We have isolated a member of the HNF-3/forkhead gene family in ascidians as a means to determine the role of winged-helix genes in chordate development. The MocuFH1 gene, isolated from a Molgula oculata cDNA library, exhibits a forkhead DNA-binding domain most similar to zebrafish axial and rodent HNF-3β. MocuFH1 is a single copy gene but there is at least one other related forkhead gene in the M. oculata genome. The MocuFH1 gene is expressed in the presumptive endoderm, mesenchyme and notochord cells beginning during the late cleavage stages. During gastrulation, MocuFH1 expression occurs in the prospective endoderm cells, which invaginate at the vegetal pole, and in the presumptive notochord and mesenchyme cells, which involute over the anterior and lateral lips of the blastopore, respectively. However, this gene is not expressed in the presumptive muscle cells, which involute over the posterior lip of the blastopore. MocuFH1 expression continues in the same cell lineages during neurulation and axis formation, however, during the tailbud stage, MocuFH1 is also expressed in ventral cells of the brain and spinal cord. The functional role of the MocuFH1 gene was studied using antisense oligodeoxynucleotides (ODNs), which transiently reduce MocuFH1 transcript levels during gastrulation. Embryos treated with antisense ODNs cleave normally and initiate gastrulation. However, gastrulation is incomplete, some of the endoderm and notochord cells do not enter the embryo and undergo subsequent movements, and axis formation is abnormal. In contrast, the prospective muscle cells, which do not express MocuFH1, undergo involution and later express muscle actin and acetylcholinesterase, markers of muscle cell differentiation. The results suggest that MocuFH1 is required for morphogenetic movements of the endoderm and notochord precursor cells during gastrulation and axis formation. The effects of inhibiting MocuFH1 expression on embryonic axis formation in ascidians are similar to those reported for knockout mutations of HNF-3β in the mouse, suggesting that HNF-3/forkhead genes have an ancient and fundamental role in organizing the body plan in chordates.

Our understanding of the molecular mechanisms of embryonic development has increased dramatically with recent discoveries of developmentally important multigene families such as those encoding homeobox, helix-loop-helix, POU domain, and paired domain proteins (reviewed by Kenyon, 1994; Garrell and Campuzano, 1991; Herr et al., 1988; Burri et al., 1989). The proteins encoded by these gene families possess highly conserved DNA-binding domains and have been shown to function as transcription factors. Within these multigene families, similarity in amino acid sequences across species has made possible the cloning of related genes, many of which are involved in embryonic patterning and control of development in a variety of phyla (e.g., McGinnis and Krumlauf, 1992).

Over the past several years, a new family of transcription factors has emerged: the HNF-3/forkhead or winged-helix family. Winged-helix genes have been described in organisms from yeast to humans and share a highly conserved DNA-binding domain of about 110 amino acids (reviewed by Kaufmann and Knöchel, 1996). Since the identification of the first member of the family, the Drosophila gene forkhead, whose mutant phenotype is homeotic transformation of the terminal regions of the embryo (Weigel et al., 1989), a number of HNF-3/forkhead genes have been identified, many of them developmentally expressed. Vertebrate HNF-3/forkhead genes, such as mouse HNF-3β (Sasaki and Hogan, 1994), Xenopus XFKH1 (Dirksen and Jamrich, 1992) and zebrafish axial (Strahle et al., 1993), display expression patterns implying a role in axial patterning and dorsoventral development of the notochord, central nervous system (CNS) and gut. HNF-3β, for example, is first expressed in the mouse node, an area corresponding to the organizer region of Xenopus, which is important in initial patterning of the body axes (Nieuwkoop et al., 1985). Later it is expressed in the notochord and floor plate of the neural tube, where it functions in dorsoventral patterning of the CNS, and in the endoderm, where it is important for liver differentiation. Mouse embryos homozygous for a targeted mutation in the HNF-3β gene show defects in organization of the node, paraxial mesoderm and neural tube, and fail to develop a gut tube (Weinstein et al., 1994; Ang and Rossant, 1994). The role of HNF-3/forkhead genes in cellular processes during gastrulation and axis formation is poorly understood, due to gene multiplicity and the complexity of gastrulation in the vertebrate embryo.

We have used ascidians to investigate the role of forkhead genes in embryogenesis. Ascidians have the smallest genome of any known chordate (Lambert and Laird, 1971), and exhibit only one or two copies of some of the regulatory genes that are present in multiple copies in vertebrates (Holland et al., 1994; di Gregorio et al., 1995; Ma et al., 1996). The ascidian tadpole larva has a simple body plan consisting of six different tissues and about 2500 cells and is considered a prototype of the ancestral chordate (Garstang, 1928; Satoh and Jeffery, 1995). There are 110 cells at the beginning of gastrulation in the ascidian embryo, whereas there are 10,000 cells in the Xenopus embryo. Therefore, the simplicity of the ascidian embryo permits complicated processes, such as cell movements during gastrulation, neurulation and axis formation, to be examined during embryonic development. Despite this reduced complexity, the ascidian tadpole larva exhibits the hallmarks of a chordate: a dorsal CNS, a notochord and a ventral gut.

Here we describe MocuFH1, a member of the HNF-3/forkhead gene family in the ascidian Molgula oculata. The single copy MocuFH1 gene contains a winged-helix domain most closely related to that of axial and HNF-3β, is expressed in mid-line organizer tissues during gastrulation and axis formation, and is necessary for endoderm and notochord cell movements during gastrulation and axis formation.

Biological materials

The ascidian Molgula oculata was collected and maintained at Station Biologique, Roscoff, France. The procedures used to prepare gametes for insemination and embryo culture have been described by Swalla and Jeffery (1990).

RNA isolation, cDNA synthesis and PCR

Total RNA from M. oculata gastrulae was isolated using the guanidinium isothiocyanate method (March et al., 1985). Then, cDNA was synthesized using the First-Strand cDNA Synthesis Kit according to the directions supplied by the manufacturer (Pharmacia Biosystems, Inc., Piscataway, NJ), and polymerase chain reaction (PCR) was performed on this cDNA at an annealing temperature of 47°C. Degenerate primers were designed against the amino acid sequences ITMAIQ (5′-ATHACNATGGCNCAG-3′) and GNMFEN (5′ CCRT-TRTCRAACATRTTNCC-3′), which are highly conserved in the DNA-binding domains of various forkhead genes (Dirksen and Jamrich, 1992). An amplified product of 211 base-pairs (bp) was subcloned into the pCRII vector (TA Cloning Kit, Invitrogen, San Diego, CA) and sequenced using the dideoxy chain termination method (Sanger et al., 1977) with Sequenase (USB; United States Biochemical Corp., Cleveland, OH).

Library screening

A 32P-labeled, random-primed PCR product was used as a probe to screen an M. oculata gastrula cDNA library prepared in the Uni-Zap vector (Stratagene, La Jolla, CA). The probe was hybridized to phage lifts on Biodyne A nylon filters (Pall BioSupport, East Hills, NY) at high stringency. Positive clones were in vivo excised to obtain the pBluescript phagemid with the cDNA insert (ExAssist helper phage kit, Stratagene). One phagemid, designated af3, contained a 2.2-kb insert, which was sequenced and shown to be a full-length ascidian forkhead cDNA. The af3 cDNA clone was designated MocuFH1 (M. oculata ForkHead 1).

Southern and northern blots

Genomic DNA isolated from M. oculata sperm (Davis et al., 1986) was digested with EcoRI and HindIII, subjected to electrophoresis through agarose gels and transferred to nylon membranes (MSI; Fisher Scientific, San Francisco, CA). Blots were probed at high and low stringencies with the random-primed 32P-labeled MocuFH1 cDNA insert, prepared from af3 DNA by EcoRI and XhoI digestion. Blots were also probed at low stringency with the random-primed, 32P-labeled 211-base-pair PCR product described above. The conditions for high and low stringency were as described by Swalla et al. (1993).

RNA for northern blots was isolated from embryos at different developmental stages, subjected to electrophoresis through formaldehyde gels and transferred to nylon membranes (MSI; Fisher Scientific). The phagemid af3, which contained the MocuFH1cDNA insert, was linearized with EcoRI and served as a template for the synthesis of an antisense RNA probe using T7 polymerase (Stratagene, La Jolla, CA) and [32P]UTP (800 Ci/mmol; Amersham, Arlington Heights, IL). Probes were hybridized to the blots and washed at high stringency (Swalla et al., 1993).

In situ hybridizations

For whole-mount in situ hybridizations, M. oculata embryos were dechorionated by treatment with 0.09% Pronase E (Sigma Chemical Company, St Louis, MO), fixed in 4% paraformaldehyde in MOPS buffer for 12 hours at 4°C and stored in 80% ethanol at −20°C. Some embryos were dechorionated manually after fixation using tungsten needles. Whole-mount in situ hybridization was performed according to Ma et al. (1996). Digoxigenin-labeled MocuFH1 antisense probes were synthesized using T7 RNA polymerase and DIG-labeled UTP (BMB; Boehringer Mannheim Biochemicals, Indianapolis, IN), with EcoRI-digested af3 as the template. MocuFH1 sense probes were made using T3 RNA polymerase (Stratagene, La Jolla, CA), with XhoI-digested af3 as the template. The signal was detected using an alkaline phosphatase-coupled secondary antibody and BCIP/X-phosphate reaction (Boehringer-Mannheim), which was allowed to proceed for 1 to 3 hours. Embryos to be photographed were then transferred into 100% ethanol and cleared in a 1:2 mixture of benzyl alcohol: benzyl benzoate (Ma et al., 1996). Embryos to be sectioned were postfixed in 4% paraformaldehyde in PBS, fixed in methanol at −20°C for 20 minutes and then in ethanol at −20°C for 20 minutes, incubated in 1:1 ethanol: polyester wax at 42°C for 1 hour and incubated in polyester wax at 42°C for 1 hour (Swalla et al., 1991). They were then placed in an embedding mold, cooled and sectioned at 8 μm.

For section in situ hybridizations, the method described by Jeffery (1989) was followed using a muscle actin probe synthesized from the MocuMA1 genomic clone (Kusakabe et al., 1996) or a cytoskeletal actin probe synthesized from the MocuCA4 cDNA clone (Swalla et al., 1993).

Oligodeoxynucleotide treatment

The following 18-mer phosphorothiolate-substituted oligodeoxyri-bonucleotides (ODNs) were synthesized by Oligos Etc., Inc. (Wilsonville, OR). Antisense Forkhead 1 (5′-CATAGTGATGTGGA-CAAA-3′) corresponds to nucleotides 29 to 42 of the MocuFH1 cDNA sequence, whereas antisense Forkhead 2 (5′-AGAAGGTG-GCGACGAAAG-3′) and sense Forkhead 2 (5′-CTTTCGTCGCCAC-CTTCT-3′) correspond to nucleotides 46 to 63 of the MocuFH1 cDNA sequence (Fig. 1). The ODNs were stored lyophilized at −20°C and a 30 nmole/μl stock solution was prepared in water prior to use in the experiments. The ODN treatments were carried out as described by Swalla and Jeffery (1996). Briefly, M. oculata embryos (100-150 embryos/ml) were suspended in Millipore-filtered sea water containing 30 μM ODN beginning just after first cleavage (about 60 minutes after insemination) and incubated at 16-20°C until hatching (10-12 hours after insemination). The morphology of the ODN-treated and control embryos was determined by light microscopy (Swalla and Jeffery, 1990). Some of the ODN-treated embryos were fixed in 4% paraformaldehyde or Bouin’s fixative, embedded in Paraplast and sectioned, and the sections were stained with hematoxylin-eosin (Jeffery, 1989). Others were fixed in 4% paraformaldehyde in MOPS buffer at 4°C, then transferred into 100% ethanol and stored at −20°C in preperation for whole-mount in situ hybridization.

Fig. 1.

The nucleotide and deduced amino acid sequences of MocuFH1. The positions of ODN-1 (nucleotides 97-114) and ODN-2 (nucleotides 118-135) are underlined in the 5′ region of the cDNA sequence. The putative forkhead DNA-binding domain is shaded and a putative polyadenylation signal is underlined. The GenBank accession number for MocuFH1 is AF007905.

Fig. 1.

The nucleotide and deduced amino acid sequences of MocuFH1. The positions of ODN-1 (nucleotides 97-114) and ODN-2 (nucleotides 118-135) are underlined in the 5′ region of the cDNA sequence. The putative forkhead DNA-binding domain is shaded and a putative polyadenylation signal is underlined. The GenBank accession number for MocuFH1 is AF007905.

Enzyme assays

Embryos were fixed with 5% formalin-sea water for 30 minutes at 4°C, washed in 0.1 M sodium phosphate buffer. Acetylcholinestase (AChE) activity was assayed as described by Jeffery and Swalla (1991). Alkaline phosphatase activity was assayed as described by Bates and Jeffery (1987a).

Isolation and characterization of MocuFH1

Sequence similarity among the DNA-binding domains of various forkhead genes was used to design degenerate primers to amplify a conserved region from the ascidian M. oculata. A 211 bp PCR product was amplified from gastrulae cDNA, sequenced and found to exhibit a high degree of similarity to the corresponding region of known forkhead genes. The PCR product was used as a probe to screen an M. oculata gastrula cDNA library. The longest cDNA clone obtained, designated MocuFH1, was sequenced and found to contain a single open reading frame (ORF) of 1699 nucleotides (nts) flanked by 5′ and 3′ untranslated regions (UTRs) of 110 and 405 nts, respectively. Of two consecutive ATG triplets at the beginning of the ORF, the first is in a context most similar to the consensus eukaryotic translation initiation site (Kozak, 1991). The 3′ UTR of the MocuFH1 cDNA contains a putative polyadenylation signal followed 14 nts downstream by a poly(A) tail (Fig. 1). The length of the MocuFH1 cDNA clone is consistent with the size of the single 2.3 kb transcript detected in northern blots (see Fig. 4A).

The MocuFH1 nucleotide sequence predicts a protein of 567 amino acids containing a DNA-binding domain (Fig. 1) with a high degree of similarity to those of other forkhead proteins (Fig. 2). When the MocuFH1 sequence was used to search the protein databases by BLAST, the best matches were to class I forkhead proteins (Sasaki and Hogan, 1993), such as zebrafish axial, Xenopus pintallavis/XFKH-1 and the rodent HNF-3 proteins. The alignments of MocuFH1 and some of these proteins are shown in Fig. 2. The forkhead domain of MocuFH1 is 92% identical to those of zebrafish axial and mouse HNF3β and 87% identical to those of Drosophila forkhead and Xenopus pintallavis/ XFKH-1. The forkhead domains of mouse HNF-3α and HNF-3γ (Lai et al., 1991) are only 82% and 84% identical to that of MocuFH1, respectively. The predicted MocuFH1 protein shows additional similarity to other forkhead proteins in two conserved regions near the C terminus (Fig. 2), which have been implicated in transcriptional activation (Pani et al., 1992). However, MocuFH1 shares no sequence similarity with other forkhead proteins outside the putative DNA-binding and transcriptional activation domains. The results suggest that the MocuFH1 gene encodes a member of the class I subgroup of winged-helix proteins and is most closely related to the zebrafish axial and rodent HNF3β genes.

Fig. 2.

Alignment of MocuFH1 with other class I forkhead proteins. MocuFH1 is aligned with zebrafish axial (Strahle et al., 1993), mouse HNF3β (Sasaki and Hogan, 1993), Xenopus pintallavis (Ruiz I Altaba and Jessell, 1992, Xenopus XFKH1 (Dirksen and Jamrich, 1992) and Drosophila forkhead (fkh) (Weigel et al., 1989). The dots represent identical amino acids and the dashes indicate gaps with respect to the MocuFH1 sequence. Light shading indicates the putative forkhead DNA-binding domains and dark shading indicates the putative transcription activation domains.

Fig. 2.

Alignment of MocuFH1 with other class I forkhead proteins. MocuFH1 is aligned with zebrafish axial (Strahle et al., 1993), mouse HNF3β (Sasaki and Hogan, 1993), Xenopus pintallavis (Ruiz I Altaba and Jessell, 1992, Xenopus XFKH1 (Dirksen and Jamrich, 1992) and Drosophila forkhead (fkh) (Weigel et al., 1989). The dots represent identical amino acids and the dashes indicate gaps with respect to the MocuFH1 sequence. Light shading indicates the putative forkhead DNA-binding domains and dark shading indicates the putative transcription activation domains.

MocuFH1 is a single copy gene

The forkhead genes are members of a multigene family (Kaufmann and Knöchel, 1996). Southern blots of M. oculata genomic DNA, hybridized with the full-length MocuFH1 cDNA probe and washed at high stringency, showed only one or two bands (Fig. 3A), suggesting that MocuFH1 is a single-copy gene. Another band appeared in the HindIII digests when the blot was washed at low stringency (Fig. 3B). When the same Southern blot was hybridized with a probe made from the conserved MocuFH1 DNA-binding domain, however, additional bands were observed (Fig. 3C). These bands did not include the additional band seen in HindIII digests when MocuFH1 was used as a probe, suggesting the existence of another gene in M. oculata with homology to the non-DNA-binding region of MocuFH1. The results suggest that MocuFH1 is a single copy gene but that other related genes are present in the M. oculata genome.

Fig. 3.

Southern blot hybridization of MocuFH1 full length and forkhead domain probes to M. oculata genomic DNA. (A,B) A blot hybridized to the full length MocuFH1 probe and washed at high (A) or (B) low stringency. (C) The same blot as in A,B hybridized to the forkhead domain probe and washed at low stringency. Left lane, Hind III digest; Right lane, EcoR1 digest.

Fig. 3.

Southern blot hybridization of MocuFH1 full length and forkhead domain probes to M. oculata genomic DNA. (A,B) A blot hybridized to the full length MocuFH1 probe and washed at high (A) or (B) low stringency. (C) The same blot as in A,B hybridized to the forkhead domain probe and washed at low stringency. Left lane, Hind III digest; Right lane, EcoR1 digest.

Fig. 4.

Temporal and spatial accumulation of MocuFH1 mRNA during M. oculata development. (A) A northern blot containing RNA from (1) gonads, (2) 8-to 16-cell embryos, (3) 32-to 64-cell embryos, (4) mid-gastrulae, (5) neurulae, (6) mid-tailbud embryos and (7) late tailbud embryos. Top: MocuFH1 probe. The single 2.3 kb MocuFH1 transcript is indicated by the arrow. Bottom: 18S rRNA loading control. (B-H) Embryos subjected to whole-mount in situ hybridization with MocuFH1 antisense RNA probe. (B-D,F) Whole mounts. (E,G,H) Sections of whole mounts. (B) A 44-/64-cell embryo showing staining in presumptive notochord (N) and endoderm (E) cells. (C) An early gastrula showing staining of presumptive notochord (N) cells at the anterior lip of the blastopore (Bl), the mesenchyme cells (M) at the lateral lips of the blastopore and endoderm cells (E) that have entered the blastopore. (D,E) Neurulae showing staining in the notochord (N), mesenchyme (M) and endoderm (E) cells. (F-H) Mid-tailbud stage embryos showing staining in the notochord (N) and spinal cord (S) cells in the tail and in the endoderm (E) and brain floor (Bfc) cells in the head. (B,C,H) Embryos are shown with anterior poles at the top; (D-G) embryos are shown with their anterior poles at the left. Scale bars in B,E, 20 μm; magnification is the same in B,C,G, and in D,E,F, respectively.

Fig. 4.

Temporal and spatial accumulation of MocuFH1 mRNA during M. oculata development. (A) A northern blot containing RNA from (1) gonads, (2) 8-to 16-cell embryos, (3) 32-to 64-cell embryos, (4) mid-gastrulae, (5) neurulae, (6) mid-tailbud embryos and (7) late tailbud embryos. Top: MocuFH1 probe. The single 2.3 kb MocuFH1 transcript is indicated by the arrow. Bottom: 18S rRNA loading control. (B-H) Embryos subjected to whole-mount in situ hybridization with MocuFH1 antisense RNA probe. (B-D,F) Whole mounts. (E,G,H) Sections of whole mounts. (B) A 44-/64-cell embryo showing staining in presumptive notochord (N) and endoderm (E) cells. (C) An early gastrula showing staining of presumptive notochord (N) cells at the anterior lip of the blastopore (Bl), the mesenchyme cells (M) at the lateral lips of the blastopore and endoderm cells (E) that have entered the blastopore. (D,E) Neurulae showing staining in the notochord (N), mesenchyme (M) and endoderm (E) cells. (F-H) Mid-tailbud stage embryos showing staining in the notochord (N) and spinal cord (S) cells in the tail and in the endoderm (E) and brain floor (Bfc) cells in the head. (B,C,H) Embryos are shown with anterior poles at the top; (D-G) embryos are shown with their anterior poles at the left. Scale bars in B,E, 20 μm; magnification is the same in B,C,G, and in D,E,F, respectively.

MocuFH1 is expressed during gastrulation and axis formation

The temporal and spatial expression of MocuFH1 during M. oculata embryogenesis was examined by northern blots and in situ hybridization. Northern blots containing gonad and embryo RNA showed a single 2.3 kb transcript, which was first detected during gastrulation, peaked during neurulation and decreased in amount during the tailbud stages (Fig. 4A). MocuFH1 transcripts could not be detected by northern hybridization in gonads, 8-to 16-cell embryos, or 64-cell embryos (although they were detected at the 44-64-cell stage by in situ hybridization; see below). The results suggest that MocuFH1 does not have a maternal expression period and is expressed zygotically.

M. oculata embryos were fixed for whole-mount in situ hybridization at various stages of embryogenesis and some of these embryos were sectioned after hybridization to verify the pattern of MocuFH1 expression (Fig. 4B-H). In 44-64 cell embryos, MocuFH1 transcripts were confined to presumptive endoderm and notochord cells in the vegetal (future dorsal) hemisphere (Fig. 4B). Staining was more concentrated in the notochord than the endoderm cells at this stage. No staining was detected in embryos hybridized with a sense MocuFH1 probe. In gastrulae, MocuFH1 transcripts are present in presumptive endoderm cells, including the large vegetal cells, which initiate invagination (Conklin, 1905), in presumptive notochord cells, which involute over the anterior lip of the blastopore, and in mesenchyme cells, which involute over the lateral lips of the blastopore (Fig. 4C; also see Fig. 7A). MocuFH1 transcripts were not detected in the presumptive muscle cells, which involute over posterior lip of the blastopore, or in the ectodermal cells, which spread over the vegetal hemisphere by epiboly. In neurulae, staining was restricted to the notochord, mesenchyme and endoderm cells (Fig. 4D,E). In tailbud stage embryos, MocuFH1 transcripts were present in endoderm and mesenchyme cells, which enter the developing trunk (head) and presumptive notochord cells, which undergo convergence and extension to form the notochord and larval tail (Fig. 4F-H). MocuFH1 transcripts were also detected in a row of cells on the basal side of the brain (Fig. 4H) and in the spinal cord (Fig. 4G). No transcripts were detected in the endodermal strand or muscle precursor cells at the tailbud stage (Fig. 4D-G). MocuFH1 transcripts persisted in the endoderm and notochord cells through the late tailbud stage, although the staining in the notochord was gradually reduced. The results suggest that the MocuFH1 gene is expressed in endoderm, mesenchyme and notochord cells during gastrulation and in these cells, as well as a restricted number of neural cells, during larval axis formation.

MocuFH1 is necessary for gastrulation and patterning of the tadpole larva

The absence of maternal expression suggested the use of antisense procedures to study the role of the MocuFH1 gene in embryogenesis. Recent experiments with phosphorothiolate-substituted oligodeoxynucleotides (ODNs) have shown that the Manx gene is required for development of the tadpole larva (Swalla and Jeffery, 1996). Therefore, antisense ODNs were used to examine the developmental role of MocuFH1.

Embryos were treated with ODNs beginning at first cleavage and incubated until hatching, about 10-12 hours after fertilization. Antisense ODN-1 had no effect on embryogenesis and was not used in subsequent experiments. Antisense ODN-2, however, showed reproducible effects on gastrulation and axis formation. The phenotypes of ODN-2-treated and control embryos are shown in Fig. 5. Sense and antisense ODN-2-treated embryos were able to cleave and initiate gastrulation normally and showed no morphological differences with respect to the controls. Patterning defects were first observed in antisense ODN-2-treated embryos at the mid-gastrula stage. Ascidian gastrulation is initiated by the invagination of large endoderm precursor cells at the vegetal pole (Conklin, 1905). After the archenteron is formed, additional endoderm cells involute over the lips of the blastopore, and later the presumptive notochord, mesenchyme and muscle cells involute over the anterior, lateral and posterior lips of the blastopore, respectively (see Satoh, 1978; Jeffery, 1992; Swalla, 1993 for descriptions of ascidian gastrulation). Control (Fig. 5A,D) and sense (Fig. 5C,F) ODN-2-treated embryos exhibited normal invagination and involution, but gastrulation was defective in most of the antisense ODN-2-treated embryos (Fig. 5B,E). The effects of antisense ODN-2 varied in severity between different clutches of embryos. However, three classes of embryos could be distinguished in every experiment (Fig. 5M-O). In severely affected embryos, the internalization of the presumptive endoderm cells was incomplete, leaving a large endodermal mass outside the embryo and the presumptive notochord cells were arrested as a cluster anterior to the endodermal mass (Fig. 5E,O). The endodermal mass and internalized endoderm cells stained positively for alkaline phosphatase (Fig. 6E), an endoderm-specific marker in ascidians (Bates and Jeffery, 1987b). However, the cytoskeletal actin gene MocuCA4 (Swalla et al. 1993), which is normally expressed in notochord and muscle cells of M. oculata embryos (Fig. 6C), was expressed in muscle cells but not in notochord cells of ODN-2-treated embryos (Fig. 6D). In moderately affected embryos, an increased number of endodermal cells were able to invaginate, reducing the size of the endodermal mass and some of the notochord cells appeared to involute, but did not undergo further movements inside the embryo (Fig. 5N). The moderately affected embryos also did not express MocuCA4 in notochord cells (data not shown). Even in the most severely affected embryos, however, the presumptive muscle cells were able to involute (Figs 5E, 6A,B), although subsequent movements of the myoblasts were affected and these cells remained as stationary ring around the posterior base of the endodermal mass (see Fig. 6A). Finally, a few embryos in each antisense ODN-2 experiment were able to complete an apparently normal gastrulation (Fig. 5M). The results suggest that antisense ODN-2 inhibits cell movements that begin during mid-gastrulation.

Fig. 5.

The effect of antisense ODNs on M. oculata embryogenesis. (A,D,G,J) Untreated embryos; (B,E,H,K,M-O) embryos treated with antisense ODN-2; (C,F,I,L) embryos treated with sense ODN-2. (A-C) Mid-gastrulae viewed from a lateral side. (D-F) Sagittal sections of mid-gastrulae treated as in A-C. The gastrulae shown in A,D,C,F have normal morphology with endoderm cells (E) invaginated into the archenteron, notochord cells (N) involuted at the anterior lip and muscle cells (M) involuted at the posterior lip of the blastopore (B). The endodermal mass (EM) is shown protruding from the vegetal pole region in embryos treated with antisense ODN-2 (B) but not in untreated embryos (A) or embryos treated with sense ODN-2 (C). The muscle cells involute in untreated and ODN-treated embryos. (G-I) Mid-tail bud stage (11 hour) embryos viewed from a lateral side; (J-L) sections through the head and tail of mid-tailbud (11 hour) stage embryos treated as in G-I. The mid-tailbud embryos shown in G,J and I,L have normal heads containing an otolith (O) and tails (T). Heads and tails are not distinguishable in 11 hour embryos treated with antisense ODN-2 (H,K), which still contain a protruding endodermal mass (EM). (M-O) Sections of mid-gastrulae with normal (M), moderately affected (N) and severely affected morphology after treatment with antisense ODN-2. Scale bars in A,D, 20 μm; magnification is the same in A-C, G-I, D-F, J-L and M-O, respectively.

Fig. 5.

The effect of antisense ODNs on M. oculata embryogenesis. (A,D,G,J) Untreated embryos; (B,E,H,K,M-O) embryos treated with antisense ODN-2; (C,F,I,L) embryos treated with sense ODN-2. (A-C) Mid-gastrulae viewed from a lateral side. (D-F) Sagittal sections of mid-gastrulae treated as in A-C. The gastrulae shown in A,D,C,F have normal morphology with endoderm cells (E) invaginated into the archenteron, notochord cells (N) involuted at the anterior lip and muscle cells (M) involuted at the posterior lip of the blastopore (B). The endodermal mass (EM) is shown protruding from the vegetal pole region in embryos treated with antisense ODN-2 (B) but not in untreated embryos (A) or embryos treated with sense ODN-2 (C). The muscle cells involute in untreated and ODN-treated embryos. (G-I) Mid-tail bud stage (11 hour) embryos viewed from a lateral side; (J-L) sections through the head and tail of mid-tailbud (11 hour) stage embryos treated as in G-I. The mid-tailbud embryos shown in G,J and I,L have normal heads containing an otolith (O) and tails (T). Heads and tails are not distinguishable in 11 hour embryos treated with antisense ODN-2 (H,K), which still contain a protruding endodermal mass (EM). (M-O) Sections of mid-gastrulae with normal (M), moderately affected (N) and severely affected morphology after treatment with antisense ODN-2. Scale bars in A,D, 20 μm; magnification is the same in A-C, G-I, D-F, J-L and M-O, respectively.

Fig. 6.

Expression of muscle, notochord, endoderm markers in M. oculata embryos treated with antisense ODN-2. (A) A whole-mount 11 hour embryo showing AChE activity (red-brown stain) in a ring of muscle cells (MC) around the endodermal mass (EM). The anterior pole of the embryo is at the top. (B) A section of an 11 hour embryo subjected to in situ hybridization with a muscle actin probe showing muscle actin mRNA in muscle cells. The anterior pole of the embryo is at the left. The scale bar in A is 50 μm; magnification is the same in A-E. (C,D) Sections of (C) an sense ODN-2-treated midtailbud embryo and (D) an antisense ODN-2-treated embryo subjected to in situ hybridization with a cytoskeletal actin probe. (E) A whole-mount 11 hour embryo showing AP activity (green) in the internal endoderm and endodermal mass. N, notochord cells; H, larval head; T, larval tail.

Fig. 6.

Expression of muscle, notochord, endoderm markers in M. oculata embryos treated with antisense ODN-2. (A) A whole-mount 11 hour embryo showing AChE activity (red-brown stain) in a ring of muscle cells (MC) around the endodermal mass (EM). The anterior pole of the embryo is at the top. (B) A section of an 11 hour embryo subjected to in situ hybridization with a muscle actin probe showing muscle actin mRNA in muscle cells. The anterior pole of the embryo is at the left. The scale bar in A is 50 μm; magnification is the same in A-E. (C,D) Sections of (C) an sense ODN-2-treated midtailbud embryo and (D) an antisense ODN-2-treated embryo subjected to in situ hybridization with a cytoskeletal actin probe. (E) A whole-mount 11 hour embryo showing AP activity (green) in the internal endoderm and endodermal mass. N, notochord cells; H, larval head; T, larval tail.

Fig. 7.

Expression of MocuFH1 mRNA in antisense ODN-2-treated and control M. oculata embryos. Whole-mount in situ hybridization was done with an MocuFH1 antisense RNA probe. (A-C) Mid-gastrulae that developed from untreated embryos (A) or embryos treated with antisense (B) or sense (C) ODNs. (D-F) Early tailbud embryos that developed from embryos untreated embryos (D) or embryos treated with antisense (E) or sense (F) ODNs. N, notochord; E, endoderm; Bl, blastopore; EM, endodermal mass. Embryos are shown with their anterior poles at the top and photographed from the dorsal (vegetal) side. Scale bar in A is 50 μm; magnification is the same in each frame.

Fig. 7.

Expression of MocuFH1 mRNA in antisense ODN-2-treated and control M. oculata embryos. Whole-mount in situ hybridization was done with an MocuFH1 antisense RNA probe. (A-C) Mid-gastrulae that developed from untreated embryos (A) or embryos treated with antisense (B) or sense (C) ODNs. (D-F) Early tailbud embryos that developed from embryos untreated embryos (D) or embryos treated with antisense (E) or sense (F) ODNs. N, notochord; E, endoderm; Bl, blastopore; EM, endodermal mass. Embryos are shown with their anterior poles at the top and photographed from the dorsal (vegetal) side. Scale bar in A is 50 μm; magnification is the same in each frame.

Some of the ODN-2-treated and control embryos were fixed at the gastrula and early tailbud stages and subjected to in situ hybridization to monitor the accumulation of MocuFH1 mRNA. The gastrulae that developed from controls and sense ODN-2-treated embryos showed normal distributions of MocuFH1mRNA (Fig. 7A,C). In contrast, most moderately and severely affected gastrulae that developed from embryos treated with antisense ODN-2 showed lower levels of staining in the endodermal mass (Fig. 7B). When early tailbud embryos were examined, however, the level of staining was similar in the control and ODN-2-treated embryos (Fig. 7D-F), although the embryos treated with antisense ODN-2 showed an endodermal mass and presumptive notochord. Thus, the initial reduction in MocuFH1mRNA at the gastrula stage in antisense ODN-2-treated embryos appears to be temporary. A transient reduction of mRNA in chick embryos treated with antisense ODNs corresponding to the Slug gene has also been reported (Nieto et al., 1994). The normal gastrulae that developed from embryos treated with antisense ODN-2 stained more intensely for MocuFH1 mRNA after in situ hybridization than those with defective gastrulation (data not shown), suggesting that variations in the effects of antisense ODN-2 are based on differential suppression of target mRNA. Similar variability in experiments with Manx antisense ODNs was shown to be related to ODN penetration (Swalla and Jeffery, 1996). The results suggest that antisense ODN-2 reduces the level of MocuFH1 transcripts early during gastrulation.

The inability of most antisense ODN-2-treated embryos to gastrulate also affected subsequent axis formation during larval development (Table 1). After gastrulation, most of the endoderm cells move toward the anterior pole of the embryo to form the larval head, and the notochord, muscle and some endoderm cells (the endodermal strand) move posteriorly to form the larval tail. The CNS is formed during neurulation by folding of ectodermal cells at the dorsal midline of the embryo to form the neural tube. Controls and embryos treated with sense ODN-2 developed into normal tadpole larvae (Fig. 5G,I,J,L). In contrast, embryos treated with antisense ODN-2 failed to develop an embryonic axis at frequencies comparable to the controls (Table 1). The severely affected embryos formed mushroom-shaped structures with a bulging endodermal mass, but lacking a definitive head, a CNS with an otolith and a tail (Fig. 5H,K). Sections through these embryos showed that they lacked a neural tube and that the endodermal mass contained endoderm and notochord cells (Fig. 5K), but the latter were not arranged in single file, as they appear in the tail of tadpole larvae. The moderately affected embryos did not form mushroom-shaped bodies, but appeared as permanent gastrulae, not extending a notochord or developing an apparent anteroposterior axis. The appearance of these embryos and the expression of AChE in non-migratory myocytes resemble the phenotype of Molgula occulta, an ascidian species closely related to M. oculata that does not undergo notochord and tail formation during normal larval development (Swalla and Jeffery, 1990; Jeffery and Swalla, 1991). The defects in larval development did not include the muscle cells, which do not express the MocuFH1 gene (Fig. 4), and were able to express muscle actin mRNA and the muscle-specific enzyme AChE (Fig. 6A,B). The results suggest that MocuFH1 expression is required for morphogenetic movements of the notochord and endoderm cells during gastrulation and axis formation.

Table 1.

Effect of MocuFH1 ODNs on development of the tadpole larva

Effect of MocuFH1 ODNs on development of the tadpole larva
Effect of MocuFH1 ODNs on development of the tadpole larva

The HNF-3/forkhead genes are members of a family of winged-helix transcription factors expressed in dorsal mid-line organizing centers in vertebrate embryos. We describe here the ascidian gene MocuFH1, which encodes a winged-helix protein with a DNA-binding domain most closely related to zebrafish axial and rodent HNF-3β proteins. The MocuFH1 gene is expressed in mid-line tissues of the ascidian embryo, including the presumptive notochord, endoderm and ventral cells of the CNS. Functional studies show that MocuFH1 is required for morphogenetic movements of the notochord and endoderm cells during gastrulation and axis formation, indicating that the HNF-3 /forkhead genes have a fundamental role in establishing the chordate body plan.

Ascidians have the smallest genome of any chordate (Lambert and Laird, 1971). Accordingly, genes that exist in multiple copies in vertebrates, such as the forkhead genes (Lai et al., 1991; Sasaki and Hogan, 1993), may be fewer and function in a less complicated fashion in ascidians (Holland et al., 1994; di Gregorio et al., 1995; 1994; Ma et al., 1996). Southern blot analysis showed that MocuFH1 is a single copy gene. Nonetheless, hybridization with a probe restricted to the MocuFH1 DNA-binding domain suggests that there are other related winged-helix genes in the M. oculata genome. The conserved DNA-binding domain of MocuFH1 is most similar to the forkhead domain of zebrafish axial and the rodent HNF3β proteins. Except for this region and two putative transcription activation domains, MocuFH1 is divergent from all known winged-helix proteins. Thus, until additional ascidian winged-helix genes are identified, the evolutionary relationship between the MocuFH1 and HNF3β genes cannot be determined with certainty. Considering the low number of bands in Southern blots, however, it is likely that ascidians have fewer winged-helix genes than higher vertebrates.

Despite its divergence from other winged-helix proteins, the embryonic expression pattern of MocuFH1 is remarkably similar to that of the vertebrate HNF-3 genes. In Xenopus and the mouse, one or more HNF-3 genes are first expressed in the organizer/node region, and later in the notochord, endoderm and floor plate (Lai et al., 1991; Dirksen and Jamrich, 1992; Sasaki and Hogan, 1993; Ruiz i Altaba et al., 1993). The expression of MocuFH1 in vegetal blastomeres of 44-/64-cell embryos, in the presumptive endoderm, mesenchyme and notochord cells during gastrulation and axis formation, and in ventral mid-line cells in the CNS of ascidian embryos resembles the expression patterns seen in vertebrate embryos. The similarity of MocuFH1 expression in ascidians and vertebrates, members of different chordate subphyla, suggests that their embryonic midline tissues may deploy the same genes to organize the body plan.

The possible homology of the ascidian and vertebrate notochord has been appreciated for some time, and is supported by expression of Brachyury (T) (Yasuo and Satoh, 1994; Corbo et al., 1997) and the HNF-3/ forkhead related gene in the ascidian notochord. The expression pattern of the MocuFH1 gene described here also provides evidence that ventral cells of the ascidian CNS, including both the brain and spinal cord, have a common evolutionary origin with the floor plate cells of the vertebrate CNS. During vertebrate body axis formation, HNF-3β seems to induce the floor plate by activating the expression of sonic hedgehog in the notochord and floor plate cells (Echelard et al., 1993; Sasaki and Hogan, 1994). Thus, it will be interesting to determine whether a similar relationship exists between MocuFH1 and sonic hedgehog in the notochord and CNS during axis formation in the ascidian embryo. There are also similarities between the MocuFH1 and HNF-3 genes in the endoderm and its derivatives. The HNF-3 genes are expressed in the liver, lungs, intestine and stomach of adult mice (reviewed by Kaufman and Knöchel, 1996). Likewise, the MocuFH1 gene is expressed in the differentiating endoderm throughout the development of the tadpole larva. Although the ascidian larva does not exhibit tissues resembling the endodermally derived organs of vertebrates, a pharynx with gill slits, stomach and intestine are prominent features of the adult, which differentiate from MocuFH1-expressing cells in the head endoderm during metamorphosis.

Although definitive transplantation experiments have not been done in ascidians, fate-mapping studies have shown that the presumptive notochord cells reside at the anterior lip of the blastopore (Conklin, 1905; Nishida, 1987), suggesting that this region may be the functional equivalent to the vertebrate organizer (Nieuwkoop et al., 1985). Consistent with this idea, the presumptive notochord cells exhibit relatively high levels of MocuFH1 expression in the ascidian embryo. The expression domain of MocuFH1 also includes prospective endoderm at the vegetal pole, the first cells to invaginate during gastrulation (Conklin, 1905). The sensitivity of the vegetal pole region to ultraviolet irradiation during ooplasmic segregation (Jeffery, 1990) suggests that this region may function similarly to the Nieuwkoop Center in Xenopus embryos. Endoderm cells that arise from the vegetal pole region of the ascidian embryo contain cytoplasmic factors responsible for initiating gastrulation (Bates and Jeffery, 1987b; Nishida, 1996). These factors are candidates for direct or indirect regulators of MocuFH1 expression in the invaginating endoderm.

Another similarity between MocuFH1 and HNF-3β genes is lack of expression in mesodermal derivatives responsible for muscle cell formation. In the ascidian embryo, the presumptive larval muscle cells involute over the posterior lip of the blastopore and are devoid of MocuFH1 transcripts at this stage as well as later stages of development when myoblasts differentiate into larval tail muscle cells. Similarly, mammalian embryos do not express HNF-3β in paraxial mesoderm, although related winged-helix genes are expressed in the somites and other non-notochordal mesodermal derivatives (Sasaki and Hogan, 1993; Miura et al., 1993; Kaestner et al., 1993). Despite the lack of HNF-3β expression in paraxial mesoderm, knockout mice exhibit defects in somite organization during later development, presumably due to the absence of a notochord (Weinstein et al, 1994). Although ascidian embryos deficient in MocuFH1 transcripts are able to express muscle cell markers, the myoblasts are not organized properly into muscle bands, as they are in the tail of normal larvae. The inability of myoblasts to differentiate into muscle cells in antisensetreated ascidian embryos could also be caused by the absence of a notochord.

In mouse embryos, loss of HNF-3β gene function through targeted mutagenesis leads to defective development of the node, notochord and floorplate (Ang and Rossant, 1994; Weinstein et al., 1994). Earlier effects on gastrulation have been more difficult to interpret, primarily because of the relative complexity of cell movements. In contrast, ascidian embryos begin gastrulation with only 110 cells (Conklin, 1905), the cell movements initiated at this stage of development are well known (Satoh, 1978; Jeffery, 1992; Swalla, 1993) and a fate map of the gastrula is available (Nishida, 1987). When M. oculata embryos are treated with antisense ODNs at first cleavage, the succeeding cleavages are normal and gastrulation is initiated by invagination of the endoderm cells, but the ensuing movements of endoderm and notochord cells are inhibited. Later in development, the notochord, CNS and embryonic axis do not appear, resulting in a grossly abnormal embryo. The effects of antisense ODN-2 are thought to be specific for MocuFH1 because (1) antisense ODN-2 suppresses the accumulation of MocuFH1 transcripts during gastrulation, (2) sense ODN-2 and other antisense ODNs have no affect on development, (3) antisense ODN-2 does not affect other zygotic genes, such as the muscle actin and AChE genes, (4) there are no direct effects on gastrulating tissues that do not express the MocuFH1 gene, such as the prospective larval muscle cells, and (5) the effects of antisense MocuFH1 ODNs on embryonic development are different from those of antisense ODNs corresponding to other genes (Swalla and Jeffery, 1996). We conclude that MocuFH1 is required for cell movements beginning at the midgastrula stage, including involution of endoderm and notochord cells during gastrulation, and interdigitation and posterior movements of notochord precursor cells during axis formation. The effects of inhibiting MocuFH1 on the presumptive mesenchyme cells, which express MocuFH1 and involute over the lateral lips of the blastopore, cannot be evaluated at present because of the lack of appropriate markers. Finally, it seems likely that failure of CNS development is due to the inability of the notochord/endoderm cells to signal to neuroectoderm and/or to attain the appropriate spatial organization to interact with the neuroectoderm.

The downstream targets of the winged-helix transcription factors have not been elucidated in great detail, although some evidence suggests that Brachyury (T) may act downstream of HNF3β in the mouse (Weinstein et al., 1994). The fact that alkaline phosphatase activity is present in ascidian embryos treated with forkhead antisense ODNs suggests that expression of this enzyme is not dependent on MocuFH1 in the endoderm. Unfortunately, we were unable to examine the relationship between the MouFH1 and T genes in the notochord because the ascidian Brachyury cDNA, cloned from Halocynthia roretzi (Yasuo and Satoh, 1994), does not react with Molgula oculata, which is phylogenetically distant from other ascidians (Hadfield et al., 1995). However, expression of the cytoskeletal actin gene MocuCA4 in the notochord, but not the muscle cells, was inhibited by treatment with forkhead antisense ODNs. This suggests that MocuCA4 is a direct or indirect downstream target of MocuFH1 in the developing notochord.

It is significant that the effects of MocuFH1 inhibition in some M. oculata embryos resemble the natural phenotype of M. occulta, a closely related species that lacks postgastrulation morphogenetic movements resulting in notochord and tail development (Swalla and Jeffery, 1990). Several M. oculata genes have been identified that are downregulated during M. occulta development (Swalla et al, 1993), including Manx, which is required for tail formation (Swalla and Jeffery, 1996). It will be interesting to investigate MocuFH1 expression and function in M. occulta embryos, and to determine whether there are interactions with the Manx gene.

The ascidians (along with the salp and larvacean tunicates) are the sister group of a clade containing the cephalochordates (amphioxus) and the vertebrates (Wada and Satoh, 1994). As such, the ascidians represent a basal group within the Phylum Chordata. The deployment of HNF-3/ forkhead genes to organize cell movements during gastrulation and axis formation in ascidians and mammals suggest that these genes have an ancient and fundamental role in organizing the body plan during chordate development.

We thank D. Martasian and J. Machula for technical assistance, the collectors at Station Biologique de Roscoff for providing M. oculata, Dr. B. J. Swalla for assistance in obtaining M. oculata embryos and providing RNA and DNA samples, and Dr. J. Natzle for providing laboratory facilities. This research was supported by NSF (IBN-94-17799) and NIH (HD-13970) research grants to W. R. J.

Ang
,
S.-L.
and
Rossant
,
J.
(
1994
).
HNF-3β is essential for node and notochord formation in mouse development
.
Cell
78
,
561
574
.
Bates
,
W. R.
and
Jeffery
,
W. R.
(
1987a
).
Alkaline phosphatase expression in ascidian egg fragments and andromerogons
.
Dev. Biol
.
119
,
382
389
.
Bates
,
W. R.
and
Jeffery
,
W. R.
(
1987b
).
Localization of axial determinants in the vegetal pole region of ascidian eggs
.
Dev. Biol
.
124
,
65
76
.
Burri
,
M.
Tromvoukis
,
Y.
,
Bopp
,
D.
,
Frigerio
,
G.
and
Noll
,
M.
(
1989
).
Conservation of the paired domain in metazoans and its structure in three isolated human genes
.
EMBO J
.
8
,
1183
1190
.
Conklin
,
E. G.
(
1905
).
The organization and cell lineage of the ascidian egg
.
J. Acad. Nat. Sci. (Philadelphia)
13
,
1
119
.
Corbo
,
J. C.
,
Levine
,
M.
and
Zeller
,
R. W.
(
1997
).
Characterization of a notochord-specific enhancer from the Brachyury promoter region of the ascidian, Ciona intestinalis
.
Development
124
,
589
602
.
Davis
,
L. G.
,
Dibner
,
M. D.
and
Batty
,
J. F.
(
1986
).
Basic Methods in Molecular Biology
.
New York
:
Elsevier Press
.
di Gregorio
,
A.
,
Spagnuolo
,
A.
,
Ristoratore
,
F.
,
Pischetola
,
M.
,
Aniello
,
F.
,
Branno
,
M.
,
Cariello
,
L.
and
Di Lauro
,
R.
(
1995
).
Cloning of ascidian homeobox gene provides evidence for a primordial chordate cluster
.
Gene
156
,
253
257
.
Dirksen
,
M. L.
and
Jamrich
,
M.
(
1992
).
A novel, activin-inducible, blastopore lip-specific gene of Xenopus laevis contains a fork head DNA-binding domain
.
Genes Dev
.
6
,
599
608
.
Echelard
,
Y.
,
Epstein
,
D. J.
,
St-Jacques
,
B.
,
Shen
,
L.
,
Mohler
,
J.
,
McMahon
,
J. A.
and
McMahon
,
A. P.
(
1993
).
Sonic hedgehog, a member of a family of putative signalling molecules, is implicated in the regulation of CNS polarity
.
Cell
75
,
1417
1430
.
Garrell
,
J.
and
Campuzano
,
S.
(
1991
).
The helix-loop-helix domain: a common motif for bristles, muscles and sex
.
BioEssays
13
,
493
498
.
Garstang
,
W.
(
1928
).
The morphology of the tunicata, and its bearings on the phylogeny of the chordata
.
Q. J. Microsc. Sci
.
72
,
51
87
.
Hadfield
,
K. A.
,
Swalla
,
B. J.
and
Jeffery
,
W. R.
(
1995
).
Multiple origins of anural development in ascidians inferred from rDNA sequences
.
J. Mol. Evol
.
40
:
413
427
.
Herr
,
W.
,
Sturm
,
R. A.
,
Clerc
,
R. G.
,
Corcoran
,
L. M.
,
Baltimore
,
D.
,
Sharp
,
P. A.
,
Ingham
,
H. A.
,
Rosenfeld
M. G.
,
Finney
,
M.
,
Ruvkun
,
G.
and
Horvitz
,
H. R.
(
1988
).
The POU domain: A large conserved region in the mammalian pit-l, oct-l. oct-2, and Caenorhabditis elegans unc-86 gene products
.
Genes Dev
.
2
,
1513
1516
.
Holland
,
P. W. H.
,
Garcia-Fernandez
,
J.
,
Williams
,
N. A.
and
Sidrow
,
A.
(
1994
).
Gene duplications and the origin of vertebrate development
.
Development
116
,
125
133
.
Jeffery
,
W. R.
(
1989
).
Requirement of cell division for muscle actin expression in the primary muscle cell lineage of ascidian embryos
.
Development
105
,
75
84
.
Jeffery
,
W. R.
(
1990
).
Ultraviolet irradiation during ooplasmic segregation prevents gastrulation, sensory cell induction, and axis formation in the ascidian embryo
.
Dev. Biol
.
140
,
388
400
.
Jeffery
,
W. R.
(
1992
).
A gastrulation center in the ascidian egg
.
Development
1992
Supplement,
53
63
.
Jeffery
,
W. R.
and
Swalla
,
B. J.
(
1991
).
An evolutionary change in the muscle cell lineage of an anural ascidian embryo is restored by interspecific hybridization with a urodele ascidian
.
Dev. Biol
.
145
,
328
337
.
Kaestner
,
K. H.
,
Lee
,
K.-H.
,
Schlöndorff
,
J.
,
Hiemisch
,
H.
,
Monaghan
,
A.. P.
and
Schütz
,
G.
(
1993
).
Six members of the mouse forkhead gene family are developmentally regulated
.
Proc. Natl. Acad. Sci. USA
90
,
7628
7631
.
Kaufmann
,
E.
and
Knöchel
,
W.
(
1996
).
Five years on the wings of fork head
.
Mech. Dev
.
57
,
3
20
.
Kenyon
,
C.
(
1994
).
If birds can fly, why can’t we? Homeotic genes and evolution
.
Cell
78
,
175
180
.
Kozak
,
M.
(
1991
).
Structural features in eucaryotic mRNAs that modulate the initiation of translation
.
J. Biol. Chem
.
266
,
19867
19870
.
Kusakabe
,
T.
,
Swalla
,
B. J.
,
Satoh
,
N.
and
Jeffery
,
W. R.
(
1996
).
Mechanism of an evolutionary change in muscle cell differentiation in ascidians with different modes of development
.
Dev. Biol
.
174
,
379
392
.
Lai
,
E.
,
Prezioso
,
Y. R.
,
Tao
,
W.
,
Chen
,
W. S.
and
Darnell
,
J. E.
(
1991
).
Hepatocyte nuclear factor 3α belongs to a gene family in mammals that is homologous to the Drosophila homeotic gene fork head
.
Genes Dev
.
5
,
416
427
.
Lambert
,
C. C.
and
Laird
,
C.
(
1971
).
Molecular properties of tunicate DNA
.
Biochim. Biophys. Acta
240
,
39
45
.
Ma
,
L.
,
Swalla
,
B. J.
,
Zhou
,
Z.
,
Dobias
,
S. L.
,
Bell
,
J. R.
,
Chen
,
J.
,
Maxson
,
R. E.
and
Jeffery
,
W. R.
(
1996
).
Expression of an Msx homeobox gene in ascidians: insights into the archetypal chordate expression pattern
.
Dev. Dynam
.
205
,
308
318
.
March
,
C. J.
,
Mosley
,
B.
,
Larsen
,
A.
,
Cerretti
,
D. P.
,
Braedt
,
G.
,
Price
,
V.
,
Gillis
,
S.
,
Henney
,
C. S.
,
Kronheim
,
S. R.
,
Grabstein
,
K.
,
Conlon
,
P. J.
,
Hopp
,
T. P.
and
Cosman
,
D.
(
1985
).
Cloning, sequence and expression of two distinct human interleukin-1 complementary DNAs
.
Nature
315
,
641
647
.
McGinnis
,
W.
and
Krumlauf
,
R.
(
1992
).
Homeobox genes and axial patterning
.
Cell
68
,
283
302
.
Miura
,
N.
,
Wanaka
,
A.
,
Tohyama
,
M.
and
Tanaka
,
K.
(
1993
).
MFH-1, a new member of the fork head domain family, is expressed in developing mesenchyme
.
FEBS Lett
.
326
,
171
176
.
Nieto
,
M. A.
,
Sargent
,
M. G.
,
Wilkinson
,
D. G.
and
Cooke
,
J.
(
1994
).
Control of cell behavior during vertebrate development by Slug, a zinc finger gene
.
Science
264
,
835
839
.
Nishida
,
H.
(
1987
).
Cell lineage analysis in ascidian embryos by intracellular injection of a tracer enzyme. III. Up to the tissue restricted stage
.
Dev. Biol
.
121
,
526
541
.
Nishida
,
H.
(
1996
).
Vegetal egg cytoplasm promotes gastrulation and is responsible for specification of vegetal blastomeres in embryos of the ascidian Halocynthia roretzi
.
Development
122
,
1271
1279
.
Nieuwkoop
,
P. D.
,
Johnen
,
A. G.
and
Albers
,
B.
(
1985
).
The Epigenetic Nature of Chordate Development
.
Cambridge
:
Cambridge University Press
.
Pani
,
L.
,
Overdier
,
D. G.
,
Porcella
,
A.
,
Qian
,
X.
,
Lai
,
E.
and
Costa
,
R. H.
(
1992
).
Hepatocyte nuclear factor 3β contains two transcription activation domains, one of which is novel and conserved with Drosophila fork head protein
.
Mol. Cell. Biol
.
12
,
3723
3732
.
Ruiz i Altaba
,
A.
and
Jessell
,
T. M.
(
1992
).
Pintallavis, a gene expressed in the organizer and dorsal midline cells of frog embryos: Involvement in the development of the neural axis
.
Development
11
,
81
93
.
Ruiz i Altaba
,
A.
,
Prezioso
,
V. R.
,
Darnell
,
J. E.
and
Jessell
,
T.
(
1993
).
Sequential expression of HNF-3β and HNF-3α by embryonic organizing centers: the dorsal lip/node, notochord, and floor plate
.
Mech. Dev
.
44
,
91
108
.
Sanger
,
F.
,
Nicklen
,
S.
and
Coulson
,
A. R.
(
1977
).
DNA sequencing with chain terminating inhibitors
.
Proc. Nat. Acad. Sci. USA
74
,
5463
5467
.
Sasaki
,
H.
and
Hogan
,
B. L. M.
(
1993
).
Differential expression of multiple fork head related genes during gastrulation and axial pattern formation in the mouse embryo
.
Development
118
,
47
59
.
Sasaki
,
H.
and
Hogan
,
B. L. M.
(
1994
).
HNF-3β as a regulator of floor plate development
.
Cell
76
,
103
115
.
Satoh
,
N.
(
1978
).
Cellular morphology and architecture during early morphogenesis of the ascidian egg: an SEM study
.
Biol. Bull. Mar. Biol. Lab. Woods Hole
155
,
608
614
.
Satoh
,
N
and
Jeffery
,
W. R.
(
1995
).
Chasing tails in ascidians: developmental insights into the origin and evolution of chordates
.
Trends Genet
.
11
,
354
359
.
Strahle
,
U.
,
Blader
.
P.
,
Henrique
,
D.
and
Ingham
,
P. W.
(
1993
).
Axial, a zebrafish gene expressed along the developing body axis, shows altered expression in cyclops mutant embryos
.
Genes Dev
.
7
,
1436
1446
.
Swalla
,
B. J.
(
1993
).
Mechanisms of gastrulation and tail formation in ascidians
.
Microsc. Res. Tech
.
26
,
274
284
.
Swalla
,
B. J.
and
Jeffery
,
W. R.
(
1990
).
Interspecific hybridization between an anural and urodele ascidian: differential expression of urodele features suggests multiple mechanisms control anural development
.
Dev. Biol
.
142
,
319
334
.
Swalla
,
B. J.
and
Jeffery
,
W. R.
(
1996
).
Requirement of the Manx gene for restoration of chordate features in a tailless ascidian embryo
.
Science
274
,
1205
1208
.
Swalla
,
B. J.
,
Badgett
,
M. R.
and
Jeffery
,
W. R.
(
1991
).
Identification of a cytoskeletal protein localized in the myoplasm of ascidian eggs: Localization is modified during anural development
.
Development
111
,
425
436
.
Swalla
,
B. J.
,
Makabe
,
K. W.
,
Satoh
,
N.
and
Jeffery
,
W. R.
(
1993
).
Novel genes expressed differentially in ascidians with alternate modes of development
.
Development
119
,
307
318
.
Wada
,
H.
and
Satoh
,
N
(
1994
).
Details of the evolutionary history from invertebrates to vertebrates, as deduced from the sequence of 18S rDNA
.
Proc. Natl. Acad. Sci. USA
91
,
1801
1804
.
Weigel
,
D.
,
Jurgens
,
G.
,
Kuttner
,
F.
,
Seifert
,
E.
and
Jäckle
,
H.
(
1989
).
The homeotic gene fork head encodes a nuclear protein and is expressed in the terminal regions of the Drosophila embryo
.
Cell
57
,
645
658
.
Weinstein
.
D. C.
,
Ruiz i Altaba
.
A.
.
Chen
,
W. S.
,
Hoodless
,
P.
,
Prezioso
,
V. R.
,
Jessell
,
T. M.
and
Darnell
,
J. E.
(
1994
).
The winged-helix transcription factor HNF-3β is required for notochord development in the mouse embryo
.
Cell
78
,
575
588
.
Yasuo
,
H
and
Satoh
,
N.
(
1994
).
An ascidian homolog of the mouse Brachyury (T) gene is expressed exclusively in notochord cells at the fate restricted stage
.
Dev. Growth Diff
.
36
,
9
18
.

While this article was in proof, Corbo et al. (Development 124, 2335-2344, 1997) reported the isolation of a Ci-frch, a forkhead gene in the ascidian Ciona intestinalis. The Ci-frch and MocuFH1 proteins show 96% conservation in their forkhead domains but no significant amino acid conservation outside this region, and are possibly encoded by different forkhead genes. The Ci-frch and MocuFH1 genes also show slightly different expression patterns.