Factors from two growth factor families have been identified as having mesoderm inducing activity. These include activin and TGFβ from the TGFβ superfamily, and all members of the fibroblast growth factor (FGF) family.

When isolated ectoderm explants are treated with any of these factors, a proportion of their cells are caused to differentiate into mesodermal tissue types instead of epidermis. There are several differences in the biological activities which can broadly be summarized by saying that activin yields dorsal type inductions and FGF ventral type inductions.

Both bFGF and an FGF receptor have been detected in Xenopus blastulae, but it has not been shown that bFGF is normally secreted from vegetal cells. Various TGF/Mike mRNAs have also been detected and it is expected that an activin-like molecule will prove to be responsible for induction of the dorsal mesoderm in vivo.

Embryonic inducing factors have been shadowy substances existing at the margins of respectable biochemistry for many decades. In the last few years a combination of better protein purification methods with the characterization of many hormones and cytokines in other branches of cell biology have made it clear that inducing factors are not a special class of substances but are factors already known to be responsible for other biological activities in later life.

This paper will deal with mesoderm induction in Xenopus. This is a process which has now been intensively studied biologically and the moderate, although not complete, degree of understanding we have at the biological level has been crucial for the work on the inducing factors. I shall deal not only with the work of my own laboratory, which has concentrated on the role of the fibroblast growth factors (FGFs), but also with results obtained by others working with factors from the TGFβ superfamily.

Biology of mesoderm induction

Amphibian embryos have long been favoured for experimental embryology because their large size and accessibility makes them much more favourable for micromanipulation than mammals. Several species have been used in the past but Xenopus laevis, the African clawed frog, is now the world standard organism for this type of work. Although the situation is now improving, the account given in many textbooks is rather misleading, describing neural induction as ‘primary embryonic induction’ and confusing several processes which are logically distinct, such as formation of the mesoderm and determination of anteroposterior levels within the mesoderm.

Our current understanding of mesoderm induction is based on work by Nieuwkoop and of Nakamura in the early seventies (Nieuwkoop, 1969; Nakamura et al. 1971), which has been refined and extended but not fundamentally challenged in recent years by ourselves and others (Dale et al. 1985; Gurdon et al. 1985; Jones and Woodland, 1987). The results of these studies are very briefly as follows.

The mesoderm arises during the blastula stages, when the embryo consists of a hollow ball of cells. It is formed as a ring of cells around the equator called the marginal zone. At least half and perhaps all of the mesoderm is induced from the animal hemisphere as a result of inductive signals from the vegetal hemisphere. It also cannot be excluded that part of the mesoderm is formed by inheritance of a cytoplasmic determinant in the fertilized egg. The whole animal hemisphere is competent to become mesoderm but less than half of it becomes induced to do so in vivo, this being the ring of tissue abutting the vegetal hemisphere. The vegetal tissue is signalling and the animal tissue is competent to respond during the blastula stages but both signalling and competence decline shortly after gastrulation has commenced. Most of these facts have been established by variations on the combination experiment first introduced by Nieuwkoop and shown in Fig. 1, in which tissue from the animal pole region (an animal cap) is combined in vitro with an explant from the vegetal hemisphere.

Fig. 1.

Basic protocol for experiments on mesoderm induction. An animal cap from a Xenopus blastula is combined with part or all of the vegetal region. After a culture period of 1–3 days the combination is assayed for mesoderm formation by molecular or histological methods.

Fig. 1.

Basic protocol for experiments on mesoderm induction. An animal cap from a Xenopus blastula is combined with part or all of the vegetal region. After a culture period of 1–3 days the combination is assayed for mesoderm formation by molecular or histological methods.

The mesoderm will later become several tissues: in the trunk region notochord, somite, kidney, lateral plate and blood islands appear in a dorsal to ventral sequence. It is thought that the initial mesodermal induction creates perhaps only two zones, a small zone about 60° in circumference called the organizer and a larger zone comprising the remaining 300° of circumference, which initially has a ventral type of specification. The organizer region has several important properties which are quite difficult to disentangle mechanistically: it is the first region to involute during gastrulation and undergoes the most profound extension movements; it becomes the dorsal midline structure of the mesoderm of the entire body from head to tail; it dorsalizes the surrounding mesoderm to form the somites and kidney; it induces the overlying ectoderm to form the central nervous system. The ventral mesoderm undergoes later and less pronounced extension movements: its cells tend to move toward the dorsal midline during gastrulation and those which end up near it become dorsalized to form the somites. Several microsurgical experiments have shown that the specificity for the induction of organizer versus ventral type mesoderm lies with the signalling and not with the responding tissue. So the signal is complex, consisting at least of one substance at two concentrations or perhaps of more than one substance. The most generally accepted model of the inductive signals responsible for the formation and dorsoventral specification of the mesoderm is called the ‘three signal model’ (Dale and Slack, 1987) and is shown schematically in Fig. 2.

Fig. 2.

The three signal model for mesoderm induction. It is proposed that distinct signals are responsible for inducing the organizer (O) and ventral mesodermal (M3) regions, possibly activin A and bFGF respectively. Specification of other territories within the mesoderm occurs in response to a later signal emitted from the organizer. A, animal; VV, ventrovegetal; DV, dorsovegetal.

Fig. 2.

The three signal model for mesoderm induction. It is proposed that distinct signals are responsible for inducing the organizer (O) and ventral mesodermal (M3) regions, possibly activin A and bFGF respectively. Specification of other territories within the mesoderm occurs in response to a later signal emitted from the organizer. A, animal; VV, ventrovegetal; DV, dorsovegetal.

Mesoderm induction is not a cell contact-mediated process. This has been shown by transfilter experiments in which the vegetal and animal tissues are placed on opposite sides of a nucleopore membrane with pores too small to admit cellular processes (Grunz and Tacke, 1986). The inducing signal(s) can cross the liquid gap quite effectively. It has also been shown that the signals do not require the presence of functional gap junctions (Warner and Gurdon, 1987) and that the range of the effect is about 80 microns which is a few cell diameters at the embryonic stages in question (Gurdon, 1989).

These experiments tell us quite a lot about what to expect of the inductive signals. We are looking for one or more substances which are capable of causing mesodermal differentiation in isolated ectoderm. They should not be intracellular but capable of secretion by the vegetal cells. They should be diffusible, but not so diffusible that they would spread all over the embryo.

The candidate mesoderm inducing factors

Those who do not work with amphibian embryos should note that progress in this area has not been made by treating intact embryos with cytokines or by injecting cytokines into them. Such experiments are very difficult to interpret because of uneven access, a multiplicity of effects of different parts of the embryo and multiple secondary events which may involve alterations in cell movements as well as cell differentiation. The advantage of Xenopus compared to the mouse or the chick lies in the availability of a test tissue which can be used for bioassays in vitro. This is the ‘animal cap’ (Fig. 1) which forms 100% epidermis when cultured in vitro but will respond to mesoderm inducing factors (MIFs) in solution by forming some mesoderm as well. Explants in which some mesoderm is present will elongate during the first day of culture and form characteristic vesicles on the second or third day (Fig. 3). It is therefore possible to score an explant as induced or uninduced simply by visual inspection down the dissecting microscope. This is made into a quantitative assay by testing serial dilutions and observing the threshold concentration above which induction occurs, which is defined as 1 unit ml−1 (Godsave et al. 1988; Fig. 4). A competent operator can dissect about 100 animal caps in one afternoon which is enough for 10–15 titrations, so this bioassay is simple enough to be used for protein purification.

Fig. 3.

Mesoderm induction by bFGF. (A) Untreated animal caps after overnight culture have rounded up into balls. (B) Caps treated with bFGF have elongated. (C) After 3 days culture, treated explants have become fluid filled vesicles. (D) Section of untreated explant after 3 days, showing 100% epidermal differentiation. (E) Explant treated with 4 units ml−1Xenopus bFGF, showing formation of mesenchyme and mesothelium. (F) Explant treated with 32 units ml−1, showing formation of a large muscle mass.

Fig. 3.

Mesoderm induction by bFGF. (A) Untreated animal caps after overnight culture have rounded up into balls. (B) Caps treated with bFGF have elongated. (C) After 3 days culture, treated explants have become fluid filled vesicles. (D) Section of untreated explant after 3 days, showing 100% epidermal differentiation. (E) Explant treated with 4 units ml−1Xenopus bFGF, showing formation of mesenchyme and mesothelium. (F) Explant treated with 32 units ml−1, showing formation of a large muscle mass.

Fig. 4.

Animal cap titrations. The top row are explants treated with serial twofold dilutions of Xenopus bFGF starting at 10 ng ml−1. The first four are positive. The bottom row shows explants which were all treated with 10 ng ml−1 bFGF but also with twofold dilutions of a neutralizing IgG. The last three are positive, meaning that 3μgml−1 of IgG can just neutralize 10 ng ml−1 of bFGF.

Fig. 4.

Animal cap titrations. The top row are explants treated with serial twofold dilutions of Xenopus bFGF starting at 10 ng ml−1. The first four are positive. The bottom row shows explants which were all treated with 10 ng ml−1 bFGF but also with twofold dilutions of a neutralizing IgG. The last three are positive, meaning that 3μgml−1 of IgG can just neutralize 10 ng ml−1 of bFGF.

The factors shown to be active are, at the time of writing, activin A (XTC-MIF), TGFβ-2, and all members of the fibroblast growth factor family (aFGF, bFGF, kFGF, FGF-5, ECDGF and int-2). XTC-MIF was purified from a Xenopus cell line by Smith et al. (1988), activin A was found to be active by Asashima et al. (1990) and it has recently been shown that XTC-MIF is the Xenopus homologue of mammalian activin A (Smith et al. 1990). TGFβ-2 was shown to be active by Rosa et al. (1988). TGFβ-1 and -3 are not active on their own, although as we shall see they have synergistic effects with the FGFs. bFGF, ECDGF and aFGF were shown to be active by Slack et al. (1987), and were later joined by kFGF, FGF-5 and int-2, although int-2 has much lower specific activity than the others (Paterno et al. 1989).

All these MIFs were already known through work in other areas of cell biology. Activin is a hormone secreted by the ovary which promotes release of follicle stimulating hormone from the pituitary (Ling et al. 1988). There are two gene products called βA and βB which can assemble into homo and heterodimers. They are quite similar in sequence and belong to the TGF//-superfamily. Activin A is the homodimer of βA and it is probable, although not yet known for sure, that the other forms will also have mesoderm inducing activity. The TGFβs are well known cytokines which have a variety of biological activities in different systems, sometimes promoting and sometimes inhibiting cell division, sometimes promoting and sometimes inhibiting differentiation (Massagué, 1987). The FGFs are mitogens, particularly active on capillary endothelial cells, and characterised (except int-2) by tight binding to heparin (Gospodarowicz et al. 1987). The prototype members of the family, a and bFGF, were identified by protein purification, while the others considered here, kFGF, FGF-5 and int-2, were initially identified as oncogenes and later shown to code for mitogenic proteins.

Biological activity of MIFs

Detailed studies have been carried out for activin A and for a and bFGF (Slack et al. 1988; Green et al. 1990), and less information is available about the rest. However, it seems that there are a number of differences in biological properties between activin A on the one hand and all the other factors, including TGFβ -2, on the other.

The dose-response curves show that activin A has a specific activity of about 5×106unitsmg−1, meaning that it is active above 0.2ngml−1. The FGFs have specific activities about 5× lower, at about 1×106 units mg−1. There is little difference between the effectiveness of bFGF from bovine and Xenopus sources. For intact explants the dose response curves are very steep, and this is the basis of the serial dilution assay mentioned above. But for isolated cells, at least with the FGFs the dose response is quite extended, indicating that individual cells differ in their response thresholds (S. F. Godsave and J. M. W. Slack, unpublished observations).

Different doses also elicit the formation of different mesodermal tissues. For activin A, low concentrations induce ventral patterns of differentiation consisting of mesenchyme and mesothelium, higher concentrations induce muscle (Fig. 5), and the highest concentrations induce notochord and endoderm-like tissue. The full range of responses is spanned by about a 10 × change in concentration above the threshold. For the FGFs low concentrations also induce ventral structures, but then there is an extended concentration range going right up to toxic levels (micrograms ml−1) in which there is a progressively greater proportion of muscle induced. When cultures of isolated cells are induced (for method, see Godsave and Slack, 1989), 100% can be converted to muscle, suggesting that the mixtures of cell types typically found in induced explants arise because only the exposed blastocoelic surface is accessible to added factors. But it is very rare to see notochord in FGF inductions at whatever dose. It should also be noted that neither type of factor induces blood cells or pronephros although these are induced by vegetal tissue, either in combination or in transfilter experiments.

Fig. 5.

A clone of muscle cells arising from a single, isolated animal cap cell treated with murine activin A. Stained with a muscle-specific antibody (12/101).

Fig. 5.

A clone of muscle cells arising from a single, isolated animal cap cell treated with murine activin A. Stained with a muscle-specific antibody (12/101).

The time of competence of the animal cap can be assessed by short term treatment of tissue isolated from different stage embryos. This has shown that competence to respond to activin A extends into the early gastrula stages, while competence to respond to the FGFs is lost just before the onset of gastrulation.

These four differences, the specific activity, the steepness of the dose-response curve, the formation of notochord, and the stages of competence could all be interpreted as saying that the FGFs are just weak inducers, evoking the same response as activin A but less effectively. But the last difference argues against this. The FGFs are more effective than activin A at turning on a homeobox-containing gene called xhox-3, which is normally expressed in the posteroventral regions of the mesoderm (Ruiz i Altaba and Melton, 1989a). This makes the two groups of factor seem to differ in a qualitative way with activin A having an organizer-inducing capability and the FGFs having a ventral mesoderm-inducing capability.

Modifications of biological activity

It was early discovered that TGFβ-1 could enhance the effect of a given dose of FGF and provoke a degree of muscle formation equivalent to a higher dose (Kimelman and Kirschner, 1987). We have found that the same is also true for TGFβ-3. However, the quality of the induction is not altered from FGF-like to activin-like, and in particular the induction of notochord remains rare, as is found for FGF alone.

A similar enhancement of FGF effect is seen with lithium ion. Lithium has been shown to cause hyperdorsalization of whole embryos, which is interpreted as an expansion in the size of the organizer territory (Kao et al. 1986; Kao and Elinson, 1988). It has no effect on isolated animal explants but will enhance the effect of FGF treatment and will also greatly increase the amount of muscle formed by explants from the ventral marginal zone, which normally form predominantly blood cells and mesenchyme (Slack et al. 1988).

A negative effect on bFGF is shown by heparin. This binds tightly to all forms of FGF but its inhibitory action is most apparent on bFGF. Inhibition is seen in the microgram ml−1 range meaning that a molar excess of >1000 is required (Slack et al. 1987). This suggests that those complexes of heparin and bFGF which are presumably formed at lower heparin concentrations are still active and that inhibition may result from the occupation not of the high affinity heparin sites on bFGF but of the low affinity sites. It is often thought that FGFs in vivo are immobilized on cell surface heparin sulphate, so it is important that immobilization is not necessarily incompatible with biological activity.

Factors and receptors present in the embryo

A number of laboratories are energetically cloning likely factors from Xenopus cDNA libraries. Several TGFβ-like clones have been obtained but it is not yet known which of them produce biologically active proteins or which produce proteins at the appropriate embryonic stages. A TGFβ-like mRNA which is present in the early embryo is called vg-1 and was discovered by differential screening of an oocyte library with animal and vegetal cDNA (Weeks and Melton, 1987). The vg-1 message becomes localized to the vegetal hemisphere during oogenesis. Protein is synthesized before and after fertilization and because the message is inherited by the vegetal cells it is in these cells that the zygotic synthesis occurs (Tannahill and Melton, 1989). However, it has been difficult to demonstrate biological activity on the part of the protein, and although it may have synergistic activity with other factors it probably does not have mesoderm inducing activity in its own right. The present consensus is that vg-1 probably is involved in mesoderm induction in some capacity, but it is probably not itself an inducing factor.

mRNA for bFGF was found in embryos by Kimelman et al. (1987, 1988). The levels are low and decrease from the oocyte to the early embryo. There is also an excess of an ‘anti-message’ which is derived from the other DNA strand and codes for an entirely different protein (Volk et al. 1989). Although the actual protein coding sequences do not overlap, the untranslated regions of each mRNA overlap the coding sequences of the other, so it seems unlikely that both mRNAs could be transcribed from the same genes at the same time. There are also low levels of heparin-binding inducing proteins present in early embryos (Kimelman et al. 1988; Slack and Isaacs, 1989). Two protein bands have been identified, of about 20 and 14×103Mr. The 20×103Mr form is certainly bFGF and the 14×103Mr form reacts with two antibodies prepared against bFGF, but since these particular antibodies may cross react with other types of FGF the identification of the 14×103Mr species is not unambiguous. The purification of these proteins from Xenopus embryos so far represents the only recovery of MIFs from the embryos themselves.

It has already been mentioned that the mesoderm inducing signals can cross a liquid gap in a transfilter apparatus. Using a modification of the transfilter method, in which the tissues are held about 100 microns apart, I have been able to show that both dorsal and ventral vegetal tissues can provoke inductions transfilter, and must therefore be secreting active factors. However, all attempts to condition even tiny volumes of medium by vegetal calls have failed, so the levels must be very low or the factors very unstable. It is reproducibly found that heparin in solution can inhibit the transmission of these transfilter signals, from whole or ventral half vegetal pieces (Slack et al. 1987). Heparin also inhibits the action of bFGF in vitro and so this may seem a good piece of evidence for the secretion of bFGF by vegetal cells. However, neutralizing antibodies to bFGF do not inhibit the transmission of the transfilter signals, and the specificity of heparin for bFGF, although reasonable, is unlikely to be absolute.

A receptor for FGF (probably binding all forms) has been detected in early Xenopus embryos by chemical cross-linking (Gillespie et al. 1989). It is a protein of about 130×103Mr and seems similar in its properties to the FGF receptors identified on mammalian cells. The levels detectable on the surfaces of animal cap cells rise and fall in a very similar way to the competence of explants to respond to FGFs, reaching a peak in the midblastula and falling by a factor of 10 by the beginning of gastrulation. Binding studies on dissected explants suggest that there are not significant differences in receptor density in different parts of the embryo, so it is present both in the region that normally responds (the marginal zone) and in the remainder of the animal hemisphere that does not respond in vivo but is capable of doing so in vitro. This is consistent with the inference from the embryological studies that the range of the response is controlled by the distribution of the signal, rather than by differential competence. While bFGF will compete efficiently with aFGF for binding to this receptor, TGFβ-2 will not. In mammalian cells, TGFβ receptors are distinct from those for the FGFs, and in view of the various differences in the responses to activin A and bFGF it also seems likely that in Xenopus embryos each group of factors will have its own receptor.

Later events of mesoderm induction

As in so many other systems, there is evidence that the immediate events following binding of the factors involve protein phosphorylation of tyrosine. The mammalian FGF receptor is a tyrosine kinase, and we have data to show that the same is true for the Xenopus receptor (L. L. Gillespie, unpublished observations). Injection of synthetic mRNA for the polyoma middle T protein has a mesoderm inducing effect (Whitman and Melton, 1989). This binds to and activates the cellular protein pp60-c-src which is a tyrosine kinase. From a technical standpoint it should be noted once again that these experiments were not based on observation of effects on whole embryos. They used the more discriminating method of injecting the mRNA into zygotes, raising the embryos to blastulae and then explanting animal pole tissue and looking for auto-induction of the explants. As far as other signal transduction pathways are concerned it seems that cyclic AMP and cyclic GMP are not involved, nor is protein kinase C activated, although it does seem to be involved in neural induction at somewhat later stages (Otte et al. 1989).

Several genes are turned on as a consequence of exposure of animal caps to MIFs (Rosa, 1989). One, turned on by activin A but not by FGFs, is a homeobox gene called mix-1 which is normally expressed in the endoderm. Whether parts of the endoderm are formed in vivo by induction is not known, but it does seem to happen in vitro. Another homeobox-containing gene which is preferentially turned on by FGFs is xhox-3 (Ruiz i Altaba and Melton, 1989a). This is normally expressed in the posteroventral parts of the mesoderm. Ectopic over-expression in whole embryos, provoked by injection of excess synthetic mRNA into the zygotes, causes suppression of the head (Ruiz i Altaba and Melton, 19896) and it has been argued by these authors that xhox-3 is a coding factor for the posterior part of the body. But there may of course be other events intervening between FGF treatment and xhox-3 elevation. There are certainly other events before the activation of muscle specific markers, such as a-actin, which have been much used for studies on mesoderm induction.

How well do we now understand the mechanism of mesoderm induction?

This field has received a disproportionate share of publicity in recent years and there have been almost more mini-reviews than original papers. Outsiders might be forgiven for thinking that it was all over. But actually it has only just begun. What we are dealing with is not the activation of one gene in a bottle of fairly homogeneous tissue culture cells. Regional specification in the early embryo is a complex, three dimensional problem, and at any given embryonic stage there are several different processes going on in the same cells. If we look at the much more profound revolution of understanding that has occurred in Drosophila we find that events are indeed complex and each process requires a number of components. Sometimes the same phenotype is obtained by mutating components to inactivity at different stages of a causal chain. Therefore it should not be surprising that similar effects can be obtained by different treatments at different stages, for example dorsalization provoked by lithium in cleavage stages, and by an organizer graft in the early gastrula. In Drosophila we also find that specification of territories along the two principal body axes occurs more or less independently. In Xenopus it is hard to imagine that one or two substances control the whole of early development, as is implied in some accounts. It is much more likely that each identifiable biological process has its own set of factors, responses and genes, and indeed that many more components are at work than would be strictly necessary if the maximum combinatorial economy was sought. We are presently still very short of sufficiently discriminating markers to dissect the individual processes, for example when we see more muscle we do not know if we are looking at a dorsalization, a posteriorization or simply more responding cells in the population.

The strongest candidate for a mesoderm inducing factor in vivo is probably still bFGF. It is present at the relevant embryonic stages, it is there in sufficient quantity for biological activity, and its receptor rises and falls during the blastula stages in a manner suggesting some role for the ligand. But we still do not know its localization in the embryo; there is no evidence that it is secreted by vegetal cells; and its biological activity only covers a partial range of the responses seen in vivo. Since there is some serious doubt about whether bFGF can be secreted at all, due to its lack of a signal sequence, it is possible that it plays some intracellular role in the responding tissue rather than being an inducing factor per se. Everyone in the field expects activin A to be the organizer-inducing signal, but it remains to be shown that it is present in the embryo and that it is secreted by vegetal cells.

What we have done in the last three years is not so much to provide definitive answers as to bring together the fields of experimental embryology and cytokine research. This is proving to be a very fruitful relationship for both sides. For the embryologists it provides explanations of previously mysterious phenomena, crudely summarised by the slogans: ‘inducing factors are growth factors; competence means receptors for growth factors; determination means activation of particular signal transduction pathways’. It also provides us with a basket of relevant molecular probes which should eventually turn out to be the key to unlock the embryonic box. For cell biologists working with cytokines, the discovery of their role in early development opens new possibilities for research on their in vivo function, always the weak link in a field previously dominated by research using tissue culture cells.

Asashima
,
M.
,
Nakano
,
H.
,
Shimada
,
K.
,
Kinoshita
,
K.
,
Ishii
,
K.
,
Shibai
,
H.
and
Ueno
,
N.
(
1990
).
Mesoderm induction in early amphibian embryos by activin A (erythroid differentiation factor)
.
‘Wilhelm Roux Arch. Devi Biol.
198
,
330
335
.
Dale
,
L.
and
Slack
,
J. M. W.
(
1987
).
Regional specification within the mesoderm of early embryos of Xenopus laevis. Development
100
,
279
295
.
Dale
,
L.
,
Smith
,
J. C.
and
Slack
,
J. M. W.
(
1985
).
Mesoderm induction in Xenopus laevis. A quantitative study using a cell lineage label and tissue specific antibodies
.
J. Embryol. exp. Morph.
89
,
289
313
.
Gillespie
,
L. L.
,
Paterno
,
G. D.
and
Slack
,
J. M. W.
(
1989
).
Analysis of competence: Receptors for fibroblast growth factor in early Xenopus embryos
.
Development
106
,
203
208
.
Godsave
,
S. F.
,
Isaacs
,
H.
and
Slack
,
J. M. W.
(
1988
).
Mesoderm inducing factors: a small class of molecules
.
Development
102
,
555
566
.
Godsave
,
S. F.
and
Slack
,
J. M. W.
(
1989
).
Clonal analysis of mesoderm induction
.
Devi Biol.
134
,
486
490
.
Gospodarowicz
,
D.
,
Ferrara
,
N.
,
Schweigerer
,
L.
and
Neufeld
,
G.
(
1987
).
Structural characterization and biological functions of fibroblast growth factor
.
Endocrine Rev.
8
,
95
114
.
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
183
.
Grunz
,
H.
and
Tacke
,
L.
(
1986
).
The inducing capacity of the presumptive endoderm of Xenopus laevis studied by transfilter experiments
.
Wilhelm Roux Arch. Devi Biol.
195
,
467
473
.
Gurdon
,
J. B.
(
1989
).
The localization of an inductive response
.
Development
105
,
27
33
.
Gurdon
,
J. B.
,
Fairman
,
S.
,
Mohun
,
T. J.
and
Brennan
,
S.
(
1985
).
The activation of muscle specific action genes in Xenopus development by an induction between animal and vegetal cells of a blastula
.
Cell
41
,
913
922
.
Jones
,
E. A.
and
Woodland
,
H. R.
(
1987
).
The development of animal cap cells in Xenopus: a measure of the start of animal cap competence to form mesoderm
.
Development
101
,
557
563
.
Kao
,
K. R.
and
Elinson
,
R. P.
(
1988
).
The entire mesodermal mantle behaves as Spemann’s organizer in dorsoanterior enhanced Xenopus laevis embryos
.
Devi Biol.
127
,
64
77
.
Kao
,
K. R.
,
Masui
,
Y.
and
Elinson
,
R. P.
(
1986
).
Lithium induced respecification of pattern in Xenopus laevis embryos
.
Nature
322
,
371
373
.
Kimelman
,
D.
,
Abraham
,
J. A.
,
Haaparanta
,
T.
,
Palisi
,
T. M.
and
Kirschner
,
M. W.
(
1988
).
The presence of fibroblast growth factor in the frog egg: its role as a natural mesoderm inducer
.
Science
242
,
1053
1056
.
Kimelman
,
D.
and
Kirschner
,
M.
(
1987
).
Synergistic induction of mesoderm by FGF and TGF-/3 and the identification of an mRNA coding for FGF in the early Xenopus embryo
.
Cell
51
,
869
877
.
Ling
,
N.
,
Ueno
,
N.
,
Ying
,
S. Y.
,
Esch
,
F.
,
Shimasaki
,
S.
,
Hotta
,
M.
,
Cuevas
,
P.
and
Guillemin
,
R.
(
1988
).
Inhibins and activins
.
Vitamins and Hormones
44
,
1
46
.
Massagué
,
J.
(
1987
).
The TGF-/I family of growth and differentiation factors
.
Cell
49
,
437
438
.
Nakamura
,
O.
,
Takasaki
,
H.
and
Ishihara
,
M.
(
1971
).
Formation of the organizer by combinations of presumptive ectoderm and endoderm
.
Proc. Jap. Acad.
47
,
313
318
.
Nieuwkoop
,
P. D.
(
1969
).
The formation of the mesoderm in urodelean amphibians I. Induction by the endoderm
.
Wilhelm Roux Arch. EntwMech. Org.
162
,
341
373
.
Otte
,
A. P.
,
Van Run
,
P.
,
Heidevald
,
M.
,
Van Driel
,
R.
and
Durston
,
A. J.
(
1989
)
Neural induction is mediated by cross talk between the protein kinase C and cyclic AMP pathways
.
Cell
58
,
641
648
.
Paterno
,
G. D.
,
Gillespie
,
L. L.
,
Dixon
,
M. S.
,
Slack
,
J. M. W.
and
Heath
,
J. K.
(
1989
).
Mesoderm inducing properties of int-2 and kFGF: two oncogene encoded growth factors related to FGF
.
Development
106
,
79
83
.
Rosa
,
F.
,
Roberts
,
A. B.
,
Danielpour
,
D.
,
Dart
,
L. L.
,
Sporn
,
M. B.
and
Dawid
,
I. B.
(
1988
).
Mesoderm induction in amphibians: The role of TGF-/J2-like factors
.
Science
239
,
783
785
.
Rosa
,
F. M.
(
1989
).
Mix-1, a homeobox mRNA inducible by mesoderm inducers, is expressed mostly in the presumptive endodermal cells of Xenopus embryos
.
Cell
57
,
965
974
.
Ruiz i Altaba
,
A.
and
Melton
,
D. A.
(
1989a
).
Interaction between peptide growth factors and homeobox genes in the establishment of anteroposterior polarity in frog embryos
.
Nature
341
,
33
38
.
Ruiz i Altaba
,
A.
and
Melton
,
D. A.
(
1989b
).
Involvement of the Xenopus gene xhox-3 in pattern formation along the anteroposterior axis
.
Cell
57
,
317
326
.
Slack
,
J. M. W.
,
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
.
Slack
,
J. M. W.
and
Isaacs
,
H. V.
(
1989
).
Presence of basic fibroblast growth factor in the early Xenopus embryo
.
Development
105
,
147
154
.
Slack
,
J. M. W.
,
Isaacs
,
H. V.
and
Darlington
,
B. G.
(
1988
).
Inductive effects of fibroblast growth factor and lithium ion on Xenopus blastula ectoderm
.
Development
103
,
581
590
.
Smith
,
J. C.
,
Price
,
B. M. J.
,
Van Nimmen
,
K.
and
Huylebroek
,
D.
(
1990
).
XTC-MIF: a potent Xenopus mesoderm inducing factor, is a homologue of activin A
.
Nature
345
,
729
731
.
Smith
,
J. C.
,
Yaqoob
,
M.
and
Symes
,
K.
(
1988
).
Purification, partial characterisation and biological effects of the XTC mesoderm-inducing factor
.
Development
103
,
591
600
.
Tannahill
,
D.
and
Melton
,
D. A.
(
1989
).
Localized synthesis of the vg-1 protein during early Xenopus development
.
Development
106
,
775
785
.
Volk
,
R.
,
Kôster
,
M.
,
Poting
,
A.
,
Hartman
,
L.
and
Knochel
,
W.
(
1989
).
An antisense transcript from the Xenopus laevis bFGF gene coding for an evolutionarily conserved 24 kD protein
.
EMBO J.
8
,
2983
2988
.
Warner
,
A. E.
and
Gurdon
,
J. B.
(
1987
).
Functional gap junctions are not required for muscle gene activation by induction in Xenopus embryos
.
J. Cell Biol.
104
,
557
564
.
Weeks
,
D. L.
and
Melton
,
D. A.
(
1987
).
A maternal messenger RNA localised to the vegetal hemisphere in Xenopus eggs codes for a growth factor related to TGF-/3
.
Cell
51
,
861
867
.
Whitman
,
M.
and
Melton
,
D.
(
1989
).
Induction of mesoderm by a viral oncogene in early Xenopus embryos
.
Science
244
,
803
806
.