Mesoderm formation is critical for the establishment of the animal body plan and in Drosophila requires the snail gene. This report concerns the cloning and expression pattern of the structurally similar gene snail1 from zebrafish. In situ hybridization shows that the quantity of snail1 RNA increases at the margin of the blastoderm in cells that involute during gastrulation. As gastrulation begins, snail1 RNA disappears from the dorsal axial mesoderm and becomes restricted to the paraxial mesoderm and the tail bud. snail1 RNA increases in cells that define the posterior border of each somite and then disappears when somitic cells differentiate. Later in development, expression appears in cephalic neural crest derivatives. Many snail1-expressing cells were missing from mutant spadetail embryos and the quantity of snail1 RNA was greatly reduced in mutant no tail embryos. The work presented here suggests that snail1 is involved in morphogenetic events during gastrulation, somitogenesis and development of the cephalic neural crest, and that no tail may act as a positive regulator of snail1.

Gastrulation in a variety of organisms is a critical event that leads to the formation of the three embryonic layers endoderm, mesoderm and ectoderm. In vertebrates, much work has focussed on mesoderm since it plays a pivotal role in organizing the body axis. Although the morphogenetic events resulting in mesoderm formation vary among different species, a number of genes and gene products involved in mesoderm formation have been conserved during evolution. Studies of Xenopus embryos have shown, for example, that molecules related to peptide growth factors (for review see Hopwood, 1990; Jessell and Melton, 1992) and activin (for review see Moon and Christian, 1992) play a role in initiating mesoderm formation in vivo. One of their target genes is the frog homologue of the mouse Brachyury or T gene (Smith et al., 1991), which is crucial for mesoderm development (Herrmann et al., 1990).

Another way to identify genes that might be involved in mesoderm formation in vertebrates is to search for genes similar to those that program a comparable process in Drosophila melanogaster. Genetic and molecular analyses have identified a regulatory cascade of maternally expressed genes that leads to a gradient of the transcription factor dorsal in cell nuclei and thus define pattern along the dorsoventral axis (Roth et al., 1989; Rushlow et al., 1989; Steward, 1989). The dorsal protein then initiates the expression of at least two genes, twist (Thisse et al., 1991; Jiang et al., 1991; Pan et al., 1991; Thisse and Thisse, 1992) and snail (Ip et al., 1991) that are required for mesoderm formation (Thisse et al., 1988; Alberga et al., 1991). Both twist and snail genes have been shown to be conserved in vertebrates: a gene homologous to twist has been identified in Xenopus (Xtwi, Hopwood et al., 1989) and in mouse (Mtwist, Wolf et al., 1991), and snail homologues have been identified in Xenopus (Xsna, Sargent and Bennett, 1990) and mouse (Sna, Nieto et al., 1992; Smith et al., 1992).

The zebrafish (Brachydanio rerio) has advantages for studying vertebrate development because it combines the benefits of amphibians as an embryological system and mice as a genetic system (Streisinger et al., 1981, 1986; Kimmel, 1989). Here, we report the cloning and structural analysis of a zebrafish snail gene (called snail1). Our results show that snail1 encodes a zinc finger protein that is structurally more related to the mouse Sna than the Xenopus Xsna. We then describe the distribution of snail1 RNA in more detail than has been possible in other vertebrates because of the transparency of zebrafish embryos and improvements in the method of in situ hybridization. We show that snail1 in zebrafish is expressed in paraxial mesoderm during gastrulation and then in somites. Later on, it is detected in head mesoderm and neural crest cell derivatives.

We also examined the expression pattern of snail1 in two mutants affecting mesoderm formation. We show a shift of snail1-expressing cells to the tail bud in spadetail (Kimmel et al., 1989; Ho and Kane, 1990), consistent with a failure in the convergence of paraxial mesoderm precursors in this mutant. We also studied the expression pattern of snail1 in mutant no tail (ntl) embryos, which are defective in axial mesoderm development (Schulte-Merker et al., 1992, 1994a; Halpern et al., 1993) similar to Brachyury in mice (Chesley, 1935; Gluecksohn-Schoenheimer, 1944; Yanagisawa et al., 1981). We show that no tail, the zebrafish homologue of the mice T gene (Halpern et al., 1993; Schulte-Merker et al., 1994b), may act as a positive regulator of snail1.

The results suggest the hypothesis that snail1 is involved in morphogenetic movements during epiboly and gastrulation, cell condensation during formation of somites and the formation of cranial cartilages.

Cloning and sequencing of snail1 cDNA

To clone the full-length snail1 cDNA, a gastrula cDNA library (gift from B. Riggleman) was screened at low stringency as published in Wolf et al. (1991). The probe used was the entire cDNA of the Drosophila snail gene (kindly provided by A. Alberga, described in Boulay et al., 1987). Deletions were made according to Lin et al. (1985) and sequencing was performed from single-strand templates by the dideoxy-termination method using Sequenase (USB Inc.) following the manufacturer’s directions.

In situ hybridization on whole-mount embryos

The procedure described by Harland (1991) was used with the following modifications. The digoxigenin RNA probes were synthesized and subjected to alkaline hydrolysis according to Boehringer Mannheim Biochemica recommendations (Cat 1175041) to provide of an average length of 100-200 nucleotides. Embryos staged as described by Kimmel et al. (1993) were fixed 24 hours in 4% paraformaldehyde 1× PBS (phosphate-buffered saline), hand dechorionated and dehydrated overnight in methanol at −20°C. Then the embryos were rehydrated stepwise in methanol/PBS and finally put back in 100% PBT (1× PBS 0.1% Tween 20). Embryos older than the beginning of somitogenesis were treated 10 minutes with proteinase K (10 μg/ml in PBT). The reaction was stopped by rinsing in glycine (2 mg/ml in PBT). Embryos were postfixed in 4% paraformaldehyde 1× PBS for 20 minutes and then rinsed in PBT 5 times for 5 minutes each. The embryos were prehybridized at least 1 hour at 70°C in hybridization buffer [50% formamide, 5× SSC, 50 μg/ml heparin, 500 μg/ml tRNA, 0.1% Tween 20, 9 mM citric acid]. The hybridization was done in the same buffer containing 50 ng to 100 ng of probe overnight at 70°C. Then the embryos were washed at 70°C for 10 minutes in [75% hybridization buffer, 25% 2× SSC], 10 minutes in [50% hybridization buffer, 50% 2× SSC], 10 minutes in [25% hybridization mix, 75% 2× SSC], 10 minutes in 2× SSC, 2 times 30 minutes in 0.2× SSC. Further washes were performed at room temperature for 5 minutes in [75% 0.2× SSC, 25% PBT], 5 minutes in [50% 0.2× SSC, 50% PBT], 5 minutes in [25% 0.2× SSC, 75% PBT], 5 minutes in PBT, and then 1 hour in [PBT with 2 mg/ml BSA (bovine serum albumin), 2% sheep serum]. Then the embryos were incubated overnight at 4°C with the preabsorbed alkalinephosphatase-coupled anti-digoxigenin antiserum (described in Boehringer instruction manual) at a 1/5000 dilution in a PBT buffer containing 2 mg/ml BSA, 2% sheep serum. Finally the embryos were washed 6 times for 15 minutes each in PBT at room temperature. Detection was performed in alkaline phosphatase reaction buffer described in the Boehringer instruction manual. When the color was developed, the reaction was stopped in 1× PBS. The embryos were then dehydrated, clarified in methylsalicylate and mounted in Permount.

Isolation and sequence analysis of zebrafish snail1 cDNA

To isolate zebrafish snail cDNA, a gastrula cDNA library (kindly provided by B. Riggleman) was screened at low stringency using as probe the entire Drosophila snail cDNA (a gift from A. Alberga). Sequence data showed that the initial clone encoded a zinc finger protein, similar to the Drosophila snail gene. Since this 1.4 kb (kilobase) clone did not include the entire coding region, it was used to rescreen the cDNA library. Several clones were isolated and sequenced and one of them contains the entire 3′ untranslated region, a complete protein coding region and a portion of the 5′ untranslated leader sequence of a gene that we call snail1.

The snail1 cDNA is 1950 base pairs long and contains an open reading frame of 789 nucleotides encoding a protein of 263 amino-acids (Fig. 1). 65 base pairs 5′ to the ATG translation initiation codon is an open reading frame 42 nucleotides long; this short open reading frame might play a role in the regulation of snail1 expression at the level of translation. snail1 encodes four zinc fingers. Fingers 1 to 3 share the classical structure CXXC(12X)HXXXH found in diverse nucleic acid binding proteins (reviewed in El Baradi and Pieler, 1991). Finger 4 is a variant on the zinc finger motif with the structure CXXC(12X)HXXXXC.

Fig. 1.

Nucleotide sequence of snail1 and deduced amino-acid sequence of the Snail1 protein. The cDNA is 1950 base pairs long and codes for a 263 amino acid residue protein. A short 5′ open reading frame 42 nucleotides long is underlined. Each zinc finger is boxed. Some useful restriction sites are indicated: SphI, AccI, HindIII, XbaI and XmnI. Nucleotide numbering is shown to the right. The presumed polyadenylation signal at the 3′ end of the transcript is underlined. The stop codon is indicated with a star.

Fig. 1.

Nucleotide sequence of snail1 and deduced amino-acid sequence of the Snail1 protein. The cDNA is 1950 base pairs long and codes for a 263 amino acid residue protein. A short 5′ open reading frame 42 nucleotides long is underlined. Each zinc finger is boxed. Some useful restriction sites are indicated: SphI, AccI, HindIII, XbaI and XmnI. Nucleotide numbering is shown to the right. The presumed polyadenylation signal at the 3′ end of the transcript is underlined. The stop codon is indicated with a star.

The amino-acid sequences of different members of the Snail family were aligned and compared (Fig. 2). The Snail family currently includes two Drosophila genes, snail and escargot (Whiteley et al., 1992), and three vertebrate genes. The fly and frog genes each contain five zinc fingers; in contrast, the mouse and fish genes both lack the first finger. The carboxy-termini of all five Snail family proteins are at about the same position, and the last four fingers are very similar. In all known members of the Snail family, the sequence of the last finger differs from the classical zinc finger motif. A search of the Swissprot. databank for the CXXC(12X)HXXXXC zinc finger structure uncovered only the Snail sequences discussed above. This motif can therefore be considered to be ‘Snail’ specific.

Fig. 2.

Comparison of the predicted protein sequences of the snail1 gene family. The zebrafish snail1 protein is compared to the mouse Sna protein (Nieto et al., 1992; Smith et al., 1992), Xenopus Xsna (Sargent and Bennett, 1990), and Snail and Escargot from Drosophila (Whiteley et al., 1992; Boulay et al., 1987). Zinc fingers are indicated. A dash indicates identity to the snail1 sequence. Dots indicate gaps inserted in the sequences to maximize homology. Cysteine (C) and histidine (H) in the zinc fingers are bold faced. A star signifies the translation stop signal at the carboxy terminus of the proteins.

Fig. 2.

Comparison of the predicted protein sequences of the snail1 gene family. The zebrafish snail1 protein is compared to the mouse Sna protein (Nieto et al., 1992; Smith et al., 1992), Xenopus Xsna (Sargent and Bennett, 1990), and Snail and Escargot from Drosophila (Whiteley et al., 1992; Boulay et al., 1987). Zinc fingers are indicated. A dash indicates identity to the snail1 sequence. Dots indicate gaps inserted in the sequences to maximize homology. Cysteine (C) and histidine (H) in the zinc fingers are bold faced. A star signifies the translation stop signal at the carboxy terminus of the proteins.

Developmental pattern of snail1 RNA

Transcripts of snail1 were localized during development by whole-mount in situ hybridization on zebrafish embryos. Digoxigenin-labelled sense and antisense RNA probes were hybridized to embryos of various stages. The pattern of snail1 expression described in the following paragraphs was obtained using a snail1-specific RNA probe corresponding to the 5′ part of the snail1 cDNA (a fragment containing the 5′ untranslated leader sequences and coding sequences up to the AccI restriction site before the first zinc finger).

(1) Zygote, cleavage stage and blastula

snail1 RNA was first detected as a maternal RNA in the zygote (0-0.7 hours postfertilization at 28.5°C) and during the cleavage period (0.7-2.2 hours). Blastomeres were homogeneously labelled with the antisense RNA probe but no signal was detected in the yolk cell (Fig. 3A). No labelling was observed with the sense RNA probe (data not shown).

Fig. 3.

Distribution of snail1 RNA during cleavage, blastula and gastrulation. Transcripts from the snail1 gene were revealed by whole-mount in situ hybridization. Lateral views of embryos are oriented with their animal pole up and dorsal side to the right. Animal pole views also have dorsal sides to the right. (A) 2-cell stage, lateral view. Blastomeres are labelled with snail1 antisense probe. ap, animal pole; b, blastomeres; y, yolk cell.(B) Beginning of dome stage, lateral view. The zygotic expression of snail1 has begun on one side of the embryo. Double labelling with the zebrafish goosecoid RNA shows that this is the future dorsal side of the embryo (our unpublished data).(C) Same stage as B, animal pole view. The zygotic RNA occupies an arc on one side of the embryo. The rest of the embryo is labelled more faintly with maternal RNA.(D) 40% epiboly, lateral view. snail1 RNA is localized all around the margin and maternal RNA has mostly disappeared. The two arrows show the position of the margin. (E) 70% epiboly, dorsal view. snail1 transcript disappears from the central part of the embryonic shield (es). (F) 90% epiboly, posterior or vegetal pole view. snail1 RNA is excluded from the axial mesoderm (a) and restricted to the paraxial mesoderm (p) and marginal region.

Fig. 3.

Distribution of snail1 RNA during cleavage, blastula and gastrulation. Transcripts from the snail1 gene were revealed by whole-mount in situ hybridization. Lateral views of embryos are oriented with their animal pole up and dorsal side to the right. Animal pole views also have dorsal sides to the right. (A) 2-cell stage, lateral view. Blastomeres are labelled with snail1 antisense probe. ap, animal pole; b, blastomeres; y, yolk cell.(B) Beginning of dome stage, lateral view. The zygotic expression of snail1 has begun on one side of the embryo. Double labelling with the zebrafish goosecoid RNA shows that this is the future dorsal side of the embryo (our unpublished data).(C) Same stage as B, animal pole view. The zygotic RNA occupies an arc on one side of the embryo. The rest of the embryo is labelled more faintly with maternal RNA.(D) 40% epiboly, lateral view. snail1 RNA is localized all around the margin and maternal RNA has mostly disappeared. The two arrows show the position of the margin. (E) 70% epiboly, dorsal view. snail1 transcript disappears from the central part of the embryonic shield (es). (F) 90% epiboly, posterior or vegetal pole view. snail1 RNA is excluded from the axial mesoderm (a) and restricted to the paraxial mesoderm (p) and marginal region.

During the early blastula period, maternal RNA persisted in all of the blastomeres. However, at the end of the sphere stage and beginning of dome stage (see Kimmel et al., 1993), strong labelling appeared at the margin in a restricted area corresponding to about 60° of arc on the circumference (Fig. 3B,C). In cells immediately surrounding this region, the amount of maternal RNA was clearly reduced (Fig. 3B). Then during the dome stage, zygotic expression spread progressively and snail1 RNA became localized all around the margin (Fig. 3D) in the deep cell layer (DEL).

(2) Gastrula

At the beginning of gastrulation, snail1 RNA was detected as a homogeneous ring around the margin of the embryo. At 60% epiboly, when the embryonic shield begins to elongate along the anteroposterior axis by convergence and extension movements (Kimmel et al., 1993), snail1 RNA disappears from the central region of the embryonic shield. This region corresponds to the axial hypoblast (Fig. 3E), including the presumptive notochord (chordamesoderm) caudally, and the prechordal plate rostrally. Paraxial hypoblast gives rise to muscles of the eye, jaw and gill, anteriorly and the segmental plate, which forms somites, posteriorly. Throughout gastrulation, snail1 transcripts continued to be detected in the paraxial hypoblast of the segmental plate and around the margin except for the axial hypoblast (Fig. 3F).

(3) Segmentation (10–24 hours)

During segmentation, the embryo elongates along the anteroposterior axis and somites appear sequentially from anterior to posterior. At the beginning of somitogenesis, labelling was very strong around the blastopore, except in the axial hypoblast. Anteriorly, snail1 RNA staining was intense in a cell sheet immediately adjacent to the axial hypoblast (Fig. 4A). We call these special cells adaxial cells. Lateral to the adaxial cells, snail1 RNA was more diffuse. snail1 RNA was not observed in the hypoblast rostral to the first somite.

Fig. 4.

Change in the snail1 expression pattern during somitogenesis. (A) 3-somite stage. snail1 RNA is intensively stained in adaxial cells (ad), a single row of paraxial hypoblast cells (p) adjacent to axial mesoderm (a). Lateral cells are also labelled but in a diffuse manner. The tail bud (tb) is strongly stained. (B) 4-somite stage. Labelling adjacent to the notochord starts to fade posterior to the furrow separating the third and fourth somite. Staining increases just anterior to each newly formed somitic furrow. (C) 7-somite stage. In the segmented mesoderm, signal disappears from adaxial cells. Note the epithelial character of adaxial cells more posteriorly in the unsegmented mesoderm. snail1 RNA is detected anterior to the somitic furrow in a single row of cells. See arrow showing snail1 RNA transiently expressed in the most lateral part of the segmental paraxial hypoblast. (D,E) 10and 12somite stages, respectively. In the unsegmented paraxial mesoderm (the segmental plate), there are two stripes of snail1 RNA at segment periodicity posterior to the most recently formed somite (see arrows). As each somite matures, the territory of snail1 expression spreads anteriorly within the somite. (F) 17-somite stage. In the most recently formed somite, snail1 RNA accumulates in a unique line of cells just anterior to the newly formed furrow. (G) Details of snail1 expression pattern in the somite (s) at 17somite stage, dorsal view. Note the position of the notochord (n) in the middle and the location of the somitic furrow (arrows). snail1 RNA is detected in the posterior compartment of the somite. In the most recently formed somite, a single sheet of cells accumulates snail1 transcript. In older somites, the labelling occupies more of the posterior portion of the somite. The star indicates the position of muscle pioneer precursors which are not labelled with snail1 RNA. (H) 17-somite stage. Dorsal view of an embryo probed with α-tropomyosin. The arrow indicates the position of the somitic furrow. αtropomyosin is detected in adaxial cells and in particular, in muscle pioneer precursors (indicated by a star). Note morphogenetic changes in α−tropomyosin-expressing cells as somites mature.

Fig. 4.

Change in the snail1 expression pattern during somitogenesis. (A) 3-somite stage. snail1 RNA is intensively stained in adaxial cells (ad), a single row of paraxial hypoblast cells (p) adjacent to axial mesoderm (a). Lateral cells are also labelled but in a diffuse manner. The tail bud (tb) is strongly stained. (B) 4-somite stage. Labelling adjacent to the notochord starts to fade posterior to the furrow separating the third and fourth somite. Staining increases just anterior to each newly formed somitic furrow. (C) 7-somite stage. In the segmented mesoderm, signal disappears from adaxial cells. Note the epithelial character of adaxial cells more posteriorly in the unsegmented mesoderm. snail1 RNA is detected anterior to the somitic furrow in a single row of cells. See arrow showing snail1 RNA transiently expressed in the most lateral part of the segmental paraxial hypoblast. (D,E) 10and 12somite stages, respectively. In the unsegmented paraxial mesoderm (the segmental plate), there are two stripes of snail1 RNA at segment periodicity posterior to the most recently formed somite (see arrows). As each somite matures, the territory of snail1 expression spreads anteriorly within the somite. (F) 17-somite stage. In the most recently formed somite, snail1 RNA accumulates in a unique line of cells just anterior to the newly formed furrow. (G) Details of snail1 expression pattern in the somite (s) at 17somite stage, dorsal view. Note the position of the notochord (n) in the middle and the location of the somitic furrow (arrows). snail1 RNA is detected in the posterior compartment of the somite. In the most recently formed somite, a single sheet of cells accumulates snail1 transcript. In older somites, the labelling occupies more of the posterior portion of the somite. The star indicates the position of muscle pioneer precursors which are not labelled with snail1 RNA. (H) 17-somite stage. Dorsal view of an embryo probed with α-tropomyosin. The arrow indicates the position of the somitic furrow. αtropomyosin is detected in adaxial cells and in particular, in muscle pioneer precursors (indicated by a star). Note morphogenetic changes in α−tropomyosin-expressing cells as somites mature.

As cells intercalate and converge toward the dorsal midline, the embryo elongates and narrows. During these events, the quantity of snail1 RNA increased in the prospective tail bud and in the paraxial hypoblast, which showed well-demarcated lateral limits of expression (Fig. 4B-D). As segmentation progresses, the territory of snail1 RNA in the tail bud and caudal paraxial mesoderm continues to narrow until the end of tail formation (Fig. 4).

In 6to 9-somite embryos, the quantity of snail1 transcripts increased transiently in the most lateral part of the segmented paraxial hypoblast in a row of cells parallel to the adaxial cells in the three anterior somites (arrow Fig. 4C). Expression in this territory decreased from posterior to anterior and completely disappeared by the 12-somite stage. The future identity of these cells is unknown.

Development of snail1 RNA pattern in adaxial cells

At the beginning of somitogenesis, adaxial cells were intensely labelled up to the level of the first somite. The signal in this sheet of cells was homogeneous along the anteroposterior axis (Fig. 4A). At about the 4-somite stage, the signal started to fade posterior to the furrow separating somites 3 and 4 (Fig. 4B). Expression in these cells decreased progressively for the three anterior somites and was no longer observed in this territory after the 7or 8somite stage (Fig. 4C). More caudally, snail1 RNA in adaxial cells disappeared progressively following the advancing wave of somite formation. The anterior extent of adaxial cell labelling reached the level of the most recently formed somitic furrow (Fig. 4C-F).

In the segmental plate, adaxial cells narrowed to a sheet 1-cell wide by the 6-somite stage. Each cell’s nucleus occupied an asymmetric position near the notochord. This pattern remained unchanged until the last somite had been added (Fig. 4D-G).

Since adaxial cells include muscle pioneer precursors (Felsenfeld et al., 1991), we tested whether the adaxial cells accumulated transcripts from a zebrafish myotome marker: the α-tropomyosin gene (probe kindly provided by B. Riggleman, see also Ohara et al., 1989; Riggleman et al., 1993). Along the anterior-posterior axis, α-tropomyosin RNA was detected in adaxial cells, in particular in cells that form muscle pioneer precursors (Fig. 4H). In newly formed posterior-most somites, α-tropomyosin RNA was maintained in muscle pioneer precursors. In contrast, snail1 RNA disappeared from these muscle pioneer precursors at this time (Fig. 4G). A comparison of the expression of these two genes shows that snail1 RNA and α-tropomyosin RNA are both found in adaxial cells before the formation of the somite, at which time there is a specific extinction of snail1 expression in adaxial cells.

snail1 RNA along the somitic furrows

When the first few somitic furrows form, snail1 RNA increases in cells just anterior to the furrow (Fig. 4B,C). Later, as each subsequent furrow forms, snail1 RNA increases in the lateral part of the paraxial mesoderm in cells just anterior to the furrow (Fig. 4D,E). Until the 14-somite stage, the level of snail1 transcripts increases nearly simultaneously in all somites. As each somite matures, snail1 RNA spreads anteriorly from a single row of cells at the posterior border of the somite (Fig. 4G). Transcripts of the snail1 gene were not detected in the anterior quarter of the somite.

Posterior to the most-recent somitic furrow, snail1 labelling was not homogeneously distributed. In the unsegmented segmental plate, transcripts of snail1 RNA appeared more strongly in two stripes spaced at segmental intervals posterior to the most recent somite (arrows in Fig. 4D,E). This pattern persisted until the end of somitogenesis.

From the 14-somite stage, as the tail elongates, snail1 RNA accumulates in posterior newly formed somites and disappears progressively in anterior differentiated somites (data not shown).

Disappearance of snail1 RNA

RNA from the snail1 gene was present until 36 hours in the extremity of the tail, but disappeared when the tail was completely formed. Transcripts of snail1 disappeared from somites dynamically. Cells close to the notochord stained in the most posterior somites, but as muscles differentiated in a wave from medial to lateral in each somite, snail1 labelling decreased in a similar fashion. Differentiated muscle cells never contained detectable amounts of snail1 transcript. The latest cells to continue expressing snail1 in each somite were localized superficially next to the body wall in a position expected for dermatome cells. In addition, some cells labelled with the snail1 RNA probe were also observed along the notochord at a position expected for sclerotome cells (data not shown).

(4) Pattern of snail1 RNA in the head

Transcript from the snail1 gene was first detected in the head at the 16-somite (17 hour) stage in a region posterior and ventral to the eye. These cells give rise to the primordia of the most anterior pharyngeal arches. Labelled cells were also observed dorsal to the eyes in the space between eyes and forebrain. These territories of expression gradually increased in size and intensity until 24 hours of development.

At 24 hours, a zebrafish embryo has well-developed eyes, several distinct brain regions, a prominent otic vesicle and six pharyngeal arch primordia. Cells located in mesenchyme surrounding and within some of these structures contained snail1 RNA (Fig. 5A, B). Labelled cells encapsulated the eyes and appeared to fill the region between the eyes and the neighboring midbrain and forebrain. Similarly, stained mesenchymal cells were found in spaces between the olfactory placodes and the forebrain (Fig. 5B), and at the midbrainhindbrain junction. The otic vesicle was also surrounded with label, except along its dorsal and medial borders.

Fig. 5.

Late expression pattern of snail1 in the head. Embryos were dissected and the yolk was removed. (A) 24 hour embryo, dorsolateral view and (B) 24 hour embryo, dorsal view. snail1 RNA is localized in various cells around the eyes (e) and otic capsule (ot), and in pharyngeal arch primordia (mandibular arch (m), hyoid arch (h), and the first (1) and second (2) gill segments). D, Dorsal; V, ventral; R, right; L, left. (C) 36 hour embryo, lateral view. The extent of labelling has expanded in the branchial arches and is easily distinguishable in the mandibular (m) arch and the hyoid (h) arch, and in the first four (1, 2, 3 and 4) gill arches. Pectoral fin buds (pf) also accumulate snail1 transcript. (D) High magnification of the ear region (ot, otic capsule) and caudal branchial arches focusing on snail1 RNA expression in gill arches. The mesenchyme of each arch is composed of neural crest cells and paraxial mesodermal cells. (E) 60 hour embryo, lateral view. snail1 RNA disappears from the middle of the arches where chondrocytes are beginning to differentiate. (F) Details of the embryo shown in E, focusing on branchial arches.

Fig. 5.

Late expression pattern of snail1 in the head. Embryos were dissected and the yolk was removed. (A) 24 hour embryo, dorsolateral view and (B) 24 hour embryo, dorsal view. snail1 RNA is localized in various cells around the eyes (e) and otic capsule (ot), and in pharyngeal arch primordia (mandibular arch (m), hyoid arch (h), and the first (1) and second (2) gill segments). D, Dorsal; V, ventral; R, right; L, left. (C) 36 hour embryo, lateral view. The extent of labelling has expanded in the branchial arches and is easily distinguishable in the mandibular (m) arch and the hyoid (h) arch, and in the first four (1, 2, 3 and 4) gill arches. Pectoral fin buds (pf) also accumulate snail1 transcript. (D) High magnification of the ear region (ot, otic capsule) and caudal branchial arches focusing on snail1 RNA expression in gill arches. The mesenchyme of each arch is composed of neural crest cells and paraxial mesodermal cells. (E) 60 hour embryo, lateral view. snail1 RNA disappears from the middle of the arches where chondrocytes are beginning to differentiate. (F) Details of the embryo shown in E, focusing on branchial arches.

By 24 hours, snail1 was expressed in at least the first four pharyngeal arch primordia. These primordia appeared as a reiterated series of darkly stained cells that extended anteriorly and ventrally below the developing hindbrain and otic capsule. From rostral to caudal, they constitute the precursors of the mandibular, hyoid and five gill segments (Schilling, 1993). The mesenchyme of each arch is composed predominantly of neural crest cells as well as some paraxial mesoderm. Most if not all of the mesenchymal cells within an arch expressed snail1 (Fig. 5A). Epithelium of the endoderm that forms the boundaries between arches showed no expression. Mandibular and hyoid segments appeared as larger labelled cell groups. The first two gill segments are smaller beneath the otic capsule and the most caudal segments appeared as a fused mass of staining. A bilateral string of labelled cells extended from the caudal boundary of the otic capsule (Fig. 5A,B) caudally along the notochord into the rostral trunk.

By 36 hours, the head has shortened and eyes and body pigmentation are well developed. At this time, snail1 RNA is located in a few cells surrounding the developing lens and the eye capsule. The labelled regions in the pharyngeal arches have enlarged and become more distinct (Fig. 5C,D). Staining in the mandibular arch extends across the ventral midline but, in more caudal arches, the two sides remain widely separated.

After 2 days of development, snail1 expression is largely confined to the pharyngeal region (Fig. 5E,F). By 3 days of development, pharyngeal cartilages and muscles have begun to differentiate. snail1 RNA fades from newly differentiated chondrocytes (Fig. 5E,F) and becomes restricted to cells that surround the differentiated cartilages. Labelling is stronger at the dorsal and ventral extremities of the arch, and weaker in the center where the first cells appear to differentiate. By 96 hours, cells throughout the embryo stop accumulating snail1 transcript.

In the pectoral fin buds, after a uniform signal at 24 hours, staining became localized in two distinct regions: one in the mesenchyme beneath the apical ectodermal ridge and another located in dorsal and ventral cells in the center of the fin. At 48 hours, snail1 RNA was not observed in differentiated chondrocytes of the fin (data not shown).

snail1, spadetail and the process of convergence

The in situ hybridization experiments showed that snail1 expression occurs in hypoblast cells undergoing convergence and forming paraxial mesoderm. This result suggests the hypothesis that snail1 might be involved in the process of convergence. This process is disrupted in homozygous spadetail embryos (Kimmel et al., 1989; Ho and Kane, 1990). Mutant spadetail (spt) embryos develop apparently normally until gastrulation, but then paraxial mesoderm fails to converge normally towards the dorsal axis of the embryo. Instead, paraxial mesoderm congregates in the tail bud, giving embryos deficient in derivatives of paraxial mesoderm. The notochord is morphologically normal in spadetail embryos (Kimmel et al., 1989; Ho and Kane, 1990). Interestingly, cells that express snail1 in wild-type gastrulas develop abnormally in spadetail embryos. Does spadetail inhibit convergence because it blocks the action of snail1? Or do snail1-expressing cells simply fail to converge in spadetail embryos?

To investigate the relationship of snail1 to the process of convergence, we examined the pattern of snail1 RNA in spadetail mutant embryos. Expression of snail1 RNA in spadetail embryos appears normal before gastrulation. Then, in the early gastrula (after 60% epiboly), snail1 RNA fails to occupy the lateral edges of the embryonic shield in presumed spadetail embryos (i.e., in a quarter of the offspring of a mating of two spadetail heterozygotes) in contrast to wild-type embryos (Fig. 6A,B). This is long before the mutant phenotype becomes morphologically visible otherwise. During somitogenesis in mutant embryos, snail1 RNA was concentrated in the tail bud, where nonconverging cells accumulate (Fig. 6C-F). While somites fail to form in spadetail embryos, a few snail1-positive cells were found scattered along the notochord (Fig. 6C-F). These results show that snail1-expressing cells are mislocalized in spadetail mutant gastrulas.

Fig. 6.

Expression pattern of snail1 in spadetail embryos. (A-D) Posterior views; (E,F) lateral views. (A,B) 80% epiboly; (C,D) 10-somite stage; (E,F) 12-somite stage. (A,C,E) Wild-type embryos; (B,D,F) spadetail embryos. While snail1 RNA lines the lateral border of the embryonic shield in wild-type (A), it is missing in spadetail embryos (B). The axial mesoderm is broader in a spadetail embryo (C) than in a wild-type embryo (D). Wild-type embryos have extensive areas containing snail1 transcripts in the paraxial mesoderm and somites (E), while spadetail embryos have much less snail1 transcript in the paraxial region (F). In the mutant, snail1 transcript is present only in the tail bud and in a few cells scattered along the axial mesoderm.

Fig. 6.

Expression pattern of snail1 in spadetail embryos. (A-D) Posterior views; (E,F) lateral views. (A,B) 80% epiboly; (C,D) 10-somite stage; (E,F) 12-somite stage. (A,C,E) Wild-type embryos; (B,D,F) spadetail embryos. While snail1 RNA lines the lateral border of the embryonic shield in wild-type (A), it is missing in spadetail embryos (B). The axial mesoderm is broader in a spadetail embryo (C) than in a wild-type embryo (D). Wild-type embryos have extensive areas containing snail1 transcripts in the paraxial mesoderm and somites (E), while spadetail embryos have much less snail1 transcript in the paraxial region (F). In the mutant, snail1 transcript is present only in the tail bud and in a few cells scattered along the axial mesoderm.

snail1 expression in no tail embryos

The expression pattern of snail1 just before gastrulation is similar to that of the no tail gene, which is homologous to Brachyury in mouse (Chesley et al., 1935; Schulte-Merker et al., 1992; Wilkinson et al., 1990). RNA from both genes accumulates in a ring of cells completely encircling the margin of the late blastula stage embryo. As gastrulation begins, the situation suddenly changes: axial cells continue to accumulate no tail message, but no longer contain snail1 transcript, while, reciprocally, the rest of the margin continues to contain snail1 RNA, but loses no tail RNA. Thus, in the anterior part of the embryo, snail1 RNA becomes restricted to segmental paraxial mesoderm and no tail RNA becomes limited to the axial mesoderm. In contrast, in the posterior, cells of the presumptive tail bud express both genes throughout somitogenesis. To investigate further the interactions of these two genes, we studied the localization of snail1 transcripts in mutant no tail embryos, which contain axial mesoderm but do not form differentiated notochord cells (Halpern et al., 1993).

During early gastrulation, when the no tail phenotype can not be recognized morphologically, all offspring of matings between two no tail heterozygotes had a normal spatial distribution of snail1 RNA. While the spatial distribution was normal, the quantity of RNA in paraxial mesoderm cells was considerably reduced in about a quarter of the embryos. When the mutant phenotype could be distinguished at the beginning of somitogenesis, snail1 RNA was observed in the paraxial mesoderm of no tail mutants but the amount of labelling was much less than in phenotypically wild-type sibling embryos stained for the same amount of time. Adaxial cells were more strongly labelled in the anterior part than in the posterior region of no tail embryos. Tail bud cells in no tail embryos failed to show the intense staining that they display in wild-type embryos; in Fig. 7A,B, wild-type embryos were stained for 2 hours and mutant embryos were stained for 6 hours in order to fully visualize the weak snail1 staining in no tail embryos. The lateral limit of snail1 expression in no tail paraxial mesoderm was less well defined than in wild type (Fig. 7A,B). The spatial distribution of snail1 RNA in somites resembled that of wild-type embryos, but the amount of RNA was substantially reduced.

Fig. 7.

Expression pattern of snail1 in no tail embryos. (A,B) Posterior views of 5-somite stage; (C,D) lateral views of 12-somite stage; (A,C) wild-type embryos; (B,D) no tail embryos. Compared to a wild-type embryo (A), note the reduction of snail1 signal in no tail, especially in the tail bud, even though this no tail embryo was stained for 6 hours to be able to see a clear signal, and the wild-type control was stained for just 2 hours. The axial mesoderm territory is broader in no tail embryos. In contrast to wild type (C), an overstained no tail embryo shows an absence of snail1 staining in the tail. Since the mutant embryo was stained for a long time, background staining started to become visible.

Fig. 7.

Expression pattern of snail1 in no tail embryos. (A,B) Posterior views of 5-somite stage; (C,D) lateral views of 12-somite stage; (A,C) wild-type embryos; (B,D) no tail embryos. Compared to a wild-type embryo (A), note the reduction of snail1 signal in no tail, especially in the tail bud, even though this no tail embryo was stained for 6 hours to be able to see a clear signal, and the wild-type control was stained for just 2 hours. The axial mesoderm territory is broader in no tail embryos. In contrast to wild type (C), an overstained no tail embryo shows an absence of snail1 staining in the tail. Since the mutant embryo was stained for a long time, background staining started to become visible.

These studies show that in the absence of no tail gene function, the spatial distribution of snail1 RNA is similar to wild type in segmented paraxial mesoderm, but the quantity of snail1 RNA is drastically reduced. Likewise, the strong accumulation of snail1 RNA observed in adaxial and tail bud cells of wild-type embryos disappears in mutant embryos. These results suggest that no tail gene function is required for a high level of expression of snail1.

The snail gene family

Sequence analysis suggests that there are two sub-types of snail-family genes: some with five fingers (described in Drosophila and Xenopus), and some with four (described in zebrafish and mouse). We have recently found (in progress) that zebrafish has a second snail gene (snail2) with five fingers like the Xenopus protein. This raises the possibility that mouse and Xenopus may also have two snail-family genes, but that the four-finger gene is the only one yet described in mouse and the five-finger gene is the only one yet described in Xenopus. For the zebrafish experiments reported here, a snail1-specific RNA probe was generated from the 5′ untranslated leader and the coding portion of the gene excluding the zinc finger region.

The expression pattern of Sna in mouse embryos appears to be a combination of the snail1 and snail2 pattern in zebrafish; Sna is expressed not only in mesoderm as is snail1, but also in presumptive pre-migratory neural crest cells (Nieto et al., 1992; Smith et al., 1992) like snail2 (our unpublished data). Two hypotheses might explain these observations: either the expression pattern of snail1 and Sna are different in these two species or the expression of Sna in presumptive migratory neural crest cells is a result of cross hybridization to an undescribed snail2-like gene in mouse.

Expression and role of snail1

The optical clarity and ease of obtaining zebrafish embryos permits an analysis of snail gene expression in greater detail than was possible for mouse or Xenopus. In addition, we have substantially improved the in situ hybridization protocol to preserve morphology and increase specificity. Improvements include fixing embryos for prolonged times, deleting the proteinase K step for young embryos, hybridizing at very high stringency, dissecting embryos from the yolk and flattening them between coverslips, which allows single cell resolution.

(1) Early snail1 RNA pattern

Expression of snail1 is one of the earliest known asymmetries in the zebrafish embryo. Just before epiboly (4 hours), at the end of the sphere stage, zygotic expression of snail1 RNA begins in a small crescent on one side of the embryo margin. Only a few zebrafish genes are known to be expressed this early, including no tail (Schulte-Merker et al., 1992) and goosecoid (our own observations). In situ hybridization using snail1 and goosecoid RNA probes in the same embryos demonstrated that both genes are expressed in the same cells (our unpublished data). In Xenopus, goosecoid expression marks the dorsal side of the embryo (Cho et al., 1991; Blumberg et al., 1991). This suggests that zygotic expression of snail1 initiates on the dorsal side of the embryo shortly after the mid-blastula transition.

The initial crescent of snail1 expression spreads to form a ring encompassing the entire blastoderm margin at the beginning of epiboly. These cells also express the no tail gene (Schulte-Merker et al., 1992), and become lateral, paraxial and axial mesoderm, as well as heart and blood (Kimmel and Warga, 1987; Kimmel et al., 1990). The ring of snail1 RNA becomes broken at the beginning of gastrulation, as staining disappears from the central part of the embryonic shield. This suggests that snail1 is repressed initially in the axial, presumptive notochordal and prechordal territories, and thus becomes restricted to the paraxial hypoblast. At the same time, no tail transcripts disappear from hypoblast cells except in the axial territory and continue to be detected at high levels in presumptive notochord (Schulte-Merker et al., 1992). Thus, snail1 and no tail are initially expressed in the same set of cells, then as the shield forms, snail1 turns off in one subset of cells and no tail turns off in the complementary subset of cells. Factors regulating this abrupt, complementary shift in gene expression have yet to be identified.

After gastrulation, snail1 is no longer expressed in the lateral and ventral mesoderm. It is possible that lateral and ventral mesoderm might not get a dorsalizing signal emanating from axial cells that is needed to specify the paraxial hypoblast (Jessell and Melton, 1992; Moon and Christian, 1992).

(2) snail1 and somitogenesis Snail1 RNA in adaxial cells

At the end of gastrulation, adaxial cells express snail1 intensively while the adjacent paraxial cells contain reduced quantities of snail1 transcript. The mechanism that controls intense snail1 expression in adaxial cells is unknown. One reasonable hypothesis, to be discussed below, is that the axial mesoderm, which is in contact with adaxial cells, stimulates snail1 expression in adjacent cells at this time. Extinction of snail1 expression in adaxial cells is related to cell maturation: except for the most anterior somites, for which a delay is observed, the loss of expression of snail1 in adaxial cells follows the anteriorposterior sequence of somite maturation. Adaxial cells form the muscle pioneer precursors, the first myotome cells to differentiate into muscle (Felsenfeld et al., 1991). Since snail1 expression is only observed in undifferentiated cells, the repression of snail1 expression in muscle pioneer precursors might reflect an obligatory step in muscle differentiation. This hypothesis predicts that ectopic expression of snail1 in these cells might suppress muscle differentiation.

snail1 and segmentation

The segmentally repeated pattern of snail1 expression appears to represent an up-regulation of snail1 in cells of unsegmented paraxial mesoderm just anterior to the new somitic furrow, and a down-regulation of snail1 in the anterior of the newly formed somite. Since this change in expression has already begun in the unsegmented segmental plate, snail1 might be involved in the definition of the posterior border of newly forming somites. While snail1 expression identifies the presumptive somite border two segments before the segmental furrow appears, a heat-shock sensitive process necessary for proper furrow formation occurs four segments before the furrow becomes visible (Kimmel et al., 1988). Since changes in snail1 expression occur about an hour later than the heat-shock-sensitive step, snail1 may not trigger segment specification; instead, it may be involved in an early but subsequent step in this process, perhaps the regulation of genes that cause morphogenetic events.

snail1 defines a posterior domain in the somite

Since snail1 is expressed in the posterior three quarters of the somite and most of the somite is myotome (Kimmel et al., 1993), snail1 is clearly expressed in at least a portion of the myotome. In more mature somites, snail1 RNA is detected in a group of superficial cells assuming a position expected for dermis, and in a second group of cells along the notochord that are probably sclerotome (data not shown). Thus, snail1 seems to be expressed in a portion of all three somite compartments.

Rather than distinguishing myotome, dermatome and sclerotome, snail1 seems to define a posterior domain in the somite. This suggests that mesoderm segmentation in vertebrate embryos may be analogous in some ways to segmentation in Drosophila embryos, where gene expression patterns define anterior and posterior compartments in each segmental unit, the parasegment (Martinez-Ariaz and Lawrence, 1985; Lawrence, 1989). The proposed analogy is that snail1 defines the posterior domain of each somite and a different gene or genes define the complementary anterior domain. Such genes have not yet been identified. We predict that anterior domain genes would repress snail1 in anterior cells and perhaps stimulate snail1 in adjacent cells. Other genes would define segment identity (i.e., Wilkinson and Krumlauf, 1990).

Expression of snail1 in mutant embryos

The spadetail phenotype involves an inability of paraxial mesoderm (i.e., snail1-expressing cells) to converge dorsally during gastrulation (Kimmel et al., 1989; Ho and Kane, 1990). This raises the hypothesis that spadetail is defective in snail1 function. The earliest detected phenotype in spadetail embryos is in the pattern of snail1 expression. Work is in progress to see if spadetail embryos lack normal snail1 function; if so, then injecting snail1 RNA into spadetail embryos could rescue the mutant phenotype.

Analysis of snail1 expression in no tail embryos suggests that mesoderm formation in zebrafish has an anterior and a posterior phase as it does in Xenopus (Cunliffe and Smith, 1992). The anterior (head and thorax) distribution of snail1 RNA is normal in no tail embryos, although its quantity is reduced. That result suggests that a no tail-dependent snailstimulating signal may come from notochordal cells or, alternatively, that a no tail-snail1 interaction occurs earlier in the tail bud at the time when both genes are expressed in the same cells. In contrast, the posterior (tail) expression of snail1 is abolished in no tail embryos. These genetic results suggest that no tail may act upstream of snail1 as a positive regulator. This conclusion supports the finding that injecting Xenopus embryos with RNA from the Xenopus no tail homologue Xbra stimulates the frog’s snail family homologue (Cunliffe and Smith, 1992).

snail1 function

Analysis of wild-type and mutant embryos suggests that snail1 is involved in several morphogenetic events, including involution and convergence during gastrulation, invagination of the somitic furrow and the condensation of pharyngeal cartilages. snail1 may regulate the expression of genes needed for cell mobility or adhesivity during these morphogenetic processes. Alberga et al. (1991) have proposed that snail is involved in morphogenetic movements during Drosophila embryogenesis. Ectopic expression of snail1 is being used to test this hypothesis.

We thank A. Alberga for the Drosophila cDNA probe, B. Riggleman for kindly providing the zebrafish library and the zebrafish α-tropomyosin probe before publication, M. Halpern for gifts of no tail embryos and sharing results before publication, R. Kimmel for providing spadetail embryos, T. Jowett for helping us in our first in situ experiments and C. Kimmel for useful comments on the manuscript. We thank all ‘zebrafish people’ in Eugene for valuable help in various aspects of this work. C. Thisse is a fellow of the EMBO (ALTF 229-1991). B. Thisse is supported by CNRS, NATO and by the FOGARTY organization. T. F. Schilling is supported by NIH (1T32HDO7348). Research was supported by NIH grant (1RO1AI26734) and a Medical Research Fund grant awarded to J. H. Postlethwait and by NIH grant (1PO1HD22486) to the University of Oregon Zebrafish Program Project (J. Weston, P.I.).

Alberga
,
A.
,
Boulay
,
J. L.
,
Kempe
,
E.
,
Dennefeld
,
C.
and
Haenlin
,
M.
(
1991
).
The snail gene required for mesoderm formation in Drosophila is expressed dynamically in derivatives of all three germ-layers
.
Development
111
,
983
992
.
Boulay
,
J. L.
,
Dennenfeld
,
C.
and
Alberga
,
A.
(
1987
).
The Drosophila developmental gene snail encodes a protein with nucleic acid binding fingers
.
Nature
330
,
392
395
.
Blumberg
,
B.
,
Wright
,
C. V. E.
,
De Robertis
,
E. M.
and
Cho
,
K. W. Y.
(
1991
).
Organizer-specific homeobox genes in Xenopus laevis embryos
.
Science
253
,
194
196
.
Chesley
,
P.
(
1935
).
Development of the short-tailed mutation in the house mouse
.
J. Exp. Zool
.
70
,
429
459
.
Cho
,
K.
,
Blumberg
,
B.
,
Steinbeisser
,
H.
and
De Robertis
,
E.
(
1991
).
Molecular nature of the Spemann’s organizer: the role of the Xenopus homeobox gene goosecoid
.
Cell
67
,
1111
1120
.
Cunliffe
,
V.
and
Smith
,
J. C.
(
1992
).
Ectopic mesoderm formation in Xenopus embryos caused by widespread expression of a Brachyury homologue
.
Nature
358
,
427
430
.
El Baradi
,
T.
and
Pieler
,
T.
(
1991
).
Zinc finger proteins: what we know and what we would like to know
.
Mech. Dev
.
35
,
155
169
.
Felsenfeld
,
A. L.
,
Curry
,
M.
and
Kimmel
,
C. B.
(
1991
).
The fub-1 mutation blocks initial myofibril formation in zebrafish muscle pioneer cells
.
Dev. Biol
.
148
,
23
30
.
Gluecksohn-Schoenheimer
,
S.
(
1944
).
The development of the normal and homozygous Brachyury (T/T) mouse embryos in the extraembryonic coelom of the chick
.
Proc. Nat. Acad. Sci.USA
30
,
134
140
.
Halpern
,
M. E.
,
Ho
,
R. K.
,
Walker
,
C.
and
Kimmel
,
C. B.
(
1993
).
Induction of muscle pioneers and floor plate is distinguished by the zebrafish no tail mutation
.
Cell, In press
.
Harland
,
R. M.
(
1991
).
In situ hybridization: an improved whole-mount method for Xenopus embryos
.
Meth. Cell Biol
.
36
,
685
695
.
Herrmann
,
B. G.
,
Labeit
,
S.
,
Poustka
,
A.
,
King
,
T. R.
and
Lehrach
,
H.
(
1990
).
Cloning of the T gene required in the mesoderm formation in the mouse
.
Nature
343
,
617
622
.
Ho
,
R. K.
and
Kane
,
D. A.
(
1990
).
Cell-autonomous action of the zebrafish spadetail-1 mutation in specific mesodermal precursors
.
Nature
348
,
728
730
.
Hopwood
,
N. D.
(
1990
).
Cellular and genetic responses to mesoderm induction in Xenopus
.
BioEssays
12
,
465
471
.
Hopwood
,
V. D.
,
Pluck
,
A.
and
Gurdon
,
J. B.
(
1989
).
A Xenopus mRNA related to Drosophila twist is expressed in response to induction in the mesoderm and the neural crest cells
.
Cell
59
,
893
903
.
Ip
,
Y. T.
,
Kraut
,
R.
,
Levine
,
M.
and
Rushlow
,
C. A.
(
1991
).
The dorsal morphogen is a sequence-specific DNA-binding protein that interacts with a long-range repression element in Drosophila
.
Cell
64
,
439
446
.
Jessell
,
T. M.
and
Melton
,
D. A.
(
1992
).
Diffusible factors in vertebrate embryonic induction
.
Cell
68
,
257
270
.
Jiang
,
J.
,
Kosman
,
D.
,
Ip
,
Y. T.
and
Levine
,
M.
(
1991
).
The dorsal morphogen gradient regulates the mesoderm determinant twist in early Drosophila embryo
.
Genes Dev
.
5
,
1881
1891
.
Kimmel
,
C. B.
(
1989
).
Genetic and early development of the zebrafish embryos
.
Genetics
103
,
125
136
.
Kimmel
,
C. B.
and
Warga
,
R. M.
(
1987
).
Cell lineages generating axial muscle in the Zebrafish embryo
.
Nature
327
,
234
237
.
Kimmel
,
C. B.
Sepich
,
D. S.
and
Trevarrow
,
B.
(
1988
).
Development of segmentation in zebrafish
.
Development
104
Supplement
,
197
207
.
Kimmel
,
C. B.
,
Warga
,
R. M.
and
Schilling
,
T. F.
(
1990
).
Origin and organization of the zebrafish fate map
.
Development
108
,
581
594
.
Kimmel
,
C. B.
,
Kane
,
D. A.
,
Walker
,
C.
,
Warga
,
C. M.
and
Rothman
,
M. B.
(
1989
).
A mutation that changes cell movement and cell fate in zebrafish embryo
.
Nature
337
,
358
362
.
Kimmel
,
C. B.
,
Ballard
,
W. W.
,
Kimmel
,
S. R.
,
Ullmann
,
B.
and
Schilling
,
T. F.
(
1993
).
Stages of embryonic development of the Zebrafish
.
The Zebrafish Book
p
1
25
.
Lawrence
,
P. A.
(
1989
).
Cell lineage and cell states in the Drosophila embryo
.
CIBA Found. Symp
.
144
,
130
149
.
Lin
,
H. C.
,
Lei
,
S.
and
Wilcox
,
G.
(
1985
).
An improved DNA sequencing strategy
.
Anal. Biochem
.
147
,
114
119
.
Martinez-Ariaz
,
A.
and
Lawrence
,
P. A.
(
1985
).
Parasegments and compartments in the Drosophila embryo
.
Nature
313
,
639
642
.
Moon
,
R. T.
and
Christian
J. L.
(
1992
).
Competence modifiers synergize with growth factors during mesoderm induction and patterning in Xenopus
.
Cell
71
,
719
712
.
Nieto
,
M. A.
,
Bennett
,
M. F.
,
Sargent
,
M. G.
and
Wilkinson
,
D. G.
(
1992
).
Cloning and developmental expression of Sna, a murine homologue of the Drosophila snail gene
.
Development
116
,
227
237
.
Ohara
,
O.
,
Dorit
,
R. L.
and
Gilbert
,
W.
(
1989
).
One-sided polymerase chain reaction: the amplification of cDNA
.
Proc. Nat. Acad. Sci. USA
86
,
5673
5677
.
Pan
,
D.
,
Huang
,
J. D.
and
Courey
,
A. J.
(
1991
).
Functional analyses of the Drosophila twist promoter reveals a dorsal binding ventral activator region
.
Genes Dev
.
5
,
1892
1901
.
Riggleman
,
B.
,
Allende
,
M.
,
Grunewald
,
D.
and
Weinberg
,
E.
(
1993
).
Expression of two myogenic HLH family genes in zebrafish. Developmental regulation and potential roles in myogenesis
.
Submitted for publication
.
Roth
,
S.
,
Stein
,
D.
and
Nüsslein-Volhard
,
C.
(
1989
).
A gradient of nuclear localization of the dorsal protein determines dorso-ventral pattern in the Drosophila embryo
.
Cell
59
,
1189
1202
.
Rushlow
,
C.
,
Frasch
,
M.
,
Doyle
,
H.
and
Levine
,
M.
(
1989
).
Maternal regulation of zerknüllt: a homeobox gene controlling differentiation of dorsal tissues in Drosophila
.
Nature
330
,
583
586
.
Sargent
,
M. G.
and
Bennett
,
M. F.
(
1990
).
Identification in Xenopus of a structural homologue of the Drosophila gene snail
.
Development
109
,
967
973
.
Schilling
,
T. F.
(
1993
).
Cell lineage and mutational analyses of cranial neural crest development in the zebrafish embryo
.
Ph. D. Thesis
,
University of Oregon Press, Eugene, Oregon
.
Schulte-Merker
,
S.
,
Ho
,
R.
,
Herrmann
,
B.
and
Nüsslein-Volhard
,
C.
(
1992
).
The protein product of the zebrafish homologue of the mouse T gene is expressed in nuclei of the germ-ring and the notochord of the early embryo
.
Development
116
,
1021
1023
.
Schulte-Merker
,
S.
,
Hammerschmidt
,
M.
,
Beuchle
,
D.
,
Cho
,
K. W.
,
De Robertis
,
E. M.
and
Nüsslein-Volhard
,
C.
(
1994a
).
Expression of the zebrafish goosecoid and no tail gene products in wild-type and mutant ntl embryos
.
Development (in press)
.
Schulte-Merker
,
S.
,
van Eeden
,
F.
,
Halpern
,
M. E.
,
Kimmel
,
C. B.
and
Nüsslein-Volhard
,
C.
(
1994b
).
No tail (ntl) is the zebrafish homologue of the mouse T (Brachyury) gene
.
Development (in press)
.
Smith
,
D. E.
,
Franco Del Amo
,
F.
and
Gridley
,
T.
(
1992
).
Isolation of Sna, a mouse gene homologous to the Drosophila genes snail and escargot: its expression pattern suggests multiple roles during post-implantation
.
Development
116
,
1033
1039
.
Smith
,
W. C.
,
Price
,
B. M. J.
,
Green
,
J. B. A.
,
Weigel
,
D.
and
Herrmann
,
B. G.
(
1991
).
Expression of a Xenopus homolog of Xbra (T) is an immediate-early response to mesoderm induction
.
Cell
67
,
79
87
.
Steward
,
R.
(
1989
).
Relocalization of the dorsal protein from the cytoplasm to the nuclei correlates with its function
.
Cell
59
,
1179
1188
.
Streisinger
,
G. F.
,
Singer
,
C.
,
Walker
,
C.
,
Knauber
,
D.
and
Dover
,
N.
(
1986
).
Segregation analysis and gene-centromere distances in zebrafish
.
Genetics
112
,
311
319
.
Streisinger
,
G.
,
Walker
,
C.
,
Dower
,
N.
Knauber
,
D.
and
Singer
,
F.
(
1981
).
Production of clones of homozygous diploid zebrafish (Brachydanio rerio)
.
Nature
291
,
293
296
.
Thisse
,
C.
and
Thisse
,
B.
(
1992
).
Formation of the mesoderm in Drosophila involves a cascade of transcriptional regulation
.
Development
1992 Supplement 173-181.
Thisse
,
C.
,
Perrin-Schmitt
,
F.
,
Stoetzel
,
C.
and
Thisse
,
B.
(
1991
).
Sequence specific transactivation of the Drosophila twist gene by the dorsal gene product
.
Cell
65
,
1191
1201
.
Thisse
,
B.
,
Stoetzel
,
C.
,
Gorostiza-Thisse
,
C.
and
Perrin-Schmitt
,
F.
(
1988
).
Sequence of the twist gene and nuclear localization of its protein in endomesodermal cells of early Drosophila embryos
.
EMBO J
.
7
,
2175
2183
.
Whiteley
,
M.
,
Noguchi
,
P. D
,
Sensabaigh
,
S. M.
,
Odenwald
,
W. F.
and
Kassis
,
K. A.
(
1992
).
The Drosophila gene escargot encodes a zinc finger motif found in snail-related genes
.
Mech. Dev
.
36
,
117
127
.
Wilkinson
,
D.
and
Krumlauf
,
R.
(
1990
).
Molecular approaches to the segmentation of the hindbrain
.
Trends Neurosci
.
13
,
235
240
.
Wilkinson
,
D. G.
,
Bhatt
,
S.
and
Herrmann
,
B. G.
(
1990
).
Expression pattern of the mouse T gene and its role in mesoderm formation
.
Nature
343
,
657
659
.
Wolf
,
C.
,
Thisse
,
C.
,
Stoetzel
,
C.
,
Thisse
,
B.
,
Gerlinger
,
P.
and
PerrinSchmitt
,
F.
(
1991
).
The M-twist gene of Mus is expressed in subsets of mesodermal cells and is closely related to the Xenopus X-twi and the Drosophila twist genes
.
Dev. Biol
.
143
,
363
373
.
Yanagisawa
,
K. O.
,
Fujimoto
,
H.
and
Urushihara
,
H.
(
1981
).
Effects of the Brachyury (T) mutation on morphogenetic movement in the mouse embryo
.
Dev. Biol
.
87
,
242
248
.