Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Development 113, 1495-1505 (1991) Printed in Great Britain © The Company of Biologists Limited 1991 1495 Hensen's node induces neural tissue in Xenopus ectoderm. Implications for the action of the organizer in neural induction C. R. KINTNER1 and J. DODD 2 ^Molecular Neurobiology Laboratory, The Salk Institute, San Diego, CA 92186, USA Department of Physiology and Cellular Biophysics, Center for Neurobiology and Behavior, Columbia University, New York, NY 10032, USA 2 Summary The development of the vertebrate nervous system is initiated in amphibia by inductive interactions between ectoderm and a region of the embryo called the organizer. The organizer tissue in the dorsal lip of the blastopore of Xenopus and Hensen's node in chick embryos have similar neural inducing properties when transplanted into ectopic sites in then" respective embryos. To begin to determine the nature of the inducing signals of the organizer and whether they are conserved across species we have examined the ability of Hensen's node to induce neural tissue in Xenopus ectoderm. We show that Hensen's node induces large amounts of neural tissue in Xenopus ectoderm. Neural induction proceeds in the absence of mesodermal differentiation and is accompanied by tissue movements which may reflect notoplate induction. The competence of the ectoderm to respond to Hensen's node extends much later in development than that to activin-A or to induction by vegetal cells, and parallels the extended competence to neural induction by axial mesoderm. The actions of activin-A and Hensen's node are further distinguished by their effects on lithium-treated ectoderm. These results suggest that neural induction can occur efficiently in response to inducing signals from organizer tissue arrested at a stage prior to gastrulation, and that such early interactions in the blastula may be an important component of neural induction in vertebrate embryos. Introduction tissue when transplanted to ectopic sites in chick or amphibian embryos, respectively. In addition, both tissues pattern the induced neural tissue along the anterior-posterior axis (Spemann, 1931; Tsung et al., 1965; see Hara, 1978). The similarities in the properties of the organizer in different vertebrate embryos suggests that the inducing signals underlying neural induction may be conserved. To examine this idea, we have asked whether Xenopus ectoderm responds to the organizer taken from the chick embryo as it does to its own organizer. Chick Hensen's node was isolated and cultured with Xenopus ectoderm in a standard animal cap assay. The assay was carried out at room temperature (approximately 22°C) such that the ectoderm can respond to inducing signals while the chick tissue is developmentally static. Thus tissue differentiation and morphogenetic movements observed represent the response of Xenopus ectoderm, while the inducing signals detected in the assay can be attributed to Hensen's node tissue as it existed at the stage of isolation from the embryo. We show that Xenopus ectoderm responds to Hensen's node by forming large amounts of neural tissue without detectably forming dorsal mesodermal Vertebrate neural development begins when ectoderm on the dorsal side of the embryo is induced to form neural tissue. The neural inducer is a region of tissue, known as the organizer, whose properties have been defined by transplantation studies in several vertebrate species. When the organizer is transplanted from one blastula to the ventral side of another, the host embryo forms a second dorsal axis (Spemann and Mangold, 1924). The nervous system in this second axis is derived from host ectoderm which, in the absence of the graft, would have formed ventral epidermis. This result indicates that ectoderm can be induced to form neural tissue upon appropriate interaction with the organizer, an event that normally occurs only on the dorsal side of the embryo. In amphibia, the organizer maps to the dorsal lip of the blastopore (DLB), a region of the marginal zone of the blastula just above the first site of invagination at gastrulation. In birds, the organizer maps to the anterior end of the primitive streak, a region called Hensen's node (Waddington and Schmidt, 1933; see Hara, 1978). Both Hensen's node and the DLB induce neural Key words: neural induction, organizer, activin-A, Hensen's node, notoplate. 1496 C. R. Kintner and J. Dodd derivatives. The inducing activity of Hensen's node is different from the potent embryonic inducer, activin-A (Asashima et al., 1990; Eijnden-Van Raaij et al., 1990; Smith et al., 1990; Thompsen et al., 1990). In addition, Hensen's node induces ectoderm to undergo morphological movements resulting in axial elongation in the absence of mesoderm. These results are discussed in the light of the model in which an early interaction between ectoderm and organizer tissue in the blastula is an important component of neural induction in vertebrate embryos. Methods Animals and reagents Embryos were obtained from Xenopus laevis adult frogs (NASCO and Xenopus 1) by hormone-induced egg-laying and in vitro fertilization using standard protocols. Animal caps were dissected from appropriately staged (Nieuwkoop and Faber, 1967) embryos in 0.5x MMR containing penicillin and streptomycin, as previously described (Dixon and Kintner, 1989). For lithium treatment, embryos were incubated at the 64 cell stage with 0.25 M LiCl for 10 minutes and then washed extensively with 0.5 x MMR. To measure competence, ectodermal caps were isolated from embryos at appropriate times after the start of gastrulation by removing ectoderm that was not yet contacted by involuting tissue. Hensen's node tissue was dissected from chick embryos (White Leghorns from Spafas, CT, Truslow Farms, VA, and Mclntyre Poultry, CA) at stages 3-10 (Hamburger and Hamilton, 1951) in ice cold L15 medium. The chick tissue was placed between two pieces of animal cap tissue as shown in Fig. 1, and the resulting recombinants maintained on agarosecoated dishes in 0.5 x MMR with penicillin and streptomycin at room temperature (approximately 22°C). Recombinants were harvested for RNA analysis after 20 hours or were maintained for 2.5 days with frequent changes of the culture medium for observation and histological analysis. A highly purified preparation of porcine activin-A was kindly provided by Drs W. Vale and J. Vaughan in the Peptide Biology Laboratory at the Salk Institute for Biological Studies. RNA analysis RNA was isolated from embryos or explants and assayed for the expression of specific RNA transcripts using an RNAase protection assay described previously (Melton et al., 1984; Kintner and Melton, 1987). The hybridization probes used to detect NCAM, NF3, muscle-specific actin and EF-1 alpha RNA are described elsewhere (Kintner and Melton, 1987; Dixon and Kintner, 1989). The hybridization probe for XlHbox6 RNA is a fragment of the XlHbox6 cDNA (Sharpe et al., 1987) kindly provided by Drs C. Wright and E. DeRobertis. The hybridization probe for en-2 RNA is a portion of the Xenopus engrailed-2 cDNA that was isolated and provided for us by Drs R. Harland and A. HemmatiBrivanlou (Hemmati-Brivanlou et al., 1991). The amounts of RNA in each sample were normalized by monitoring the levels of EF-1 alpha RNA (Krieg et al., 1989). RNA samples prepared from 4 recombinants or embryos were assayed simultaneously with all probes except for the muscle-specific actin probe, in which case 10% of the RNA sample was assayed separately. The Xenopus hybridization probes were fully protected by transcripts isolated from frog tissue but were not protected by transcripts isolated from chick embryos. In every assay the size of undigested probes for each transcript used was measured on the same gel as the experimental samples. This permitted us to determine whether probe escaped RNAase protection in any experimental samples. Histology and immunohistochemistry Recombinants were fixed in 3% trichloracetic acid at 4°C for 30 min. After extensive washing, they were dehydrated through an ethanol series and embedded in paraffin using standard procedures (Kintner, 1988). Tissue was sectioned at 10 fsm on a rotary microtome and mounted on gelatin-subbed slides. For histology, sections were stained with haematoxylin and eosin or Giemsa's stain. For immunohistochemistry, the sections were dewaxed, labelled with monoclonal antibodies, using a fluorescein-labelled second antibody and counterstained with Hoechst dye. Monoclonal antibodies used were as follows: anti-NCAM (gift of Drs K. Sakaguchi and W. Harris, UCSD), 12/101 (Kintner and Brockes, 1985) and Not1 (Placzek et al., 1990; Yamada et al., 1991). Results Hensen's node induces neural tissue Fate mapping and transplantation studies have led to the idea that Hensen's node in the chick is functionally equivalent to the dorsal lip of the blastopore (DLB) in amphibia (see Hara, 1978). We therefore examined whether Hensen's node could substitute for DLB tissue in a Xenopus animal cap assay by combining Hensen's nodes, dissected from stage 3.5 and stage 4 chick embryos, with pairs of ectodermal caps dissected from stage 9 Xenopus laevis embryos (Fig. 1). The recombinants were allowed to develop for 18-24 hours at room temperature and then assayed for the expression of tissue-specific RNA transcripts. The recombinants were examined for the presence of neural tissue by measuring the expression of two neural-specific transcripts, NCAM and NF-3. NCAM is expressed in Xenopus ectoderm soon after neural induction and is restricted to neural tissue (Jacobson and Rutishauser, 1986; Kintner and Melton, 1987) providing an early general marker of neural differentiation. NP-3 encodes a neuronal intermediate filament protein (Charnas et al., 1987) that is expressed in neurons beginning about 4 hours after neural tube closure (Dixon and Kintner, 1989). All transcripts were detected with an RNAase protection assay in which the Xenopus tissue but not the chick tissue generated fulllength protected fragments. Recombinants formed between Xenopus ectoderm and Hensen's node expressed both NCAM and NF-3 transcripts (Fig. 2, lane 4). The levels of the two transcripts expressed in the ectodermal caps after induction by Hensen's node were approximately equivalent to the levels observed in stage-matched control embryos normalized for the expression of EF-1 alpha RNA (Fig. 2, lane 6). Control explants of ectoderm cultured alone expressed neither NCAM nor NF-3 (Fig. 2, lane 2). Of the tissue in the chick blastula, only Hensen's node was able to induce large amounts of neural tissue in Xenopus ectoderm, although very low Neural induction in Xenopus laevis 1497 ectoderm Chicken embryo levels of NCAM RNA were induced by primitive streak immediately posterior to the node (Fig. 2, lane 5). Equal-sized pieces of tissue dissected from lateral or anterior epiblast (see Fig. 1) or caudal primitive streak did not induce neural transcripts in isolated frog ectoderm (Fig. 2, lane 3). Thus, neural inducing capacity was restricted to the region of the chick that is thought to be equivalent to the amphibian DLB (Hara, 1978). Hensen's node does not appear to differentiate within the recombinates An important aspect of these experiments is that the recombinants were maintained at room temperature (approximately 22°C) in 0.5x MMR, conditions in which development of Hensen's node should be arrested. The inducing signals detected in this assay are therefore likely to be present in the node at the time of explantation and unlikely to be generated by tissues arising from the node during the course of the assay. To determine whether Hensen's node arrests within the recombinant, or alone at 22°C in 0.5x MMR, we examined sections of the node for the expression of the differentiation antigen, Not-1. Not-1 is expressed by chick notochord, beginning at stage 5, as Hensen's node begins to regress and notochord is laid down (Placzek et al., 1990; Yamada et al., 1991). Hensen's node incubated in the presence or absence of ectodermal caps in 0.5 x MMR overnight at room temperature did not express Not-1. The antigen was expressed by stage 5 or 6 chick notochord before and after incubation of the tissue in 0.5x MMR overnight, indicating that the antigen itself is not destroyed by incubation under these conditions (not shown). Not-1 was expressed in stage 4 Hensen's node that had been incubated at 37°C in isotonic medium overnight. These results suggest that Hensen's node is arrested developmentally at room temperature and that its effects on Xenopus ectodermal tissue do not depend on differentiation into its mesodermal derivatives. The inducing properties of Hensen's node change with developmental stage Transplantation of the amphibian DLB taken from different developmental stages results in the induction Fig. 1. Pieces of chick epiblast with underlying hypoblast were dissected as shown from Hensen's node (HN), the primitive streak (PS) and other regions of the embryo (e). These were sandwiched between two animal ectodermal caps dissected from stage 9-10 Xenopus blastulae. Xenopus blastula chick tissue and frog ectoderm UNPROTECTED PROBES muscle actin 8 « 1 a O 2 Q z f) Q. ^ 8 PROTECTED PROBES w to muscle actin I 8 NF-3 NF-3 NCAM NCAM EF-1A EF-1A 1 2 3 4 6 7 Fig. 2. RNA transcripts expressed in Xenopus ectoderm in response to Hensen's node. Total RNA samples prepared from 4 recombinants or control tissues were assayed simultaneously for the presence of NF-3, NCAM and EF1alpha transcripts by RNAase protection. 10% of each sample was assayed separately for the presence of musclespecific actin transcripts. Lane 1 shows the position of the unprotected probes, indicated by arrows. Arrows on the extreme right show the position of the probes after protection with the appropriate RNA transcripts. Ectoderm cultured with stage 4 Hensen's node expresses large amounts of NCAM and NF-3 transcripts (lane 4) at levels found in stage-matched control embryos (lane 6). Ectoderm cultured with anterior primitive streak tissue expressed NCAM RNA at just detectable levels (lane 5). Note that ectoderm cultured alone (lane 2) or combined with chick epiblast (lane 3) expresses EFl-alpha but not NCAM RNA. None of the probes are protected by tRNA (lane 7). 1498 C. R. Kintner and J. Dodd of tissues with different regional characteristics (Spemann, 1931). Dorsal lip tissue taken from an early gastrula stage induces anterior structures while dorsal lip taken from later-staged embryos induces posterior tissues such as spinal cord. A similar phenomenon has been observed in the chick embryo in that secondary neural tubes induced by a subepiblast transplant of Hensen's node became progressively more posterior in character as donor nodes were taken from later staged embryos (Tsung et al., 1965; but see Dias and Schoenwolf, 1990). These observations suggest that the inducing properties of Hensen's node and the DLB change similarly with developmental age. We therefore tested the ability of Hensen's nodes, dissected from chick over a range of ages, to induce neural tissues of anterior and posterior character in Xenopus ectoderm. To do this we measured the expression of two homeobox-containing genes, en-2 and XlHbox6, that are expressed at different levels of the A-P axis in Xenopus ectoderm. En-2 is expressed in the neural plate at the level of the midbrain (Hemmati-Brivanlou and Harland, 1989; Hemmati-Brivanlou et al., 1991). XlHbox6 transcripts are expressed in the posterior region of the neural plate (Sharpe et al., 1987) and in lateral plate mesoderm (Wright et al., 1990). In animal caps induced with stage 4 Hensen's nodes the en-2 transcript was observed at levels comparable to those observed in stage-matched control embryos (Fig. 3, lanes 1,4). However, when Hensen's nodes from late stage 5 to stage 9 were used, the expression of en-2 ceased (Fig. 3, lanes 2,3). Conversely, stage 4 Hensen's node did not induce expression of XlHbox6 (Fig. 3, lane 1), whereas later-staged Hensen's nodes did induce XlHbox6 (Fig. 3, lanes 2,3). The results described above indicate that Hensen's node tissue mimics at least two characteristics of tissue in the DLB. First, Hensen's node induces large amounts of neural tissue in Xenopus ectoderm. This effect is observed with both stage 3 and stage 4 Hensen's nodes, indicating that Hensen's node contains neural inducing signals before gastrulation. Second, the neural tissue induced has different regional characteristics that depend on the age of the embryos from which Hensen's node is isolated. These results suggest that Hensen's node can be used as a model to study the properties of the signals that are present in the DLB. We have therefore analysed the response of ectoderm to Hensen's node and compared the responses of Xenopus ectoderm to Hensen's node and another potent embryonic inducer, activin-A. Hensen's node does not induce axial mesoderm Ectoderm treated with embryonic inducers such as activin-A and FGF forms mesodermal derivatives such as notochord and muscle (Slack et al., 1987; Smith, 1987). In order to determine whether ectoderm also responds to Hensen's node by forming axial mesoderm, we examined recombinants for the expression of somitic mesoderm using two assays. We first measured the expression of muscle-specific actin RNA in an RNAase protection assay. Hensen's node did not HENSEN'S NODE c in o CO CO CO c/D 8 x en-2 r NCAM EF-1A •HI? XIHBOX 6 1 2 3 4 5 Fig. 3. Expression of RNA transcripts in Xenopus ectoderm in response to Hensen's node from embryos at different developmental stages. Hensen's node was isolated from chick embryos of different stages and combined with Xenopus ectoderm from stage 10 embryos. After 24 hours in culture, total RNA samples were prepared and assayed simultaneously for en-2, NCAM, XlHbox6 and EFl-alpha RNA, using RNAase protection. The position of the protected probe for each transcript is indicated by arrows on the left. Stage 30 embryos expressed all four transcripts (lane 4) while ectoderm cultured alone expressed only EF1 alpha RNA (lane 5). Ectoderm combined with stage 4 Hensen's node expressed NCAM and en-2 but not XlHbox6 transcripts (lane 1). Ectoderm combined with late stage 5 (lane 2) or later (lane 3) expressed NCAM and XlHbox6 but not en-2 RNA. induce the expression of the muscle-specific actin transcript (Fig. 2, lane 4), indicating that the responding ectoderm did not form somitic mesoderm. In contrast, whenever neural tissue was induced by activin-A, it was accompanied by the induction of large amounts of somitic mesoderm (Fig. 5, lanes 2,3). As an alternative assay for somitic muscle, recombinants were examined for the presence of the muscle specific antigen, 12/101, by immunohistochemical labelling of tissue sections. While patches of muscle were brilliantly labelled with this antibody in control explants contain- Neural induction in Xenopus laevis 1499 Fig. 4. Histological analysis of recombinants. (A,B) Haematoxylin and eosin-stained paraffin sections of recombinants formed between, Xenopus ectoderm and Xenopus organizer (A) and Xenopus ectoderm and chick Hensen's node (B). Notochord cells (n) are present in A but not in B. (C) Immunofluorescence micrograph showing labelling of neural tissue with anti-NCAM antibodies. The neural tissue (large arrowhead) is found adjacent to the chick tissue (small arrowheads) which can be identified by Hoechst staining in serial sections (D) (small arrowheads). ing Xenopus DLB tissue, no labelling with the 12/101 antibody was observed in recombinants between ectoderm and Hensen's node (not shown). Although muscle-specific actin and 12/101-reactive tissue were not induced by Hensen's node, it is possible that notochord tissue was induced in the absence of somitic mesoderm. To determine whether chordamesoderm was present we examined sections of recombinants for the presence of notochord and other mesodermal derivatives. Notochordal tissue, recognizable by the presence of large vacuolated cells in recombinants containing DLB tissue (Fig. 4A), could not be detected in recombinates containing Hensen's node (Fig. 4B). The lack of detectable mesoderm was independent of the stage of embryo (3-10) from which Hensen's node was taken (not shown). While this analysis does not exclude the presence of a few isolated notochordal cells, it provides strong evidence that the formation of axial mesoderm is not a major response of Xenopus ectoderm to stage 4 Hensen's node. Recombinants containing stage 4 Hensen's node were also examined immunohistologically for the formation of neural tissue, using a monoclonal antibody against NCAM. Recombinants contained large amounts of NCAM-immunoreactive tissue (Fig. 4C), confirming the results obtained measuring neural transcripts. NCAM labelling was primarily associated with tissue in direct contact with Hensen's node (Fig. 4C,D), suggesting that the ectoderm formed neural tissue in the absence of intervening cells or tissues. Together, these results indicate that Hensen's node can induce large amounts of neural tissue to form from Xenopus ectoderm and that this induction occurs in the absence of axial mesodermal derivatives such as notochord and somitic muscle. In addition, since Hensen's node induces neural tissue in the absence of axial mesoderm, its properties appear to be different from those of factors identified as embryonic mesodermal inducers, such as activin-A and FGF. Hensen's node inducing activity is distinct from activin-A activity Several studies indicate that different parameters can alter the responsiveness of ectoderm to induction by polypeptide growth factors. In the case of activin-A, the ectodermal response can change as a function of the age of the embryo from which the ectoderm is isolated and by treatment of embryos with a dorsalizing agent, lithium. To characterize the properties of the inducing signals in Hensen's node further, we examined the effect of these parameters on responses of ectodermal caps to Hensen's node. 1500 C. R. Kintner and J. Dodd (i) Effects of ectodermal competence on the response to Hensen's node The ability of the ectoderm to respond to mesoderminducing signals is known to decline at the beginning of gastrulation and to be lost by midgastrula stages. We therefore tested the response of ectoderm, taken from progressively older Xenopus embryos, to activin-A and Hensen's node. In accordance with the results of others (Green et al., 1990), ectoderm taken from embryos before gastrulation (stages 8.5-10) responded to activinA (0.07 nM) by expressing both NCAM and musclespecific actin transcripts (Fig. 5A, lanes 1,2). No response was observed at 7 pM activin-A (not shown). By midgastrulation (stage 11), ectoderm no longer expressed NCAM RNA in response to activin-A (Fig. 5A, lanes 3,4) but continued to express muscle-specific actin in response to high concentrations of activin-A (1.4 nM) (Fig. 5A, lane 4). Ectoderm taken from embryos older than stage 11 did not express NCAM or muscle-specific actin in response to activin-A (not shown). Surprisingly, ectoderm continued to respond to activin-A by expressing XlHbox6 even when its ability to form neural tissue was lost. The XlHbox6 signal in ectoderm treated with activin-A after stage 11 may reflect the induction of posterior lateral mesoderm that is known to express XlHbox6 in embryos (Wright et al., 1990). B k The response of ectoderm to Hensen's node also declined with developmental age, but the decline occurred at later stages than that observed with activinA. Ectoderm continued to respond to stage 4 Hensen's node by expressing large amounts of NCAM RNA until late gastrula stages (Fig. 5B, lanes 1-3), considerably later than the developmental stages at which activin-A is still able to induce NCAM expression. Moreover, even when the ability of the ectoderm to respond to Hensen's node declined, Hensen's node did not induce the expression of muscle-specific actin (Fig. 5B, lane 3). Thus the competence of ectoderm to form neural tissue in response to Hensen's node persisted for a longer period after gastrulation than that to activin-A. At stage 11, the ectoderm has lost competence to form neural tissue in response to activin-A despite the fact that muscle-specific actin RNA was expressed. In contrast, Hensen's node continues to induce neural tissue at late stages of gastrulation. (ii) Effects of lithium on the responses of ectoderm to Hensen's node The response to activin-A is enhanced in ectoderm taken from embryos treated with lithium at the 64 cell stage. The ratio of dorsal to ventral induced mesoderm increases and the competence of the ectoderm to respond is extended (Cooke et al., 1989; Kao and Fig. 5. Effects of competence on the ectodermal response to Ect + Act (nM) activin-A and stage 4 Hensen's Ect + HN node. Animal caps were 8 isolated from embryos at O i-r^ o different ages (stages 9-11.5) muscle muscle and exposed either to activinactin actin A (A) or to Hensen's node (B). Total RNA samples were prepared from 8 animal caps, 4 NCAM NCAM recombinants or 4 control tissues, and assayed by RNAase protection for the expression of NCAM, en-2, XlHbox6 and EFl-alpha RNA. EF-1A EF-1A 10% of each sample was assayed separately for the presence of muscle-specific actin transcripts. The position of the protected probe for each transcript is shown to the XIHBOX 6 XIHBOX 6 left of each panel. Stage 30 m embryos (St 30 con) express all en-2 en-2 *" Av -•-• ~~* ^m "" ^* """' ^ five transcripts (lane 5 in A; 1 2 3 4 5 1 2 3 4 ' a n e 4 in B). (A) Ectoderm from stage 10 embryos expressed muscle actin, NCAM and XlHbox6 RNA in response to both 0.07 nM (lane 1) and 1.4 nM (lane 2) activin-A. When ectoderm was isolated from embryos at a later stage (stage 11, early gastrulation), the response to activin-A was greatly diminished. Stage 11 ectoderm did not express NCAM RNA in response to 0.07 nM (lane 3) or 1.4 nM (lane 4) activin-A. Stage 11 ectoderm could be induced to express muscle-specific actin RNA but in response only to high concentrations of activin-A (1.4 nM, lane 4). The low levels of en-2 signal in lane 4 represent undigested probe (see methods). (B) Recombinants of Hensen's node and ectoderm from stage 10 (lane 1), stage 11 (lane 2) and stage 11.5 (lane 3) embryos express NCAM and en-2 RNA but not muscle-specific actin or XlHbox6 RNA. The levels of NCAM and en-2 transcripts are approximately the same in ectoderm from stage 10 and stage 11 embryos, but decrease in ectoderm from stage 11.5. Although the stage 11.5 recombinant has less NCAM RNA expression, it does not express muscle-specific actin or XlHbox6 RNA- I Neural induction in Xenopus laevis 1501 B Fig. 6. Effect of lithium on the response of ectoderm to activin-A and Hensen's node. Animal caps were isolated from stage 9-11.5 lithiummuscle muscle 8 8 treated or control embryos and actin actin were treated with activin-A or combined with stage 4 Hensen's nodes. Total RNA NCAM NCAM samples were prepared from 8 animal caps, 4 recombinants or 4 control tissues and assayed by RNAase protection for the expression of NCAM, en-2, XlHbox6 and EFl-alpha RNA. EF-1A EF-1A 10% of each sample was used to assay for the muscle-specific actin transcript. The position of the protected probe for each transcript is marked to the left of each panel with XIHBOX 6 XIHBOX 6 arrows. Stage 30 embryos expressed all five transcripts (lane 5 in A; lane 5 in B). (A) en-2 __ „ _ Ectoderm from stage 11 embryos does not express NCAM RNA in response to t -a 3 4 activin-A and requires high concentrations of activin-A (1.4 nM) to express muscle-specific actin and XlHbox6 transcripts (lane 2). The low levels of en-2 in lane 2 represent unprotected probe. In contrast, stage 11 ectoderm from lithium-treated embryos responds to activin-A at concentrations of 0.07 nM and 1.4 nM by expressing NCAM, muscle-specific actin and XlHbox6 transcripts. Note that en-2 transcripts are not expressed in ectodenn treated with activin-A even when isolated from lithium-treated embryos. (B) The response of stage 11 (lanes 1,2) and stage 11.5 ectoderm (lanes 3,4) to Hensen's node was the same in animal caps isolated from lithium-treated (lanes 2,4) and untreated (lanes 1,3) embryos. The band in the muscle-specific actin assay (lane 4) represents unprotected probe (see methods). St11 Ect + Act (nM) st 11 Ect st 11.5 Ect + HN +HN c o o o co I Elinson, 1989). We therefore compared the effects of lithium treatment on the responses of ectodenn to Hensen's node and activin-A. As described above, ectoderm from control embryos older than stage 11 does not express NCAM in response to activin-A and expresses muscle-specific actm transcripts only in response to high concentrations of activin-A (1.4 nM) (Fig. 6A, lanes 1,2). In marked contrast, lithiumtreated stage 11 ectoderm expressed both NCAM and muscle-specific actin RNA (Fig. 6A, lanes 3,4) even when exposed to low concentrations of activin-A (0.07 nM). Thus, as previously reported (Cooke et al., 1989), lithium treatment potentiates the response of ectodenn to activin-A. The generation of neural tissue by ectoderm in response to stage 4 Hensen's node was not altered by lithium treatment (Fig. 6B, lanes 1-4). Both the levels of NCAM transcripts and the competence of the ectoderm to respond to Hensen's node were the same in lithium-treated and control embryos. Because the Hensen's node signal cannot be titred, we do not know whether it was already optimized in these experiments. Thus we cannot definitively conclude that lithium is not capable of potentiating the response of early ectoderm to Hensen's node. However, the results do show that lithium treatment is not able to extend the competence of older ectoderm to respond to Hensen's node (Fig. 6B, lanes 1-4), in contrast to its effect on the activin-A response. The effects of competence and lithium treatment on the expression of region specific markers, en-2 and XlHbox6 in Xenopus ectoderm in response to Hensen's node Stage 4 Hensen's node induces the expression of en-2 but not of XlHbox6 in stage 10 Xenopus ectoderm, as described above. Conversely, activin-A induces XlHbox6 but not en-2 in stage 10 ectoderm. The differential expression of en-2 and XlHbox6 RNA in response to activin-A and Hensen's node was unaffected by lithium treatment or by the age of the ectoderm. As in control ectoderm, en-2 RNA was not induced by activin-A in lithium-treated ectodenn (Fig. 6A, lanes 14) but was expressed at control levels in response to Hensen's node (Fig. 6B, lanes 1-4). XlHbox6 RNA was not expressed in response to stage 4 Hensen's node but was expressed in ectoderm that responded to activin-A (Fig. 6A, lanes 2-4). Extension movements are induced by Hensen's node In the course of these experiments, we consistently observed that ectodenn underwent extensive morphological movements in response to Hensen's node. In some cases these movements resulted in extreme axial 1502 C. R. Kintner and J. Dodd A B neurulation (Fig. 7 B-D). Movements were not observed in isolated Hensen's node placed alone in 0.5x MMR at room temperature (not shown) or in isolated ectoderm from Xenopus embryos, indicating that they were initiated in the ectoderm in response to Hensen's node. It is unlikely that ectoderm induced the movements in Hensen's node, since in sections of the recombinants the chick cells appeared to have remained in a local cluster. Elongation movements of the magnitude observed in the Hensen's node recombinants are known to occur in Xenopus embryos only in cells of the involuting marginal zone at the DLB as they form the midline of the prospective dorsal mesoderm, or in cells of the noninvoluting marginal zone as they form the midline of the neural plate (Keller, 1985). Since the recombinants do not form mesodermal derivatives, the observed movements are likely to be associated with the formation of neural tissue. Discussion Both Hensen's node of the chick and the DLB of amphibia induce host ectoderm, in chick and Xenopus respectively, to form neural tissue upon transplantation to an ectopic site (Spemann and Mangold, 1924; Waddington and Schmidt, 1933; Smith and Slack, 1983; Vakaet, 1965; McCallion and Shinde, 1973; Hara, 1978; Hornbruch et al., 1989; Dias and Schoenwolf, 1990). The results reported here show that Xenopus ectoderm also responds in vitro to Hensen's node. The ectoderm forms neural tissue at levels similar to those formed in normal embryos and expresses distinct regional neural markers in response to nodes taken from embryos at different developmental stages. The induction of neural tissue by Hensen's node is not accompanied by the formation of any detectable mesodermal derivatives and occurs under conditions in which development of the node itself is arrested. Thus organizer tissue appears to be able to act as a potent neural inducer before it gastrulates and forms mesodermal derivatives. Fig. 7. Xenopus ectoderm/Hensen's node recombinants show gastrulation-like movements. (A) Several recombinants showing extreme elongation and concentration of pigment are shown 18 hours after the tissues were placed together. (B-D) A recombinant at 2 (B), 6 (C) and 18 (D) hours after sandwiching of the Hensen's node with Xenopus ectoderm. elongation of the recombinants (Fig. 7A), during which the inner ectodermal cells were extruded. In other cases, cell movements were extensive and resulted in a complex pigmentation patterns, but seemed not to be oriented and the explants did not elongate. The timing of the movements occurred over the time course of axial elongation in embryos undergoing gastrulation and The organizer as an early inducer of neural tissue The finding that avian organizer tissue is a potent inducer before gastrulation and need not develop into mesodermal derivatives to act as an effective neural inducer suggests that neural induction in vivo may occur in the blastula. This induction may be due to an interaction between the organizer and ectoderm across the boundary they share in the blastula and may be carried by signals travelling in the plane of the ectoderm. This implies that organizer tissue need not involute beneath the ectoderm for neural induction to occur. This idea has already been suggested from results with Xenopus embryos (exogastrulae) and explants (Keller sandwiches) in which DLB tissue does not involute beneath the prospective neural ectoderm during gastrulation. In this situation in Xenopus the ectoderm in edgewise contact with the DLB tissue Neural induction in Xenopus laevis 1503 forms large amounts of neural tissue (Kintner and Melton, 1987; Keller and Danilchik, 1988; Dixon and Kintner, 1989). The idea of an early induction between the organizer and ectoderm is also supported by the observation that one of the first responses of ectoderm to Hensen's node is elongation movements. Elongation movements are known to occur in only two regions of the Xenopus embryo. One region corresponds to the cells within the involuting marginal zone, a region which converges, involutes, comes to lie beneath the neural plate and forms notochord. The other region maps to cells in the noninvoluting marginal zone, called notoplate cells, which extend within the ectodermal sheet along the midline of the prospective neural plate, during gastrulation and neurulation (Jacobson and Gordon, 1976; Gordon and Jacobson, 1978; Keller, 1985; Jessell et al., 1989). The notoplate arises from the ectoderm that is closest to the organizer in the blastula fate map and its movements have been shown to occur in the absence of underlying mesoderm and when dissected free of prospective mesoderm after stage 11 (Keller and Danilchik, 1988). This suggests that the notoplate is specified by an interaction between the organizer and adjacent ectoderm or by an inductive signal from vegetal cells (Jacobson and Sater, 1988) prior to the onset of gastrulation. Our results support this notion in that Xenopus ectoderm responds to Hensen's node with striking movements that can be observed within two hours of contact. These movements cause the explants to elongate over a time course similar to that of gastrulation and neurulation in whole embryos (Symes and Smith, 1987), and occur in the absence of detectable mesoderm. The simplest explanation of these movements is that they reflect the movements of notoplate cells which have been induced in the ectoderm by signals from organizer tissue. Induction of the notoplate by an interaction between the organizer and ectoderm before gastrulation may be an important component of neural induction. As they converge and extend along the anterior-posterior axis, notoplate cells may be the source of neural inducing signals that act on surrounding dorsal ectoderm to form lateral regions of the neural plate. This model predicts that notoplate extension is necessary for efficient neural induction and that inducing signals travel from the midline of the prospective neural plate as proposed by others (Nieuwkoop et al., 1952; Leussink, 1970; Gordon and Brodland, 1987). This model may also explain the formation of neural tissue in cases of edgewise induction in Keller sandwiches and exogastrulae as a two step process in which the organizer first induces the notoplate cells in the noninvoluting marginal zone. Our results, interpreted in the light of this model, suggest that Hensen's node may owe its neural inducing properties to its ability to induce a notoplate. We cannot rule out the alternative model, however, that induction of the notoplate and the rest of the neural plate are both mediated directly, and in parallel, by interactions between the organizer and ectoderm. Induction with Hensen's node is distinct from that of growth factors such as activin-A To begin to characterize the inducing signals generated by organizer tissue, we compared induction by Hensen's node to that by another potent embryonic inducer, activin-A. Three lines of evidence suggest that the induction of neural tissue by Hensen's node and activinA involve different pathways. First, recognizable mesodermal derivatives are not observed when neural tissue is induced by Hensen's node. Second, the competence of the ectoderm for neural induction extends much later than that for induction by activin-A. Third, the response of the ectoderm to activin-A is enhanced by lithium treatment while induction by Hensen's node is unaffected. In Xenopus, neural tissue has been shown to differentiate from animal cap ectoderm exposed to activin-A. The induction of neural tissue by these factors is always accompanied by the induction of axial mesoderm structures, in particular notochordal tissue and muscle. The neural tissue formed in response to growth factors such as activin-A may therefore be generated in response to signals derived from the induced mesoderm rather than from direct actions of activin-A on the undifferentiated ectoderm. In support of this, recent studies have shown that dissociated animal cap ectodermal cells treated with activin-A do not themselves acquire neural characteristics but do have neural inducing ability when assayed by the response of naive animal cap ectodermal cells (Green and Smith, 1990). In contrast, it is unlikely that Hensen's node induces neural tissue indirectly through the induction of mesoderm. Axial mesoderm is undetectable in the recombinants by RNAase protection assays and by morphological and immunohistochemical analysis. Furthermore, in sections of the recombinants the induced neural tissue is often found to be in contact with the node. It is possible that Hensen's node induces prechordal mesoderm, for which there is no reliable marker in animal cap conjugates and that this mesodermal cell type is responsible for neural induction. This is unlikely for two reasons. First, prechordal mesoderm is relatively ineffective in inducing neural tissue in Xenopus ectodermal caps (Dixon and Kintner, 1989; Savage and Phillips, 1989; Sive et al., 1989). Second, the induction of neural tissue with posterior character by late stage Hensen's nodes suggests that prechordal mesoderm is not involved. The different time courses over which ectoderm responds to activin-A and Hensen's node provide compelling evidence that the actions of these two inducers differ. Hensen's node induced neural tissue in ectoderm taken from late gastrulation stage embryos. This time course is similar to the competence of the ectoderm to form neural tissue in response to axial mesoderm (Sharpe and Gurdon, 1990) and extends much later than that for mesodermal induction. Ectoderm responded to activin-A only until just after the onset of gastrulation. These observations support the idea that more than one induction step is required 1504 C. R. Kintner and J. Dodd for the induction of neural tissue by activin-A while, in contrast, Hensen's node induces neural tissue more directly. The evidence described above shows that the neural inducing signal in Hensen's node cannot be ascribed solely to an activin-A-like molecule, suggesting that the Hensen's node-derived signal may be different from activin-A. Alternatively, the neural inducing properties of Hensen's node may result from the presence of an activin-A-like molecule together with a second agent that alters the response properties of the ectoderm to the growth factor. In fact, experiments with lithium have previously shown that the response properties of ectodermal cells can be altered (Cooke et al., 1989; Kao and Elinson, 1989). In addition, a dorso-ventral difference in the response of ectoderm to activin-A has been observed: activin-A induces anterior dorsal structures in dorsal ectoderm and ventral mesodermal derivatives in ventral ectoderm (Ruiz i Altaba and Jessell, 1991; Sokol and Melton, 1991). In contrast, the response of ectoderm to Hensen's node was not altered by lithium treatment. Furthermore, in preliminary experiments we have found that Hensen's node induced equal amounts of neural tissue in stage 10 dorsal and ventral ectoderm (not shown). Thus, if activin-A were produced by Hensen's node, a second factor would have to be invoked that would abolish the dorso-ventral difference and maximize the response of the ectoderm to an activin-A-like molecule so that lithium cannot potentiate the response further. Regionalization of neural tissue by Hensen's node Hensen's nodes taken from embryos at different developmental stages induced neural tissue with different anterior-posterior character. Nodes dissected from stage 4 and early stage 5 chicks induced neural tissue with anterior properties (en-2 positive, XlHbox6 negative) while later-staged Hensen's nodes induced posterior neural tissue (en-2 negative, XlHbox6 positive). Stage-dependent changes in the signals emanating from Hensen's node may result from a changing cell population in the node or from a change in the properties of the same population of cells within the node over time. Signals emanating from the organizer may therefore act early to regionalize the prospective neural ectoderm along the anterior-posterior axis. This patterning information may also be retained in differentiated mesoderm, including notochord. This possibility is supported by recent experiments in which the A-P distribution of en-2 in Xenopus can be altered by transplanting underlying notochord originating from different A-P levels (Hemmati-Brivanlou et al., 1990). Taken together, our results provide evidence that the inducing signals underlying neural induction and early axial patterning are conserved in evolution. Furthermore, organizer tissue acts as a potent neural inducer even when it fails to differentiate into dorsal mesodermal derivatives. These findings support the model that early interactions between the organizer and animal cap ectoderm in the blastula are important components of neural induction. We thank Cliff Hume, Tom Jesse 11, Nancy Papalopulu, Marysia Placzek and Ariel Ruiz i Altaba for helpful, if heated, discussions about the work and comments on the manuscript. We are also extremely grateful to Drs Ali Hemmati-Brivanlou and Harland for providing us with an en-2 cDNA before publication and to Ray Keller for insights into the gastrulation movements of neural tissue. Eric Hubel provided excellent photographic assistance and Ira Schieren generatedfigure1. The work was supported by grants from the NIH, to C.K. and to J.D., from The McKnight Fund for Neuroscience, to C.K. and to J.D., and The Esther A. and Joseph Klingenstein Fund, to J.D. References ASASHIMA, M., NAKANO, H . , SHIMADA, K., KJNOSHITA, K., ISHn, K., SHIBAI, H. AND UENO, N. (1990). Mesodermal induction in early amphibian embryos by activin A (erythroid differentiation factor). WUhelm Roux's Arch. Dev. Biol. 198, 330-335. CHARNAS, L., RICHTER, K., SARGENT, T. AND DAWID, I. (1987). Complementary DNA cloning of a nervous system specific intermediate filament from Xenopus laevis with homology to mammalian neurofilament. Soc. Neurosci Abstr. 450, 15. COOKE, J., SYMES, K. AND SMITH, E. J. (1989). Potentiation by the lithium ion of morphogenetic responses to a Xenopus inducing factor. Development 105, 549-558. DIAS, M. AND SCHOENWOLF, G. S. (1990). Formation of ectopic neuroepithelium in chick blastoderm: age-related capacities for induction and self-differentiation following transplantation of quail Hensen's nodes. Anat. Rec. 229, 437-448. DIXON, J. AND KINTNER, C. R. (1989). Cellular contacts requires for neural induction in Xenopus embryos: evidence for two signals. Development 106, 749-757. EUNDEN-VAN RAAII, A. J. M., ZOELENT, E. J. J., NIMMEN, K., KOSTER, C. H., SNOEK, G. T., DURSTON, A. J. AND HUYLEBROECK, D. (1990). Activin-like factor from a Xenopus laevis cell line responsible for mesoderm induction. Nature 345, 732-734. GORDON, R. AND BRODLAND, G. R. (1987). The cytoskeletal mechanics of brain morphogenesis. I. Cell state splitters causes primary neural induction. Cell Biophysics 11, 177-238. GORDON, R. AND JACOBSON, A. G. (1978). The shaping of tissues in embryos. Sci. Amer. 238, 106-113. GREEN, J. B. A., HOWES, G., SYMES, K., COOKE, J. AND SMITH, J. C. (1990). The biological effects of XTC-MIF: quantitative comparison with Xenopus bFGF. Development 108, 173-184. GREEN, J. B. A. AND SMITH, J. C. (1990). Graded changes in dose of a Xenopus activin A homologue elicit stepwise transitions in embryonic cell fate. Nature 347, 391-394. HAMBURGER, V. AND HAMILTON, H. (1951). A series of normal stages in the development of the chick embryo. / . Morphol. 88, 49-92. HARA, K. (1978). 'Spemann's Organizer' in birds. In Organizer - a Milestone of a Half-Century from Spemann (eds O. Nakamura and S. Toivonen). Elsevier, North Holland. pp221-265. HEMMATI-BRIVANLOU, A., DE LA TORRE, J. R., HOLT, C. AND HARLAND, R. M. (1991). Cephalic expression and molecular characterization of Xenopus En-2. Development 111, 715-724. HEMMATI-BRIVANLOU, A. AND HARLAND, R. M. (1989). Expression of an engrailed-re\atcd protein is induced in the anterior neural ectoderm of early Xenopus embryos. Development 106, 611-617. 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. HORNBRUCH, A . , SUMMERBELL, D . AND WOLPERT, L. (1979). Somite formation in the early chick embryo following grafts of Hensen's node. J. Embryol. Exp. Morphol. 51, 51-62. JACOBSON, A. G. AND GORDON, R. (1976). Changes in the shape of the developing vertebrate nervous system analyzed experimentally, mathematically and by computer simulation. / . exp. loot 197, 191-246. Neural induction in Xenopus laevis 1505 JACOBSON, A. G. AND SATER, A. K. (1988). Features of embryonic induction. Development 104, 341-359. JACOBSON, M. AND RUTISHAUSER, U. (1986). Induction of neural cell adhesion molecule (NCAM) in Xenopus embryos. Dev. Biol. 116, 524-531. JESSELL, T. M., BOVOLENTA, P., PLACZEK, M., TESSIER-LAVIGNE, M. AND DODD, J. (1989). Polarity and patterning in the neural tube: the origin and function of the floor plate. Ciba Foundation Symp. 144, 255-276. KAO, K. R. AND EUNSON, R. P. (1989). Dorsalization of mesoderm induction by lithium. Dev. Biol. 132, 81-90. KELLER, R. E. (1985). The cellular basis of amphibian gastrulation. In: Browder L. W. (ed) Developmental Biology. Plenum New York, vol 2, 241-327. KELLER, R. E. AND DANILCHIK, M. (1988). Regional expression, pattern and timing of convergence and extension during gastrulation of Xenopus laevis. Development 103, 193-209. KINTNER, C. R. (1988). Effects of altered expression of the neural cell adhesion molecule, NCAM, on early neural development in Xenopus embryos. Neuron 1, 545-555. KINTNER, C. R. AND BROCKES, J. P. (1985). Monoclonal antibodies to the cells of a regenerating limb. J. Embryol. exp. Morphol. 89, 37-55. KINTNER, C. R. AND MELTON, D. M. (1987). Expression of Xenopus NCAM RNA is an early response of ectoderm to induction. Development 99, 311-325. KJUEG, P. A., VARNUM, S. M., WORMINGTON, W. M. AND MELTON, D. A. (1989). The mRNA encoding the elongation factor-la is a major transcript at the midblasrtula transition. Dev. Biol. 133, 93-100. LEUSSINK, J. A. (1970). The spatial distribution of inductive capacities in the neural plate and archenteron roof of urodeles. Netherlands J. Zool. 20, 1-79. MCCALUON, D. J. AND SHTNDE, V. A. (1973). Induction in the chick by quail Hensen's node. Experentta 29, 321-322. MELTON, D. A., KJUEG, P. A., REBAGLIATI, M. R., MANIATIS, T., ZTNN, K. AND GREEN, M. R. (1984). Efficient in vitro synthesis of biologically active RNA and RNA hybridization probes from plasmids containing a bacteriophage SP6 promoter. Nucl. Acids Res. 12, 7035-7056. NIEUWKOOP, P. D., BLOEMSMA, F. F. S. N., BOTERENBROOD, E. C , HOESSELS, E. L. M. J., KREMER, A., MEYER, G. AND VERHEYEN, F. J. (1952). Activation and organization of the central nervous system in amphibians. Part I,II and III. / . exp. Zool. 120, 1-108. NIEUWKOOP, P. D. AND FABER, J. (1967). Normal table of Xenopus Laevis (Daudin). Amsterdam, North Holland. PLACZEK, M., TESSIER-LAVIGNE, M., YAMADA, T., JESSELL, T. M. AND DODD, J. (1990). Mesodermal control of neural cell identity: Floor plate induction by the notochord. Science 250, 985-988. Ruiz I ALTABA, A. AND JESSELL, T. M. (1991). Retinoic acid modifies mesodermal patterning in early Xenopus embryos. Genes and Dev. 5, 175-187. SAVAGE, R. AND PHILLIPS, C. R. (1989). Signals from the dorsal blastopore lip region during gastrulation bias the ectoderm toward a nonepidermal pathway of differentiation in Xenopus laevis. Dev. Biol. 133, 157-168. differentiation shows the importance of predetermination in neural induction. Cell 49, 749-758. SHARPE, C. R. AND GURDON, J. B. (1990). The induction of anterior and posterior neural genes in Xenopus laevis. Development 109, 765-774. SrvE, H. L., HATTORI, K. AND WEINTRAUB, H. (1989). Progressive determination during formation of the anteroposterior axis in Xenopus laevis. Cell 58, 171-180. SLACK, J. M., DARLINGTON, B. G., HEATH, J. K. AND GODSAVE, S. F. (1987). Mesoderm induction in early Xenopus embryos by heparin-binding growth factors. Nature 326, 197-200. SMITH, J. C. (1987). A mesoderm-inducing factor is produced by a Xenopus cell line. Development 99, 3-14. SMITH, J. C , PRICE, B. M. J., NIMMEN, K. AND HUYLEBROECK, D. (1990). Identification of a potent Xenopus mesoderm-inducing factor as a homologue of activin-A. Nature 345, 729-731. SMITH, J. C. AND SLACK, J. M. (1983). Dorsalization and neural induction: properties of the organozer in Xenopus laevis. J. Embryol. exp. Morphol. 78, 299-317. SOKOL, S. AND MELTON, D. A. (1991). Pre-existent pattern in Xenopus animal pole revealed by induction with activin. Nature 351, 409-411. SPEMANN, H. (1931). Uber den Abteil vom Implantat und Wirtskeime an der Orientierung und Beschaffenheit der induzierten Embryonalanlage. Roux's Arch. EntwMech. Org. 123, 389-517. SPEMANN, H. AND MANGOLD, H. (1924). Uber Induktion von Embryonanlage durch Implantation artfremder Organisatoren. Roux's Arch. EntwMech. Org. 100, 599-638. SYMES, K. AND SMJTH, J. (1987). Gastrulation movements provide an early marker of mesoderm induction in Xenopus laevis. Development 101, 339-349. THOMSEN, G., WOOLF, T., WHITMAN, M., SOKOL, S., VAUGHAN, J., VALE, W. AND MELTON, D. A. (1990). Activins are expressed early in Xenopus embryogenesis and can induce axial mesoderm and anterior structures. Cell 63, 485-493. TSUNG, S. D . , NING, I. L. AND SHIEH, S. P. (1965). Studies on the inductive action of the Hensen's node following its transplantation in ovo to the early chick blastoderm. II Regionally specific induction of the node region of different ages. Acta Biol. Exp. Sinica. 10, 69-80. VAKAET, L. (1965). Resultats de la greffe de noeud Hensen d'age different sur le blastoderme de poulet. C. R. Sianc. Soc. Biol. 159, 232-233. WADDINGTON, C. H. AND SCHMIDT, C. A. (1933). Induction by heteroplastic grafts of the primitive streak in birds. Roux's Arch. EntwMech. Org. 128, 522-563. WRIGHT, C. V. E., MORTTA, E. A., WILKIN, D. J. AND D E ROBERTIS, E. M. (1990). The Xenopus XlHbox 6 homeo protein, a marker of posterior induction, is expressed in proliferating neurons. Development 109, 225-234. YAMADA, T., PLACZEK, M., TANAKA, H., DODD, J. AND JESSELL, T. M. (1991). Control of cell pattern in the developing nervous system: polarizing activity of the floor plate and notochord. Cell 64, 635-648. SHARPE, C. R., FRITZ, A., DEROBERTIS, E. M. AND GURDON, J. B. (1987). A homeobox-containing marker of posterior neural (Accepted 19 September 1991)