We have used blastocoel and vegetal pole grafts to investigate the effect of environment on differentiation and movement of animal pole cells of Xenopus. In the blastocoel of embryos earlier than stage 10, fragments of animal pole primarily form mesoderm. The cells are either integrated into normal host tissues or they organize a secondary posterior dorsal axis. If either host or graft is later than stage 9 the graft forms ectoderm and its cells all migrate into the host ectoderm. Inner layer animal cells form sensorial layer; outer cells move to the epidermis. Thus considerable powers of appropriate movement are seen. In the vegetal pole no movement occurs. If the graft is stage 9 or earlier, or the host is stage 10| or earlier, the graft forms mesoderm, including striated muscle in the gut. This shows that muscle can develop in wholly the wrong environment, it suggests that the dorsal inductive signal from mesoderm is rather general in the vegetal mass and suggests that dorsal mesoderm development involves little subsequent adjustability. If the host is stage 11 or later, or the graft later than stage 9, the graft forms epidermis in the gut. This shows that the epidermal pathway of development is also insensitive to environment.

In this paper, we describe experiments using monoclonal antibodies to epidermis and muscle-specific epitopes to investigate the migration, development and subsequent differentiation of animal cap cells of Xenopus embryos when they are transplanted into unusual positions in the embryo.

In Xenopus, the ectoderm is primarily derived from the pigmented half of the embryo (Keller, 1975; Cooke & Webber, 1985; Dale & Slack, 1987) though even vegetal pole cells of the 32-cell embryo give rise to a little ectoderm at high frequency (Heasman, Wylie, Hausen & Smith, 1984). The ectoderm eventually produces two main components, epidermis and nervous system, a process involving a number of steps of commitment, first to ectoderm rather than mesoderm and subsequently to either epidermis or nervous system. Recent fate mapping shows that the animal cap region also forms much of the mesoderm (Cooke & Webber, 1985; Dale & Slack, 1987) though isolated animal hemispheres only form epidermis (Holtfreter & Hamburger, 1955; Asashima & Grunz, 1983; Slack, 1984; Jones & Woodland, 1986). It is believed that this mesoderm is formed by the inducing action of cells in the presumptive endoderm on competent ectoderm, the latter being reported to be able to respond to this induction up to gastrulation (Dale, Smith & Slack, 1985). In experimental tissue combinations, at least, the presumptive ectoderm may also form pharyngeal endoderm (Sudarwati & Nieuwkoop, 1971). Thus, in atypical sites, presumptive epidermis might be expected to form mesoderm and anterior gut, in addition to the epidermis or nervous system that it normally forms.

When single cells are placed in the blastocoel of a host embryo their descendants appear in a variety of tissues and the cells concerned apparently conform to the differentiated state of their surroundings (Wylie, Smith, Snape & Heasman, 1985; Wylie, Snape, Heasman & Smith, 1987; Snape, Wylie, Smith & Heasman, 1987). Do they differentiate in accordance with their surroundings or do they settle on a differentiation pathway and then move to the appropriate site? Indeed, how much are migratory abilities of cells responsible for maintaining and achieving the three germ layer structure of the embryo? In this paper, we show that ectodermal cells have considerable ability to migrate to their appropriate location in an embryo, but that this location is not necessary for them to form epidermis. Similarly, muscle can develop in completely unusual surroundings, although mesoderm cells probably also have migratory abilities around the general blastocoel region. The picture that emerges is first that the structure of the embryo is probably maintained by sophisticated migratory abilities in its constituent cells. Second, it seems that once certain major choices in differentiation pathways are made, cells differentiate autonomously.

We have recently isolated two monoclonal antibodies that react specifically with the epidermis and striated muscle of the amphibian embryo. The epidermal marker reacts with all of the surface epidermal cells of the neurula, even though these cells may be as different as the cement gland and ciliated cells (Jones, 1985; Jones & Woodland, 1986). This antigen, which first appears in the stage- late gastrula, is a major secreted molecule with a protein component, present in all superficial cells of the early neurula, except the future central nervous system. It can be used as a marker of the appearance of the epidermal phenotype, even when cells do not gain the morphological characteristics of epidermis. For example, the marker subsequently appears when cell division is blocked at the mid-blastula stage, even though the cells become multicellular and disorganized (Jones & Woodland, 1986). The muscle-specific marker is a monoclonal antibody (5A3.B4) raised by immunizing Balb/c mice with a homogenate of adult Xenopus muscle. It stains striated muscle from stage 20 onwards and reacts with no other tissue type (Fig. 1). We also used a further muscle-specific antibody (Kinter & Brockes, 1985). All the antibodies used in this study stain X. laevis and X. borealis in an identical way.

Fig. 1.

Characterization of cellular and differentiation markers on cryostat sections. (A) Grafted Xenopus borealis cells (arrowed) in the somites and nervous system of a host Xenopus laevis embryo showing the characteristic punctate nuclear pattern when stained with quinacrine. Host cells fluoresce diffusely.

(B) Cryostat section through a stage-19 Xenopus laevis embryo stained with the epidermal marker 2F7.C7. Only the outer layer of the epidermis is stained.

(C) Longitudinal cryostat section through a stage-46 Xenopus laevis tadpole tail stained with the muscle marker 5A3.B4. The somites only are stained, showing a characteristic striated pattern. Abbreviations: a, archenteron; ep, epidermal ectoderm; g, graft; ns, nervous system; nt, notochord; s, somitic muscle. Bar: (A,C) 75 μm; (B) 150 μm.

Fig. 1.

Characterization of cellular and differentiation markers on cryostat sections. (A) Grafted Xenopus borealis cells (arrowed) in the somites and nervous system of a host Xenopus laevis embryo showing the characteristic punctate nuclear pattern when stained with quinacrine. Host cells fluoresce diffusely.

(B) Cryostat section through a stage-19 Xenopus laevis embryo stained with the epidermal marker 2F7.C7. Only the outer layer of the epidermis is stained.

(C) Longitudinal cryostat section through a stage-46 Xenopus laevis tadpole tail stained with the muscle marker 5A3.B4. The somites only are stained, showing a characteristic striated pattern. Abbreviations: a, archenteron; ep, epidermal ectoderm; g, graft; ns, nervous system; nt, notochord; s, somitic muscle. Bar: (A,C) 75 μm; (B) 150 μm.

Embryo culture, manipulations and histology

Embryos were cultured and explants made as described by Jones & Woodland (1986). Ectodermal sandwich experiments were made with ectodermal explants from X. borealis sandwiched between complete animal caps derived from two X. laevis stage-9 blastula and incubated. They were fixed when embryos synchronous with the implant had reached stage 19. X. borealis cells were recognized by staining with quinacrine. They exhibit intensely fluorescent chromatin granules, which are absent from X. laevis (Fig. 1A; Thiébaud, 1983).

Blastocoel-grafted embryos were made by inserting rhodamine-labelled or X. borealis donor ectoderm into a small slit at the animal pole of demembranated embryos. Pieces of ectoderm were approximately one eighth of an animal cap in size. They were cultured in MBS [88mM-NaCl; ImM-KCl; 24mM-NaHCO3; 15mM-Tris-HCl; 0·33mM-Ca(NO3)2; lmM-MgSO4; lmM-NaHCO3; 2mM-sodium phosphate pH 7·4; and 0·1 mM-Na2EDTA (Gurdon, 1977)] to heal and then transferred into 1/10 MBS to gastrulate. Vegetal pole grafts were achieved by grafting similar explants into gaps teased between vegetal pole cells or into the holes left after removing whole vegetal pole blastomeres. All grafted embryos were healed in MBS. They were either maintained in this medium to produce exogas-trulae or transferred to 1/10 MBS to gastrulate normally.

Fixation, embedding, sectioning and staining with antibodies or simpler chemicals were as described by Jones & Woodland (1986). Fig. 1B,C shows the normal staining pattern of the epidermal and muscle-specific antibodies on stage-46 X. laevis embryos.

Migration and differentiation of epidermal cells in ectodermal sandwiches

In a normal embryo, the cells that form epidermis, the outermost layer of which binds 2F7.C7, bound the embryo. The same is very largely true when a morula or blastula explant of animal cap cells is cultured in saline, although in this case there is also a scattering of somewhat more lightly stained cells within the solid ball of ‘atypical epidermis’ which forms (Jones & Woodland, 1986). Can highly pigmented ectodermal cells differentiate into the strongly positive phenotype in an internal position or is it essential that they migrate to the cell surface before differentiating in this way?

To find if this was so, we made a sandwich of two animal caps from stage-9 blastulae and placed a smaller piece of animal cap from embryos of various stages in the centre (Fig. 2A). Implants were taken from embryos between stage 3 and stage 9 (8-cell to late blastula) and placed in stage-9 tissue. In every case, greater than 20 in total, a high proportion of the implant, and all of its heavily pigmented cells, bound 2F7.C7 strongly. However, the great majority did so without moving to the surface (Fig. 2B–D). An outside position is thus not necessary to form epidermis and migration to the surface of the explant does not occur. However, the overall environment of the explants is still ectodermal.

Fig. 2.

Epidermal differentiation in ectodermal sandwiches. Ectodermal fragments were dissected from three different stage-9 embryos and the sandwich constructed as in A. The internal fragment was taken from X. borealis and the outer caps from X. laevis. Sandwiches were incubated until control embryos were stage 19, fixed and processed, as described in the methods. Sections were stained with 2F7.C7 and rhodamine-conjugated FITC RAM IgG as the second step antibody. Grafted X. borealis cells were identified by quinacrine staining. (B) Cryostat section through a representative graft containing region of a sandwich made from stage-9 embryos, illuminated to show antibody binding. The arrow indicates grafted 2F7.C7-positive cells. The arrowed region is enlarged to show antibody binding (C) and the punctate quinacrine staining diagnostic of graft-derived X. borealis cells. Abbreviations as in Fig. 1. Bar (B) 180 μm; (C,D) 36 μm.

Fig. 2.

Epidermal differentiation in ectodermal sandwiches. Ectodermal fragments were dissected from three different stage-9 embryos and the sandwich constructed as in A. The internal fragment was taken from X. borealis and the outer caps from X. laevis. Sandwiches were incubated until control embryos were stage 19, fixed and processed, as described in the methods. Sections were stained with 2F7.C7 and rhodamine-conjugated FITC RAM IgG as the second step antibody. Grafted X. borealis cells were identified by quinacrine staining. (B) Cryostat section through a representative graft containing region of a sandwich made from stage-9 embryos, illuminated to show antibody binding. The arrow indicates grafted 2F7.C7-positive cells. The arrowed region is enlarged to show antibody binding (C) and the punctate quinacrine staining diagnostic of graft-derived X. borealis cells. Abbreviations as in Fig. 1. Bar (B) 180 μm; (C,D) 36 μm.

Migration of ectodermal cells in blastocoel grafts

Since ectodermal cells do not migrate in the wholly ectodermal environment of an ectodermal sandwich, we have tested the ability of ectodermal cells to migrate after grafting into different regions of the whole embryo. Initially, ectoderm was grafted into the blastocoel. Classically this operation was used as a test of the ability of the dorsal mesoderm to induce a secondary CNS, that is as a modification of the original Spemann and Mangold graft (Spemann & Mangold, 1924). As pointed out by Slack (1983), this kind of experiment introduces the graft into variable situations, with complex results, at least in terms of the overall tissue organization of the embryo. However, we have grafted animal cap and not dorsal mesoderm, and primarily ask three very simple questions: do the grafted cells migrate and. if so, what kind of tissue do the grafted cells enter and what differentiated phenotype do they display? Similar approaches have been used in Xenopus with single ectodermal and endodermal cells (Heasman et al. 1984; Wylie et al. 1985) and in axolotls with multiple disaggregated cells from a number of germ layers (Boucaut, 1974a,b). In mammals, the analogous technique is to inject cells into the blastocyst (Gardner, 1985).

When ectoderm from pregastrula embryos is grafted into the blastocoel of similar, but not necessarily identical, stages of embryos, secondary embryonic axes were often formed. Embryos with both normal and secondary embryonic axes were serially sectioned and the location of grafted cells and their differentiated phenotype recorded.

Tissue distribution of grafted cells

When ectoderm, either from X. borealis or rhodamine-labelled X. laevis donor embryos, was taken prior to stage 10 and grafted into the blastocoel of X. laevis host embryos earlier than stage 10, it was mainly found in the mesoderm of the host, although some cells were also found in the ectoderm (Fig. 3A). No cells remained in the blastocoel. The grafted cells were usually integrated into all parts of the somitic and lateral plate mesoderm, though none were found in the notochord. No grafted cells were ever seen in the endoderm, though experiments of Heasman et al. (1984), using single cells, show that ectodermal cells from these stages can be found in this location.

Fig. 3.

Tissue distribution of ectodermal cells grafted into the blastocoel. Small pieces of Xborealis ectoderm were inserted into the blastocoel of X. laevis host embryos and the grafted embryos allowed to develop to stage 25–28. The embryos were then fixed and processed as described in the Methods and stained with the muscle marker, 5A3.B4, or the epidermal marker 2F7.C7, and these antibodies were revealed with rhodamine RAM IgG and counterstained with quinacrine to identify the grafted X. borealis cells. Cells from stage-9 X. borealis grafted into stage-9 X. laevis hosts (A) were found mainly in the mesoderm, in this case the somites (arrowed), and stained strongly with the muscle marker. When ectoderm from stage-9 X. borealis was grafted into stage-10 hosts (B) grafted cells were found exclusively in the ectoderm and stained with 2F7.C7, when in the outer epidermal layer. Occasionally double epidermal folds or blisters were formed from this latter graft (C) which also showed the normal epidermal staining with 2F7.C7. D–F show quinacrine staining of the same regions, identifying X. borealis-áemeá grafted cells. Abbreviations as in Fig. 1. Bar: (A,B,D) 75 μm; (C,E,F) 150 μm.

Fig. 3.

Tissue distribution of ectodermal cells grafted into the blastocoel. Small pieces of Xborealis ectoderm were inserted into the blastocoel of X. laevis host embryos and the grafted embryos allowed to develop to stage 25–28. The embryos were then fixed and processed as described in the Methods and stained with the muscle marker, 5A3.B4, or the epidermal marker 2F7.C7, and these antibodies were revealed with rhodamine RAM IgG and counterstained with quinacrine to identify the grafted X. borealis cells. Cells from stage-9 X. borealis grafted into stage-9 X. laevis hosts (A) were found mainly in the mesoderm, in this case the somites (arrowed), and stained strongly with the muscle marker. When ectoderm from stage-9 X. borealis was grafted into stage-10 hosts (B) grafted cells were found exclusively in the ectoderm and stained with 2F7.C7, when in the outer epidermal layer. Occasionally double epidermal folds or blisters were formed from this latter graft (C) which also showed the normal epidermal staining with 2F7.C7. D–F show quinacrine staining of the same regions, identifying X. borealis-áemeá grafted cells. Abbreviations as in Fig. 1. Bar: (A,B,D) 75 μm; (C,E,F) 150 μm.

Grafted cells had the morphology typical of their surroundings, which suggested that they had adopted the appropriate differentiated phenotype. They were tested with respect to two cell types, striated muscle and epidermis. As in control embryos, 2F7.C7 bound only to the outer ectodermal layer of the embryo, and this included some of the grafted cells. We identified muscle cells using the two monoclonal reagents described in the methods. Grafted cells found in the myotomes reacted appropriately with these antibodies, even in cases where the embryos were quite abnormal, proving that they had become differentiated as muscle (Fig. 3A).

The tissue distribution in double-axis embryos was the same as in morphologically normal embryos, except that in single-axis embryos grafted cells were more frequently seen in lateral plate mesoderm or epidermis. In double-axis embryos, grafted cells were found more exclusively in the dorsal mesoderm.

In this kind of experiment, the competence of the animal cap cells to form mesoderm was lost at about stage 10 and . The capacity of the host to induce mesoderm formation also disappeared by stage 10 (Table 1). Thus, one can test the migratory capacity of cells that will later form ectoderm either by grafting stage-10 animal cap cells into a blastocoel of hosts at any stage, or by grafting pre-stage-10 cells into hosts at stage 10 or later.

Table 1.

Final destination of ectodermal cells grafted into the blastocoels of host embryos

Final destination of ectodermal cells grafted into the blastocoels of host embryos
Final destination of ectodermal cells grafted into the blastocoels of host embryos

When the implanted donor ectoderm was derived from embryos at stage 10 or later, the final positions of the grafted cells were quite different. They all moved to the surface of the embryo and became incorporated into the ectoderm, where they formed either epidermis, including both the outer 2F7.C7-positive epidermal layer and the negative, inner, sensorial layer, or else they formed nervous system (Fig. 3B). Often these grafts formed a blister, fold or pouch in the skin, usually ventrally, possibly simply because there were larger numbers of epidermal cells than usual in this region (Fig. 3C). However, the morphology of these regions was always typical of larval epidermis.

These experiments suggest two conclusions. First, migration to the correct site (the outside of the embryo) is part of the ectodermal phenotype. Since the cells that form mesodermal tissues also become incorporated into normal/mesodermal structures, this conclusion also seems to be true of mesoderm, although the very frequent occurrence of secondary mesodermal axes suggests that this ability is limited. Second, animal cap cells are determined to form ectoderm by the early gastrula stage and have lost the ability to form mesoderm. However, this conclusion applies only to tissue fragment grafts into the blastocoel.

We also tested the migration of the inner and outer ectodermal layers from donor embryos in blastocoel grafts. During normal development of Xenopus, the outer layer in non-neural regions predominantly forms epidermis, the inner layer forming the so-called ‘sensorial layer’. In neural regions, both contribute to the developing nervous system. When grafted separately into stage-9 hosts, both layers from stage-9 embryos were incorporated into mesoderm, epidermis and CNS. Both layers from stage-10 embryos entered only epidermis and CNS, predominantly returning to their original inner or outer locations in the host embryos.

Migration is absent in cells grafted into the vegetal pole, but epidermal and muscle differentiation is unaffected

Perhaps the most unusual position for ectoderm to be grafted is into the vegetal pole region of host embryos. This region normally forms the internal border of the gut. Can ectodermal cells migrate from such a position and do they differentiate? The results of these experiments are summarized in Table 2A,B. The grafted embryos were stained sequentially with the epidermal marker and then the muscle marker, since one of the likely outcomes of such a graft might be its induction to form mesoderm. Table 2A shows the summary of results of grafting donor ectoderm from stage 6- to-12 embryos into the vegetal poles of hosts at stage 9 or 10. Grafted embryos gastrulated normally resulting in the graft being internalized into the gut region. The grafts were either found as coherent tissue masses completely surrounded by endoderm, or in regions bordering the lumen of the developing gut. No cells migrated away. All grafts from stage 6- to-9 embryos were found to express the muscle marker, indicating that the ectoderm had been induced to form muscle (Fig. 4). None of these grafts expressed the epidermal marker, though some of the grafted cells were negative with both antibodies. The differentiation of grafted cells into notochord, assessed purely by morphological criteria, was not detected. In all grafts from stage 10 or later, the graft always expressed the epidermal marker (Fig. 5) and did not express any muscle-specific determinants. These results show that epidermal differentiation can take place in an environment as unusual as the centre of the gut. They also show that the extreme vegetal pole of an embryo is capable of inducing dorsal mesoderm in competent ectoderm and that stage-10 ectoderm is no longer competent to respond to this signal under the conditions of this graft.

Table 2.

The differentiation of ectoderm in vegetal pole grafts

The differentiation of ectoderm in vegetal pole grafts
The differentiation of ectoderm in vegetal pole grafts
Fig. 4.

Determination of the ectoderm tested by grafting ectoderm into the vegetal pole of host blastulae; epidermal differentiation. Small pieces of donor X. laevis ectoderm were inserted into the vegetal pole region of host X. borealis embryos and allowed to develop until stage 25. Grafted embryos were fixed, embedded and stained as described before with 2F7.C7, rhodamine RAM IgG and quinacrine. A and B show the expression of the epidermal marker on cryostat sections of grafted cells on the archenteron wall following grafting from stage-9 and -10 donors, respectively, into stage-9 hosts. C and D show quinacrine staining of the same sections, identifying the graft, in this case, by the lack of X. borealis-specific granules in the nuclei. Abbreviations as in Fig. 1. Bar: (A,B) 180 μm; (C,D) 75 μm.

Fig. 4.

Determination of the ectoderm tested by grafting ectoderm into the vegetal pole of host blastulae; epidermal differentiation. Small pieces of donor X. laevis ectoderm were inserted into the vegetal pole region of host X. borealis embryos and allowed to develop until stage 25. Grafted embryos were fixed, embedded and stained as described before with 2F7.C7, rhodamine RAM IgG and quinacrine. A and B show the expression of the epidermal marker on cryostat sections of grafted cells on the archenteron wall following grafting from stage-9 and -10 donors, respectively, into stage-9 hosts. C and D show quinacrine staining of the same sections, identifying the graft, in this case, by the lack of X. borealis-specific granules in the nuclei. Abbreviations as in Fig. 1. Bar: (A,B) 180 μm; (C,D) 75 μm.

Fig. 5.

Determination of the ectoderm tested by grafting into the vegetal pole of host blastulae. A graft from a stage-9 X. borealis animal cap was made into the vegetal pole of a stage-9 X. laevis host embryo. The grafted embryo was incubated until control embryos reached stage 30, fixed, embedded and stained with the muscle marker 5A3.B4. Characteristic striated muscle is seen in most of the graft. Abbreviations: e, endoderm; ism, induced somitic muscle.

Fig. 5.

Determination of the ectoderm tested by grafting into the vegetal pole of host blastulae. A graft from a stage-9 X. borealis animal cap was made into the vegetal pole of a stage-9 X. laevis host embryo. The grafted embryo was incubated until control embryos reached stage 30, fixed, embedded and stained with the muscle marker 5A3.B4. Characteristic striated muscle is seen in most of the graft. Abbreviations: e, endoderm; ism, induced somitic muscle.

Table 2B shows the results of a similar series of experiments in which the stage of competent donor ectoderm was kept relatively constant, but the graft was made into host embryos varying from stage 7 to stage 11. All grafts carried out into host embryos from stages 7 to expressed the muscle marker and did not express the epidermal antigen. Those grafts carried out into stage-11 host embryos developed into balls of epidermis suggesting that the inductive stimulus is no longer present in the vegetal pole of stage-11 embryos. This shows that epidermal development, as defined by 2F7.C7 binding, can proceed in a wholly inappropriate environment.

These experiments show that ectoderm grafted into the vegetal pole region of host embryos cannot migrate from this position, but differentiates within the endoderm. The final differentiated cell type depends on the stage of development of both host and donor tissue. If the ectoderm is still responsive to mesodermal induction and the endoderm still capable of producing the signal, then the graft differentiates as mesoderm and often quite a large proportion of the graft can be identified as apparently normal striated muscle. If either of these conditions is not fulfilled then the grafted cells differentiate into epidermis.

Migratory abilities of embryonic ectodermal cells

By placing the cells of the future ectoderm into the blastocoel of another embryo, we have been able to test their migratory behaviour in later development. The results are clearest where the timing is arranged so that the animal cap forms only ectoderm, rather than mesoderm. This is achieved by grafting animal cap cells of any stages into an embryo at stage 10 or later, when mesodermal induction does not ensue (Table 1). Alternatively, unresponsive ectoderm from stage 10 or later may be placed in a blastocoel at any stage (Table 1). By the tailbud tadpole stage, the grafted cells are to be found in the epidermis, or to a lesser extent the nervous system, of the host. Moreover, transplanted cells from the inner or outer layers of the animal cap are predominantly found, respectively, in the inner sensorial layer or in the outer, epidermal layer, just as they are in normal development. The appearance of a minority of inner cells in the epidermal layer is consistent with the view that the scattered ciliated cells of the epidermis originate in the inner layer (Billet & Courtenay, 1973; Steinman, 1968; our unpublished observations).

Thus ectodermal cells do find their way to their correct location in the embryo. Indeed, after grafting ectoderm into the blastocoel we never see cells expressing the epidermal marker except in the epidermis. Since epidermis can differentiate in the endoderm (see below), it seems that cells following this epidermal pathway find their correct location from the blastocoel. In addition, we do not see negative cells inside the sensorial layer.

These results contrast with the failure of the grafted animal cells to move when surrounded by animal tissue as in a Holtfreter sandwich. In contrast, an explant that contains mesoderm shows proper organization of the ectoderm (data not shown). This suggests that the normal inner components of the embryo provide something necessary for the migration of the cells. This could be extracellular matrix, positional cues as to the location of the graft, or a disaggregating environment, or a combination of these factors.

What is the role of this migration in normal development? The progeny of lineage-labelled animal cells show very considerable scattering after gastrulation (Moody, 1987; Dale & Slack, 1987). This shows that the cells are naturally very mobile within their germ layer. Our results suggest that the integrity of the layer is also actively maintained, to the extent that a cell which becomes displaced as far as the blastocoel can still regain its appropriate location. Somewhat similar migratory abilities of the ectoderm were demonstrated by Boucaut in Pleurodeles waltl (Boucaut, 1974u,b). They are also an intrinsic part of the single-cell transfers of late-stage animal cap cells of Heasman et al. (1984), although in these experiments the differentiated state of the cells was not always tested with cell-type markers.

Migration of mesodermal cells

Boucaut (1974b) came to the conclusion that disaggregated mesodermal cells when placed into the blastocoel of a recipient embryo had considerable ability to organize themselves correctly within the mesoderm when injected into the blastocoel. Our experiments would suggest that the same is true in Xenopus. When fragments of animal caps are introduced into blastocoels under timing regimes where they can and are induced to form mesoderm, we find that induced grafted cells are appropriately organized into mesodermal tissues, although their fully differentiated state can only be positively identified when the grafts are incorporated into somites and the musclespecific monoclonal 5A3.B4 can be used. This situation mainly occurs in embryos displaying secondary embryonic axes (61 % of grafts in inductive combinations) when grafted cells are much more strongly represented in dorsal mesoderm than in normal grafted embryos when the majority of grafted cells are in lateral plate and ventral mesoderm. A possible interpretation of grafted embryos with secondary embryonic axes might lie in a reduced migratory potential of dorsal mesoderm. If this were so, dorsal mesoderm formed would not move to the primary dorsal region, but instead subverts gastrulation movements and organizes a second embryonic axis from surrounding tissue. In contrast, ectodermal cells are both relatively inert at blastula and gastrula stages and more mobile. They might, consequently, have longer in which to reach their appropriate positions before they would upset development. However, since primary and secondary axes are properly organized, and since mesoderm can differentiate in an unusual environment (see below), mesodermal cells can certainly organize themselves in the short range.

The role of the environment in epidermal and mesodermal differentiation

When animal cap cells are placed in the vegetal pole under circumstances where they are unresponsive to mesodermal induction (post stage )) or the host has caused mesodermal induction (post stage ), they invariably form epidermis in the walls of the gut or within its tissue. This indicates first that neural inductive stimuli do not occur here and, second, that once either the inductive stimulus or the competence to respond to mesodermal induction is lost, development into epidermis proceeds in a way that is not at all upset by the bizarre environment.

When mesodermal induction can occur, striated muscle is always formed, even though this is not normally found in the gut. Moreover, in normal embryos, this cell interaction occurs with vegetal cells in an entirely different location, that is at the dorsal margin between vegetal and animal cells. Our results indicate that dorsal inductive stimuli are present generally through the vegetal mass, even at the extreme vegetal pole, and that once the stimulus to form muscle has occurred, the cells differentiate without reference to their environment. This correlates with the fact that blastula cells can differentiate into muscle when disaggregated (Gurdon, Brennan, Fairmans & Mohun, 1984; Sargent, Jamrich & Dawid, 1986). It also fits with the observation that dorsal mesoderm can change the fate of more vegetal regions, but is not itself influenced (Slack & Forman, 1980). All of these observations support the idea that a certain number of major steps in commitment can be made in early development and that for these subsequent reference to the environment is not made.

This work was funded by the Medical Research Council. The authors acknowledge the clerical assistance of Mrs Len Schofield and the technical assistance of P. Day.

Asashima
,
M.
&
Grunz
,
H.
(
1983
).
Effects of inducers on inner and outer gastrula ectoderm layers of Xenopus laevis.
Differentiation
23
,
206
212
.
Billet
,
F. S.
&
Courtenay
,
T. H.
(
1973
).
A stereoscan study of the origin of ciliated cells in the embryonic epidermis of Ambystoma mexicanum.
J. Embryol. exp. Morph
.
29
,
549
558
.
Boucaut
,
J.-C.
(
1974a
).
Etude autoradiographique de la distribution de cellules embryonnaires isolées, transplantées dans le blastocèle chez Pleurodeles waltlii Micah (Amphibien, Urodèle)
.
Annls Embryol. Morph
.
7
,
7
50
.
Boucaut
,
J. C.
(
1974b
).
Chimères intergénériques entre Pleurodeles waltlii Micah et Ambystoma mexicanum Shaw (Amphibiens, Urodèles)
.
Annls Embryol. Morph
.
7
,
119
139
.
Cooke
,
J.
&
Webber
,
J. A.
(
1985
).
Dynamics of the control of body pattern in Xenopus laevis. I. Timing and pattern in the development of dorsoanterior and of posterior blastomere pairs isolated at the 4-cell stage
.
J. Embryol. exp. Morph
.
88
,
85
112
.
Dale
,
L.
&
Slack
,
J. M. W.
(
1987
).
Fate map for the 32cell stage of Xenopus laevis.
Development
100
,
279
296
.
Dale
,
L.
,
Smith
,
J. C.
&
Slack
,
J. M. W.
(
1985
).
Mesoderm induction in Xenopus laevis; a quantitative study using cell lineage label and tissue specific antibodies
.
J. Embryol. exp. Morph
.
89
,
289
313
.
Gardner
,
R. L.
(
1985
).
Clonal analysis of early mammalian development
.
Phil. Trans. R. Soc. Lond. B
313
,
163
178
.
Gurdon
,
J. B.
(
1977
).
Methods for nuclear transplantation in amphibia
.
Methods Cell Biol
.
16
,
125
139
.
Gurdon
,
J. B.
,
Brennan
,
S.
,
Fairmans
,
S.
&
Mohun
,
T. J.
(
1984
).
Transcription of muscle-specific actin genes in early Xenopus development; nuclear transplantation and cell dissociation
.
Cell
38
,
691
700
.
Heasman
,
J.
,
Wylie
,
C. C.
,
Hausen
,
P.
&
Smith
,
J. C.
(
1984
).
Fates and states of determination of single vegetal pole blastomers of Xenopus laevis.
Cell
37
,
185
194
.
Holtfreter
,
J.
&
Hamburger
,
V.
(
1955
).
In Analysis of Development
(ed.
B. H.
Willier
,
P. A.
Weiss
&
V.
Hamburger
), pp.
230
296
.
New York
:
Saunders
.
Jones
,
E. A.
(
1985
).
Epidermal development in Xenopus laevis: the definition of a monoclonal antibody to an epidermal marker
.
J. Embryol. exp. Morph
.
89
Supplement
,
155
166
.
Jones
,
E. A.
&
Woodland
,
H. R.
(
1986
).
Development of the ectoderm in Xenopus: tissue specification and the role of cell association and division
.
Cell
44
,
345
355
.
Keller
,
R. E.
(
1975
).
Vital dye mapping of the gastrula and neurula of Xenopus laevis. I. Prospective areas and morphogenetic movements of the superficial layer
.
Devi Biol
.
42
,
222
241
.
Kinter
,
C. R.
&
Brockes
,
J. P.
(
1984
).
Monoclonal antibodies identify blastemal cells derived from differentiating muscle in newt limb regeneration
.
Nature, Lond
.
308
,
67
69
.
Moody
,
S. A.
(
1987
).
Fates of the blastomers of the 16cell stage Xenopus embryo
.
Devi Biol
.
119
,
560
578
.
Sargent
,
T. D.
,
Jamrich
,
M.
&
Dawid
,
I. B.
(
1986
).
Cell interactions and the control of gene activity during early development of Xenopus laevis.
Devi Biol
.
114
,
238
246
.
Slack
,
J. M. W.
(
1983
).
From Egg to Embryo. Determinative Events in Early Development. Cambridge
,
London
:
Cambridge University Press
.
Slack
,
J. M. W.
(
1984
).
In vitro development of isolated ectoderm from axolotl gastrulae
.
J. Embryol. exp. Morph
.
80
,
321
330
.
Slack
,
J. M. W.
&
Forman
,
D.
(
1980
).
An interaction between dorsal and ventral regions of the marginal zone in early amphibian embryos
.
J. Embryol. exp. Morph
.
56
,
283
289
.
Snape
,
A.
,
Wylie
,
C. C.
,
Smith
,
J. C.
&
Heasman
,
J.
(
1987
).
Changes in states of commitment of single animal pole blastomeres of Xenopus laevis.
Devi Biol
.
119
,
503
510
.
Spemann
,
H.
&
Mangold
,
H.
(
1924
).
Über Induktion von Embryonalanlagen dutch Implantation artfremder Organisatoren
.
Arch. Mikrosk. Anat. EntwMech. Org
.
100
,
599
638
.
Steinman
,
R. M.
(
1968
).
An electron microscopic study of ciliogenesis in the developing epidermis and trachea in the embryo of Xenopus laevis.
Am. J. Anat
.
122
,
19
56
.
Sudarwati
,
S.
&
Nieuwkoop
,
P. D.
(
1971
).
Mesoderm formation in the anuran Xenopus laevis (Daudin)
.
Wilhelm Roux Arch. EntwMech. Org
.
166
,
189
204
.
Hiébaud
,
C. H.
(
1983
).
A reliable new cell marker in Xenopus. Devi Biol.
98
,
245
249
.
Wylie
,
C. C.
,
Smith
,
J. C.
,
Snape
,
A.
&
Heasman
,
J.
(
1985
).
The use of single cell transplantation in the study of cell commitment in early Xenopus embryos
. In
44th Annual Symposium of the Society of Developmental Biology. Gametogenesis and the Early Embryo
.
New York
:
Alan R. Liss
.
Wylie
,
C. C.
,
Snape
,
A.
,
Heasman
,
J.
&
Smith
,
J. C.
(
1987
).
Vegetal pole cells and commitment to form endoderm in Xenopus laevis.
Devi Biol
.
119
,
496
502
.