We describe the isolation and early developmental expression of three novel zebrafish genes (rtk1-3) that encode members of the eph family of receptor tyrosine kinases. At the onset of gastrulation, rtk1 is expressed in the shield region corresponding to the future dorsal side of the embryo. As gastrulation proceeds, both rtk1 and rtk2 are expressed within the axial hypoblast along the entire axis of the embryo. After the gastrula stage is complete, expression of both genes is maintained in precursor cells of the notochord in the tail bud but is downregulated in other regions of the axial hypoblast. rtk3 is expressed in anterior axial hypoblast including the ‘pillow’ at the anterior tip of the hypoblast and in paraxial tissue in posterior regions of the embryo.

We show that the precise spatial regulation of expression of rtk genes, ntl and goosecoid along the anteroposterior axis is maintained in embryos that have no dorsoventral axis. This indicates that the mechanisms that regulate gene expression along the anteroposterior and dorsoventral axes of the hypoblast may be independent.

The inductive interactions and extensive cell movements that occur prior to and during gastrulation generate the three distinct germ layers of the vertebrate embryo. In Xenopus, the induction of mesoderm by underlying endoderm has been intensively studied and several molecules believed to be involved in this process have now been identified. Experimental evidence suggests that a combination of fibroblast growth factors (FGFs) and members of the transforming growth factor beta (TGFβ) family (such as activin) are involved in both mesoderm induction and early mesodermal patterning (Green and Smith, 1991). For instance, targeted removal of functional receptors for activin (Hemmati-Brivanlou and Melton, 1992) or FGF (Amaya et al., 1992) prevent or disrupt mesoderm formation, respectively. Furthermore, incubation of ectodermal animal caps in different concentrations of activin activates genes expressed in the mesoderm in a dose-dependent manner (Green et al., 1992). As FGF modulates the response of cells to activin (Green et al., 1992), it seems likely that in vivo these, and other, factors may act in combination during early mesodermal induction and patterning.

The majority of identified genes known to be expressed in response to mesoderm-inducing signals encode transcription factors. It is thought that these proteins act as developmental switches initiating specific patterns of cell behaviour and differentiation. However, cell-to-cell signalling is also a crucial component of early regional patterning of the mesoderm. For instance, cells expressing goosecoid can recruit neighbouring cells to form an embryonic axis (Niehrs et al., 1993). It is therefore likely that specific cell surface receptors and their ligands may function as regulators of embryonic mesodermal patterning. As yet however, little is known of the molecular components of cell-to-cell signalling pathways that are initiated following mesoderm induction.

In order to identify novel receptor molecules that may be involved in early patterning of the vertebrate embryo, we undertook a search for receptor tyrosine kinase (RTK) genes that are expressed in restricted domains of the zebrafish gastrula. RTKs are characterised by a variable extracellular ligand-binding domain, a membrane-spanning domain and an intracellular tyrosine kinase domain (Pazin and Williams, 1992; Schlessinger and Ullrich, 1992). They can be divided into distinct families, including insulin-like growth factor, epidermal growth factor and platelet-derived growth factor-related RTKs, based on the structures of their extracellular and intracellular domains. RTKs have been directly implicated in the regulation of several developmental patterning events in both vertebrates and invertebrates (Pawson and Bernstein, 1990). For instance, development of the R7 photoreceptor in the Drosophila eye is critically dependent upon activation of the sevenless RTK (review; Rubin, 1991) and mutations in the c-kit RTK result in abnormal development of neural crest derivatives (eg. Chabot et al., 1988).

We chose to screen a zebrafish cDNA library for genes of the eph family of RTKs. The ligands for this family of receptors are currently unknown, but recent descriptions of spatially restricted expression patterns of individual family members suggest that eph-related RTKs are likely to play important roles in cell-to-cell communication during development (Nieto et al., 1992; Pasquale et al., 1992). In this study, we report the cloning of three cDNAs for members of this family. All three genes are expressed during gastrulation in the fish and all show spatially restricted expression along the axial midline.

We also examine the expression of rtk genes in embryos disrupted by lithium treatment. Exposure of blastulae to lithium ions prior to the mid-blastula transition has recently been shown to cause hyperdorsal development of zebrafish embryos (Stachel et al., 1993). This is believed to occur due to most, or all, of the germ ring acquiring shield properties, which results in the generation of multiple axes. We find that, in lithium-treated embryos which lack a DV axis, there is, nevertheless, precise regulation of rtk gene expression along the AP axis.

Maintenance of fish

Breeding fish were maintained at 28.5°C on a 14 hour light/10 hour dark cycle. Embryos were collected from the colony by natural spawning. They were raised in 10% Hank’s saline (137 mM NaCl, 5.4 mM KCl, 1.3 mM CaCl2, 1.0 mM MgSO4, 0.25 mM Na2 HPO4, 4.2 mM NaHCO3) at 28.5°C and staged according to ‘The Zebrafish Book’ (Westerfield, 1993).

Isolation of cDNA clones

Genes encoding zebrafish RTKs were selected by screening a λZAP II neurula stage cDNA library, kindly provided by Dr David Grunwald. Approximately 0.5×106 recombinant phages were screened by plaque hybridisation under moderate stringency conditions using a 32P-labelled random-primed probe, which included the kinase domain and was derived from a cDNA fragment of the Xenopus homologue of Sek (Wilkinson, unpublished data) kindly provided by Dr David Wilkinson. Prehybridisation, hybridisation and post-hybridisation washings were performed by standard methods and as previously described (Xu and Tata, 1992). Positive clones were selected and purified after three rounds of hybridisation. Sequence information of the three partial cDNAs that were isolated was obtained by the dideoxy-mediated chain termination method.

Whole-mount in situ hybridisation

Several protocols were tried for in situ-hybridization protocols. Methods giving the best results were based on those of Oxtoby and Jowett (1993) and Wilkinson (1992). Zebrafish embryos were fixed overnight in 4% paraformaldehyde/PBS at 4°C and transferred to 100% methanol after manual dechorionation. Embryos were stored in methanol at −20°C or were rehydrated through 75%, 50%, 25% methanol in PBT (PBS/0.1% Tween-20), three times PBT, each for 5 minutes at room temperature. All further steps were carried out at room temperature unless stated otherwise. Embryos were treated with proteinase K (10 μg/ml) for 20 minutes and then refixed in 4% paraformaldehyde/PBT for 20 minutes. After washing five times with PBT (5 minutes each), the embryos were preincubated in the hybridisation mix (HM) containing 50% formamide, 5× SSC, pH 6.0 (adjusted with 1 M citric acid), 50 μg/ml yeast RNA, 50 μg/ml heparin, 0.1% Tween-20 at 65°C for 2 hours and were then hybridised overnight in the fresh HM containing 50–200 ng of digoxigenin-labelled RNA at the same temperature. RNA probes were synthesised according to the manufacturer’s instruction (Boehringer) and large probes were partially hydrolised under alkaline conditions to average sizes of 400–500 nucleotides.

Following hybridisation, the embryos were rinsed twice in 2× SSC, washed (10 minutes) three times with 25% hybridisation mix in 2× SSC, once in 2× SSC and twice with 0.2× SSC at 65°C (30 minutes each time). The subsequent washings were 5 minutes each in 75%, 50%, 25% 0.2× SSC in PBT then PBT. After pretreatment in PBT containing 2% sheep serum and 2 mg/ml BSA for 1 hour with gentle agitation, embryos were incubated with preabsorbed anti-DIG-AP antibody (Boehringer) in the same solution for 2 hours. To preabsorb, antibody was diluted in the above solution mixed with zebrafish embryo powder and incubated at 4°C overnight. Zebrafish embryo powder was prepared by homogenisation of 24 hour old embryos in minimum volume of PBS and precipitation twice in ice-cold acetone. Following the antibody reaction, embryos were rinsed twice in PBT, washed 6-8 times in PBT with gentle agitation for 15 minutes each time and then three times in alkaline phosphatase reaction buffer (0.1 M Tris/HCl, pH 9.5, 50 mM MgCl2, 0.1 M NaCl, 0.1% Tween-20), 5 minutes each time. The colour reaction was carried out by incubation of embryos in the above solution containing 5-bromo-4-chloro-3-indolyl-phosphate (BCIP) and 4-nitro blue tetrazolium chloride (NBT). When colour developed to the desired extent the reaction was stopped by washing the embryos three times with PBT. Specimens were mounted in 70% glycerol and photographed on Nikon microscopes using Kodak Ektar 25 film.

Whole-mount antibody staining

Embryos were refixed in 4% paraformaldehyde/PBT for 2 hours at 4°C after whole-mount in situ hybridisation. Following four washes in PBT, embryos were blocked in 10% goat serum in PBT for 2 hours at room temperature and then incubated overnight with anti-Ntl anti-serum (kindly provided by Dr Stefan Schulte-Merker) at 4°C. (Ntl is the new zebrafish name for zf-T or brachyury - Schulte-Merker, personal communication). After washing several times with PBT, secondary antibody (goat anti-rabbit horseradish peroxidase conjugated) was applied to the embryos and subsequent procedures were performed according to standard methods (Wilson et al., 1990).

Lithium treatment of embryos

Embryos at 256–1000 cell stage were immersed in their chorions for 12 minutes in 0.3 M LiCl at 28.5°C as described in Stachel et al. (1993). Following immersion, they were washed several times and allowed to develop at 28.5°C. Most embryos were fixed between 10 and 14 hours, though some were allowed to develop further. Development of lithium-treated embryos appeared to be slightly retarded compared to untreated embryos and so embryos fixed between 10 and 12 hours were approximately equivalent to normal embryos at the end of epiboly.

Observations on living embryos

Observations of living embryos enabled us to examine structural details of the hypoblast that were much less apparent following in situ hybridization protocols. Living embryos were viewed in their chorions under differential interference contrast optics. Embryo viewing chambers consisted of two coverslips separated by spacers. The distance between coverslips was such that light pressure was put on the chorion of embryos in the chambers. Gentle movement of the top coverslip moved the chorion and allowed the embryo to be rolled into any orientation for viewing.

Cloning and characterisation of the RTK cDNAs

Using a cDNA fragment, which includes part of the intracellular kinase domain, from the Xenopus homologue of Sek, we isolated eighteen clones from a neurula stage zebrafish cDNA library. Three separate cDNA classes were identified after restriction mapping and Southern blot analyses. Comparison of the deduced amino acid sequences over various regions of the zebrafish rtks, mouse Sek (Nieto et al., 1992), chick Cek5 and Cek8 (Pasquale, 1991; Sajjadi and Pasquale, 1993) and eph (Hirai et al., 1987) reveals extensive conservation (Fig. 1).

Fig. 1.

Comparison of the derived amino acid sequence of rtk1, rtk2 and rtk3 with other members of the eph family of RTK genes. (A) Diagram depicting RTK structure. The extent of sequence comparisons shown in B-E is shown by black lines. (B) Sequence comparison within the cytoplasmic domain including the kinase domain. (C) Carboxy terminal amino acid sequence comparison of rtk1, Sek and Cek8 reveal homology of zebrafish rtk1 to mouse Sek and chick Cek8. (D,E) Sequence comparison within extracellular domains of rtk2, rtk3 and other eph family members reveals some homology suggesting that rtk2 and rtk3 are members of the eph family. Sek, Cek5, Cek8 and eph sequences were obtained from published data (see references in the text).Numbers indicate the positions of the amino acids and gaps are indicated by dots. Conserved residues are underlined.

Fig. 1.

Comparison of the derived amino acid sequence of rtk1, rtk2 and rtk3 with other members of the eph family of RTK genes. (A) Diagram depicting RTK structure. The extent of sequence comparisons shown in B-E is shown by black lines. (B) Sequence comparison within the cytoplasmic domain including the kinase domain. (C) Carboxy terminal amino acid sequence comparison of rtk1, Sek and Cek8 reveal homology of zebrafish rtk1 to mouse Sek and chick Cek8. (D,E) Sequence comparison within extracellular domains of rtk2, rtk3 and other eph family members reveals some homology suggesting that rtk2 and rtk3 are members of the eph family. Sek, Cek5, Cek8 and eph sequences were obtained from published data (see references in the text).Numbers indicate the positions of the amino acids and gaps are indicated by dots. Conserved residues are underlined.

There is very high homology between the zebrafish rtks at the amino acid level, most notably rtk1 and rtk2, which have 92.3% identity over their kinase domains. All three rtks also share about 64% amino acid identity in their kinase domain with the prototype member of the family, eph. As rtk1 and murine Sek are 94.5% homologous over the kinase domain and also show high homology in the C-terminal region (Fig. 1C), it is likely that rtk1 is the fish homologue of Sek. It is also likely that chick Cek8 (Sajjadi and Pasquale, 1993) is homologous to these genes. In addition to homology within the kinase domain, further sequence data (Fig. 1D,E) shows that the extracellular domains of rtk2 and rtk3 share some homology with equivalent domains of other eph family members. There is a characteristic distribution of 20 conserved cysteines in the extracellular domains of eph family members (Pasquale, 1991) and Fig. 1D shows that four of these residues are conserved in rtk2, Sek, Cek5 and eph over the region sequenced. Conserved residues are also found in the extracellular domain of rtk3; for instance, a stretch of eight residues is conserved between rtk3, Sek and Cek5 (Fig. 1E). These sequence comparisons suggest that rtk2 and rtk3 are also members of eph family of RTKs. However, the degree of homology of rtk2 and rtk3 with eph family members of other species is not high enough to be certain that either receptor is the homolog of any currently known eph family member.

Gastrulation and axial patterning in the zebrafish

As all three rtk genes have spatially restricted expression domains during gastrulation, we will briefly describe this stage of development.A more detailed description of gastrulation in zebrafish may be found in Warga and Kimmel (1990).

At 50% epiboly (5.25 hours postfertilization (h)), blastoderm cells around the margin of the yolk cell start to involute leading to a thickening of the marginal zone termed the germ ring. Involuting cells also converge towards the dorsal side of the embryo creating a localised enlargement of the germ ring termed the embryonic shield. The embryonic shield is comparable to the dorsal lip of the frog and is the site from which cells undergo convergent extension movements that will generate the AP axis of the embryo. Cells that have involuted form the hypoblast layer, which consists of both prospective endoderm and mesoderm (Kimmel et al., 1990) (note that the term hypoblast in fish has a different meaning than the term hypoblast in some other species, such as the chick).

Many of the cells involuting at the dorsal margin form axial mesoderm that extends from the head to the future tail of the embryo and consists of presumptive notochord and, more anteriorly, prechordal axial mesoderm. In Xenopus, anterior axial mesoderm is present within a broad extension of the notochord termed the prechordal plate. In zebrafish, the development of anterior axial mesoderm has not been well studied though some features are apparent in living embryos. Towards the end of epiboly, presumptive notochord cells segregate from paraxial somitic mesoderm and an anterior continuation of the presumptive notochord extends into the head region (Fig. 2A,B). At about this stage, a further hypoblastic structure is discernible at the very anterior tip of the embryo (Fig. 2C). This accumulation of axial hypoblast has been described in fish embryos for over 100 years and has been variously termed the anterior mass, orales zellpolster, polster and pillow (Wilson, 1891; Ballard, 1982; Kimmel et al., 1990). In slightly older embryos, the pillow appears in part to lie in front of the developing brain although in its posterior part it clearly cushions the forebrain (Fig. 2D). Thus by the end of epiboly, observations of living embryos suggest that the axial hypoblast continues anterior to the notochord and ends in a broad pillow-like structure at the anteriormost tip of the embryo.

Fig. 2.

Axial hypoblast at the end of epiboly. Living zebrafish embryos at 9.5 h (A-C) and 11.5 h (D). Anterior is to the left in A and B; C and D are vegetal views looking at the head region. (A) Dorsal view of an embryo focused on the hypoblast. Presumptive notochord has segregated from paraxial mesoderm. (B) Axial hypoblast beneath the future hindbrain/midbrain. Anterior to the presumptive notochord, prechordal axial hypoblast continues as a strip towards the front of the embryo. (C,D) The ‘pillow’ of anterior hypoblast. At the end of epiboly there is a disc-like accumulation of cells at the very anteriormost tip of the embryo (C). A few hours later, the forebrain has formed a depression in the anterior hypoblast giving the pillow its characteristic shape (D). Abbreviations: fb, forebrain; pn, presumptive notochord; pch, prechordal hypoblast; pm, paraxial mesoderm. Scale bars: 50 μm.

Fig. 2.

Axial hypoblast at the end of epiboly. Living zebrafish embryos at 9.5 h (A-C) and 11.5 h (D). Anterior is to the left in A and B; C and D are vegetal views looking at the head region. (A) Dorsal view of an embryo focused on the hypoblast. Presumptive notochord has segregated from paraxial mesoderm. (B) Axial hypoblast beneath the future hindbrain/midbrain. Anterior to the presumptive notochord, prechordal axial hypoblast continues as a strip towards the front of the embryo. (C,D) The ‘pillow’ of anterior hypoblast. At the end of epiboly there is a disc-like accumulation of cells at the very anteriormost tip of the embryo (C). A few hours later, the forebrain has formed a depression in the anterior hypoblast giving the pillow its characteristic shape (D). Abbreviations: fb, forebrain; pn, presumptive notochord; pch, prechordal hypoblast; pm, paraxial mesoderm. Scale bars: 50 μm.

rtk1, 2 and 3 are expressed in regionally restricted domains of the gastrula

To investigate the developmental time course of expression of the three genes, we performed in situ hybridizations on whole embryos from the early blastula stage of development onwards. Prior to the onset of gastrulation, we did not detect transcripts of any of the genes (data not shown) indicating that there are no maternal transcripts of rtk genes. In this paper, we describe rtk gene expression patterns during gastrula stages of development. All three genes have complex spatially restricted expression domains in neural tissues at later developmental stages, which will be presented elsewhere.

As all of the rtk genes have high degrees of homology, we performed in situ hybridizations with two or more different probes for each gene to ensure there was no cross-hybridization between probes. At least one probe for each gene exclusively contained 3′-untranslated sequence. All probes gave specific characteristic labelling and we conclude that there is no cross-hybridization between probes. Sense probes gave no specific labelling.

rtk1

rtk1 transcripts are first detectable at a localised region of the marginal blastoderm at the onset of gastrulation (∼ 5 h) (Fig. 3A). As we show below, this region corresponds to the future dorsal side of the embryo, and is the site at which involuting cells are fated to give rise to axial and head mesodermal structures (Kimmel et al., 1990). As gastrulation proceeds, rtk1 transcripts accumulate along the axial hypoblast in regions that correspond to presumptive notochord and prechordal hypoblast (Fig. 3B). (As we do not know the fate of axial hypoblast anterior to the notochord, we refer to this region as prechordal hypoblast rather than prechordal mesoderm.)

Fig. 3.

Expression of rtk1 during gastrulation. Whole-mounted embryos labelled with an antisense RNA probe to rtk1. Viewed from animal pole (except D) with the future anterior end of the embryo to the left. (A) 50% epiboly. Transcripts (arrowheads) are detected at a restricted region of the germ ring. (B) 50–60% epiboly. Transcripts begin to accumulate along the developing axis of the embryo (arrowhead). (C,D) 75–85% epiboly. Transcripts are detectable along the elongating axis. Labelling at the animal pole in C is out of the plane of focus of the photograph. In D, the embryo is tilted to show the relationship of rtk1 transcripts to the yolk cell.(E)100% epiboly. Transcripts are still detectable along the entire axis but labelling intensity is much reduced at the level of the future midbrain/anterior hindbrain (large arrowhead). The small arrowheads indicate transcripts within the neural plate. Abbreviations: ap, animal pole; pn, presumptive notochord; yp, yolk plug. Scale bars: 50 μm.

Fig. 3.

Expression of rtk1 during gastrulation. Whole-mounted embryos labelled with an antisense RNA probe to rtk1. Viewed from animal pole (except D) with the future anterior end of the embryo to the left. (A) 50% epiboly. Transcripts (arrowheads) are detected at a restricted region of the germ ring. (B) 50–60% epiboly. Transcripts begin to accumulate along the developing axis of the embryo (arrowhead). (C,D) 75–85% epiboly. Transcripts are detectable along the elongating axis. Labelling at the animal pole in C is out of the plane of focus of the photograph. In D, the embryo is tilted to show the relationship of rtk1 transcripts to the yolk cell.(E)100% epiboly. Transcripts are still detectable along the entire axis but labelling intensity is much reduced at the level of the future midbrain/anterior hindbrain (large arrowhead). The small arrowheads indicate transcripts within the neural plate. Abbreviations: ap, animal pole; pn, presumptive notochord; yp, yolk plug. Scale bars: 50 μm.

Towards the end of epiboly (8-10 h), rtk1-expressing cells form a continuous strip of hypoblast from the head to the site of yolk plug closure (Fig. 3C-E). Transcripts are not detectable in marginal zone cells prior to their involution, rather it appears that the gene is only activated after cells enter the axial hypoblast (Fig. 3D). By the end of epiboly, there is a lowering in the level of rtk1 transcripts in the anterior axial hypoblast in the region that will come to underlie the midbrain and the anterior hindbrain, though high levels of transcripts are maintained in the presumptive notochord (Fig. 3E). Analysis of older embryos and of sectioned tissue (data not shown) indicates that the laterally located transcripts that start to appear at the end of epiboly are within the presumptive neural plate of the ectoderm (Fig. 3E). As development proceeds, the levels of transcripts in the axial hypoblast gradually decrease so that they are only present in the presumptive notochord and mesenchyme in the elongating tail (Fig. 4C).

Fig. 4.

Relationship of rtk1 expression to the developing notochord. Whole-mounted embryos labelled with an antisense RNA probe to rtk1 (A-C) and antibody to Ntl (A,B). Anterior is to the left in A and C and dorsal is up in B and C. (A) 60% epiboly. rtk1 transcripts (blue) and Ntl protein (brown) are colocalised at the shield, although rtk1 expression is detectable along the axis slightly anterior (arrowheads) to Ntl-containing cells. Ntl is also within germ ring cells lateral to the shield. (B) Transverse section through the presumptive notochord. rtk1 is expressed in the same cells that contain Ntl protein. (C) rtk1 expression is downregulated as the notochord differentiates but continues to be expressed in presumptive notochord cells of the extending tail bud. Abbreviations: n, notochord, ne, neuroepithelium; pn, presumptive notochord; y, yolk. Scale bars: 25 μm.

Fig. 4.

Relationship of rtk1 expression to the developing notochord. Whole-mounted embryos labelled with an antisense RNA probe to rtk1 (A-C) and antibody to Ntl (A,B). Anterior is to the left in A and C and dorsal is up in B and C. (A) 60% epiboly. rtk1 transcripts (blue) and Ntl protein (brown) are colocalised at the shield, although rtk1 expression is detectable along the axis slightly anterior (arrowheads) to Ntl-containing cells. Ntl is also within germ ring cells lateral to the shield. (B) Transverse section through the presumptive notochord. rtk1 is expressed in the same cells that contain Ntl protein. (C) rtk1 expression is downregulated as the notochord differentiates but continues to be expressed in presumptive notochord cells of the extending tail bud. Abbreviations: n, notochord, ne, neuroepithelium; pn, presumptive notochord; y, yolk. Scale bars: 25 μm.

Labelling of embryos both with an rtk1 RNA probe and an antibody to Ntl (previously termed zf-T or brachyury) confirmed that the early expression of rtk1 was confined to axial hypoblast (Fig. 4A,B). Ntl protein is present in all involuting cells of the epiblast and hypoblast but after involution only remains in axial hypoblast cells that will form the notochord (Schulte-Merker et al., 1992). rtk1 transcripts were detectable anterior to the extent of ntl-expressing cells (Fig. 4A), consistent with the observation that ntl is not expressed in prechordal hypoblast (Schulte-Merker et al., 1992).

rtk2

Spatially restricted rtk2 transcripts were first detectable at around 80-90% epiboly (Fig. 5A). Expression was observed in the axial midline and in a circumferential ring around the yolk plug. Levels of transcripts were much lower than for rtk-1 and, because of this, we cannot be sure that the expression is limited to the hypoblast. The anterior extent of rtk2 expression in the axial midline is very similar to rtk1, in that transcripts are present anterior to the presumptive notochord (Fig. 5B).

Fig. 5.

Expression of rtk2 during gastrulation. Whole-mounted embryos labelled with an antisense RNA probe to rtk2. Anterior is to the left. (A) 80–90% epiboly. rtk2 transcripts (arrowheads) are present along the entire axis and in cells surrounding the yolk cell. (B) Flat-mounted preparation of the presumptive head region at the end of epiboly. Slightly higher levels of rtk2 transcripts are present anteriorly (large arrowheads) than are present posteriorly (arrow) beneath the presumptive hindbrain. rtk2 expression is also apparent in lateral regions of the neural plate (small arrowheads). (C) 22 hour embryo. rtk2 expression is downregulated as the notochord differentiates but continues to be expressed in presumptive notochord cells of the extending tail bud. Transcripts are also present in the neural epithelium and in paraxial mesoderm (sectioned tissue suggests that this staining is in the sclerotome). Abbreviations: n, notochord; ne, neural epithelium; s, somite; yp, yolk plug. Scale bars: 50 μm.

Fig. 5.

Expression of rtk2 during gastrulation. Whole-mounted embryos labelled with an antisense RNA probe to rtk2. Anterior is to the left. (A) 80–90% epiboly. rtk2 transcripts (arrowheads) are present along the entire axis and in cells surrounding the yolk cell. (B) Flat-mounted preparation of the presumptive head region at the end of epiboly. Slightly higher levels of rtk2 transcripts are present anteriorly (large arrowheads) than are present posteriorly (arrow) beneath the presumptive hindbrain. rtk2 expression is also apparent in lateral regions of the neural plate (small arrowheads). (C) 22 hour embryo. rtk2 expression is downregulated as the notochord differentiates but continues to be expressed in presumptive notochord cells of the extending tail bud. Transcripts are also present in the neural epithelium and in paraxial mesoderm (sectioned tissue suggests that this staining is in the sclerotome). Abbreviations: n, notochord; ne, neural epithelium; s, somite; yp, yolk plug. Scale bars: 50 μm.

By the end of epiboly, expression is still detectable along the axial midline with slightly higher levels of transcripts present in the head (Fig. 5B). In addition, analysis of older embryos and of sectioned tissue (data not shown) indicates that the laterally located transcripts that start to appear at this stage are within the presumptive neural plate (Fig. 5B). As development proceeds, transcripts continue to be detectable within the developing notochord in the tail bud (Fig. 5C).

rtk3

Spatially restricted rtk3 transcripts were first detectable at 50–60% epiboly. By 80% epiboly, it was apparent that these transcripts were at the leading edge of the involuting axial hypoblast (Fig. 6A). Low levels of staining over all of the embryo except the mesoderm-free zone at the anterior ventral side suggest that this gene may have low levels of expression throughout the hypoblast in addition to stronger expression within the anterior axial hypoblast.

Fig. 6.

Expression of rtk3 during gastrulation. Whole-mounted embryos labelled with an antisense RNA probe to rtk3. The future anterior end of embryos is to the left. (A) Two views of the same embryo at around 80% epiboly. The first panel shows an anterior end view, the second panel is a view of the dorsal side of the embryo. Spatially restricted rtk3 transcripts are detectable at the leading edge of the involuting hypoblast (white arrowheads). (B-D) 100% epiboly. (B) View of the dorsal side of the embryo. rtk3 expression is present in paraxial mesoderm posterior to a sharp boundary (arrowheads) and in axial regions of the head. (C) View of the head region of the embryo. rtk3 expression is present in prechordal axial hypoblast and the pillow. (D) View of the posterior end of the embryo. rtk3 transcripts are apparent in cells around the posterior tip of the presumptive notochord (arrowhead) but are not detectable in the presumptive notochord itself. Abbreviations: mfz, mesoderm-free zone; p, pillow; pch, prechordal hypoblast; pm, paraxial mesoderm; pn, presumptive notochord; yp, yolk plug. Scale bars: 50 μm.

Fig. 6.

Expression of rtk3 during gastrulation. Whole-mounted embryos labelled with an antisense RNA probe to rtk3. The future anterior end of embryos is to the left. (A) Two views of the same embryo at around 80% epiboly. The first panel shows an anterior end view, the second panel is a view of the dorsal side of the embryo. Spatially restricted rtk3 transcripts are detectable at the leading edge of the involuting hypoblast (white arrowheads). (B-D) 100% epiboly. (B) View of the dorsal side of the embryo. rtk3 expression is present in paraxial mesoderm posterior to a sharp boundary (arrowheads) and in axial regions of the head. (C) View of the head region of the embryo. rtk3 expression is present in prechordal axial hypoblast and the pillow. (D) View of the posterior end of the embryo. rtk3 transcripts are apparent in cells around the posterior tip of the presumptive notochord (arrowhead) but are not detectable in the presumptive notochord itself. Abbreviations: mfz, mesoderm-free zone; p, pillow; pch, prechordal hypoblast; pm, paraxial mesoderm; pn, presumptive notochord; yp, yolk plug. Scale bars: 50 μm.

Towards the end of epiboly, transcripts were detectable along the axial midline in the anterior part of the embryo and, more posteriorly, in regions lateral to the presumptive notochord (Fig. 6B-D). Expression of rtk3 along the axial midline was strongest anteriorly and decreased towards the posterior end of the embryo (Figs 6B, 7A). We do not know whether this gradient is explained by individual cells expressing the gene at different levels along the AP axis or whether it reflects the fact that less cells express the gene in more posterior axial tissue. The most anterior rtk3 expression was in a broad patch of cells (Fig. 6C). By comparison with living embryos at similar stages of development (compare to Fig. 2C), we conclude that this patch of cells is the pillow.

Fig. 7.

Relationship of rtk3 expression to the presumptive notochord.Whole-mounted embryos at 100% epiboly labelled with an antisense RNA probe to rtk3 (A-C) and an antibody to Ntl (B,C). Anterior is to the left. (A) Flat-mounted preparation. rtk3 transcripts are present at high levels in axial tissue anteriorly and paraxial tissue posteriorly. Some expression is also detectable in paraxial tissue of the head (arrowhead).(B)Posterior end of the embryo. rtk3 is expressed in paraxial mesoderm adjacent to the Ntl-containing cells of the presumptive notochord.(C) Hindbrain region. The most anterior Ntl-containing cell (arrowhead) is within the posterior axial expression domain of rtk3. Abbreviations: p, pillow; pch, prechordal hypoblast; pm, paraxial mesoderm; pn, presumptive notochord; yp, yolk plug. Scale bars: 25 μm.

Fig. 7.

Relationship of rtk3 expression to the presumptive notochord.Whole-mounted embryos at 100% epiboly labelled with an antisense RNA probe to rtk3 (A-C) and an antibody to Ntl (B,C). Anterior is to the left. (A) Flat-mounted preparation. rtk3 transcripts are present at high levels in axial tissue anteriorly and paraxial tissue posteriorly. Some expression is also detectable in paraxial tissue of the head (arrowhead).(B)Posterior end of the embryo. rtk3 is expressed in paraxial mesoderm adjacent to the Ntl-containing cells of the presumptive notochord.(C) Hindbrain region. The most anterior Ntl-containing cell (arrowhead) is within the posterior axial expression domain of rtk3. Abbreviations: p, pillow; pch, prechordal hypoblast; pm, paraxial mesoderm; pn, presumptive notochord; yp, yolk plug. Scale bars: 25 μm.

To estimate the AP extent of the axial expression, we labelled embryos both with a probe to detect rtk3 expression and also an antibody to Ntl protein (which is present in all cells that will contribute to the notochord but is not present in pre-chordal hypoblast (Schulte-Merker et al., 1992). In double-labelled embryos, it was apparent that the posterior boundary of axial rtk3 expression overlapped with the anterior boundary of Ntl protein (Fig. 7C). Therefore, it is likely that rtk3 is expressed in some cells in anterior presumptive notochord.

At approximately the same AP level at which axial expression of rtk3 ceased, transcripts were detected in regions lateral to the presumptive notochord (Fig. 6B). The anterior boundary of paraxial expression was a sharp line angled caudally from the midline (Fig. 6B). Posterior to the boundary, rtk3 was expressed to the caudal end of the embryo with levels of transcripts being highest in the paraxial region adjacent to the notochord and decreasing further from the midline (Figs 6B,D, 7B).

Summary of expression patterns

Fig. 8 summarises the domains of expression of the three rtk genes at the end of epiboly. For comparison, we also show the expression domains of ntl, goosecoid and Axial, genes that are also expressed along the axial midline during gastrulation in the zebrafish (Schulte-Merker et al., 1992; Schulte-Merker, personal communication; Stachel et al., 1993; Strähle et al., 1993). As stated above, ntl is expressed in all involuting cells but expression is only maintained in presumptive notochord. Axial is expressed along the axis as far anterior as the diencephalon. Goosecoid is expressed in prechordal axial cells.

Fig. 8.

Schematic diagrams of embryos at the end of epiboly showing the approximate expression domains of rtk1-3, ntl, Axial and goosecoid. The expression pattern of ntl is adapted from Schulte-Merker et al. (1992), that of goosecoid from Stachel et al. (1993) and unpublished observations of Schulte-Merker and Xu, and that of Axial is adapted from Strähle et al. (1993). Intensity of shading is indicative of the levels of expression of the genes.

Fig. 8.

Schematic diagrams of embryos at the end of epiboly showing the approximate expression domains of rtk1-3, ntl, Axial and goosecoid. The expression pattern of ntl is adapted from Schulte-Merker et al. (1992), that of goosecoid from Stachel et al. (1993) and unpublished observations of Schulte-Merker and Xu, and that of Axial is adapted from Strähle et al. (1993). Intensity of shading is indicative of the levels of expression of the genes.

Spatial regulation of gene expression along the AP axis is maintained in embryos lacking a dorsoventral axis

In Xenopus (Kao and Elinson, 1986) and more recently zebrafish (Stachel et al., 1993), lithium treatment of blastulae has proved to be a very successful method with which to disrupt dorsoventral patterning of the embryo. We examined rtk gene expression in lithium-treated embryos to see if the genes maintain specific domains of expression in embryos with severe defects in DV patterning.

The most severely affected lithium-treated embryos are completely radialized such that hypoblast around the entire germ ring assumes a dorsal axial phenotype, and paraxial hypoblast is much reduced or absent (Stachel et al., 1993). Completely radialized embryos therefore lack a dorsoventral axis and all hypoblast assumes axial character. Using a slightly more severe protocol than that described in Stachel et al. (1993), we found that the majority of lithium-treated embryos assumed a completely radialized phenotype (Fig. 9A). Most of the completely radialized embryos failed to complete epiboly, instead, forming a mound of cells atop the yolk. Less than 20% of these embryos survived to 24 h. A minority of embryos formed multiple, discrete, axes as described in Stachel et al. (1993).

Fig. 9.

Effect of lithium treatment upon rtk gene expression. (A) Schematic of a severely dorsalized (radialized) embryo at about 12 hours of development. The embryo is symmetrical around its circumference and forms a mound of cells on top of the yolk. The dotted line indicates the extent of the yolk inside the domed embryo. (B-F) Completely radialized embryos labelled with antibody to Ntl (brown) and probes to rtk genes or goosecoid (blue). The approximate anterior limits of Ntl protein and posterior limits of rtk gene expression are indicated with white and black arrowheads respectively. (B) Animal pole view of an embryo labelled with antibody to Ntl plus an RNA probe to goosecoid. Goosecoid is expressed in the most anterior hypoblast at the top of the dome, Ntl protein is present in a ring around the goosecoid-expressing cells. The dotted line indicates the cuts made in such an embryo to produce a flat mount similar to that in C. D-F show similar flat mounts.(C)Ntl plus goosecoid. (D) Ntl plus rtk1. (E) Ntl plus rtk2. (F) Ntl plus rtk3. Scale bars, 50 μm. Abbreviations: a, anterior; p, posterior; y, yolk.

Fig. 9.

Effect of lithium treatment upon rtk gene expression. (A) Schematic of a severely dorsalized (radialized) embryo at about 12 hours of development. The embryo is symmetrical around its circumference and forms a mound of cells on top of the yolk. The dotted line indicates the extent of the yolk inside the domed embryo. (B-F) Completely radialized embryos labelled with antibody to Ntl (brown) and probes to rtk genes or goosecoid (blue). The approximate anterior limits of Ntl protein and posterior limits of rtk gene expression are indicated with white and black arrowheads respectively. (B) Animal pole view of an embryo labelled with antibody to Ntl plus an RNA probe to goosecoid. Goosecoid is expressed in the most anterior hypoblast at the top of the dome, Ntl protein is present in a ring around the goosecoid-expressing cells. The dotted line indicates the cuts made in such an embryo to produce a flat mount similar to that in C. D-F show similar flat mounts.(C)Ntl plus goosecoid. (D) Ntl plus rtk1. (E) Ntl plus rtk2. (F) Ntl plus rtk3. Scale bars, 50 μm. Abbreviations: a, anterior; p, posterior; y, yolk.

Radialized embryos fixed between 7 and 10 h and labelled with an RNA probe to goosecoid confirmed that goosecoid-expressing cells involuted around the entire margin of the germ ring and eventually accumulated at the top of the dome of cells (our data is not shown but is comparable to Stachel et al. (1993)) Later involuting cells accumulated behind the goosecoid-expressing cells. Thus by 10–12 hours, radialized embryos possess polarity with the earliest involuting hypoblast at the top of the dome and late involuted hypoblast around the circumference of the margin (Fig. 9A).

Examination of goosecoid and ntl expression in 10-12h radialized embryos revealed that anteroposterior polarity was maintained within the dome of cells (Fig. 9B,C). ntl was expressed in hypoblast cells up to, but not overlapping with goosecoid-expressing cells. This is the same distribution of expression of these two genes as is seen along the axial hypoblast of normal embryos.

Examination of the expression of rtk genes with respect to Ntl protein in completely radialized embryos revealed that all three genes maintained precise spatially restricted expression domains at appropriate levels of the anteroposterior axis. At 10–12h, rtk1 was expressed in cells up to the same A-P boundary as Ntl protein (Fig. 9D). In embryos fixed soon after the onset of gastrulation, rtk1 was expressed in cells more anterior than those expressing ntl (data not shown). We presume that the absence of rtk1 transcripts anterior to Ntl protein at 10–12 hours reflects down regulation of rtk1 in anterior hypoblast, as is observed along the axis of normal animals. rtk2 was frequently expressed at such low levels in lithium-treated animals that we could not be confident of its precise expression domains. However in those embryos in which transcripts could be unequivocally identified, rtk2 was expressed to a boundary anterior to ntl-expressing cells (Fig. 9E).

rtk3 was expressed in the most anterior hypoblast in a domain that overlapped caudally with the leading edge of ntl expression (Fig. 9F). This is precisely the same relationship as is seen along the axis of normal embryos (compare to Fig. 7C). Furthermore rtk3 expression was absent from posterior regions, consistent with the complete lack of any paraxial tissue in completely radialized embryos.

A novel role for eph family RTK genes?

The spatially and temporally restricted expression of rtk1, rtk2 and rtk3 during gastrulation suggests that the receptors encoded by these genes are involved in early patterning of the embryo. All three genes are members of the same subfamily of RTKs, the prototype member of which is the protooncogene eph (Hirai et al., 1987). Genes closely related to eph include elk (Lhotak et al., 1991) and eck (Lindberg and Hunter, 1990) and indeed this class of protein kinases is sometimes referred to as the eph/elk/eck subfamily (eg. Lai and Lemke, 1991). It is the largest, yet one of the least characterized families of RTKs; however, recent isolation of novel eph-related genes has revealed several members that potentially have roles in early embryonic development. For instance, Sek may be involved in segmental patterning of the hindbrain and somites (Nieto et al., 1992) and cell-specific expression of four tyro genes (Lai and Lemke, 1991) and cek5 (Pasquale et al., 1992) suggest roles in the regulation of cell differentiation in the nervous system. The results presented in this study indicate that members of the eph family may also have an earlier developmental role during patterning of the gastrula.

As yet, no ligand has been identified for any eph family members. The majority of vertebrate RTKs with identified ligands are receptors for growth factors (reviewed in Sch-lessinger and Ullrich, 1992), but it is apparent that RTKs also can be involved in direct interactions between proteins on adjacent cells. For instance, in Drosophila, the ligand for the sevenless RTK is the membrane spanning bride of sevenless protein (Kramer et al., 1991). The presence of an immunoglobulin-like loop and fibronectin type III repeats within eph family RTKs (O’Bryan et al., 1991; Pasquale, 1991) indicates that their extracellular domains share features with the extracellular domains of some cell adhesion molecules that are involved in direct cell-to-cell interactions (review: Hynes and Lander, 1992). The possibility that an eph family RTK may be activated by direct cell-to-cell contact was suggested following the observation that Cek5 protein is concentrated in the chick CNS in regions rich in axons and neuronal processes indicating that the receptor may play a role in either cell process outgrowth or adhesion (Pasquale et al., 1992).

Regional differences in the axial hypoblast of zebrafish embryos

Our observations on the hypoblast of living embryos combined with analysis of gene expression domains suggest that the axial hypoblast is a continuous structure that extends anteriorly from the presumptive notochord to the pillow. In support of this, several authors have suggested that prechordal axial mesoderm and presumptive notochord may transiently form a continuous axial structure (Dhorn, 1904; Vogt, 1929; Bjerring, 1967). Indeed, albeit solely on the basis of morphological criteria, it appears that axial mesoderm anterior to the notochord can differentiate into a notochord-like structure in vitro (Holfreter, 1938; Takaya, 1953).

There is also good evidence of a functional role for the prechordal axial hypoblast in that, like the notochord, it has the capacity to induce neural tissue (eg. Berquist and Kallen, 1954). For instance, in cycb16 mutant zebrafish embryos, an inability to receive signals from the underlying hypoblast is believed to be responsible for the abnormal development of the ventral midline of the CNS along its entire length, including those regions that lie anterior to the notochord (Hatta et al., 1991). Although these results suggest that prechordal axial hypoblast of fish does have a functional role during development, there is little known of the fate of cells from this region of the hypoblast.

Regulation of gene expression within the hypoblast

The spatially restricted expression of rtk1 in axial hypoblast and rtk3 in adjacent paraxial hypoblast demonstrates that during gastrulation, involuting axial and paraxial cells have distinct programmes of gene expression. This conclusion is supported by data on other genes, which show distinct expression patterns in axial and paraxial mesoderm during gastrulation (for instance: Schulte-Merker et al., 1992; Dirksen and Jamrich, 1992, Ruiz i Altaba and Jessel, 1992). Several possibilities could account for the sharp gene expression boundary between axial and paraxial mesoderm. For instance, there may be a barrier to cell movement between axial and paraxial tissue such that once cells have entered the hypoblast, they maintain expression of either ‘axially expressed’ or ‘paraxially expressed’ genes. Alternatively, cells might be able to move between axial and paraxial hypoblast but must alter their patterns of gene expression in their new environment. Detailed analysis of the behaviour of cells at the boundary between axial and paraxial hypoblast should resolve which of these mechanisms is correct.

Other boundaries may exist in the gastrula hypoblast as indicated by the observation that rtk3 is only expressed in paraxial tissue posterior to a line angled caudally from the presumptive notochord. As yet, we do not know the exact AP location of this boundary with respect to later developmental structures but comparison to ntl expression indicates that it lies within the hindbrain region of the embryo and may coincide with the future location of the first somite.

rtk genes are also expressed in precise domains along the AP axis of the axial hypoblast, both in normal embryos and in lithium-treated embryos, which lack a DV axis. Although this result suggests that the mechanisms underlying the regulation of gene expression along the AP and DV axes may be independent, there is some evidence to suggest that the same molecules may be involved in patterning both axes. For instance, higher concentrations of activin are needed to induce genes expressed in dorsal and anterior mesoderm rather than ventral and posterior mesoderm (Green et al., 1992; Cho et al., 1991; Ruiz i Altaba and Melton, 1989).

Potential roles of rtk genes during gastrulation

The early segregation of axial from paraxial hypoblast may reflect the acquisition of functional properties specific to the axial hypoblast. A large number of studies have demonstrated that axial mesoderm has a profound influence on the patterning of the overlying neural tube (eg. Dodd, 1992). Furthermore, differences in the inductive signals that emanate from the axial mesoderm have been proposed to exist along its AP axis (Mangold, 1933; Ter Horst, 1948; Hemmati-Brivanlou et al., 1990; Ang and Rossant, 1993; Placzek et al., 1993)

Although the molecular nature of pathways involved in the inductive interactions remains unknown, recent studies have suggested the possible involvement of members of the forkhead family of transcription factors (Dirksen and Jamrich, 1992, Ruiz i Altaba and Jessel, 1992, Strähle et al., 1993). In embryos in which signaling between mesoderm and overlying ectoderm is perturbed, forkhead-related genes fail to be expressed in the axial midline of the neural plate (Knöchel et al., 1992; Ruiz i Altaba and Jessel, 1992; Strähle et al., 1993). Other as yet unidentified genes, which presumably encode receptors and signalling molecules, must actually mediate the inductive events between the mesoderm and ectoderm. The early expression of rtk genes in overlapping domains along the entire axial midline of the hypoblast raises the possibility that these receptors may function in signalling pathways between axial hypoblast and adjacent neural plate or paraxial mesoderm. However, as the genes encode receptors, it is probable that the axial hypoblast receives, rather than sends, signals via the rtks. There are several lines of experimental evidence supporting such reciprocal signalling mechanisms between hypoblast and ectoderm (reviewed in Slack and Tannahill, 1992).

An alternative possibility is that rtks are involved in signalling pathways within the axial hypoblast. Transplantation experiments have shown that cells, which normally do not express ntl, can initiate expression of the gene if relocated to an environment of ntl-expressing cells (Schulte-Merker et al., 1993). This result suggests that local cell-to-cell signalling is an important factor in determining which genes a hypoblast cell will express and hence a potential role for the rtk genes would be in the signalling pathways that maintain the identity of neighbouring axial cells.

We thank David Wilkinson and Stefan Schulte-Merker for antibodies and probes, for sharing results prior to publication and for helpful comments on the manuscript, and David Grunwald, Robert Riggleman and Kathryn Helde for the cDNA library. This work was supported by a grant from the Medical Research Council. S. W. is an SERC Advanced Research Fellow.

Amaya
,
E.
,
Musci
,
T. J.
and
Kirschner
,
M. W.
(
1992
).
Expression of a dominant negative mutant receptor of the FGF receptor disrupts mesoderm formation in Xenopus embryos
.
Cell
66
,
257
270
.
Ang
,
S. L.
and
Rossant
,
J.
(
1993
).
Anterior mesendoderm induces Engrailed genes in explant cultures
.
Development
118
,
139
149
.
Ballard
,
W. W.
(
1982
).
Morphogenetic movements and fate map of the cypriniform teleost, Catostomus commersoni (Lacepede)
.
J. Exp. Zool
.
219
,
301
321
.
Berquist
,
H.
and
Kallen
,
B.
(
1954
).
Notes on the early histogenesis and morphogenesis of the central nervous system in vertebrates
.
J. Comp. Neurol
.
100
,
627
660
.
Bjerring
,
H. C.
(
1967
).
Does a homology exist between the basicranial muscle and the polar cartilage?
Colloques Int. Cent. Natn. Scient
.
163
,
223
267
.
Chabot
,
B.
,
Stephenson
,
D. A.
,
Chapman
,
V. M.
,
Besmer
,
P.
and Bernstein, A
. (
1988
).
The proto-oncogene c-kit encoding a transmembrane tyrosine kinase receptor maps to the mouse W locus
.
Nature
335
,
88
89
.
Cho
,
K. W. Y.
,
Blumberg
,
B.
,
Steinbeisser
,
H.
and
De Robertis
,
E. M.
(
1991
).
Molecular nature of Spemann’s organizer: the role of the Xenopus homeobox gene goosecoid. Cell
67
,
1111
1120
.
Dhorn
,
A.
(
1904
).
Studien zur Urgeschichte des Wirbelthierkorpers. 23. Die Mandibularhohle der Selachier. 24. Die Premandibularhohle
.
Mitt. Zool. Stn. Neapel
.
17
,
1
294
.
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 Devel
.
6
,
599
608
.
Dodd
,
J.
(
1992
).
Mesodermal control of cell identity in vertebrates
.
Curr. Op. Neurobiol
.
2
,
3
8
.
Green
,
J.
and
Smith
,
J. C.
(
1991
).
Growth factors as morphogens: do gradients and thresholds establish body plan? Trends Genet
.
7
,
245
250
.
Green
,
J. B. A.
,
New
,
H. V.
and
Smith
,
J. C.
(
1992
).
Responses of embryonic Xenopus cells to activin and FGF are separted by multiple dose thresholds and correspond to distinct axes of the mesoderm
.
Cell
71
,
731
739
.
Hatta
,
K.
,
Kimmel
,
C. B.
,
Ho
,
R. K.
and
Walker
,
C.
(
1991
).
The cyclops mutation blocks specification of the floorplate of the zebrafish central nervous system
.
Nature
350
,
339
341
.
Hemmati-Brivanlou
,
A.
and
Melton
,
D. A.
(
1992
).
A truncated activin receptor inhiits mesoderm induction and formation of axial struactures in Xenopus embryos
.
Nature
359
,
609
614
.
Hemmati-Brivanlou
,
A.
,
Stewart
,
R. M.
and
Harland
,
R. M.
(
1990
).
Region specific neural induction of an engrailed protein by anterior notochord in Xenopus
.
Science
250
,
800
802
.
Hirai
,
H.
,
Maru
,
Y.
,
Hagiwara
,
K.
,
Nishida
,
J.
and
Takaku
,
F.
(
1987
).
A novel putative tyrosine kinase receptor encoded by the eph gene
.
Science
238
,
1717
1720
.
Holfreter
,
J.
(
1938
).
Veranderungen der Reaktionswiese im alternden isolierten Gastrulaektoderm
.
Wilhelm Roux Arch. EntwMech. Org
.
138
,
163
196
.
Hynes
,
R. O.
and
Lander
,
A. D.
(
1992
).
Contact and adhesive specificities in the associations, migrations and targeting of cells and axons
.
Cell
68
,
303
322
.
Kao
,
K. R.
and
Elinson
,
R. P.
(
1991
).
Lithium induced respecification of pattern in Xenopus laevis embryos
.
Nature
322
,
371
373
.
Kimmel
,
C. B.
,
Warga
,
R. M.
and
Schilling
,
T. F.
(
1990
).
Origin and organisation of the zebrafish fate map
.
Development
108
,
581
594
.
Knöchel
,
S.
,
Lef
,
J.
,
Clement
,
J.
,
Klocke
,
B.
,
Hille
,
S.
,
Koster
,
M.
and
Knochel
,
W.
(
1992
).
Activin A induced expression of a forkhead related gene in posterior chordamesoderm (notochord) of Xenopus laevis embryos
.
Mech. Devel
.
38
,
157
165
.
Kramer
,
H.
,
Cagan
,
R. L.
and
Zipursky
,
S. L.
(
1991
).
Interaction of bride of sevenless membrane bound ligand and the sevenless tyrosine kinase receptor
.
Nature
352
,
207
212
.
Lai
,
C.
and
Lemke
,
G.
(
1991
).
An extended family of protein tyrosine kinase genes differentially expressed in the vertebrate nervous system
.
Neuron
6
,
691
704
.
Lhotak
,
V.
,
Greer
,
P.
,
Letwin
,
K.
and
Pawson
,
T.
(
1991
).
Characterization of Elk, a brain specific receptor tyrosine kinase
.
Mol. Cell Biol
.
11
,
2496
2502
.
Lindberg
,
R. A.
and
Hunter
,
T.
(
1990
).
cDNA cloning and characterization of eck, an epithelial cell receptor tyrosine kinase in the eph/elk family of protein kinases
.
Mol. Cell Biol
.
10
,
6316
6324
.
Mangold
,
O.
(
1933
).
Uber die Induktionsfahigkeit der verscheidenen Bezirke der Neurula von Urodelen
.
Naturwissenschaften
43
,
761
766
.
Niehrs
,
C.
,
Keller
,
R.
,
Cho
,
K. W. Y.
,
DeRobertis
,
E. M.
(
1993
).
The homeobox gene goosecoid controls cell migration in Xenopus embryos
.
Cell
72
,
491
503
.
Nieto
,
M. A.
,
Gilardi-Hebenstreit
,
P.
,
Charnay
,
P.
and
Wilkinson
,
D. G.
(
1992
).
A receptor tyrosine kinase implicated in the segmental patterning of the hindbrain and mesoderm
.
Development
116
,
1137
1150
.
Oxtoby
,
E.
and
Jowett
,
T.
(
1993
).
Cloning of the zebrafish krox-20 gene (KRX-20) and its expression
.
Nucleic Acids Res
.
21
,
1087
1095
.
O’Bryan
,
J. P.
,
Frye
,
R. A.
,
Cogswell
,
P. C.
,
Neubauer
,
A.
,
Kitch
,
B.
,
Prokop
,
C.
,
Espinosa
,
R.
,
LeBeau
,
M. M.
,
Earp
,
H. S.
and
Liu
,
E. T.
(
1991
).
axl, a transforming gene isolated from primary human myeloid leukemia cells, encodes a novel receptor tyrosine kinase
.
Mol. Cell Biol
.
11
,
5016
5031
.
Pasquale
,
E. B.
(
1991
).
Identification of chicken embryo kinase 5, a developmentally regulated receptor tyrosine kinase of the Eph family
.
Cell Regul
.
2
,
523
534
.
Pasquale
,
E. B.
,
Deerinck
,
T. J.
,
Singer
,
S. J.
and
Ellisman
,
M. H.
(
1992
).
Cek5, a membrane receptor-type tyrosine kinase, is in neurons of the embryonic and postnatal avian brain
.
J. Neurosci
.
12
,
3956
3967
.
Pawson
,
T.
and
Bernstein
,
A.
(
1990
).
Receptor tyrosine kinases: genetic evidence for their role in Drosophila and mouse development
.
Trends Genet
.
6
,
350
356
.
Pazin
,
M. J.
and
Williams
,
L. T.
(
1992
).
Triggering signaling cascades by receptor tyrosine kinases
.
Trends Biochem.Sci
.
17
,
374
378
.
Placzek
,
M.
,
Jessell
,
T. M.
and
Dodd
,
J.
(
1993
).
Induction of floorplate differentiation by contact dependent homeogenetic signals
.
Development
117
,
205
218
.
Rubin
,
G. M.
(
1991
).
Signal transduction and the fate of the R7 photoreceptor in Drosophila
.
Trends Genet
.
7
,
372
377
.
Ruiz i Altaba
,
A.
and
Jessel
,
T. M.
(
1992
).
Pintallavis, a gene expressed in the organizer and midline cells of frog embryos: involvement in the development of the neural axis
.
Development
116
,
81
93
.
Ruiz i Altaba
,
A.
and
Melton
,
D. A.
(
1989
).
Interaction between peptide growth factors and homeobox genes in the establishment of anteroposterior polarity in frog embryos
.
Nature
341
,
33
38
.
Sajjadi
,
S. G.
and
Pasquale
,
E. B.
(
1993
)
Five novel avian eph related tyrosine kinases are differentially expressed
.
Oncogene
8
,
1807
1813
.
Schlessinger
,
J.
and
Ullrich
,
A.
(
1992
).
Growth factor signaling by receptor tyrosine kinases
.
Neuron
9
,
383
391
.
Schulte-Merker
,
S.
,
Ho
,
R.
,
Herrmann
,
B. G.
and
Nusslein-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
1032
.
Slack
,
J. M. W.
and
Tannahill
,
D.
(
1992
).
Mechanisms of anteroposterior axis specification in vertebrates. Lessons from the amphibians
.
Development
114
,
285
302
.
Stachel
,
S. E.
,
Grunwald
,
D. J.
,
Myers
,
P. Z.
(
1993
).
Lithium perturbation and goosecoid expression identify a dorsal specification pathway in the pregastrula zebrafish
.
Development
117
,
1261
1274
.
Strähle
,
U.
,
Blader
,
P.
,
Henrique
,
D.
and
Ingham
,
P.
(
1993
).
Axial, a zebrafish gene expressed along the developing body axis, shows altered expression in cyclops mutant embryos
.
Genes Devel. (In press)
Takaya
,
H.
(
1953
).
On the notochord forming potency of the prechordal plate in Triturus gastrulae
.
Proc. Japan Acad
.
29
,
374
380
.
Ter Horst
,
J.
(
1948
).
Differenzierungs- und Inductionsleistungen verschiedener Abschnitte der Medullarplatte und des Urdarmdaches von Triton im Kombinat
.
Wilhelm Rouxs Arch. EntwMech. Org
.
143
,
275
303
.
Vogt
,
W.
(
1929
).
Gestaltungsanalyse am Amphibienkiem mit örtlicher Vitalfarbüng. II. Gastrulation und Mesodermbildung bei Urodelen und Anuren
.
Wilhelm Roux Arch. Entwmech. Org
.
120
,
384
706
.
Warga
,
R. M.
and
Kimmel
,
C. B.
(
1990
).
Cell movements during epiboly and gastrulation in zebrafish
.
Development
108
,
569
580
.
Westerfield
,
M.
(
1993
)
The Zebrafish Book
.
University of Oregon Press
Wilkinson
,
D.
(
1992
)
In Situ Hybridization. A Practical Approach
.
Oxford
:
IRL Press
.
Wilson
,
H. V.
(
1891
).
The embryology of the sea bass
.
Bull U.S. Fish Comm
.
9
,
209
278
.
Wilson
,
S. W.
,
Ross
,
L.
,
Parrett
,
T.
and
Easter
,
S. S.
, Jr
.
(
1990
).
The development of a simple scaffold of axon tracts in the brain of the embryonic zebrafish, Brachydanio rerio
.
Development
110
,
121
145
.
Xu
,
Q.
and
Tata
,
J. R.
(
1992
).
Characterization and developmental expression of Xenopus C/EBP gene
.
Mech. Devel
.
38
,
69
81
.