Ascidian eggs and embryos have provided an appropriate experimental system to explore the cellular and molecular mechanisms involved in the embryonic cell specification and pattern formation of the embryo. In Japan, most of the studies of ascidian embryology have been carried out with the large eggs of Halocynthia roretzi. However, for future studies, Ciona species may provide a better experimental system, in particular with respect to the incorporation of genetic approaches. In order to establish Ciona as an experimental system, molecular markers with which to examine cellular differentiation are required. In the present study, we isolated and characterized cDNA clones for two epidermis-specific genes (CsEpi-1 and CsEpi-2) and for two muscle-specific genes (CsMA-1 and CsMu-1). CsEpi-1 encodes a polypeptide with three trefoil domains, while CsMA-1 encodes a muscle-type actin from C. savignyi. Although CsEpi-2 and CsMu-1 transcripts seem to have a poly(A) tail at the 3′ end, we could not find a distinct open reading frame in the sequences. Probes for CsEpi-1, CsMA-1 and CsMu-1 cross-reacted with C. intestinalis embryos. These cDNAs are useful as molecular markers for the specification of epidermis and muscle of Ciona embryos.
INTRODUCTION
Ascidian eggs and embryos have provided an appropriate experimental system to explore the molecular nature of localized maternal factors and their roles in cell specification and pattern formation (for reviews see Satoh, 1994; Satoh et al., 1996). The fertilized egg develops quickly into a tadpole larva, which consists of a small number of tissues including the epidermis, central nervous system with two sensory organs, nerve cord, endoderm, mesenchyme, notochord and muscle. The lineage of these embryonic cells is completely described up to the gastrula stage (Conklin, 1905; Nishida, 1987).
Recent molecular embryological studies have isolated and characterized cDNA clones for genes that are expressed in a tissue-specific manner, genes encoding transcriptional factors, and genes encoding signal molecules (reviewed by Chiba and Nishikata, 1998). In addition, recent studies have succeeded in the characterization of maternal genes with localized mRNAs, including posterior end mark (pem; Yoshida et al., 1996), pem-2, pem-4, pem-5, and pem-6 (Satou and Satoh, 1997), and HrWnt-5 (Sasakura et al., 1998). One of the difficulties in ascidian molecular embryology is the need for techniques that deduce the function of these developmentally important genes. In some of the genes, the overexpression of the proteins produced by a microinjection of synthetic mRNA (Yoshida et al., 1996; Yasuo and Satoh, 1998) and the inhibition of the mRNA function by treatment with antisense oligo-nucleotides (Swalla and Jeffery, 1996; Olsen and Jeffery, 1997) resulted in distinct effects, providing cues to infer the gene functions. However, these techniques are not always successful.
We have speculated that genetic approaches such as those used for Drosophila, C. elegans and zebrafish could be applied to ascidians to identify genes with developmentally important functions. In Japan, most of the studies of ascidian embryology have been carried out with the large and transparent eggs of Halocynthia roretzi. However, H. roretzi may not be an appropriate system for future studies with genetic approaches. The spawning season is limited to winter, and the generation time may be more than two years. We propose Ciona eggs and embryos as an experimental system for further studies, because their spawning season is basically all year-round, and their generation time appears to be about 3 months (Kano and Amemiya, personal communication). Ciona savignyi or C. intestinalis could thus be useful as an experimental system for future studies, including those using genetic approaches.
In the present study, we therefore attempted to isolate cDNA clones for genes that are useful as molecular markers for the specification of embryonic cells. With such an aim, we used the subtractive hybridization of mRNAs of tailbud embryos with those of fertilized eggs. Taking advantage of the well-known lineage and segregation pattern of developmental fates as well as the in situ hybridization of whole-mount specimens, we were able to isolate cDNA clones for two epidermis-specific and two muscle-specific genes.
MATERIALS AND METHODS
Ascidian eggs and embryos
Ciona savignyi and C. intestinalis adults were collected near the Otsuchi Marine Research Center, Ocean Research Institute of the University of Tokyo, Iwate, Japan, and maintained under constant light to induce oocyte maturation. Eggs and sperm were obtained surgically from the gonoduct. After insemination, eggs were reared at about 18°C in Millipore-filtered seawater (MFSW) containing 50 μg/ ml streptomycin sulfate.
RNA isolation and cDNA library construction
Total RNA was isolated from fertilized eggs or tailbud embryos by the acid guanidinium thiocyanate-phenol-chloroform method (Chomczynski and Sacchi, 1987). Poly(A)+ RNA was purified using Oligotex beads (Roche Japan, Tokyo). cDNA libraries of fertilized eggs (FE-library) and tailbud embryos (TB-library) were constructed in Uni-ZAP XR using a ZAP-cDNA synthesis kit (Stratagene, La Jolla, CA, USA). A tailbud-mRNA concentrated subtractive library was constructed from TB- and FE-libraries as described by Satou and Satoh (1997).
Screening of the subtracted cDNA library and sequencing
From the library, clones were randomly picked up and partially sequenced from poly(A) tail to avoid analyzing the same clones any further. After partial sequencing, each clone was examined for the localization of corresponding mRNA by whole-mount in situ hybridization using digoxigenin-labeled antisense RNA probes. Gastrulae and tailbud embryos were used as specimens for the in situ hybridization screening. cDNA clones exhibiting the localization of corresponding mRNAs were selected for further analyses.
Nucleotide sequences were determined for both strands with a dye primer cycle sequencing FS ready reaction kit and ABI PRISM 377 DNA sequencer (Perkin Elmer, Norwalk, CT, USA).
Northern analysis
Poly(A)+ RNA was isolated as described above and fractionated by agarose gel electrophoresis, and transferred to a Hybond-N(+) membrane (Amersham, Buckinghamshire, UK). Blots were hybridized with 32P-random-labeled DNA probes in 6 X SSPE, 0.5% SDS, 5 X Denhardt's solution, 100 μg/ml salmon sperm DNA, and 50% formamide. The filter was washed twice in 2 X SSC/0.1% SDS, and twice in 0.2 X SSC/0.1% SDS at 65°C, and exposed to X-ray film.
Whole-mount in situ hybridization
RNA probes were prepared with a DIG RNA labeling kit (Boehringer Mannheim, Heidelberg, Germany). Whole-mount in situ hybridization was performed as described previously (Satou et al., 1995). The control specimens hybridized with sense probes did not show signals above the background.
RESULTS AND DISCUSSION
Isolation of cDNA clones for tissue-specific genes in C. savignyi embryos
In order to obtain cDNA clones for genes that are expressed in a tissue-specific manner, we constructed a cDNA library of tailbud-embryo mRNAs subtracted with the fertilized-egg mRNAs of C. savignyi. The library was estimated to contain about 90,000 clones. From the library, clones were randomly selected and their nucleotide sequences were determined from the 3′ end to prevent the further analysis of the same clones. Each clone was then examined for the localization of corresponding mRNA by whole-mount in situ hybridization. Gastrulae and tailbud embryos were subjected to in situ hybridization to determine the specific expression of the genes. We have examined 100 clones to date and were able to find cDNA clones for two epidermis-specific genes (CsEpi-1 and CsEpi-2) and two muscle-specific genes (CsMA-1 and CsMu-1) of C. savignyi embryos, which are described below.
Expression of the CsEpi-1 gene
Sequence analysis.
The nucleotide and predicted amino acid sequences of a cDNA clone for CsEpi-1 are shown in Fig. 1. The insert of the clone consisted of 2,653 nucleotides. The clone contained a single open reading frame (ORF) that predicted 741 amino acids. The calculated molecular mass (Mr) of the CsEpi-1-encoded protein (CsEpi-1) was 81.9 k. A Northern blot showing a transcript of about 3.0 kb (Fig. 3) suggested that the clone contains all the coding sequences and is close to full-length.
As shown in Fig. 1, the sequence motif search using the Block Searcher ( http://www.blocks.fhcrc.org/blocks-search.html) suggested that CsEpi-1 contains three P-type trefoil domains in the C-terminal half. Thim (1989) pointed out that four peptides present in completely different biological sources have been shown to exhibit a large degree of structural similarity. The peptides include the breast cancer-associated pS2 peptide isolated from human gastric juice and culture media of the human breast cancer cell line MCF-7 (Jakowlew et al., 1984), the pancreatic spasmolytic polypep-tide (PSP) isolated from porcine pancreas (Tomasetto et al., 1990), and the peptide predicted from a cDNA isolated from the skin of Xenopus laevis (Hoffmann, 1988). The domain contain 6 cysteine residues in nearly the same positions, and these 6 residues are linked by 3 disulphide bonds to form a characteristic “trefoil“ disulphide loop structure, as shown in Fig. 2b. Several studies have shown the presence of the trefoil domain in peptides abundantly produced at the mucousal surface of various animals (e.g., Hauser et al., 1992; Podolsky et al., 1993). Figure 2a shows a comparison ofthe amino acid sequence of the trefoil domain of CsEpi-1 with those of human intestinal trefoil hITF (Podolsky et al., 1993), human pS2 (Jakowlew et al., 1984), PSP (Tomasetto et al., 1990), the Xenopus laevis skin protein FIM-A.1 (Hoffmann, 1988), and another Xenopus laevisskin protein xP2 (Hauser et al., 1992). These domains shared the consensus sequences (Hoffmann and Hauser, 1993).
The present report may be the first report of a trefoil family protein from invertebrates, although another trefoil family protein has been isolated from a colonial ascidian (Dr. Kazuo Kawamura, personal communication). During ascidian em-bryogenesis, epidermal cells produce larval and adult tunics. The tunic consists of mucus substances. CsEpi-1 may be a component of such mucus substances.
Spatial expression of CsEpi-1.
The in situ hybridization of whole-mount specimens demonstrated that no signal was detected by the early gastrula stage (Fig. 4a) and that the first distinct signal was detected at the neurula stage (Fig. 4b). At this stage, the hybridization signal was evident in the nuclei of almost all of the epidermal cells. This signal was retained by the epidermal cells of the early tailbud embryos (Fig. 4c). A cross-section of hybridized embryos clearly showed that the CsEpi-1 expression was restricted to epidermal cells (Fig. 4d).
Cross-reactivity with Ciona intestinalis embryos.
When we examined whether the CsEpi-1 antisense probe cross-reacts with C. intestinalis embryos, it became clear that the probe cross-reacted with C. intestinalis embryos (Fig. 5a). Thus, this gene is a useful molecular marker for epidermal cell differentiation in embryos of both Ciona species.
Expression of the CsEpi-2 gene
Sequence analysis.
Nucleotide sequence of a cDNA clone for CsEpi-2 is shown in Fig. 6. The insert of the clone consisted of 1,618 nucleotides. There was a putative signal sequence for polyadenylation. In addition, the sequence included 31 adenylyl residues at the 3′ end, suggesting that the transcript has a poly(A) tail (Fig. 6). However, we could not detect any distinct ORF in the cDNA (Fig. 6). A Northern blot analysis, shown in Fig. 3, demonstrated that CsEpi-2 was not expressed in fertilized eggs but the transcript of about 1.7 kb was evident in the tailbud embryos. In addition, as described below, an in situ hybridization showed that the CsEpi-2 transcript was evident in the nuclei of the 8-cell embryos. All of these data suggest that CsEpi-2 is expressed zygotically in C. savignyiembryos. We repeated the isolation and sequence determination of three independent clones corresponding to CsEpi-2, which showed sequence identity with a few differences.
As mentioned above, the CsEpi-2 transcript has no distinct ORF. We therefore examined possible secondary structures of CsEpi-2 transcript by calculation with the version 2.3 of Mfold (Zuker, 1989; Zuker and Jacobson, 1995). The predicted secondary structures of the CsEpi-2 transcript are shown in Fig. 7a.
Spatial expression of CsEpi-2.
In most cases of zygotic expression of ascidian genes, the detection of mRNA by in situ hybridization of whole-mount specimens is more sensitive than that by Northern hybridization. This is because in situ hybridization can detect signals first in the nucleus of certain cells which frequently develop in a lineage-specific and/ or region-specific manner (Yasuo and Satoh, 1993; Satou et al., 1995). This was the case for the CsEpi-2 gene.
The in situ hybridization demonstrated that the first distinct signal was detected as early as the 8-cell stage (Fig. 8). At this stage, the hybridization signal was evident in the nuclei of pairs of the a- and b-line primordial epidermal cells (Fig. 8a). During gastrulation and neurulation, the CsEpi-2 expression was retained only by epidermal cells (Fig. 8b). This signal was evident in the epidermal cells of the early tailbud embryos (Fig. 8c). The cross-section of hybridized embryos clearly showed that the CsEpi-1 expression was restricted to epidermal cells (Fig. 8d).
In order to deduce the gene function, we treated embryos with CsEpi-2antisense oligos, but we did not obtain any meaningful results. The CsEpi-2 antisense probe did not cross-react with C. intestinalis embryos (data not shown).
The initiation of the appearance of CsEpi-2 was as early as the 8-cell stage. The first detection of a zygotic expression of ascidian genes was of the forkhead/HNF-3 gene which was reported to be at the 16-cell stage (Corbo et al., 1997; Olsen and Jeffery, 1997; Shimauchi et al., 1997). The CsEpi-2 gene may therefore represent the first zygotic expression of ascidians.
Expression of a muscle-type actin gene CsMA-1
Sequence analysis.
The nucleotide sequence of the cDNA for CsMA-1 will appear under the accession number AB008817 in the DDBJ/EMBL/GenBank database. The insert of the clone consisted of 1,300 nucleotides including 21 adenylyl residues. The clone contained a single ORF that predicted 378 amino acids. Since (as shown below) the clone encodes a muscle actin, we designated this gene CsMA-1. The calculated molecular mass (Mr) of the CsMA-1-encoded protein (CsMA-1) was 42.1 k.
Most animals exhibit multiple actin isoforms which are encoded by a small gene family. In mammals, there are four muscle isoforms (α-skeletal, α-cardiac, α-vascular, and γ-en-teric) and two nonmuscle isoforms (β;- and γ-cytoplasmic) (Vandekerckhove and Weber, 1979) The mammalian α-skeletal muscle actin is distinguishable from the β-cytoplasmic actin by about 20 diagnostic amino acid positions (Vandekerckhove and Weber, 1978, 1979). Figure 9 shows the comparison of the amino acid sequence of CsMA-1 with those of muscle-type and cytoplasmic-type actin genes of ascidians. The comparison of the amino acid residues at the diagnostic positions indicated that the CsMA-1 is a muscle actin, while CsCA-1 is a cytoplasmic actin (Y. Satou, unpublished data).
Spatial expression of CsMA-1.
The in situ hybridization demonstrated that the first distinct signal was detected at the 64-cell stage (Fig. 10a). The signals are evident in the nuclei of B7.4 and B7.8, the primordial B-line muscle cells. During gastrulation, signals became evident in B- (Fig. 10b), A-, and b-line presumptive muscle cells, and the neurulae showed signals in the primordial muscle cells (Fig. 10c). An early tailbud embryo showed distinct signal in muscle cells of the tail region of the embryo (Fig. 10d).
Cross-reactivity with Ciona intestinalis embryos.
We confirmed that the CsMA-1 antisense probe cross-reacts with C. intestinalis embryos (Fig. 5b), and thus is useful as a molecular marker in that embryo.
The isolation of a C. savignyi muscle actin may provide material for future studies. We have already isolated a genomic clone of CsMA-1 and characterized the cis-regulatory elements required for the muscle-specific expression of CsMA-1.
Expression of the CsMu-1 gene
Sequence analysis.
The nucleotide sequence of the cDNA clone of the CsMu-1 gene is shown in Fig. 11. The insert of the clone consisted of 1,364 nucleotides. There was a putative signal sequence for polyadenylation. In addition, the sequence included 51 adenylyl residues at the 3′ end, suggesting that the transcript has a poly(A) tail (Fig. 11). However, as in the case of CsEpi-2, we did not detect any distinct ORF in the CsMu-1 cDNA (Fig. 11). The Northern blot analysis shown in Fig. 3 demonstrated that CsMu-1 is not expressed in fertilized eggs, but the transcript of about 1.4 kb is evident in the tailbud embryos. In addition, the in situ hybridization showed that the CsMu-1 transcript is evident in the nuclei of the gastrula. Therefore, it is highly likely that CsMu-1 is expressed zygotically in C. savignyi embryos. We examined four independent clones corresponding to CsMu-1, which showed sequence identity with a few differences.
Similarity to the case of the CsEpi-2 transcript, we inferred possible secondary structures of the CsMu-1 transcript by calculation with the 2.3 version of Mfold. The predicted secondary structures of the CsMu-1 transcript are shown in Fig. 7b.
Spatial expression of CsMu-1.
The in situ hybridization demonstrated that the first distinct signal was detected at the late gastrula stage (Fig. 12a, b). At this stage, the hybridization signal was evident in the nuclei of pairs of primordial muscle cells (Fig. 12b). During neurulation, the CsMu-1 expression expanded (Fig. 12c). An early tailbud embryo showed distinct signal in muscle cells of the tail region of the embryo (Fig. 12d). In order to deduce the gene function, we treated embryos with CsMu-1 antisense oligos, but we did not obtain any meaningful results.
Acknowledgments
We thank Dr. Yasuaki Takagi and Mr. Koichi Morita and all staff of the Otsuchi Marine Research Center, Ocean Research Institute, University of Tokyo for their hospitality and their help in collecting ascidians. We also thank Kaz Makabe and Hiroki Takahashi for their critical reading of the manuscript.
YS was supported by a Predoctoral Fellowship from the Japan Society for the Promotion of Science (JSPS) for Japanese Junior Scientists with Research Grant #8-6806. This research was supported by Research for the Future Program from the JSPS (96L00404) to TN and was also supported in part by a Grant-in-Aid for Specially Promoted Research (#07102012) from the Ministry of Education, Science, Sports and Culture, Japan to NS.