reported for only epidermis, CNS, endoderm, muscle and notochord (Chiba et al., 1998; Takamura, 1998; Takahashi et al., 1999; Hotta et al., 2000; Imai et al., 2000). If our study could characterize such markers, they might be used in future studies to understand molecular developmental mechanisms of Ciona embryos. Second, our study should provide information about the overall gene expression profiles in Ciona tailbud embryos. Lee et al. have reported EST analysis of gene expression in early cleavage-stage sea urchin embryos, in which, for example, 29 of 153 proteins were identified as being associated with cell-cell communication (Lee et al., 1999). Similar data should be obtained in our study. In addition, our study has investigated the gene expression profiles of more than 1000 independent genes by whole-mount in situ hybridization. We discuss how many and what kind of genes are expressed in each tissue. Third, determination of the entire genome sequence is now under way. Our data could be used for annotating these genes to genome sequencing data in the future.

Ciona intestinalis were cultivated at the Maizuru Fisheries Research Station of the Kyoto University, Maizuru, and the Education and Research Center of Marine Bio-Resources of Tohoku University, Miyagi, Japan. Adults were maintained under constant light to induce oocyte maturation. Eggs and sperm were obtained surgically from the gonoduct. After insemination, eggs were dechorionated by immersing them in seawater that contain 1.3% sodium thioglycolate (Wako Pure Chemical Industries, Osaka, Japan) and 0.065% actinase E (Kaken Pharmaceutical Company, Tokyo, Japan). After washing, they were maintained in agar-coated dishes with Millipore-filtered seawater (MFSW) containing 50 μg/ml streptomycin sulfate at room temperature (18-20°C). They developed to tailbud embryos by 10-13 hours of development. These were collected for RNA isolation and whole-mount in situ hybridization.

Total RNA was isolated from C. intestinalis tailbud embryos by the acid guanidinium thiocyanate-phenol-chloroform method (Chomczynski and Sacchi, 1987). Poly (A)⁺ RNA was purified twice using Oligotex beads (Roche Japan, Tokyo). Poly (A)⁺ RNA was converted to double-stranded cDNA, which contains EcoRI site at the 5′-end and XhoI site at the 3′-end, using a cDNA synthesis kit (Stratagene). The cDNAs were ligated to pBSII-SK(−) digested with EcoRI and XhoI. The cDNAs were electroporated into XL-1 Blue MRF′ bacteria (Stratagene). The library was arrayed in 384-well plates in a Genetix Q-Pix robot. The first stage of the arrayed library preparation was done in the Division of Biology, California Institute of Technology with kind help of Dr Eric H. Davidson.

Clones were picked up from the 384-well plates and cDNA inserts were amplified using PCR. Successful amplifications were confirmed using agarose electrophoresis. After purification of the PCR products, their sequences were determined by the conventional procedures using the big-dye terminators on ABI3700 autosequencers at the Academia DNA Sequencing Center, National Institute of Genetics Japan. The primer for 3′-sequencing was the anchored oligo dT primers (5′-(T)17-V-3′) and that for 5′-sequencing was BS740 (5′-CCGCTCTAGAACTAGTG-3′).

Using 3′-most sequence tags, clones were grouped into clusters, each of which contained cDNA clones encoding the same gene. The DDBJ DNA database and protein database (DAD) were searched with 5′-most sequence tags using BLAST algorithm (BLASTN and BLASTX). We categorized clusters into several groups according to their functions that were predicted from BLASTX results (see Results).

For convenience, we applied a method of whole-mount in situ hybridization using DNA probes to ascidian embryos. DNA probes were synthesized basically as described by Tabara et al. (Tabara et al., 1996) with several modifications. First, cDNA inserts were PCR-amplified with SK and T7 primers. Each 10 μl reaction mixture contained 1.4 ng template DNA, 0.1 μM primers, 0.2 mM 4dNTP, 1.5 mM MgCl₂, 1× buffer, and 0.25 U of Taq DNA polymerase (TOYOBO). Reactions proceeded through 35 cycles (30 seconds at 94°C, 30 seconds at 55°C, 3 minutes at 72°C). Digoxigenin-labeled DNA probes were also synthesized by PCR reactions. Each 10 μl reaction mixture contained 1 μl first PCR product, 0.0525 mM DIG-dUTP (Boehringer Mannheim), 0.0975 mM dTTP, 0.15 mM dATP, 0.15 mM dCTP, 0.15 mM dGTP, 1.5 mM MgCl₂, 0.53 μM (dT)17dG, 0.53 μM (dT)17dC, 2.13 μM (dT)17dA, 1× buffer and 0.5 U of Taq DNA polymerase (TOYOBO). Reactions proceeded through 30 cycles (30 seconds at 95°C, 1 minute at 45°C, 1 minute at 72°C). After synthesis, free nucleotides and primers that were not incorporated were removed by ethanol precipitation.

Embryos for whole-mount in situ hybridization were fixed in 4% paraformaldehyde in 0.1 M MOPS (pH 7.5), 0.5 M NaCl at 4°C for 16 hours, before storage in 80% ethanol at −30°C. After thorough washes with PBST (phosphate-buffered saline containing 0.1% Tween 20), the fixed specimens were partially digested with 2 μg/ml proteinase K in PBST for 20 minutes at 37°C and boiled in a water bath for 5 minutes. After being washed with PBST, the specimens were post-fixed with 4% paraformaldehyde in PBST for 1 hour at room temperature, then washed again with PBST. After prehybridization at 50°C for 1 hour, the specimens were hybridized with digoxigenin-labeled probes at 50°C for 16 hours. The hybridization buffer contained 50% formamide, 5× SSC, 5× Denhardt’s solution, 100 μg/ml salmon sperm DNA, 0.1% Tween 20, and DIG-labeled DNA probe.

After hybridization, the specimens were washed twice for 20 minutes in the wash solution (50% formamide, 2× SSC, 0.1% Tween 20) at 50°C, and then twice in the wash solution/PBST (1:1). They were washed in PBST at 50°C four additional times. After being cooled to room temperature, the specimens were blocked with 0.5% blocking reagent (Boehringer Mannheim) in PBST for 30 minutes before 1 hour incubation with 1:2000 alkaline-phosphatase-conjugated anti-digoxigenin antibody (Boehringer Mannheim). The specimens were washed in PBST for 10 minutes four times and then in alkaline phosphatase buffer (100 mM NaCl, 50 mM MgCl₂, 100 mM Tris-HCl, pH 9.5) twice. For signal detection, the embryos were incubated with NBT/BCIP/alkaline phosphatase buffer following the supplier’s instructions (Boehringer Mannheim). The reaction was stopped in PBST. The embryos were dehydrated in a graded series of ethanol, then cleared in a 1:2 mixture of benzyl alcohol:benzyl benzoate (BABB).

In the present study, we determined sequences of both the 5′-most and 3′-most ends of a total of 3423 cDNA clones derived from Ciona intestinalis tailbud embryos. The average length of the sequences was 400∼500 nucleotides. The sequences of the 3′-most end were used to examine the overlap of clones, and this analysis categorized the 3423 clones into 1213 independent clusters (Table 1). Our consortium is now proceeding with further EST analysis of C. intestinalis tailbud embryos, and sequences of more than 10,000 clones are accessible with DDBJ (GenBank/EMBL) database. To obtain the gene expression profiles we first selected C. intestinalis tailbud embryos, as we wish to re-examine more carefully the sequence data when we eventually obtain nearly all the saturated ESTs. The present analysis is therefore partial and will be described rather briefly.

Sequences of the 5′-most end, as well as the 3′-most end, of 1213 clusters that were subjected to BLASTX analysis (Table 1). BLASTX analysis showed that, of the 1213 clusters, 502 were strong matches (P<E-15) to previously identified proteins with distinct functions. The frequency of matches was therefore 502/1213, or about 0.41. As summarized in Table 1, these 502 proteins were categorized into three major classes, basically according to previous classification (Lee et al., 1999); Class A, which contains 390 proteins associated with functions that many kinds of cell use; Class B, which involves 85 proteins associated with cell-cell communication; and Class C, which includes 27 proteins that function as transcription factors and other gene regulatory proteins. Furthermore, the 390 proteins of Class A were categorized into nine subclasses (AI∼AIX), the 85 proteins of Class B into three subclasses (BI∼BIII) and the 27 proteins of Class C into three subclasses (CI∼CIII) (Table 1). The number of genes belonging to each subclass is also shown in Table 1. The great majority of proteins were identified only once in our analysis, although a few highly abundant proteins were identified many times (data not shown). The average number of hits/protein was 3.01 (1511/502; Table 1).

Besides the 502 clusters with strong matches, 184 clusters are classified into Class DI in which sequences were matches to ESTs (mostly from Caenorhabditis elegans, Mus musculus and Homo sapiens) or reported proteins, functions of which have not enough information to classify (for example, posterior end mark gene of Ciona; Table 1). The remaining 527 clusters were categorized into Class DII with no significant sequence similarities to other known proteins (Table 1). A simple estimation of these data suggests that nearly 43% (527/1213) of genes that are expressed in Ciona tailbud embryos should be characterized in future studies.

Table 2 shows some examples of sequence analysis to identify cDNAs that encode especially strong candidates for proteins with a defined function. For example, as members of the subclass BI of signaling receptors including cytokine and hormone receptors and signal ligands, cDNAs for nuclear receptor I (cluster ID, 00334), frizzled-8 (00817) and x-Delta-1 (00914) were identified. The subclass BII for intracellular signal transduction pathway molecules including kinases and signal intermediates includes cDNAs for the receptor for activated protein kinase C (00060), β-catenin (00321), TGF-β type I receptor (00863), and protein-tyrosine kinase (02050). In addition, as members of the subclass CI of sequence-specific DNA-binding proteins, cDNAs for NK homeobox protein (00074), Cdx (00319), Sox-11B (00517) and SoxNeuro (01880) were identified. The cDNAs for transduction-like enhancer protein (00684) and nuclear protein Skip (00711) were identified as members of the subclass CII for non-DNA binding proteins that perform positive and negative roles. (Further information can be found at http://ghost.zool.kyoto-u.ac.jp)

All of the 1213 clusters were subjected to analysis by whole-mount in situ hybridization to explore the overall gene expression profiles in Ciona tailbud embryos. As summarized in Table 3, 148 clusters showed no significant hybridization signals while 329 clusters showed ambiguous signals. These categories that did not show distinct gene expression patterns are presumably due to the presence of small amount of mRNAs in the embryo, because most of them constitute a category of one clone/cluster.

However, 387 of the 1213 clusters showed spatial expression patterns that were specific to a single tissue or organ (Table 3; see Fig. 2). Within these clusters, 149 genes were expressed specifically in epidermal cells, 34 genes were specific to the nervous system, including neuronal cells in the palps (an adhesive organ located at the anterior most part of the embryo) and epithelial sensory cells. Endoderm-specific expression was observed for 29 genes, the number of mesenchyme-specific genes was 112, notochord-specific expression was detected for 32 genes, and 31 genes were expressed specifically in muscle cells. Several examples of specific gene expression for each tissue will be described in the following section.

Table 4 and Fig. 1 show the relationships between the genes with single tissue-specific expression and classes obtained from sequence data. This analysis highlighted a characteristic gene expression profile for each type of tissue (see Fig. 2). Nearly three-quarters of the epidermis-specific genes with distinct functions are in Class A (Fig. 1A). When this ratio is calculated against the clone number, nearly 90% of mRNAs expressed in the epidermis are categorized into Class A or genes that encode for proteins associated with functions that many cell types use (Fig. 1A). Of the genes that are specifically expressed in the nervous system, more than 50% are categorized into Class A (Fig. 1B). When they are examined by clone number, genes for cytoskeletal proteins (subclass AIV) and extracellular matrix (subclass BIII) are predominant (Fig. 1B).

Fig. 1C shows that, as was expected, many endoderm-specific genes are associated with basic cell metabolism (subclass AVI). Interestingly, it was also noticed that many genes for intracellular signaling molecules (subclass BII) are also expressed in the endoderm (Fig. 1C). Many mesenchyme-specific genes are associated with protein synthesis (subclass AV), and more than half of their mRNAs are for that function (Fig. 1D). Fig. 1E shows that the most abundant mRNAs of genes that are expressed specifically in the notochord are associated with metabolism (subclass AVI), suggesting an active metabolism of ascidian notochord cells. As was expected, many muscle-specific genes are associated with cytoskeleton proteins (subclass AIV), and this category of gene accounted for nearly 90% of the clones expressed there (Fig. 1F).

In addition, the present analysis demonstrates that 349 of the 1213 clusters are expressed in multiple tissues simultaneously (Table 3, Fig. 2). This multi-tissue expression pattern was conspicuous in genes that are expressed in TLCs or TVCs (Table 3). Although the present whole-mount in situ hybridization analysis failed to identify genes that are exclusively expressed in either of these two cell types, 57 genes are expressed in TLCs and other tissues, and 48 genes are expressed in TVCs and other tissues. A tissue that showed the simultaneous expression with TLCs was mesenchyme (Table 5); three genes are specifically expressed in TLCs and mesenchyme, 51 genes are expressed simultaneously in TLCs, mesenchyme and other tissues, and only six genes in TLCs and tissues other than mesenchyme. For example, 00141 cDNA encodes oxidoreductase (P is 3E-22; Table 5) and is expressed in TLCs and mesenchyme cells (Fig. 2N). 00597 cDNA is another example, which encodes M-TAXREB107 (P is 2E-72; Table 5) and is expressed in both TLCs and mesenchyme cells. A recent cell lineage analysis has demonstrated that embryonic mesenchyme cells give rise to tunic cells (free living cells within the tunic) of the adult, while embryonic TLCs give rise to coelomic cells and other mesodermal cells of the adult (Hirano and Nishida, 1997). Therefore, it is likely that both cell types express the same genes in the tailbud embryos.

In contrast, a tissue that showed the simultaneous expression with TVCs was muscle (Table 5); five genes are specifically expressed in TVCs and muscle, 35 genes are expressed simultaneously in TVCs, muscle and other tissues, and 13 genes in TVCs and tissues other than muscle. Four examples are listed in Table 5. 00327 cDNA encodes a Ciona counterpart of Halocynthia embryonic muscle myosin heavy chain (P=6E-96; Table 5), which is expressed in TVCs and muscle cells. 02049 cDNA is another example; it encodes Ca²⁺ transporting ATPase (P=1E-88; Table 5) and is expressed in both TVCs and muscle. Because TVCs are responsible for the formation of adult muscle tissues (Hirano and Nishida, 1997), it is conceivable that the same genes are expressed in muscle and TVCs of the tailbud embryos.

Simultaneous expression of specific genes in multiple tissues was also evident for every tissue. For example, the present analysis demonstrated that not only are there 34 nervous system-specific genes, but 239 genes are also expressed in the nervous system and in other tissues (Table 3); 29 genes are specifically expressed in the endoderm, but an additional 195 genes are expressed in the endoderm and other tissues (Table 3). Two interesting examples are shown in Table 5 and Fig. 2: 00794 cDNA encodes a zinc transporter hZIP2-like protein (P=7E-25), the gene for which is expressed specifically in a subset of brain cells and endodermal strand cells (Fig. 2L); 00152 cDNA encodes a protein with no significant sequence information, but this gene is expressed in endoderm, notochord, TVCs and muscle cells (Fig. 2T).

The following are descriptions of representative genes that show tissue-specific expression pattern, although the spatial expression pattern for every gene appears at http://ghost.zool.kyoto-u.ac.jp

A single layer of 800 epidermal cells constitutes the outer most layer of the ascidian tailbud embryo. In Halocynthia embryos, cDNA clones for eight different epidermis-specific genes have been isolated and characterized (Ueki et al., 1991; Ueki and Satoh, 1995; Ishida et al., 1996), whereas Cs-Epi1 and Cs-Epi2 are two genes that are specifically expressed in C. savignyi embryonic epidermal cells (Chiba et al., 1998).

The present analysis of gene expression profiles demonstrate that 149 genes are expressed specifically in the epidermal cells, and a wide variety of other genes in relation to the other functional classes are also expressed there (Table 4; Fig. 1A). However, as is evident in Table 4, 67 of the 149 genes encode for proteins that are not in the database, suggesting a need to further characterize these genes to understand their function in the development of the ascidian epidermis. Eight examples of epidermis-specific genes are listed in Table 5. Among the 149 genes, five were matched to Halocynthia epidermis-specific genes. 00014 clone encodes for a protein similar to human GDP-4-keto-6-deoxy-D-mannose epimerase-reductase (P=9E-75). As shown in Fig. 2A, gene 00104 is expressed in epidermal cells, except for those along the midline of the embryo; epidermal cells along the midline of the tailbud embryo are thought to be specialized to form the larval fin (Ishida et al., 1996). Interestingly, 00224 is expressed in epidermal cells along the dorsal and ventral midlines (Fig. 2B). This pattern of gene expression was not detected using Halocynthia epidermis-specific genes (Ishida et al., 1996). In addition, gene 01493, which shows no sequence similarity with any other genes, was expressed in epidermal cells that reside in the posterior half or tail region of the embryo (Fig. 2C).

Ishida et al. have reported that spatio-temporal expression patterns of eight epidermis-specific genes of Halocynthia embryos are categorized into four groups (Ishida et al., 1996). All four patterns were observed using Ciona epidermis-specific genes: 133 of the 149 genes are expressed in the entire surface of the embryo, 11 genes are only expressed in the tail epidermis, four are expressed in the midline, and one gene is expressed in all epidermal cells, except for midline cells.

The formation of nervous system of Ciona larva has been studied in detail (Katz, 1983; Nicol and Meinertzhagen, 1991; Takamura, 1998). Altogether, the nervous system consists of the central nervous system (CNS) and peripheral sensory nervous system, which consists of approximately 350 cells. The anterior-most of the larva is an adhesive organ or palps (papilla) in which several neuronal cells are differentiated. In the mid-dorsal trunk region is located a brain vesicle with two sensory organs called the otolith and ocellus. The CNS extends into the tail nerve cord, which consists of glial ependymal cells. Peripheral epithelial neuronal cells are scattered over the entire surface of the larva. Although several marker genes for the CNS have been characterized in Halocynthia embryos, only one monoclonal antibody has been used to monitor neuronal differentiation in Ciona embryos (Takamura, 1998).

The present comprehensive analysis of gene expression profiles in Ciona tailbud embryos revealed 34 genes that are expressed exclusively in the nervous system (Table 3). For example, 01087 cDNA encodes for a Etr-1 homolog (Ci-Etr-1; P=9E-33; Table 5), and this gene is expressed in almost all of the neuronal cells including palps, brain, nerve cord and epithelial sensory neurons (Fig. 2D). The expression of Ci-Etr-1 may be used as a pan neuronal differentiation marker in future studies. In addition, the present study revealed genes that are specifically expressed in various subpopulations of nervous system cells: six genes in the brain, eight genes in the palps and four genes in the nerve cord (Table 3). 00531 cDNA encodes for Drosophila dorsal switch protein 1 homolog (P=3E-47;Table 5), and the expression of this gene is evident only in the CNS (Fig. 2E). 01993 is PIWI homolog (P=1E-50; Table 5) and is expressed in a subset of brain cells (Fig. 2F). Two cDNAs, 00659 (Fig. 2G) and 02027 (Fig. 2H) are expressed only in palps, the former encodes for tubulin α-1 chain and the latter for a protein with no sequence similarity (Table 5). The nerve cord consists of four rows of ependymal cells (upper, two lateral and lower) and functions in the guidance of axons from neuronal cell bodies situated in the brain. To date, only one gene (HrWnt-7) of Halocynthia was identified to be expressed exclusively in the nerve cord (Sasakura and Makabe, 2000), although several genes are reported to be expressed in the nerve cord and other tissues. 00124 encodes a protein with no sequence similarity, but this gene is expressed in the upper and lower rows of nerve cord cells only (Fig. 2I). There is no database information for 01427; however, this gene is expressed in the nerve cord cells only (Fig. 2J).

The nervous system of Ciona provides researchers with a simple and primitive model with which to analyze the functional complexities of the vertebrate CNS. Genes that have been identified in the present study should be very useful in future studies.

The endodermal tissue of the Ciona tadpole larva constitutes the trunk endoderm and tail endodermal strand. The endoderm consists of about 500 cells. Because the tadpole of most ascidian species is the dispersal phase of their life cycle, they do not open their mouth before attachment to the substrate, which is an early event of metamorphosis. The histochemical detection of alkaline phosphatase (AP) activity is a conventional way to study endoderm differentiation (Whittaker, 1990), and a cDNA clone for the AP gene was isolated in H. roretzi (Kumano and Nishida, 1998), and C. intestinalis and C. savignyi (Imai et al., 2000). The present analysis demonstrates 29 independent genes that are expressed specifically in endodermal cells (Table 3); three examples are listed in Table 5. For example, 00453, which encodes for AFRAPTIN 2 (P=3E-53), and 00783, which encodes for a protein with no sequence similarity, are specifically expressed in the endodermal cells (Fig. 2K).

In Ciona early embryos, nuclear accumulation of β-catenin is most probably the first step of endodermal cell specification (Imai et al., 2000). The nuclear localization of β-catenin appears to trigger the activation of many transcription factor genes, as well as genes for signal transduction molecules involved in endodermal cell differentiation (Y. Satou et al., unpublished). Future studies should ask how these upstream genes control the expression of endoderm-specific genes identified by the present analysis.

The Ciona tailbud embryo contains four pockets (two pairs) of mesenchyme cells situated in the posterior part of the trunk (Katz, 1983). To date, no mesenchyme-specific genes have been reported in ascidian embryos. Although a cytoplasmic-type actin gene (HrCA1) is expressed predominantly in mesenchyme cells in Halocynthia embryos, this gene is also expressed in specific neuronal cells in the dorsal trunk region, as well as in notochord cells and in muscle cells (Araki et al., 1996). The present analysis has scored 112 genes that are expressed specifically in mesenchyme cells. As shown in Table 3, many of the different classes of genes are expressed specifically in the mesenchyme cells, and a representative of eight genes are listed in Table 5. For example, 00142 cDNA encodes for a protein with no sequence similarity, which gene is expressed exclusively in mesenchyme cells (Fig. 2M).

It is believed that the mesenchyme of an ascidian embryo is a source of adult mesodermal tissues and therefore this embryonic tissue is rather quiescent in gene activity during embryogenesis. However, as shown in the present study, the second largest number of specific genes were observed in this embryonic tissue, the first largest number of genes were expressed in the epidermis. This suggests that nuclei of embryonic mesenchyme cells are not always quiescent, a new finding that should be explored in future studies.

The notochord of an ascidian larva is composed of just 40 cells, and the entire lineage has been described completely (Nishida, 1987). Previous studies have revealed that ascidian Brachyury genes, HrBra (As-T) in H. roretzi (Yasuo and Satoh, 1993) and Ci-Bra in C. intestinalis (Corbo et al., 1997), are expressed exclusively in the notochord precursor cells and play a pivotal role in notochord formation (Yasuo and Satoh, 1998). In addition, isolation and characterization of Ci-Bra downstream genes has demonstrated that nearly 40 genes are expressed specifically and predominantly in notochord cells (Takahashi et al., 1999). Twenty cDNAs were completely sequenced and found to be ascidian homologs of Drosophila prickled genes, genes for antithrombin III-like protein, netrin protein, tropomyosin-like protein, ATP citrate lyase, cdc45, fibrinogen-like protein, pellino-like protein, ATP sulfurylase kinase, tropomyosin-2-like protein, leprecan, ERM, protein tyrosine phosphatase, mad4-like protein and β4-Gal-transferase, and seven other genes had no sequence similarity (Hotta et al., 2000).

The present analysis has revealed that 32 genes are expressed in a notochord-specific manner (Table 3; Fig. 2O,P). For example, 00051 (Table 5; Fig. 2O) is a Ci-trop gene that has already been characterized as a target gene of Ci-Bra (Di Gregorio and Levine, 1999). 00347 (Table 5; Fig. 2O) encodes a fibrinogen-like protein that was identified as a Ci-Bra target gene (Hotta et al., 2000). Interestingly, although 20 of the 32 genes were previously identified as notochord-specific genes, the other 12 genes are newly identified in the present analysis. This raises an intriguing question of whether or not these 12 genes are targets of Ci-Bra transcription factor.

The Ciona tailbud embryo develops 18 unicellular, striated muscle cells on each side of the tail (total 36 cells); of them, 28 cells of the anterior and middle part of the tail are derived from B4.1 pair (primary lineage), while four cells at the posterior part and four cells at the tip of the tail originate from the A4.1 and b4.2 pair, respectively (Nishida, 1987). Several muscle-specific genes, including genes for actin (HrMA4) and myosin heavy chain (HrMHC), have been characterized in Halocynthia embryos (reviewed by Satou and Satoh, 1999). In addition, cDNA clones for troponin I (MacLean et al., 1997) and tropomyosin (Meedel and Hastings, 1993) have been isolated from C. intestinalis.

The present study has revealed that 31 genes are specifically expressed in muscle cells; 17 examples are listed in Table 5. In addition, 28 of these muscle-specific genes encode structural proteins for muscle cell function; for example, 00031 encodes myosin regulatory light chain (MRLC) and is expressed in all of the muscle cells in the tailbud embryo (Fig. 2Q), 00173 encodes troponin I, 00189 encodes fast-twitch myosin light chain I, 00435 encodes creatine kinase-M and 00608 encodes tropomyosin (Table 5).

Two interesting clones are 00069 (Fig. 2R) and 01633 (Fig. 2S). Both genes encode proteins with no sequence similarity. In Halocynthia embryos, HrMA4 and HrMHC begin to be expressed as early as the 32-cell stage (Satou et al., 1995), and HrMRLC gene is also expressed at the 32-cell stage (Y. Satou et al., unpublished). Therefore, transcripts of these muscle-specific genes are distributed in the entire cytoplasm of muscle cells at the tailbud stage (Fig. 2Q). However, in situ signals of 00069 (Fig. 2R) are seen mainly in nuclei of muscle cells in the tailbud embryo, suggesting that this gene begins to be expressed around the neurula stage or later. 01633 clone seems to be more interesting, because this gene is expressed in only six pairs of muscle cells located in the posterior region of the tail (Fig. 2S). Cell lineage studies suggest that they are two pairs of b-line muscle cells at the tip of the tail, two pairs of A-line cells at the posterior region and two pairs of B-line muscle cells (Nishida, 1987). All of the muscle-specific genes so far identified are expressed in every muscle cell, although some genes show differences in temporal expressions between the primary (B-line) and secondary (A- and b-lines) lineage muscle cells. The spatial expression of 01633 is therefore unique, and determining the molecular mechanisms that underlie the spatial expression of the 01633 gene should be an intriguing goal of future studies.

In conclusion, the present analysis of a set of 1213 independent clusters derived from the Ciona intestinalis tailbud embryos reveals that 502 of them show significant matches to reported proteins, while 184 do not have enough information to be categorized and 527 do not show significant similarities to any known proteins. Sequence similarity analyses of the 502 clusters suggest that 390 of them are associated with functions that many types of cells use, 85 with cell-cell communication, and 27 with transcription factor functions and other gene regulatory proteins. Whole-mount in situ hybridization analysis of all of the 1213 clusters demonstrates that a total of 387 clusters show expressions that are specific to a tissue or organ; 149 show epidermis-specific expression, 34 are specific to the nervous system, 29 to endoderm, 112 to mesenchyme, 32 to notochord and 31 to muscle. Many genes are also specifically expressed in multiple tissues. The present analysis also highlights characteristic gene expression profiles that are dependent on the type of tailbud tissue and provides new information on many molecular markers for every tissue and organ constituting the Ciona tailbud embryo. The sequence information will also be used for future whole genome analysis to explore molecular mechanisms involved in the development of one of the most primitive chordate body plans.

We thank Kazuko Hirayama, Chikako Imaizumi and Reiko Nakamura for technical support. This research was supported by a Grant-in-Aid for Priority Area C (No. 12202001) from the Ministry of Education, Science, Sports and Culture, Japan to N. S. The Academia DNA Sequencing Center is supported by a Grant-in-aid for Scientific Research on Priority Area C (No. 12201001) from the Ministry of Education, Science, Sports and Culture, Japan to Y. K.