Search

In this article, we describe the mechanisms by which morphogens in the Xenopus embryo exert their long-range effects. Our results are consistent with the idea that signalling molecules such as activin and the nodal-related proteins traverse responding tissue not by transcytosis or by cytonemes but by movement through the extracellular space. We suggest, however, that additional experiments, involving real-time imaging of morphogens, are required for a real understanding of what influences signalling range and the shape of a morphogen gradient.

In the mouse, embryological and genetic studies have indicated that two spatially distinct signalling centres, the anterior visceral endoderm and the node and its derivatives, are required for the correct patterning of the anterior neural ectoderm. The divergent homeobox gene Hex is expressed in the anterior visceral endoderm, in the node (transiently), and in the anterior definitive endoderm. Other sites of Hex expression include the liver and thyroid primordia and the endothelial cell precursors. We have used transgenic analysis to map the cis-acting regulatory elements controlling Hex expression during early mouse development. A 4.2-kb upstream region is important for Hex expression in the endothelial cell precursors, liver, and thyroid, and a 633-bp intronic fragment is both necessary and sufficient for Hex expression in the anterior visceral endoderm and the anterior definitive endoderm. These same regions drive expression in homologous structures in Xenopus laevis, indicating conservation of these regulatory regions in vertebrates. Analysis of the anterior visceral endoderm/anterior definitive endoderm enhancer identifies a repressor region that is required to downregulate Hex expression in the node once the anterior definitive endoderm has formed. This analysis also reveals that the initiation of Hex expression in the anterior visceral endoderm and axial mesendoderm requires common elements, but maintenance of expression is regulated independently in these tissues., (Copyright 2001 Academic Press.)

Brachyury, a member of the T-box gene family, is required for posterior mesoderm and notochord differentiation in vertebrate development, and mis-expression of Xenopus Brachyury causes ectopic mesoderm formation. Brachyury is a transcription activator, and its ability to activate transcription is essential for its biological function, but Brachyury target genes have proved difficult to identify. Here we employ a hormone-inducible Brachyury construct and subtractive hybridization to search for such targets. Using this approach we have isolated Bix1, a homeobox gene expressed both in the marginal zone of Xenopus and in the vegetal hemisphere. Expression of Bix1 is induced in an immediate-early fashion by mesoderm-inducing factors such as activin as well as by the products of the T-box genes Xbra and VegT (also known as Antipodean, Brat and Xombi). Activation of Bix1 in response to Xbra is direct in the sense that it does not require protein synthesis, and both Xbra and VegT activate expression of a reporter gene driven by the Bix 5' regulatory region, which contains an Xbra/VegT binding site. Mis-expression of low levels of Bix1 causes formation of ventral mesoderm, while high levels induce endodermal differentiation. These results suggest that Bix1 acts downstream of both VegT and Xbra to induce formation of mesoderm and endoderm.

The TGFbeta family member activin induces different mesodermal cell types in a dose-dependent fashion in the Xenopus animal cap assay. High concentrations of activin induce dorsal and anterior cell types such as notochord and muscle, while low concentrations induce ventral and posterior tissues such as mesenchyme and mesothelium. In this paper we investigate whether this threshold phenomenon involves the differential effects of the two type I activin receptors ALK-2 and ALK-4. Injection of RNA encoding constitutively active forms of the receptors (here designated ALK-2* and ALK-4*) reveals that ALK-4* strongly induces the more posterior mesodermal marker Xbra and the dorsoanterior marker goosecoid in animal cap explants. Maximal levels of Xbra expression are attained using lower concentrations of RNA than are required for the strongest activation of goosecoid, and at the highest doses of ALK-4*, levels of Xbra transcription decrease, as is seen with high concentrations of activin. By contrast, the ALK-2* receptor activates Xbra but fails to induce goosecoid to significant levels. Analysis at later stages reveals that ALK-4* signalling induces the formation of a variety of mesodermal derivatives, including dorsal cell types, in a dose-dependent fashion, and that high levels also induce endoderm. By contrast, the ALK-2* receptor induces only ventral mesodermal markers. Consistent with these observations, ALK-4* is capable of inducing a secondary axis when injected into the ventral side of 32-cell stage embryos whilst ALK-2* cannot. Co-injection of RNAs encoding constitutively active forms of both receptors reveals that ventralising signals from ALK-2* antagonise the dorsal mesoderm-inducing signal derived from ALK-4*, suggesting that the two receptors use distinct and interfering signalling pathways. Together, these results show that although ALK-2* and ALK-4* transduce distinct signals, the threshold responses characteristic of activin cannot be due to interactions between these two pathways; rather, thresholds can be established by ALK-4* alone. Furthermore, the effects of ALK-2* signalling are at odds with it behaving as an activin receptor in the early Xenopus embryo.

Injection of RNA encoding BMP-4 into the early Xenopus embryo suppresses formation of dorsal and anterior cell types. To understand this phenomenon, it is necessary to know the stage at which BMP-4 acts. In this paper, we present three lines of evidence showing that BMP-4 misexpression has no effect on the initial steps of mesoderm induction, either dorsal or ventral, but instead causes ventralization during gastrulation. Firstly, activation of organizer-specific genes such as goosecoid, Xnot, pintallavis and noggin occurs normally in embryos injected with BMP-4 RNA, but transcript levels are then rapidly down-regulated as gastrulation proceeds. Similarly, BMP-4 does not affect the initial activation of goosecoid by activin in animal caps, but expression then declines precipitously. Secondly, embryos made ventral by injection with BMP-4 RNA cannot be rescued by grafts of Spemann's organizer at gastrula stages. Such embryos therefore differ from those made ventral by UV-irradiation, where the defect occurs early and rescue can be effected by the organizer. Finally, the dorsalizing effects of the organizer, and of the candidate dorsalizing signal noggin, both of which exert their effects during gastrulation, can be counteracted by BMP-4. Together, these experiments demonstrate that BMP-4 can act during gastrulation both to promote ventral mesoderm differentiation and to attenuate dorsalizing signals derived from the organizer.

Identification of the signalling molecules involved in mesoderm formation in amphibian embryos still presents problems. None of the original candidates, such as activin, have been definitively ruled out, and new factors, such as the nodal-related genes, have come on to the scene. Of the original candidates, activin has been definitively shown to act as a morphogen, whereas bone morphogenetic protein (BMP)-4 has emerged as a ventral inducer and an inhibitor of neural differentiation. The effects of BMP-4 are antagonized by chordin, a molecule related to the product of the Drosophila gene short gastrulation.

Mouse embryos homozygous for a null mutation in nodal arrest development at early gastrulation and contain little or no embryonic mesoderm. Here, two Xenopus nodal-related genes (Xnr-1 and Xnr-2) are identified and shown to be expressed transiently during embryogenesis, first within the vegetal region of late blastulae and later in the marginal zone during gastrulation, with enrichment in the dorsal lip. Xnrs and mouse nodal function as dose-dependent dorsoanterior and ventral mesoderm inducers in whole embryos and explanted animal caps. Using a plasmid vector to produce Xnr proteins during gastrulation, we show that, in contrast to activin and other TGF beta-like molecules, Xnr-1 and Xnr-2 can dorsalize ventral marginal zone explants and induce muscle differentiation. Xnr signalling also rescues a complete embryonic axis in UV-ventralized embryos. The patterns of Xnr expression, the activities of the proteins and the phenotype of mouse nodal mutants, all argue strongly that a signaling pathway involving nodal, or nodal-related peptides, is an essential conserved element in mesoderm differentiation associated with vertebrate gastrulation and axial patterning.

Mesoderm induction is a critical early step in vertebrate development, involving changes in gene expression and morphogenesis. In Xenopus, normal mesoderm formation depends on signalling through the fibroblast growth factor (FGF) tyrosine kinase receptor. One important signalling pathway from receptor tyrosine kinases involves p21ras (ref. 5). Ras associates with the serine kinase c-Raf-1 in a GTP-dependent manner, and this complex phosphorylates and activates MAPK/ERK kinase (MEK), a protein kinase with dual specificity. MEK then activates p42mapk and (at least in mammals) p44mapk, members of the mitogen-activated protein (MAP) kinase family. FGF activates MAP kinase during mesoderm induction, and the use of dominant-negative constructs suggests that mesoderm induction by FGF requires both Ras and Raf. However, these experiments do not reveal whether Ras and Raf do act through MAP kinase to induce mesoderm or whether another pathway, such as the phosphatidylinositol 3-kinase cascade, is involved. Here we show that expression of active forms of MEK or of MAP kinase induces ventral mesoderm of the kind elicited by FGF. Overexpression of a Xenopus MAP kinase phosphatase blocks mesoderm induction by FGF, and causes characteristic defects in mesoderm formation in intact embryos, whereas inhibition of the P13 kinase and p70 S6 kinase pathways has no effect on mesoderm induction by FGF. FGF induces different types of mesoderm in a dose-dependent manner; strikingly, this is mimicked by expressing different levels of activated MEK. Together, these experiments demonstrate that activation of MAP kinases is necessary and sufficient for mesoderm formation.

An old idea about the relationship between arthropod and vertebrate body plans has been given new life by studies of the signalling genes controlling dorsal and ventral development in Drosophila and Xenopus.

Widespread expression of the DNA-binding protein Brachyury in Xenopus animal caps causes ectopic mesoderm formation. In this paper, we first show that two types of mesoderm are induced by different concentrations of Brachyury. Animal pole explants from embryos injected with low doses of Xbra RNA differentiate into vesicles containing mesothelial smooth muscle and mesenchyme. At higher concentrations somitic muscle is formed. The transition from smooth muscle formation to that of somitic muscle occurs over a two-fold increase in Brachyury concentration. Brachyury is required for differentiation of notochord in mouse and fish embryos, but even the highest concentrations of Brachyury do not induce this tissue in Xenopus animal caps. Co-expression of Brachyury with the secreted glycoprotein noggin does cause notochord formation, but it is difficult to understand the molecular basis of this phenomenon without knowing more about the noggin signal transduction pathway. To overcome this difficulty, we have now tested mesoderm-specific transcription factors for the ability to synergize with Brachyury. We find that co-expression of Pintallavis, but not goosecoid, with Brachyury causes formation of dorsal mesoderm, including notochord. Furthermore, the effect of Pintallavis, like that of Brachyury, is dose-dependent: a two-fold increase in Pintallavis RNA causes a transition from ventral mesoderm formation to that of muscle, and a further two-fold increase induces notochord and neural tissue. These results suggest that Pintallavis cooperates with Brachyury to pattern the mesoderm in Xenopus.

Members of the activin family are believed to act as mesoderm-inducing factors during early amphibian development. Little is known, however, about mesoderm formation in the mammalian embryo, and as one approach to investigating this we have studied activin and follistatin expression during early mouse development. Activins are homo- or heterodimers of the beta A or beta B subunits of inhibin, itself a heterodimer consisting of one of the beta subunits together with an alpha subunit. Follistatin is a single-chain polypeptide which inhibits activin function. Expression of the inhibin alpha chain could not be detected in embryonic or extraembryonic tissues at any of the stages studied (5.5 to 8.5 days) and expression of the beta A and beta B subunits could only be observed in the deciduum in cells surrounding the embryo. Expression of follistatin could also be detected in the deciduum, but in a pattern complementary to that of the beta subunits. Embryonic expression of follistatin first occurred in the primitive streak, and at later stages transcripts were detectable in the somites and in rhombomeres 2, 4 and 6 of the hindbrain. These results are consistent with a role for activin in mesoderm formation in the mouse embryo, and suggest functions for follistatin in addition to its role as an inhibitor of activin.

In Xenopus embryos, the extracellular matrix (ECM) protein tenascin (TN) is expressed dorsally in a very restricted pattern. We have studied the spatial and temporal expression of TN mRNA in tailbud-stage embryos by RNAase protection and in situ hybridization using a cDNA probe for Xenopus TN obtained by PCR amplification. We report that TN transcripts are principally expressed in cells dispersed around the neural tube and notochord as well as in myotome and sclerotome cells. No TN mRNA could be detected in lateral plate mesoderm, but expression was detectable beneath tail fin epidermis. In a second series of experiments, we studied the expression of TN mRNA and protein in combinations between animal and vegetal stage-6 blastomeres and in stage-8 blastula animal caps treated with activin A or basic fibroblastic growth factor (b-FGF). Isolated animal cap tissue cultured alone differentiates into epidermis, which expresses neither TN protein nor TN mRNA. TN expression is, however, elicited in response to isolated dorsal vegetal blastomeres and in response to high concentrations of activin, both of which treatments lead to formation of muscle and/or notochord. Low concentrations of activin, and ventral vegetal blastomeres, treatments that induce mesoderm of ventral character, are poor inducers of TN. However, b-FGF, which also induces ventral mesoderm, elicits strong expression. These results indicate that TN regionalization is a complex process, dependent both on the pattern of differentiation of mesodermal tissues and on the agent with which they are induced. The data further show that "ventral mesoderm" induced by low concentrations of activin is distinct from that induced by b-FGF, and imply that activin induces ventral mesoderm of the trunk while b-FGF induces posterior mesoderm of the tailbud.

The mesoderm of amphibian embryos such as Xenopus laevis arises through an inductive interaction in which cells of the vegetal hemisphere of the embryo act on overlying equatorial and animal pole cells. Three classes of 'mesoderm-inducing factor' (MIF) that might be responsible for this interaction in vivo have been discovered. These are members of the transforming growth factor type beta (TGF-beta), fibroblast growth factor (FGF) and Wnt families. Among the most potent MIFs are the activins, members of the TGF-beta family, but RNA for activin A and B is not detectable in the Xenopus embryo until neurula and late blastula stages, respectively, and this is probably too late for the molecules to act as natural inducers. In this paper, we use the polymerase chain reaction to clone additional members of the TGF-beta family that might possess mesoderm-inducing activity. We show that transcripts encoding Xenopus bone morphogenetic protein 4 (XBMP-4) are detectable in the unfertilized egg, and that injection of XBMP-4 RNA into the animal hemisphere of Xenopus eggs causes animal caps isolated from the resulting blastulae to express mesoderm-specific markers. Surprisingly, however, XBMP-4 preferentially induces ventral mesoderm, whereas the closely related activin induces axial tissues. Furthermore, the action of XBMP-4 is 'dominant' over that of activin. In this respect, XBMP-4 differs from basic FGF, another ventral inducer, where simultaneous treatment with FGF and activin results in activin-like responses. The dominance of XBMP-4 over activin may account for the ability of injected XBMP-4 RNA to 'ventralize' whole Xenopus embryos. It is interesting, however, that blastopore formation in such embryos can occur perfectly normally. This contrasts with embryos ventralized by UV-irradiation and suggests that XBMP-4-induced ventralization occurs after the onset of gastrulation.

One of the reasons that we know so little about the control of vertebrate gastrulation is that there are very few systems available in which the process can be studied in vitro. In this paper, we suggest that one suitable system might be provided by the use of mesoderm-inducing factors. In amphibian embryos such as Xenopus laevis, gastrulation is driven by cells of the mesoderm, and the mesoderm itself arises through an inductive interaction in which cells of the vegetal hemisphere of the embryo emit a signal which acts on overlying equatorial cells. Several factors have recently been discovered that modify the pattern of mesodermal differentiation or induce mesoderm from presumptive ectoderm. Some of these mesoderm-inducing factors will also elicit gastrulation movements, which provides a powerful model system for the study of gastrulation, because a population of cells that would not normally undertake the process can be induced to do so. In this paper, we use mesoderm-inducing factors to attempt to answer four questions. How do cells know when to gastrulate? How do cells know what kind of gastrulation movement to undertake? What is the cellular basis of gastrulation? What is the molecular basis of gastrulation?

This is an update of a previous review (Current Opinion in Cell Biology 2:969-974) in which we discussed recent work attempting to understand the sequence of inductive interactions responsible for establishing the body plan of the early embryo. As before, we concentrate on inductive interactions in amphibian embryos, where significant progress has been made in the past two years. In this update, however, we also consider recent embryological data obtained with amniote embryos such as the chick, together with complementary data provided by genetic analyses of mouse and Drosophila development.

Theories of morphogenesis have centred around gradients of morphogens interpreted by cells according to preset concentration thresholds. Growth factor proteins have been candidates for such morphogens, and experiments in the frog Xenopus, among other organisms, now show that embryonic cells are capable of interpreting growth factor concentrations according to the models. This opens up questions of whether such mechanisms are sufficient, and-if there are growth factor gradients-which factors are actually used and how.

The first inductive interaction in amphibian development is mesoderm induction, during which a signal from the vegetal hemisphere of the blastula-staged embryo induces mesoderm from overlying equatorial cells. Recently, a number of 'mesoderm-inducing factors' (MIFs), which may be responsible for this interaction, have been discovered. Examples of these MIFs include members of the fibroblast growth factor family as well as members of the TGF-beta superfamily such as TGF-beta 2. In addition to these purified factors, several new sources of mesoderm-inducing activity have been described. One of the most potent of these is the murine myelomonocytic leukemia cell line WEHI-3. Even at high dilutions, conditioned medium from WEHI-3 cells induces isolated Xenopus animal pole regions to form a variety of mesodermal cell types. In this paper we show by several criteria, including N-terminal amino acid sequencing, Northern blotting and various functional assays, that the WEHI-MIF is activin A. Activins are known to modulate the release of follicle-stimulating hormone from cultured anterior pituitary cells and to cause the differentiation of two erythroleukemia cell lines. Our results, along with recent data from other laboratories, indicate that these molecules may also act in early development in the formation of the mesoderm.

Exposure of isolated Xenopus animal pole ectoderm to the XTC mesoderm-inducing factor (XTC-MIF) causes the tissue to undergo gastrulation-like movements. In this paper, we take advantage of this observation to investigate the control of various aspects of gastrulation in Xenopus. Blastomeres derived from induced animal pole regions are able, like marginal zone cells, but unlike control animal pole blastomeres, to spread and migrate on a fibronectin-coated surface. Dispersed animal pole cells are also able to respond to XTC-MIF in this way; this is one of the few mesoderm-specific responses to induction that has been observed in single cells. The ability of induced animal pole cells to spread on fibronectin is abolished by the peptide GRGDSP. However, the elongation of intact explants is unaffected by this peptide. This may indicate that fibronectin-mediated cell migration is not required for convergent extension. We have investigated the molecular basis of XTC-MIF-induced gastrulation-like movements by measuring rates of synthesis of fibronectin and of the integrin beta 1 chain in induced and control explants. No significant differences were observed, and this suggests that gastrulation is not initiated simply by control of synthesis of these molecules. In future work, we intend to investigate synthesis of other integrin subunits and to examine possible post-translational modifications to fibronectin and the integrins.

The mesoderm of Xenopus laevis and other amphibia is formed through an inductive interaction during which cells of the vegetal hemisphere act on cells of the animal hemisphere. Two groups of factors mimic the effects of the vegetal hemisphere. One group consists of members of the fibroblast growth factor (FGF) family, while the other is related to transforming growth factor type beta (TGF-beta). In this paper we discuss the evidence that the FGF family represents 'ventral' mesoderm-inducing signals, and the TGF-beta family 'dorsal' signals. The evidence includes a discussion of the cell types formed in response to each type of factor, the fact that only XTC-MIF (a member of the TGF-beta family) and not bFGF can induce animal pole ectoderm to become Spemann's organizer, and an analysis of the timing of the gastrulation movements induced by the factors.

Inductive interactions play a major role in the diversification of cell types during vertebrate development. These interactions have been extensively studied in amphibian embryos (usually Xenopus laevis) where the earliest is mesoderm induction, in which an equatorial mesodermal rudiment is induced from the animal hemisphere under the influence of signal from the vegetal hemisphere. The molecular basis of mesoderm induction is unknown, although Tiedemann has isolated a protein form 9- to 13-day chick embryos that has the properties one would expect of a mesoderm-inducing factor. However, the relevance of this molecule to the events of early amphibian development is unclear, and it is a matter of some importance to discover a Xenopus mesoderm-inducing factor. In this paper I show that the Xenopus XTC cell line secretes mesoderm-inducing activity into the culture medium. Isolated animal pole regions cultured in XTC-conditioned medium differentiate into muscle and notochord, while controls form 'atypical epidermis'. Three different cell lines -XL, XL177 and KR- secrete no such activity indicates that the active principle is heat stable, trypsin sensitive, nondialysable, and has an apparent relative molecular mass of about 16,000. Work is in progress to characterize the activity further and to discover whether the mesoderm-inducing factor is also present in normal embryos.

This paper reviews work with cell lineage labels and cell-type specific markers in the analysis of inductive interactions in early amphibian development. Our results provide clear evidence for the existence of three such interactions. Mesodermal induction occurs in the early blastula and results from the action of vegetal pole cells on the animal hemisphere. At least two mesodermal rudiments are formed, one dorsal and one ventral. During the next interaction, which we call dorsalization, the ventral mesodermal rudiment becomes subdivided into several territories under the influence of the dorsal marginal zone, or organizer. Finally, during gastrulation, the involuting organizer induces neural tissue from the overlying ectoderm. This interaction is called neural induction. Although these phenomena can readily be demonstrated under experimental conditions, direct evidence that they occur in normal development awaits an understanding of the molecular basis of induction.

We have studied the action of the organizer in Xenopus laevis using grafts labelled with horseradish peroxidase (HRP). Orthotropic grafts of the dorsal marginal zone (the organizer) from an HRP-labelled embryo into an unlabelled host showed that this region contributes to the anterior archenteron wall, to the entire craniocaudal extent of the notochord and to a few cells in the somites. Little or no contribution was made to the neural tube. Orthotopic grafts of the ventral marginal zone (the tissue that responds to a grafted organizer) indicated that it only contributes to the posterior half of the embryo. Within this region it spreads around the entire ventrolateral mesoderm, occasionally contributing a few cells to the somites. The posterior endoderm was also heavily labelled. When the dorsal marginal zone from an HRP-labelled embryo was inserted into a slit cut in the ventral marginal zone of an unlabelled host a mirror-symmetrical double-dorsal duplicated embryo resulted, in which only the notochord and a few cells in the somites of the secondary embryo were derived from the graft. The bulk of the secondary somites was, therefore, derived from host ventral marginal zone tissue which normally makes very little contribution to the somites. This indicates that host ventral marginal zone becomes dorsalized by the graft. The neural tube of the secondary embryo was also unlabelled, showing that it was induced by the influence of the graft on the overlying ectoderm, which normally forms ventral epidermis. We have also grafted ventral marginal zone tissue into a slit cut into the dorsal marginal zone of a host embryo. HRP-labelled tissue was grafted into an unlabelled embryo and vice versa. This graft did not produce a double ventral embryo and this reinforces the traditional view that the dorsal marginal zone is a special signalling region. Instead, the resulting embryos usually had a twinned notochord with the graft tissue in between, differentiated as somite. This confirms that juxtaposing ventral and dorsal marginal zone 'dorsalizes' the ventral tissue but does not affect the dorsal tissue which differentiates, as usual, as notochord. Thus, our results allow us to conclude that the organizer mediates two distinct interactions in bringing about the formation of duplicated embryos. The first is dorsalization of adjacent ventral mesoderm and the second is the induction of neuroepithelium from ectoderm overlying the new archenteron roof.

In this paper, we study the mechanism by which a Xenopus cell line-derived mesoderm-inducing factor (MIF) might establish the spatial pattern of cellular differentiation in the mesoderm. The effects of the factor on competent animal pole tissue are consistent with it being identical to the natural mesoderm-inducing factor. The signal can only act on those membrane domains of the animal pole that face the blastocoel, but it can be stably recorded there, such that axial mesoderm is formed, after 15 min exposure or less. This exposure can end some hours, or several cell cycles, before the onset of RNA synthesis yet nevertheless be fully effective, although competence to respond also extends well after the onset of transcription. Exposure of the entire blastocoel lining of intact embryos to MIF causes a synchronous and sudden transformation of the behaviour and adhesive properties of all inner animal cap cells. This transformation mimics and is contemporaneous with the involution behaviour of normal mesoderm in the early gastrula marginal zone. Although high concentrations of MIF totally disorganize gastrulation, lower concentrations permit gastrulation to proceed. However, the pattern of mesoderm in these embryos is disrupted and ectopic mesoderm is formed around the blastocoel remnant. When MIF is injected directly into blastomeres, rather than into the blastocoel, it has no effect. This suggests that the molecule is secreted from source cells and affects target cells through an extracellular receptor. Finally, we show that small pieces of animal pole tissue recently exposed to MIF go on to produce morphogenetic signals perhaps distinct from MIF. We discuss the role of these signals in establishing and modifying the spatial pattern of cellular differentiation in the mesoderm of Xenopus.

We have compared the development of the animal pole (AP) region of early Xenopus embryos in normal development, in isolation, and in combination with explants of tissue from the vegetal pole (VP) region. For the grafts and the combinations the animal pole tissue was lineage labelled with FLDx in order to ascertain the provenance of the structures formed. The normal fate of the AP region was determined by orthotopic grafts at stages 7 1/2 (early blastula), 8 (mid blastula) and 10 (early gastrula). At later stages most of the labelled cells were found in ectodermal tissues such as epidermis, head mesenchyme and neural tube (the last from stages 7 1/2 and 8 only). However, in stage-7 1/2 and stage-8 grafts some of the labelled cells were also found in the myotomes and lateral mesoderm. In isolated explants the AP region of all three stages differentiated only as epidermis assessed both histologically and by immunofluorescence using an antibody to epidermal keratin. The fate of labelled cells in AP-VP combinations was quite different and confirms the reality of mesoderm induction. In combinations made at stages 7 1/2 and 8 the proportion of AP-derived mesoderm is substantially greater than the proportion of labelled mesoderm in the equivalent fate mapping experiments. This shows that the formation of mesoderm in such combinations is the result of an instructive rather than a permissive interaction. The formation of mesodermal tissues in stage-7 1/2 combinations was confirmed by using a panel of antibodies which react with particular tissues in normal tailbud-stage embryos: anti-keratan sulphate for the notochord, anti-myosin for the muscle and anti-keratin for epidermis and notochord. Combinations made at stage 10 gave no positive cases and reciprocal heterochronic combinations between stages 7 1/2 and 10 showed that this is the result of a loss of competence by the stage-10 AP tissue. Whereas stage-7 1/2 AP tissue combined with stage-10 VP tissue gave many positive cases, the reciprocal experiment gave only a few. We have also tested the regional specificity of the induction. Stage-7 1/2 vegetal pole explants were divided into dorsal and ventral regions and then combined, separately, with stage-7 1/2 animal poles. The dorsovegetal tissue induces 'dorsal-type' mesoderm (notochord and large muscle masses) while ventrovegetal tissue induces 'ventral-type' mesoderm (blood, mesothelium and a little muscle). We conclude that mesoderm formation in combinations is an instructive event and propose a double gradient model to explain the complex character of the response.