Mesoderm induction: from caps to chips

The vertebrate mesoderm produces a wide range of tis-sues including the muscles, heart, vasculature, blood, kidney, gonads, dermis and cartilage, and it also has a major role in the morphogenetic movements of gastru-lation. The study of mesoderm formation originated with Pieter Nieuwkoop’s classical experiments in the amphibian embryo, nearly 40 years ago. Nieuwkoop showed that explanted tissue from the bottom of the blastula-stage embryo (the vegetal cap) could convert prospective ectodermal cells taken from the top of the embryo (the animal cap) into mesodermal tissues that normally reside at the equator, demonstrating that the mesoderm is formed by a mechanism called induction1. Induction refers to a process in which extracellular sig-nals bring about a change from one cell fate to another in a particular group of cells. In the classical view, induc-tion occurs when one group of cells signal to a different set of cells that respond by changing fate. In practice, it can be more complex than this. In zebrafish, for exam-ple, the prospective mesodermal cells are involved in sending and receiving the inducing signals.

Before Nieuwkoop’s seminal experiments, the first inductive event in the amphibian embryo was thought to involve gastrula-stage signals that originate in the organizer and pattern the mesoderm and ectoderm (reviewed in REF. 2). Although subsequent experiments have led to the subdivision of mesoderm induction into initiation, maintenance and patterning events, here I use mesoderm induction in a broad sense to encompass all these steps because all three take place concurrently in the late blastula/early gastrula embryo. Nieuwkoop’s original assay provided a valuable tool for studying embryonic signalling that is still in constant use today.

Because animal caps can be easily isolated and grown in a simple buffer solution, and because conversion of ani-mal cap explants to mesoderm can be readily monitored at the macroscopic level, this assay provided a valuable means for identifying the first mesoderm-inducing fac-tors (reviewed in REF. 3). With the development of more sophisticated methods for measuring gene expression and manipulating the amount of endogenous signalling in Xenopus, the utility of the animal cap-explant assay for understanding the basic biology of inductive processes has continued to expand.

Although the field of mesoderm induction began with studies in amphibians, today it is impossible to discuss this topic without including the genetic stud-ies in zebrafish that have contributed important new discoveries to this area. Comparison of the two species, which is also the primary focus of this review, provides interesting insights into conserved aspects of mesoderm induction, as well as changes in the induction mecha-nisms to accommodate different embryological features. Studies in Xenopus and zebrafish also provide a valuable foundation for understanding mesoderm induction in other species, such as birds and mammals. We can now begin to develop a clear understanding of the molecu-lar interactions that underlie this essential process in embryonic development.

Here I first introduce the principal signalling factors that are involved in mesoderm induction, with emphasis on members of the Nodal family, which function to initi-ate the formation of the mesoderm. I discuss how these factors function at different times and in various combi-nations to regulate different regions within the embryo. I conclude by examining some new approaches that are

Department of Biochemistry, Box 357350, University of Washington, Seattle, Washington 98195-7350, USAe-mail: [email protected]:10.1038/nrg1837

Blastula A stage during which the embryo undergoes cleavage to become multicellular. The late blastula stage precedes gastrulation.

Gastrula A stage during which the embryo undergoes major morphogenetic changes, which positions the endoderm on the inside, the mesoderm in the middle and the ectoderm on the outside.

Organizer A signalling centre in a vertebrate embryo comprising a group of cells that secrete signalling factors or inhibitors of signalling factors, which changes the fate of the surrounding cells.

Mesoderm induction: from caps to chipsDavid Kimelman

Abstract | Vertebrate mesoderm induction is one of the classical problems in developmental biology. Various developmental biology approaches, particularly in Xenopus and zebrafish, have identified many of the key factors that are involved in this process and have provided major insights into how these factors interact as part of a signalling and transcription-factor network. These data are beginning to be refined by high-throughput approaches such as microarray assays. Future challenges include understanding how the prospective mesodermal cells integrate the various signals they receive and how they resolve this information to regulate their morphogenetic behaviours and cell-fate decisions.

R E V I E W S

360 | MAY 2006 | VOLUME 7 www.nature.com/reviews/genetics

© 2006 Nature Publishing Group

Blood Blood HeadPronephros

NotochordTailsomites

Trunksomites

Animal

Vegetal

DorsalXenopus ZebrafishVentral

Ectoderm

EndodermYSL

Mesoderm

Atrium

Erythroid Ventricle Myeloid

Pronephros HeadNotochordTail somites Trunk somites

Notochord A rod-shaped structure that runs along the dorsal axis of the embryo, separating the muscle blocks. It is one of the defining features of the phylum Chordata, to which vertebrates belong.

Fate map A map that shows which tissues are likely to develop from different regions of the embryo.

being used to gain further insights into the mechanism of mesoderm formation and regulation.

Forming the mesodermIn both zebrafish and Xenopus embryos, mesoderm induction creates a zone of mesodermal cells at the equator of the embryo (often called the marginal zone). Whereas the mesoderm and endoderm are regionally distinct in Xenopus, in zebrafish the endoderm and mesoderm precursors are mixed (FIG. 1). In addition, whereas all the cells in the early Xenopus gastrula con-tribute to the final embryo, in zebrafish the embryo sits on top of a yolk cell that does not undergo cleavage. This difference in architecture between the two species has major consequences for the mechanism of mesoderm induction, as discussed below.

In addition to establishing the mesodermal zone, inductive processes pattern the embryo along what is commonly called ‘the dorsal–ventral axis’ as well as the animal–vegetal axis (FIG. 1). Because the ‘dorsal’ side pro-duces both anterior (head) and dorsal (notochord) fates, whereas the ‘ventral’ side produces both posterior (tail) and ventral (for example, pronephros) fates, the equa-torial axis in the pre-gastrula embryo might be better termed the ‘dorsoanterior–ventroposterior axis’ (for an excellent discussion of the complexities in labelling the pre-gastrula axis see REF. 4). Regardless of the nomen-clature, complex morphogenetic movements during gastrulation bring mesodermal cells into their correct position within the post-gastrula embryo.

Although fate maps are often drawn with the late blastula/early gastrula mesoderm demarcated into regions of defined fate, careful lineage labelling of zebrafish embryos at these stages has demonstrated that mesodermal derivatives outside the organizer region are highly intermixed5,6. Similarly, a recent study in Xenopus has shown that more cells in the early gastrula embryo

express a marker for a different domain than is predicted by the fate map7. This result probably reflects the fact that embryonic cells are exposed to multiple signals (including signalling factors and their inhibitors) as they migrate in the embryo; therefore, their eventual fate might represent the sum of these external influences, rather than their position at a particular point in time.

Signalling factors in mesoderm inductionIt is likely that all of the families of signalling factors that are important for mesoderm induction have been identi-fied, although all of the crucial individual factors might not yet have been determined. A surprising feature is the complexity of signalling factors that are used by the embryo for inducing mesoderm (TABLE 1). From the per-spective of the embryo, there might be evolutionary advan-tages to having multiple factors with overlapping functions. From the perspective of the experimentalist, however, this degree of redundancy can make loss-of-function studies much more challenging.

As a first approximation one can say that: the Nodal family is involved in initiating mesoderm formation, FGFs (Fibroblast growth factors) and Wnts are involved in maintaining the mesodermal state, and BMPs (Bone morphogenetic proteins) are involved in patterning the mesoderm. This, however, is an oversimplification of what these factors actually do. For example, in various experi-mental models, FGFs, BMPs and Wnts have been shown to be sufficient for initiating mesoderm formation8–13, and Nodal family members have been shown to be involved in patterning the mesoderm14,15. These results indicate that extracellular signals do not necessarily have rigidly separated functions in mesoderm induc-tion, and instead indicate that these signals might work in a partially overlapping way to form and pattern the mesoderm. Because the Nodal family functions as the main initiating stimulus for mesoderm induction16, it is the main focus of the discussion below (for a more thorough discussion of the roles of Wnts, BMPs and FGFs see REFS 17–19). A description of the intracellular signalling pathways that are used by each of these factors can be found in BOX 1.

Models for mesoderm induction in fish and frogsRecent studies have revealed how the mesoderm-inducing signal described by Nieuwkoop1 is localized to the vegetal hemisphere in Xenopus embryos. During oogenesis, tran-scripts that encode the T-box transcription factor VegT are localized to the vegetal pole through a complex process that involves specific RNA-binding factors and cytoskel-etal elements (reviewed in REF. 20). At fertilization, VegT transcripts are released from the vegetal pole and slowly diffuse upwards. Because the third cleavage plane passes through the equator of the embryo before the transcripts leave the vegetal hemisphere, the VegT transcripts are trapped there, and therefore the subsequently translated VegT protein is restricted to the vegetal half. At the start of zygotic transcription at the 4,000-cell stage, which is called the mid-blastula transition, VegT activates the transcrip-tion of the Xenopus nodal related (Xnr) genes, which then initiate mesoderm formation21–25 (FIG. 2).

Figure 1 | Fate maps of Xenopus and zebrafish embryos at the late blastula/early gastrula stage. Prospective mesodermal territories are shown in red. Note that in zebrafish the bottom layer of the mesoderm sits on top of the extra-embryonic yolk syncytial layer (YSL), and that mesodermal cells are intermixed with endodermal cells (green). The fate maps are based on REFS 5,125,126.

R E V I E W S

NATURE REVIEWS | GENETICS VOLUME 7 | MAY 2006 | 361


In zebrafish, the Nodal genes squint (sqt; also known as ndr1) and cyclops (cyc; also known as ndr2) (BOX 1) initiate mesoderm formation, but their transcription is activated by an unknown signal originating in the extra-embryonic yolk syncytial layer (YSL; reviewed in REF. 26) (FIG. 2). The zebrafish VegT orthologue has been identi-fied, and although it has an important role in forming the trunk musculature as discussed below, it is not maternal and it does not activate nodal gene expression27. It is still unclear how a YSL-specific signal is actually produced, although it seems unlikely to use the same mechanism as is used by Xenopus VegT. The YSL does not even form as a distinct entity until approximately the 1,000-cell stage, and the YSL nuclei and cytoplasm have the same origin as the overlying embryonic cells (called the blastomeres) that form the mesendoderm. Therefore, if a transcript were restricted to the vegetal pole during oogenesis and released at fertilization, it would probably be inherited by both the YSL and the blastomeres. It is possible that there is a mechanism that anchors a transcript (or protein) to the yolk and keeps it inactive until the YSL is formed; at that point the factor could be released into the YSL and activated to initiate mesoderm formation.

Nodal signalling: are gradients important?As described below, the embryo potentially has gradients of Nodal activity along the dorsal–ventral and animal–vegetal axes. The existence, significance and role of these Nodal gradients in mesoderm induction continue to be

hot topics of debate. In Xenopus, several lines of evidence indicate that Nodal signalling is stronger on the dorsal than on the ventral side of the pre-gastrula embryo. In Xenopus embryos, a ‘dorsalizing’ activity moves from the vegetal pole towards one side of the embryo soon after fertilization (reviewed in REF. 28), and evidence from experiments in which the early vegetal cytoplasm was removed from the embryo indicates that there is a similar mechanism in zebrafish29,30. The net result of this move-ment is the stabilization of β-catenin, an intracellular Wnt pathway component (BOX 1), on the future dorsal side of the embryo (FIG. 3). This asymmetrical stabilization of β-catenin is essential for the formation of all dorsal and anterior structures (REFS 31,32; reviewed in REFS 28,33). Among the many targets of β-catenin are members of the Nodal gene family called the Xenopus nodal-related fac-tors; for this reason, the Xnr genes are expressed earlier and/or at a higher level on the dorsal side of the embryo (FIG. 3). For example, Xnr5 and Xnr6 are transcribed on the dorsal side of the embryo in response to β-catenin as early as the 256-cell stage, well before most embry-onic genes begin to be transcribed at the mid-blastula transition34. These two Xnr genes, as well as Xnr1, Xnr2 and Xnr4, which are first transcribed at the mid-blastula transition, are expressed in a dorsal–ventral gradient in the late blastula stages35,36. Consequently, Nodal signal-ling is at least transiently enriched dorsally, as indicated by increased phosphorylation (and so activation) of the Nodal intracellular factor Smad2 (BOX 1) on the dorsal

Table 1 | Candidate zygotic signalling molecules in mesoderm induction*

Signal family Xenopus Mouse Chick Zebrafish (mutant/morpholino phenotype)

Nodal Xnr1 Nodal Nodal Sqt‡§ (cyclopia/dorsal mesoderm defects)

Xnr2 Gdf3 Vg1 Cyc‡§ (cyclopia)

Xnr4 Vg1 (N.D.)

Xnr5

Xnr6

Activin

Derriére

Vg1

FGF Fgf3 Fgf3 Fgf8? Fgf3 (N.D.)

Fgf4 Fgf9 Fgf4 Fgf8‡ (cerebellum and mild posterior defects)

Fgf8 Fgf8 Fgf24‡ (pectoral fin defect||)

BMP Bmp2 Bmp7 Bmp2 Bmp2b‡ (severe dorsalization)

Bmp4 Bmp2 Bmp4 Bmp4 (N.D.)

Bmp7 Bmp4 Bmp7 Bmp7‡ (severe dorsalization)

Admp Admp Admp (dorsalized)

Wnt¶ Wnt8 Wnt8 Wnt8c Wnt8 (posterior mesoderm and neural defects)

Wnt3 Wnt3a (no effect#)

Wnt3a

*There is no horizontal relationship between the genes listed in this table. ‡Zebrafish with mutations in these genes have been identified. §sqt;cyc double mutants have no head or trunk mesoderm, and no endoderm. ||fgf24MO in fgf8 mutants causes loss of the most posterior mesoderm. ¶Members of the canonical (β-catenin stabilizing) family of Wnts are listed. #Loss of wnt3a enhances loss of wnt8. Admp, Anti-dorsalizing morphogenetic protein; BMP, Bone morphogenetic protein; Cyc, Cyclops; FGF, Fibroblast growth factor; Gdf3, Growth/differentiation factor; N.D., not determined; Sqt, Squint; Xnr, Xenopus nodal related.

R E V I E W S



LRP

Friz

zled

Wnt FGF

Fgfr

Fgfr

DshG

β-Catenin

Tcf/LEF

Mapk

P

P

P

Smad4Smad2,3 Smad4Smad1,5,8

P

Mapk

P

Raf

Ras

P

Mek

β-Catenin

Axin

Gsk3

APC

BMP

Type

II

Type

I

Smad4Smad1,5,8

P

EGF–CFC

Smad4Smad2,3

P

Nodal

Type

II

Type

I

BMP WntNodal FGF

Box 1 | Intercellular signalling pathways

Many of the essential details of the intracellular pathways that are used by the four signalling factors discussed in this review have been identified and are shown here (see figure). Below I discuss key aspects of each of the main pathways; for more extensive discussions of these pathways see REFS 119,120,121.

Nodal and BMPBoth the Nodal and BMP (bone morphogenetic protein) pathways are activated when a ligand binds a specific heterotetramer that is composed of two type I and two type II receptors, as is the case with all members of the Transforming growth factor-β (TGFB) family. In zebrafish, mutants have been identified in two Nodal ligands: squint (sqt) and cyclops (cyc), and in two BMP ligands: swirl (bmp2b) and snailhouse (bmp7). The type II receptor phosphorylates a cytoplasmic domain on the type I receptor, activating the type I receptor to phosphorylate a Smad factor. Whereas Smad2 and Smad3 are specific mediators of Nodal (including Activin and Vg1) pathways, Smad1, Smad5 (Somitabun in zebrafish) and Smad8 are mediators of the BMP pathway. When these Smads are phosphorylated they bind Smad4 and translocate to the nucleus where they bind to specific DNA-binding factors. The Smad factors also have a weak DNA-binding ability, and it is the combination of the DNA-binding specificity of the molecular partner together with the DNA-binding specificity of the Smads that allows the transcriptional activation of specific targets. For example, in the zebrafish Nodal pathway, the binding of Smads to the trancription factor Bonnie (Bon) is an essential step in endoderm formation, whereas the binding of Smads to Schmalspur (Sur) is important for mesoderm formation. The Nodal pathway is unique in that it requires an extracellular membrane-bound EGF–CFC cofactor called One-eyed pinhead (Oep) in zebrafish, Xcr1/Frl1, Xcr2 and Xcr3 in frogs, Cryptic in chicks and Cripto and Cryptic in mice122,123. Although these factors are required for Nodal and Vg1 signalling124, it remains unclear why Nodal and Vg1 specifically require these cofactors to signal whereas related ligands such as Activin do not.

Canonical WntThe Wnt pathway is divided into at least two main branches, including the canonical (β-catenin-dependent) pathway and the non-canonical pathway(s), which is independent of β-catenin. In the canonical Wnt pathway, the Wnt ligand binds to a Frizzled/LRP heterodimer, which mediates the intracellular response. Exactly how these receptors activate the downstream pathway is still unclear, but it involves G-protein signalling, LRP phosphorylation on its cytosolic C terminus, and the activity of Dishevelled (Dsh). The net result of these factors is the disruption of a large protein machine called the β-catenin destruction complex — which phosphorylates β-catenin on its N terminus — causing it to be ubiquitylated and then rapidly degraded by the proteasome. The destruction complex is composed of many proteins, including the central scaffolding protein Axin (Masterblind in zebrafish), the Adenomatous polyposis coli (APC) protein, and the kinase Glycogen synthase kinase 3 (Gsk3). Several studies have suggested that Wnt signalling disrupts the complex by causing one or more proteins to leave the complex, preventing the phosphorylation and ubiquitylation of β-catenin. When β-catenin is not degraded, it accumulates and translocates to the nucleus, where it binds members of the Tcf/LEF1 family of DNA binding factors (including Headless (Hdl) in zebrafish), and recruits transcriptional activators to the promoter.

FGFBinding of the ligand to the Fibroblast growth factor (FGF) receptor (Fgfr) results in receptor dimerization and transphos-phorylation of the receptor’s cytosolic domain. The phosphorylated receptor recruits proteins that activate the G-protein Ras, which then activates the kinase Raf. Raf phosphorylates and activates Mek, which subsequently phosphorylates and activates MAP kinase (Mapk). Mapk enters the nucleus where it phosphorylates and activates target transcription factors.

R E V I E W S



YSL

VegT

Xenopus Zebrafish

Xnrs

sqt and cyc

Morpholinos Antisense oligonucleotides that are stable and are commonly used in zebrafish and Xenopus to inhibit either the translation or splicing of mRNAs.

Spemann’s organizer A signalling centre in amphibians that is created on the dorsal side of the late blastula embryo. The equivalent centre in fish is called the shield.

side of the late blastula embryo37,38. However, by the early gastrula stages, phosphorylated Smad2 levels are equal across the dorsal–ventral axis37,38. Therefore, there is a period of time during which Nodal signalling is asym-metrical across the dorsal–ventral axis, and this could be important for establishing asymmetrical gene expression along this axis.

In zebrafish, the Nodal gene sqt has been proposed to be directly activated by β-catenin, resulting in dorsal expression of sqt for a brief time before its expression throughout the mesendoderm39. Whether or not this causes a transient dorsal–ventral asymmetry in phos-phorylated Smad2 levels has not yet been determined. Another potential mechanism for regulating asym-metrical zebrafish nodal expression has recently been reported. Maternal sqt transcripts seem to be transported to the dorsal side of the zebrafish embryo as early as the 4-cell stage, presumably allowing the Sqt protein to accu-mulate dorsally before the onset of zygotic transcrip-tion40. Morpholino-mediated knockdown experiments indicate that a loss of maternal and zygotic Sqt causes a more severe phenotype than loss of the zygotic Sqt alone. However, embryos that genetically lack maternal and zygotic Sqt were reported to have no more severe defects than embryos that lack only zygotic Sqt41, and embryos that lack only maternal Sqt show no early pat-terning defects (S. Dougan, personal communication). Therefore, the importance of localized maternal sqt might depend on the genetic background of the embryo and/or epigenetic effects.

Several elegant experiments in Xenopus using Activin as a Nodal pathway stimulant have shown that variations in the level of Nodal activity can regulate dorsal–ventral patterning, with low levels inducing ventral fates and high levels inducing dorsal fates42–44. These results are consistent with the idea that Nodal activity is higher dor-sally than ventrally, at least in the pre-gastrula embryo. Nevertheless, it has been difficult to ascertain whether the Nodal pathway actually regulates dorsal–ventral

patterning. Studies in frogs using different doses of a syn-thetic Nodal-specific inhibitor (the truncated Cerberus protein) seemed to indicate that ventral mesoderm was most readily eliminated by small reductions in Nodal signalling, whereas dorsal mesoderm was only elimi-nated with higher doses of Nodal antagonist35. These are the only data so far that support the importance of a dorsal–ventral Nodal gradient in Xenopus. Similarly, no evidence has yet been presented that a reduction of Nodal signalling on the dorsal side results in a respeci-fication of dorsal to ventral fates. Surprisingly, reduction of Nodal signalling using various sqt and cyc mutants in zebrafish produced the opposite result; the dorsal region was the most sensitive to a loss of Nodal signal-ling45. In this case, reduced Nodal signalling resulted in a reduction in dorsal mesoderm formation but not a respecification to ventral mesodermal fates. Therefore, there is no compelling evidence at this point that a temporal/spatial gradient of Nodal signalling regulates differences in dorsal–ventral mesodermal fates in either fish or frogs.

Studies in zebrafish have indicated that Nodal gradi-ents are involved in patterning the mesoderm along the animal–vegetal axis. Differences in Nodal levels are pro-posed to separate head mesoderm from the notochord within the organizer14,45, and to separate ventricular and atrial myocardial precursors46. This concept is generally consistent with Xenopus studies showing that partial Nodal or Activin inhibition causes head defects without affecting the notochord47–50, and with the existence of a vegetal to animal gradient of phosphorylated Smad2, indicating that a Nodal gradient might also pattern the embryo along the animal–vegetal axis in frogs37,38. Further studies will need to determine to what extent differences in Nodal-signalling levels specify different mesodermal tissue fates in vivo.

The enigmatic Vg1The Xenopus Transforming growth factor-β (TGFB) fam-ily member Vg1 was originally identified as an excellent candidate to be the vegetally derived Nieuwkoop signal (REF. 51; reviewed in REF. 52). Vg1 transcripts are localized to the vegetal pole in a similar way to VegT. Moreover, the mature region of the Vg1 pro-protein attached to a dif-ferent pro-domain can function as a potent mesoderm-inducing agent. However, the originally identified Vg1 pro-protein has no activity and endogenous mature Vg1 is not detectable in Xenopus embryos. Although there is evidence that Vg1 has an important role in chick mesoderm induction53,54 and a Vg1-related pro-tein Growth/differentiation factor 3 (Gdf3) is essential for mesoderm formation in the mouse55, until recently Vg1 was shown to be crucial only for establishing left–right asymmetry in Xenopus56–58. However, Birsoy et al. showed that embryos with depleted Vg1 lack head and notochord structures, which is most likely due to a reduction in the expression of Spemann’s organizer genes, particularly the BMP and Wnt inhibitors15. Moreover, the authors resolved the issue of the inactive Vg1 pro-protein by demonstrating the existence of a second Vg1 allele, which is active as an inducing agent15.

Figure 2 | Models for activation of Nodal signalling in Xenopus and zebrafish. In Xenopus, the vegetally localized transcription factor VegT (blue) activates the transcription of the Xenopus nodal-related genes (Xnrs) in the vegetal hemisphere, which then initiate mesoderm formation. In zebrafish, signals from the yolk syncytial layer (YSL; blue) activate the transcription of the Nodal genes squint (sqt) and cyclops (cyc), which then initiate mesoderm formation.

R E V I E W S



Post-fertilization

Pre-gastrula

Blastula

Xenopus ZebrafishDorsalizingactivity

Nodalactivity

β-Catenin

Maternalsquint

?

Although this work has re-established Vg1 as an impor-tant member of the Nodal family of mesoderm induc-ers, it raises the question as to why so many members of this family are used in Xenopus (TABLE 1). In zebrafish, the vg1 transcript is maternal but not localized59, and so far there is no evidence that it is involved in either mesoderm induction or left–right patterning. Therefore, if zebrafish Vg1 is essential for Nodal signalling as in frogs, chicks and mice, it either functions as a ubiquitous factor or the processing of the pro-protein to the mature ligand is spatially regulated.

Beyond Nodals: maintenance and patterningEvidence from both fish and frogs demonstrates that the head (anterior) is regulated differently from the trunk and tail (posterior). For example, inhibition of FGF signalling in both systems eliminates the trunk and tail without causing severe head defects, indicating that these two regions of the body use different regulatory circuits60,61 (FIG. 4). Mesoderm induction in the head

activates various transcription factors and secreted inhibitors of the BMP, Wnt and Nodal pathways that are essential for patterning the embryo33. Meanwhile, trunk mesoderm induction activates signalling factors of the FGF and Wnt families (BOX 1; TABLE 1), along with several transcription factors. Among the transcription factors are at least three members of the T-box family, Xbra (in frogs) or No tail (Ntl; in fish), VegT (in frogs) or Spadetail (Spt; in fish), and Tbx6 (in fish and frogs), which func-tion combinatorially to regulate mesoderm formation62. Whereas zebrafish that lack spt (also known as tbx16) or ntl function have major defects within the trunk or tail, respectively, fish that lack both T-box genes fail to form the trunk and tail mesoderm, demonstrating that these factors function early and redundantly in the initial stages of mesoderm induction63. Among the targets of the T-box genes are FGFs, which function in an autoregulatory loop to maintain T-box gene expres-sion64–66 (FIG. 4). Zygotic Wnt signalling not only limits the size of the Spemann’s organizer67, it is also necessary for the maintained expression of the T-box genes68–70 (FIG. 4). Why both FGFs and Wnts are needed to maintain the mesoderm is still not clear.

In addition to Nodals, Wnts and FGFs, several mem-bers of the BMP family and their inhibitors are involved in mesoderm patterning (REFS 71,72; reviewed in REF. 73) (BOX 1; TABLE 1). Surprisingly, both the dorsal and ven-tral sides of the embryo simultaneously express BMPs and BMP inhibitors, indicating that the regulation of mesodermal patterning by BMPs is complex. Zebrafish embryos that are mutant for bmp2b or frog embryos with depleted Bmp4 and Bmp7 fail to form tails26,74–76, which is consistent with the proposal that BMPs are essential regulators of tail development77,78 (FIG. 4).

Many lines of evidence indicate that the embryonic mesoderm of frogs and fish is subdivided into three principal domains: the head, trunk and tail. As dis-cussed above, the trunk and tail require FGF signalling and functional T-box genes, whereas the head does not. The trunk and tail are also under separate regulatory control, although the regulation of tail formation seems to be different between fish and frogs. In Xenopus, the initiation of tail outgrowth involves the BMP and Notch signalling pathways, coupled with distinct changes in gene expression as the tail forms79–82. Although BMP signalling is also required for tail formation in zebrafish75,76,78, there is no evidence for Notch signal-ling in this process, nor are there any reported changes in gene expression when the tail begins to develop. However, zebrafish tail formation clearly involves a pathway that is distinct from head or trunk formation, because zebrafish embryos with deficient Nodal signal-ling form essentially normal tails, even though the head and trunk mesoderm are absent83.

Depletion of the dorsally expressed BMP ligand Admp (Anti-dorsalizing morphogenetic protein) in Xenopus and zebrafish also results in embryos with defects in posterior development84–86. Moreover, an important new analysis of Xenopus embryos that lack dorsal (Admp) and ventral (Bmp2, Bmp4 and Bmp7) BMPs revealed that interactions between the BMPs and

Figure 3 | Establishing the dorsal–ventral axis. In Xenopus (left panels) a dorsalizing activity moves from the vegetal pole to one side of the embryo after fertilization, and a similar mechanism seems to occur in zebrafish (right panels). During the blastula stages, this dorsalizing activity stabilizes β-catenin on what will be the future dorsal side of the embryo. When zygotic transcription of the Xenopus Nodal genes begins, the β-catenin creates an asymmetry in Nodal expression, which results in elevated phosphorylated Smad2 levels on the dorsal side of the pre-gastrula embryo. In zebrafish, maternal squint transcripts are localized to the future dorsal side by the 4-cell stage, and zygotic squint is initially transcribed dorsally. It is not yet clear whether this creates a significant asymmetry in phosphorylated Smad2 levels.

R E V I E W S



Maternalβ-catenin

T-box genes

BMP

HighNodal

Low Nodal (frog)Low Nodal/YSL signal (fish)

MediumNodal

FGFWnt

Wnt

T-box genes

TailTrunkHead

FGF

their inhibitors on both the dorsal and ventral sides of the embryo are essential for establishing normal dorsal–ventral patterning86,87.

It is still unclear how the posterior dorsal (tail mesoderm), posterior ventral (the posterior blood islands) and ventral (pronephros) fates (FIG. 1) are specified. Although BMP signalling is necessary for the formation of the ventral mesodermal fates, there is no compelling evidence that a gradient of BMP sig-nalling distinguishes between the different tissues that form on the ventral side. This remains an important area of investigation.

Temporal aspects of mesoderm inductionMesoderm induction occurs over several hours during which embryonic cells are exposed to shifting concentra-tions of inducing signals and their inhibitors. Even if a cell begins to head towards a particular fate decision, it might change fate if it receives new external influences. But once a cell becomes irreversibly committed to a specific fate it is said to be determined.

An important area of research involves understanding when cells are able to respond to the different mesoderm-inducing signals and when they become committed to a particular fate. A detailed analysis of cell commitment in Xenopus showed that cells in the early gastrula embryo are labile, expressing genes that mark multiple germ lay-ers and regions88. By the end of gastrulation, however, cells seem to be more committed to a specific fate, as judged by gene expression. How and when cells become committed is still mostly unknown. One mechanism that limits the ability of Nodal-type signals to affect the mesoderm after the mid-gastrula stage in Xenopus involves phosphorylation of residues in the middle of the Smad2 protein, which keeps Smad2 out of the nucleus and therefore prevents it from activating gene expression89.

Many studies in Xenopus and zebrafish have exam-ined how signalling pathways influence embryonic gene expression by using overexpression of specific signal-ling molecules, overexpression of dominant-negative

receptors, morpholino oligonucleotides and zebrafish embryos that are mutant for one component of a par-ticular signalling pathway. Although these results have been informative, they do not provide information about the important temporal requirement for the signalling pathway as all of these approaches alter the signalling pathway throughout the entire process of mesoderm induction, and therefore there is not much known in this area in general. Fortunately, new approaches are being developed that will allow the time requirement for different signalling factors to be addressed. Relatively specific drugs that inhibit the FGF (SU5402 (REF. 90)) and Nodal (SB505124 (REF. 91)) pathways are commercially available. For example, the FGF inhibitor has been used to examine a wide vari-ety of processes, including the regulation of the T-box genes by FGF in the zebrafish mesoderm, Xenopus neural induction, Xenopus forebrain development, and zebrafish tooth development64,92–94.

As an alternative approach, transgenic zebrafish that express inhibitors of the canonical Wnt and BMP pathways under heat-shock control have been devel-oped95,96. For example, this was used to show that the major role for BMP signalling in patterning the meso-derm occurs during the late blastula/early gastrula period, but that BMP signalling continues to operate during the mid–late gastrula stages to specify ventral tail-fin formation95. Similarly, studies with secreted natural inhibitors of FGF signalling in frogs indicated that the FGF-signalling pathway functions during the gastrula stages primarily to regulate mesodermal gene expression, whereas FGF signalling largely regulates morphogenesis after the gastrula stages97. These types of temporal analysis provide a framework for examin-ing changes in mesodermal gene expression in order to understand how and why the mesoderm changes its ability to respond to inducing signals.

Combinatorial signallingMesodermal cells respond to a cacophony of signals, but somehow they need to integrate this information to make a specific fate decision. How the different sig-nalling pathways combine to regulate gene expression in the mesoderm is still poorly understood. Different signalling pathways might to some extent regulate each other’s intracellular networks but, most likely, much of the interaction occurs through changes in gene expression. Although there is some evidence for genes that are directly regulated by multiple signal-ling pathways using promoter analysis (for example, Wnt and Nodal signalling regulating the Xenopus twin promoter98 and BMP and Wnt signalling regulating the zebrafish tbx6 promoter11), for the most part it is not known how many mesodermal genes are regulated by just one pathway and how many are regulated by more than one pathway. Promoter analysis is feasible in zebrafish and Xenopus using injected plasmid DNA, although it is best done using technology that allows transgenic Xenopus embryos to be produced in the F0 generation99; nevertheless, these methods are still laborious. Moreover, with the exception of binding

Figure 4 | A model for patterning the embryonic body in Xenopus and zebrafish. The fish and frog body (shown here as a generic embryo) is divided into three principal domains: the head, trunk and tail, and some of the key signalling events are shown. In zebrafish, the yolk syncytial layer (YSL) signal seems to be redundant, with Nodal signalling inducing the formation of the tail. Note that the role of Wnt signalling within the trunk is uncertain. BMP, Bone morphogenetic protein; FGF, Fibroblast growth factor.

R E V I E W S



Nodal

Cerberus

Hypoblast

Cerberus

Vg1 andWnt8c

Vg1 andWnt8c

Primitivestreak

Nodal

AVENodal inhibitorsWnt inhibitors

5.5 DPC6.5 DPC

PosteriorAnterior

Proximal

Distal

a

b

Brachyury

Nodal

Wnt3Wnt3

AVE

Primitive streak The site of major morphogenetic movements during gastrulation in reptiles, birds and mammals. The mesoderm, as well as the endoderm, moves through this structure as it ingresses.

sites for the Tcf factor (a transcriptional activator in the Wnt pathway; BOX 1), it is still not possible to easily recognize elements that respond to the FGF, Nodal or BMP pathways because each of these pathways regulate several DNA-binding proteins and it is typically not clear which binding protein interacts with a specific promoter. For these reasons, few mesodermal promoters have been studied in any detail, leaving much of the molecular network of mesoderm induction incomplete.

A recent paper has brought to light a new inter-esting example of combinatorial signalling in the mesoderm, which would not have been obviously revealed by promoter studies. As discussed above, the T-box transcription factor Xbra is a crucial regulator of posterior mesoderm formation. The specificity of Xbra-mediated regulation of gene expression is altered by its binding to Smad1, an intracellular mediator of the BMP pathway100 (BOX 1). Because BMP signalling is high ventrally and low dorsally, targets of Xbra-mediated activation are likely to be different on the dorsal and ventral sides of the embryo. It will be interesting

to see how general this type of mechanism is for regu-lating the activity of transcription factors in different regions of the embryo.

Mesoderm induction in chicks and miceThe basic process of mesoderm induction is generally conserved among all vertebrates, although the basic embryonic architecture is different between species (for a more extensive discussion comparing mesoderm induc-tion in chick and mouse to fish and frogs see REF. 3). A direct comparison of chicks and mice to fish and frogs reveals interesting commonalities and differences. For example, in chicks, a combination of the TGFB factor Vg1 and canonical Wnt (Wnt8c) signalling has been proposed to initiate the formation of the initial axial structure, the primitive streak, albeit with the caveat that there are no loss-of-function data yet that test the exact role of Vg1 signalling in chicks54 (FIG. 5a). Intriguingly, one of the earliest targets of Vg1 and Wnt8c is the chick Nodal gene101, which parallels the observations in fish and frogs that Nodal expression is a key step in the

Figure 5 | Mesoderm induction in the chick and mouse. a | In the chick embryo, Vg1 and Wnt8c cooperate to activate Nodal expression. Nodal activity is blocked initially by the secreted Nodal-binding protein Cerberus, which is expressed in the underlying hypoblast (green circles). The movement of the endoblast (white circles) displaces the Cerberus-expressing hypoblast, allowing Nodal to function, which causes mesoderm and endoderm cells to ingress through the primitive streak. Modified with permission from REF. 103 © (2004) Company of Biologists Ltd. b | In 5.5 day postcoitum (DPC) mouse embryos, Wnt3 is initially expressed in the posterior extra-embryonic tissue (the visceral endoderm). The anterior visceral endoderm (AVE) initially forms in the distal region of the embryo. By 6.5 DPC, the AVE, which secretes Nodal and Wnt inhibitors, has moved to the anterior side. Wnt3 is now expressed in the posterior epiblast, where the primitive streak forms, and in the posterior visceral endoderm. Wnt3 activates the expression of genes such as Nodal and Brachyury in the posterior epiblast. Nodal is also expressed throughout the epiblast at 5.5 DPC (not shown). Modified with permission from REF. 105 © (2005) Academic Press.

R E V I E W S



Epiblast The portion of the mouse embryo that will become the definitive embryo (as opposed to extra-embryonic tissues).

Amniotes Include reptiles, birds and mammals, which all have a protective membrane (the amnion) surrounding the embryo that prevents it from desiccating.

mesoderm-induction process (see above). However, an important difference is that the chick extra-embryonic structure called the hypoblast, which initially lies under the future primitive streak, secretes the Nodal inhibitor Cerberus, which blocks Nodal function102. Only when the hypoblast is physically displaced away from the pos-terior region of the embryo, where the primitive streak forms, is Nodal able to function in forming the primitive streak, together with Fgf8 and Chordin103 (FIG. 5a). This mechanism of hypoblast displacement is proposed to be essential for ensuring that the primitive streak forms only in the posterior region of the embryo102,103. It is also possible that the gradual movement of Nodal inhibitors away from the primitive streak establishes a gradient of Nodal activity within the streak itself102.

In mice, understanding mesoderm induction is com-plicated by the fact that the primitive streak forms on one side of the embryo (proximal to the site of implan-tation and in the region of the future posterior end of the embryo), whereas a separate crucial signalling centre, the anterior visceral endoderm (AVE), initially forms distal to the site of implantation and then moves to the opposite side from the primitive streak (the future anterior end of the embryo; FIG. 5b). A complex series of signalling interactions that involve Nodal, BMP and Wnt signalling sets up these two centres (reviewed in REF. 104).

Gdf3, a Vg1-like gene, is expressed in the very early mouse embryo, paralleling the very early expression of Vg1 in frogs, fish and chick. As with zebrafish vg1, but unlike Vg1 in frogs and chicks, Gdf3 expression is uni-form within the cells that will give rise to the embryo55. Surprisingly, Gdf3-null mutant mice are often viable, with approximately one-third showing a range of mor-phological defects, which can be explained by alterations in Nodal signalling55. In general terms this result fits with loss-of-function experiments in frogs, indicating that Xenopus Vg1 works together with the Nodal factors15, and it differs from studies in chicks, indicating that Vg1 is the key initiator of Nodal expression and primitive streak formation53,54,103. These differences might reflect the ways in which the use of these separate TGFB factors has evolved to establish the mesoderm in embryos that have different topologies.

A recent study challenges the view that the mouse primitive streak is initially established in a radial pat-tern and then refined to a localized site in the proximal posterior region of the epiblast, potentially through inhibitory activities of the AVE105. Instead, the authors find that Wnt3, which is essential for the formation of the primitive streak and mesoderm and for Nodal expression106, is initially expressed in tissues next to the proximal posterior region of the epiblast before the AVE has moved into the future anterior region, and not in a radial pattern throughout the proximal region as was previously thought. Wnt3 then regulates downstream genes such as Nodal and the T-box gene Brachyury (and potentially Wnt3 itself) in the adjacent proximal poste-rior epiblast (FIG. 5b). How Wnt3 expression is initially restricted to the proximal posterior region remains unanswered.

In an interesting parallel to the function of the chick hypoblast, the AVE secretes inhibitors of Nodal signalling, Cerberus-like and Lefty1, that are essential for restricting the primitive streak to the posterior end of the embryo by limiting the region in which Nodal signalling can function107,108 (FIG. 5b). Moreover, these Nodal inhibitors are essential for the normal patterning of the streak because mutants that lack both Nodal inhib-itors have an expansion of the anterior mesoderm and a concomitant loss of posterior mesodermal tissues108. The AVE also expresses inhibitors of Wnt signalling109, which also limits the domain of Wnt3 signalling. Therefore, an essential role of the AVE is to secrete inhibitors that restrict the primitive streak to one side of the embryo.

Because different experimental approaches are often used in different species, a direct comparison of meso-derm induction between vertebrates is complicated. Nonetheless, it is possible to make some basic gener-alizations. First, the same families of signalling factors are involved in all vertebrates, as are the T-box genes, although there are some significant differences in which molecular players are involved (TABLE 1). Second, localized maternal determinants have no role or a reduced role in axis specification and mesoderm induction in amniotes, and the polarity of the amniote embryo is not fixed until the beginning of gastrulation. Third, extra-embryonic structures have an important role in the early intercellular signalling processes in amniotes, whereas the fish extra-embryonic structure, the YSL, initiates only the Nodal pathway, and frogs do not have extra-embryonic struc-tures. Finally, the effect of inhibiting signalling pathways in the mouse embryo produces much more severe effects than in fish and frogs. For example, inhibition of BMP signalling110,111 or elimination of Wnt3 (REF. 106) in mouse embryos causes a failure of mesoderm formation, whereas inhibition of these pathways in fish and frogs results only in truncations of the body. Schier and Talbot have sug-gested that the differences in the roles of the signalling factors in these different species are due to a much greater role for cross-regulation between the signalling pathways in mouse embryos than in fish and frogs, so that if one pathway is inhibited, the others are affected as well26.

Future directionsMany studies on mesoderm induction have involved identifying a new gene either through an unbiased screen or homology searches that are based on what is known from other species, and analysing its regula-tion by the signalling pathways that are discussed in this review. An alternative approach involves altering one of the signalling pathways by molecular or genetic methods and then analysing the expression and func-tion of a handful of marker genes, typically by RT-PCR. Although the zebrafish and Xenopus genomes are still not completely sequenced, it has recently become possible to use genomic methods, such as microar-ray screens and genome-based expression screens, to analyse mesoderm induction. For example, Xenopus embryos that lack the T-box transcription factor VegT were examined using microarrays, identifying 99 known and novel genes that are activated by VegT, of

R E V I E W S



Heart and cartilage

Notochord

Pronephros

Muscle

Lymphocyte

Erythrocyte

Vasculartissue

0.1–1 ng ml–1 Activin + Scf

0.1–1 ng ml–1 Activin + Ang2

0.1–1 ng ml–1 Activin + Il3

+ 5–10 ng ml–1 Activin

+ 10–50 ng ml–1 Activin + RA

+ 10–50 ng ml–1 Activin

+ 100 ng ml–1 Activin

which 13 were shown to be direct targets of VegT within the mesendoderm112. Among these direct VegT targets, the transcription factors Snail, Hesr1 (hairy/enhancer-of-split related) and Esr4 (enhancer-of-split related 4) were shown to be essential for proper embryonic morphology. A similar approach using the drug SU5402 to block FGF signalling in Xenopus identified 43 FGF-regulated genes, of which 26 were novel113. One of the 43 genes (Xmig6) was shown to be required for muscle dif-ferentiation, whereas a novel G-protein coupled receptor (encoded by Xgpcr4) seems to be involved in the mor-phogenesis of gastrulation. These approaches demonstrate the value of microarrays for finding novel genes that are regulated by the signalling and transcription factors that are involved in mesoderm induction, and for establishing regulatory relationships between known genes.

Although initial experiments are likely to identify targets when a pathway is activated or blocked through-out the entire mesoderm-induction period (as with the elimination of VegT in the above example), the develop-ment of transgenic fish and frog embryos that contain inducible activators and inhibitors, as well as the use of specific pharmacological reagents, will facilitate the genome-wide examination of changes in gene expres-sion when a signalling pathway is activated or blocked at a specific time in development. Because the functions of the different signalling factors are constantly chang-ing over developmental time as discussed above, the use of microarray analysis combined with the ability to temporally regulate each of the signalling pathways will show how the transcriptional network changes to allow each of the signals to regulate different processes as embryogenesis proceeds.

Microarray analysis is not without its drawbacks. One of the problems is that when whole embryos are analysed, many non-mesodermal cells contribute to the signal. For

genes that are expressed broadly throughout the meso-derm this might not be a problem, but many mesodermal fates are represented by only a small subset of cells from the entire embryo. One probable solution is to return to Nieuwkoop’s original assay. When animal caps are removed from embryos and treated with different fac-tors such as Activin, which activates the Nodal signalling pathway, different cell fates are induced depending on the concentration of the factors used (reviewed in REF. 114) (FIG. 6). Although the cell populations in these explants typically are not made up of single cell types, they are enormously enriched in a particular tissue compared with the whole embryo. Moreover, because the explants can be maintained in culture for up to 2 months115, not only can the early events in mesoderm induction be examined, but longer-term changes in gene expression can also be analysed. Although animal cap explants have occasionally been used in zebrafish, they are technically more challenging and it is not yet clear whether the same approach will be useful. In zebrafish the use of transgenic lines that express a fluorescent protein from a cell-type specific promoter combined with fluorescent cell sorting might be a more feasible alternative.

Because mesoderm induction happens early in embryogenesis, it has been particularly easy to manipu-late this process by injecting pools of synthetic RNAs into Xenopus embryos to identify genes that regulate mesoderm formation. Previous attempts used pooled libraries of cDNA that is attached to a promoter for the SP6 RNA polymerase (reviewed in REF. 116). Although some interesting genes have been identified in this way, there were several problems with this approach. Even in the most highly normalized library many cDNAs, espe-cially abundant ones, were overrepresented. In addition, particularly for longer mRNAs, many cDNAs are not full length. Because of these problems, it was necessary to

Figure 6 | Production of different mesodermal cell types from animal cap explants. Animal caps are excised from Xenopus embryos at the mid-blastula stage and cultured in a saline solution. In the absence of inducers, the explants will form primitive epidermis. Addition of different factors will enhance the formation of specific mesodermal cell types. RA, Retinoic acid; Il3, Interleukin 3; Scf, Stem cell factor; Ang2, Angiopoietin 2. Modified with permission from REF. 114 © (2003) Elsevier Science.

R E V I E W S



inject relatively large pools of synthetic RNAs to sample enough different cDNAs. Because there is a limit to the amount of RNA that can be injected into an embryo, these methods select for genes that produce an effect at very low doses, which might explain why many of the genes found this way encode signalling factors.

With increasing information from large-scale sequencing projects, it is now possible to produce a library that contains full-length clones of many genes to test them in small pools by overexpression in the early embryo116,117. One large-scale approach of this type identified 64 Xenopus genes that affected the mesoderm, from genes that caused defects in the morphogenesis of gastrulation to those that produced an extra axis117. Although some of these genes are well-known factors such as VegT, others are either novel or have not been studied in the context of mesoderm induction before.

Given the amount of data that are generated using both classical and genomics-based approaches it is increasingly important to develop databases to con-nect and cross reference the data. One valuable effort in Xenopus called the Xenopus Mesendoderm Network shows the known molecular interactions as a gene-regulatory circuit, which can be constantly modified as new data become available118. Because data are of vari-able reliability and sometimes conflict, curating such a database is a major challenge, but it still represents the best approach for understanding the underlying cir-cuitry in the mesoderm. Other databases with searchable gene-expression patterns not only allow all genes with a mesodermal expression pattern to be found, but also provide connections to the relevant literature (Axeldb for Xenopus and ZFIN for zebrafish).

ConclusionsThe field of mesoderm induction has advanced a great deal in the almost 40 years since its origin, identifying the key signalling factors and establishing the groundwork for understanding their roles in forming the mesoderm. Genomic approaches will probably identify additional key players downstream of these signals, including those that are involved not only in cell-fate decisions but also in the equally essential process of mesodermal morpho-genesis, which transforms the spherical embryo into the final embryonic body plan.

Because the mesoderm gives rise to many cell fates and because it involves a wide variety of morphogenetic movements, it continues to provide a wealth of important

embryological problems to study. With the zebrafish and Xenopus genomes to be completed in the near future, and the use of microarrays, morpholinos, expression screens and zebrafish genetic screens, most if not all the genes involved in this process will probably soon be identified. The future challenge will be to understand how these genes interact to regulate cell fate and morphogenesis.

One of the main goals will be to understand how the signalling factors interact to activate specific patterns of gene expression within the mesoderm, and how the roles of the signalling factors can change rapidly over developmental time. The development of methods to temporally regulate each of the signalling pathways within the embryo will be useful in this regard, as will the production of transgenic lines that contain the regu-latory elements of specific marker genes that drive the expression of a fluorescent protein, which provides the ability to examine gene expression in living embryos (and using different colours of fluorescent proteins, mul-tiple marker genes can be analysed simultaneously). The large wealth of data that will come from the sequencing projects, along with improved bioinformatics methods for analysing microarray data and finding candidate transcription-factor binding sites, will also be useful in determining how combinatorial gene regulation occurs within the mesoderm.

A second goal will be to understand how the different signalling factors coordinate cell fate and morphogenesis, which will be greatly aided by the increasing knowledge of the intracellular networks used by each of the signal-ling pathways that are involved in mesoderm induction, and by the ability to precisely manipulate different aspects of these intracellular pathways in the embryo. Moreover, the ability to analyse cells in vivo using fluo-rescent proteins that are expressed within specific sub-sets of mesodermal cells, along with improved confocal microscopes, will allow the behaviour of cells within the embryo to be analysed in a much more sophisticated way than has been possible before.

Finally, by comparing the results that have been obtained in Xenopus and zebrafish to those from chick and mouse studies, it will be possible to understand which aspects of mesoderm induction are common among verte-brates, and which have changed during evolution to allow for the specific needs of each species. Building on the clas-sical foundations and using the new methodologies, the mesoderm-induction field will continue to produce new insights into many important embryological questions.

1. Nieuwkoop, P. D. The formation of the mesoderm in urodelean amphibians I. The induction by the endoderm. W. Roux’ Arch. Ent. Org. 162, 341–373 (1969).This paper provided crucial evidence that the mesoderm in amphibians is produced by an inductive event.

2. Niehrs, C. Regionally specific induction by the Spemann–Mangold organizer. Nature Rev. Genet. 5, 425–434 (2004).

3. Kimelman, D. & Bjornson, C. in Gastrulation (ed. Stern, C. D.) 363–372 (Cold Spring Harbor Laboratory Press, Cold Spring Harbor, 2004).

4. Harland, R. M. in Gastrulation (ed. Stern, C. D.) 373–388 (Cold Spring Harbor Laboratory Press, Cold Spring Harbor, 2004).

5. Warga, R. M. & Nüsslein-Volhard, C. Origin and development of the zebrafish endoderm. Development 126, 827–838 (1999).

6. Kimmel, C. B., Warga, R. M. & Schilling, T. F. Origin and organization of the zebrafish fate map. Development 108, 581–594 (1990).

7. Wardle, F. C. & Smith, J. C. Refinement of gene expression patterns in the early Xenopus embryo. Development 131, 4687–4696 (2004).

8. Slack, J. M. W., Darlington, B. G., Heath, J. K. & Godsave, S. F. Mesoderm induction in earlyXenopus embryos by heparin-binding growth factors. Nature 326, 197–200 (1987).

9. Kimelman, D. & Kirschner, M. Synergistic induction of mesoderm by FGF and TGF-β and the identification of an mRNA coding for FGF in the early Xenopus embryo. Cell 51, 869–877 (1987).Together with reference 8, these studies were the first to identify a specific protein as a mesoderm-inducing agent, in this case FGF.

10. Kelly, G. M., Greenstein, P., Erezyilmaz, D. F. & Moon, R. T. Zebrafish wnt8 and wnt8b share a common activity but are involved in distinct developmental pathways. Development 121, 1787–1799 (1995).

11. Szeto, D. P. & Kimelman, D. Combinatorial gene regulation by Bmp and Wnt in zebrafish posterior mesoderm formation. Development 131, 3751–3760 (2004).

R E V I E W S



12. Dale, L., Howes, G., Price, B. M. J. & Smith, J. C. Bone morphogenetic protein 4: a ventralizing factor in early Xenopus development. Development 115, 573–585 (1992).

13. Köster, M. et al. Bone morphogenetic protein 4 (BMP-4), a member of the TGF-β family in early embryos of Xenopus laevis: analysis of mesoderm inducing activity. Mech. Dev. 33, 191–200 (1991).

14. Gritsman, K., Talbot, W. S. & Schier, A. F. Nodal signaling patterns the organizer. Development 127, 921–932 (2000).

15. Birsoy, B., Kofron, M., Schaible, K., Wylie, C. & Heasman, J. Vg1 is an essential signaling molecule in Xenopus development. Development 133, 15–20 (2006).

16. Smith, J. C. A mesoderm-inducing factor is produced by a Xenopus cell line. Development 99, 3–14 (1987).A seminal paper in modern mesoderm-induction research, which shows that a semi-purified factor produced by a specific cell line induces mesoderm in Xenopus animal cap explants.

17. Sivak, J. & Amaya, E. in Gastrulation (ed. Stern, C. D.) 463–474 (Cold Spring Harbor Laboratory Press, Cold Spring Harbor, 2004).

18. Munoz-Sanjuan, I. & Hemmati-Brivanlou, A. H. in Gastrulation (ed. Stern, C. D.) 475–490 (Cold Spring Harbor Laboratory Press, Cold Spring Harbor, 2004).

19. Houston, D. W. & Wylie, C. in Gastrulation (ed. Stern, C. D.) 521–538 (Cold Spring Harbor Laboratory Press, Cold Spring Harbor, 2004).

20. King, M. L., Messitt, T. J. & Mowry, K. L. Putting RNAs in the right place at the right time: RNA localization in the frog oocyte. Biol. Cell 97, 19–33 (2005).

21. Hyde, C. E. & Old, R. W. Regulation of the early expression of the Xenopus nodal-related 1 gene, Xnr1. Development 127, 1221–1229 (2000).

22. Clements, D., Friday, R. V. & Woodland, H. R. Mode of action of VegT in mesoderm and endoderm formation. Development 126, 4903–4911 (1999).

23. Kofron, M. et al. Mesoderm induction in Xenopus is a zygotic event regulated by maternal VegT via TGFβ growth factors. Development 126, 5759–5770 (1999).

24. Xanthos, J. B., Kofron, M., Wylie, C. & Heasman, J. Maternal VegT is the initiator of a molecular network specifying endoderm in Xenopus laevis. Development 128, 167–180 (2001).

25. Zhang, J. et al. The role of maternal VegT in establishing the primary germ layers in Xenopus embryos. Cell 94, 515–524 (1998).Demonstration that the T-box transcription factor VegT is essential for the formation of the mesoderm (and endoderm) in Xenopus embryos.

26. Schier, A. F. & Talbot, W. S. Molecular genetics of axis formation in zebrafish. Annu. Rev. Genet. 39, 561–613 (2005).

27. Griffin, K. J. P., Amacher, S. L., Kimmel, C. B. & Kimelman, D. Molecular identification of spadetail: regulation of zebrafish trunk and tail mesoderm formation by T-box genes. Development 125, 3379–3388 (1998).

28. Weaver, C. & Kimelman, D. Move it or lose it: axis specification in Xenopus. Development 131, 3491–3499 (2004).

29. Mizuno, T., Yamaha, E., Kuroiwa, A. & Takeda, H. Removal of vegetal yolk causes dorsal deficencies and impairs dorsal-inducing ability of the yolk cell in zebrafish. Mech. Dev. 81, 51–63 (1999).

30. Ober, E. A. & Schulte-Merker, S. Signals from the yolk cell induce mesoderm, neuroectoderm, the trunk organizer, and the notochord in zebrafish. Dev. Biol. 215, 167–181 (1999).

31. McMahon, A. P. & Moon, R. T. Ectopic expression of the proto-oncogene int-1 in Xenopus embryos leads to duplication of the embryonic axis. Cell 58, 1075–1084 (1989).This work provided the first evidence that the Wnt signalling pathway has a crucial role in the formation of the Xenopus embryonic axis.

32. Heasman, J. et al. Overexpression of cadherins and underexpression of β-catenin inhibit dorsal mesoderm induction in early Xenopus embryos. Cell 79, 791–803 (1994).

33. DeRobertis, e. M., Larrain, J., Oelgeschlager, M. & Wessely, O. The establishment of Spemann’s organizer and the patterning of the vertebrate embryo. Nature Genet. 1, 171–181 (2000).

34. Yang, J., Tan, C., Darken, R. S., Wilson, P. A. & Klein, P. S. β-Catenin/Tcf-regulated transcription prior to the midblastula transition. Development 129, 5743–5752 (2002).

35. Agius, E., Oelgeschlager, M., Wessely, O., Kemp, C. & De Robertis, E. M. Endodermal Nodal-related signals and mesoderm induction in Xenopus. Development 127, 1173–1183 (2000).

36. Takahashi, S. et al. Two novel nodal-related genes initiate early inductive events in Xenopus Nieuwkoop center. Development 127, 5319–5329 (2000).

37. Schohl, A. & Fagotto, F. β-Catenin, MAPK and Smad signaling during early Xenopus development. Development 129, 37–52 (2002).

38. Faure, S., Lee, M. A., Keller, T., ten Dijke, P. & Whitman, M. Endogenous patterns of TGFβ superfamily signaling during early Xenopus development. Development 127, 2917–2931 (2000).

39. Feldman, B. et al. Zebrafish organizer development and germ-layer formation require nodal-related signals. Nature 395, 181–185 (1998).

40. Gore, A. V. et al. The zebrafish dorsal axis is apparent at the 4-cell stage. Nature 438, 1030–1035 (2005).

41. Aoki, T. O. et al. Regulation of Nodal signalling and mesendoderm formation by TARAM-A, a TGFβ-related type I receptor. Dev. Biol. 241, 273–288 (2002).

42. Green, J. B. A., Howes, G., Symes, K., Cooke, J. & Smith, J. C. The biological effects of XTC-MIF: quantitative comparison with Xenopus bFGF. Development 108, 173–183 (1990).

43. Green, J. B. A., New, H. V. & Smith, J. C. Responses of embryonic Xenopus cells to activin and FGF are separated by multiple dose thresholds and correspond to distinct axes of the mesoderm. Cell 71, 731–739 (1992).

44. McDowell, N. & Gurdon, J. B. Activin as a morphogen in Xenopus mesoderm induction. Semin. Cell Dev. Biol. 10, 311–317 (1999).A review of the important studies, particularly from the laboratories of John Gurdon, Jim Smith and Doug Melton, using Activin to study the mechanisms of mesoderm induction, including the role of Nodal family members as morphogens.

45. Dougan, S. T., Warga, R. M., Kane, D. A., Schier, A. F. & Talbot, W. S. The role of the zebrafish nodal-related genes squint and cyclops in patterning of mesendoderm. Development 130, 1837–1851 (2003).

46. Keegan, B. R., Meyer, D. & Yelon, D. Organization of cardiac chamber progenitors in the zebrafish blastula. Development 131, 3081–3091 (2004).

47. Piepenburg, O., Grimmer, D., Williams, P. H. & Smith, J. C. Activin redux: specification of mesodermal pattern in Xenopus by graded concentrations of endogenous activin B. Development 131, 4977–4986 (2004).

48. Osada, S. I. & Wright, C. V. Xenopus nodal-related signaling is essential for mesendodermal patterning during early embryogenesis. Development 126, 3229–3240 (1999).

49. Piccolo, S. et al. The head inducer Cerberus is a multifunctional antagonist of Nodal, BMP and Wnt signals. Nature 397, 707–710 (1999).

50. Houston, D. W. & Wylie, C. Maternal Xenopus Zic2 negatively regulates Nodal-related gene expression during anteroposterior patterning. Development 132, 4845–4855 (2005).

51. Weeks, D. L. & Melton, D. A. A maternal mRNA localized to the vegetal hemisphere in Xenopus eggs codes for a growth factor related to TGF-β. Cell 51, 861–867 (1987).The first identification of an endogenous TGFB factor in mesoderm induction in Xenopus.

52. Kessler, D. S. in Gastrulation (ed. Stern, C. D.) 505–520 (Cold Spring Harbor Laboratory Press, Cold Spring Harbor, 2004).

53. Shah, S. B. et al. Misexpression of chick Vg1 in the marginal zone induces primitive streak formation. Development 124, 5127–5138 (1997).

54. Skromne, I. & Stern, C. D. Interactions between Wnt and Vg1 signalling pathways initiate primitive streak formation in the chick embryo. Development 128, 2915–2927 (2001).

55. Chen, C. et al. The Vg1-related protein Gdf3 acts in a Nodal signaling pathway in the pre-gastrulation mouse embryo. Development 133, 319–329 (2005).

56. Hyatt, B. A. & Yost, H. J. The left–right coordinator: the role of Vg1 in organizing left–right axis formation. Cell 93, 37–46 (1998).

57. Ramsdell, A. F. & Yost, H. J. Cardiac looping and the vertebrate left–right axis: antagonism of left-sided Vg1 activity by a right-sided ALK2-dependent BMP pathway. Development 126, 5195–5205 (1999).

58. Kramer, K. L. & Yost, H. J. Ectodermal syndecan-2 mediates left–right axis formation in migrating mesoderm as a cell-nonautonomous Vg1 cofactor. Dev. Cell 2, 115–124 (2002).

59. Helde, K. A. & Grunwald, D. J. The DVR-1 (Vg1) transcript of zebrafish is maternally supplied and distributed throughout the embryo. Dev. Biol. 159, 418–426 (1993).

60. Amaya, E., Musci, T. J. & Kirschner, M. W. Expression of a dominant negative mutant of the FGF receptor disrupts mesoderm formation in Xenopus embryos. Cell 66, 257–270 (1991).The first use of a dominant-negative receptor to study mesoderm induction provided crucial evidence for a role of FGF signalling in forming the posterior body.

61. Griffin, K. J. P., Patient, R. K. & Holder, N. H. K. Analysis of FGF function in normal and no tail zebrafish embryos reveals separate mechanisms for formation of the trunk and tail. Development 121, 2983–2994 (1995).

62. Goering, L. M. et al. An interacting network of T-box genes directs gene expression and fate in the zebrafish mesoderm. Proc. Natl Acad. Sci. USA 100, 9410–9415 (2003).

63. Amacher, S. L., Draper, B. W., Summers, B. R. & Kimmel, C. B. The zebrafish T-box genes no tail and spadetail are required for development of trunk and tail mesoderm and medial floor plate. Development 129, 3311–3323 (2002).

64. Griffin, K. J. P. & Kimelman, D. Interplay between FGF, One-eyed pinhead and T-box transcription factors during zebrafish posterior development. Dev. Biol. 264, 456–466 (2003).

65. Draper, B. W., Stock, D. W. & Kimmel, C. B. Zebrafish fgf24 functions with fgf8 to promote posterior mesodermal development. Development 130, 4639–4654 (2003).

66. Isaacs, H. V., Pownal, M. E. & Slack, J. M. W. eFGF regulates Xbra expression during Xenopus gastrulation. EMBO J. 13, 4469–4481 (1994).

67. Ramel, M. C. & Lekven, A. C. Repression of the vertebrate organizer by Wnt8 is mediated by Vent and Vox. Development 131, 3991–4000 (2004).

68. Vonica, A. & Gumbiner, B. M. Zygotic Wnt activity is required for Brachyury expression in the early Xenopus laevis embryo. Dev. Biol. 250, 112–127 (2002).

69. Thorpe, C. J., Weidinger, G. & Moon, R. T. Wnt/β-catenin regulation of the Sp1-related transcription factor sp5l promotes tail development in zebrafish. Development 132, 1763–1772 (2005).

70. Lekven, A. C., Thorpe, C. J., Waxman, J. S. & Moon, R. T. Zebrafish wnt8 encodes two wnt8 proteins on a bicistronic transcript and is required for mesoderm and neurectoderm patterning. Dev. Cell 1, 103–114 (2001).

71. Smith, W. C. & Harland, R. M. Expression cloning of noggin, a new dorsalizing factor localized in the Spemann organizer in Xenopus embryos. Cell 70, 829–840 (1992).

72. Sasai, Y. et al. Xenopus chordin: a novel dorsalizing factor activated by organizer-specific homeobox genes. Cell 79, 779–790 (1994).Together with reference 71, these papers report the cloning of two novel BMP inhibitors that are expressed in the organizer (Noggin and Chordin), which are essential for the function of the organizer.

73. De Robertis, E. M. & Kuroda, H. Dorsal–ventral patterning and neural induction in Xenopus embryos. Annu. Rev. Cell Dev. Biol. 20, 285–308 (2004).

74. Reversade, B., Kuroda, H., Lee, H., Mays, A. & De Robertis, E. M. Depletion of Bmp2, Bmp4, Bmp7 and Spemann organizer signals induces massive brain formation in Xenopus embryos. Development 132, 3381–3392 (2005).

75. Nguyen, V. H. et al. Ventral and lateral regions of the zebrafish gastrula, including the neural crest progenitors, are established by a bmp2b/swirl pathway of genes. Dev. Biol. 199, 93–110 (1998).

76. Kishimoto, Y., Lee, K. H., Zon, L., Hammerschmidt, M. & Schulte-Merker, S. The molecular nature of zebrafish swirl: BMP2 function is essential during early dorsoventral patterning. Development 124, 4457–4466 (1997).

77. Beck, C. W, Whitman, M., Slack J. M. The role of BMP signaling in outgrowth and patterning of the Xenopus tail bud. Dev. Biol. 238, 303–314 (2001).

78. Agathon, A., Thisse, C. & Thisse, B. The molecular nature of the zebrafish tail organizer. Nature 424, 448–452 (2003).

R E V I E W S



79. Beck, C. W., Whitman, M. & Slack, J. M. The role of BMP signaling in outgrowth and patterning of the Xenopus tail bud. Dev. Biol. 238, 303–314 (2001).

80. Beck, C. W. & Slack, J. M. Analysis of the developing Xenopus tail bud reveals separate phases of gene expression during determination and outgrowth. Mech. Dev. 72, 41–52 (1998).

81. Beck, C. W. & Slack, J. M. A developmental pathway controlling outgrowth of the Xenopus tail bud. Development 126, 1611–1620 (1999).

82. Beck, C. W. & Slack, J. M. Notch is required for outgrowth of the Xenopus tail bud. Int. J. Dev. Biol. 46, 255–258 (2002).

83. Gritsman, K. et al. The EGF–CFC protein one-eyed pinhead is essential for Nodal signaling. Cell 97, 121–132 (1999).This study demonstrates that the EGF–CFC protein Oep is essential for the formation of all endoderm and most mesoderm because Oep is required for Nodal signalling.

84. Lele, Z., Nowak, M. & Hammerschmidt, M. Zebrafish admp is required to restrict the size of the organizer and to promote posterior and ventral development. Dev. Dyn. 222, 681–687 (2001).

85. Willot, V. et al. Cooperative action of ADMP- and BMP-mediated pathways in regulating cell fates in the zebrafish gastrula. Dev. Biol. 241, 59–78 (2002).

86. Reversade, B. & De Robertis, E. M. Reciprocal regulation of Admp and Bmp2/4/7 at opposite embryonic poles generates a self-regulating morphogenetic field. Cell 123, 1147–1160 (2005).

87. Kimelman, D. & Pyati, U. J. Bmp signaling: turning a half into a whole. Cell 123, 982–984 (2005).

88. Wardle, F. C. & Smith, J. C. Refinement of gene expression patterns in the early Xenopus embryo. Development 131, 4687–4696 (2004).

89. Grimm, O. H. & Gurdon, J. B. Nuclear exclusion of Smad2 is a mechanism leading to loss of competence. Nature Cell Biol. 4, 519–522 (2002).

90. Mohammadi, M. et al. Structures of the tyrosine kinase domain of fibroblast growth factor receptor in complex with inhibitors. Science 276, 955–960 (1997).

91. DaCosta Byfield, S., Major, C., Laping, N. J. & Roberts, A. B. SB-505124 is a selective inhibitor of transforming growth factor β type I receptors ALK4, ALK5, and ALK7. Mol Pharmacol 65, 744–752 (2004).

92. Delaune, E., Lemaire, P. & Kodjabachian, L. Neural induction in Xenopus requires early FGF signalling in addition to BMP inhibition. Development 132, 299–310 (2005).

93. Galli, A., Roure, A., Zeller, R. & Dono, R. Glypican 4 modulates FGF signalling and regulates dorsoventral forebrain patterning in Xenopus embryos. Development 130, 4919–4929 (2003).

94. Jackman, W. R., Draper, B. W. & Stock, D. W. Fgf signaling is required for zebrafish tooth development. Dev. Biol. 274, 139–157 (2004).

95. Pyati, U. J., Webb, A. E. & Kimelman, D. Transgenic zebrafish reveal stage-specific roles for Bmp signaling in ventral and posterior mesoderm development. Development 132, 2333–2343 (2005).

96. Lewis, J. L. et al. Reiterated Wnt signaling during zebrafish neural crest development. Development 131, 1299–1308 (2004).

97. Sivak, J. M., Petersen, L. F. & Amaya, E. FGF signal interpretation is directed by Sprouty and Spred proteins during mesoderm formation. Dev. Cell 8, 689–701 (2005).

98. Nishita, M. et al. Interaction between Wnt and TGF-β signalling pathways during formation of Spemann’s organizer. Nature 403, 781–785 (2000).

99. Kroll, K. L. & Amaya, E. Transgenic Xenopus embryos from sperm nuclear transplantations reveal FGF signaling requirements during gastrulation. Development 122, 3173–3183 (1996).

100. Messenger, N. J. et al. Functional specificity of the Xenopus T-domain protein Brachyury is conferred by its ability to interact with Smad1. Dev. Cell 8, 599–610 (2005).

101. Skromne, I. & Stern, C. D. A hierarchy of gene expression accompanying induction of the primitive streak by Vg1 in the chick embryo. Mech. Dev. 114, 115–118 (2002).

102. Bertocchini, F. & Stern, C. D. The hypoblast of the chick embryo positions the primitive streak by antagonizing nodal signaling. Dev. Cell 3, 735–744 (2002).

103. Bertocchini, F., Skromne, I., Wolpert, L. & Stern, C. D. Determination of embryonic polarity in a regulative system: evidence for endogenous inhibitors acting sequentially during primitive streak formation in the chick embryo. Development 131, 3381–3390 (2004).

104. Tam, P. P. L. & Gad, J. M. in Gastrulation (ed. Stern, C. D.) 233–262 (Cold Spring Harbor Laboratory Press, Cold Spring Harbor, 2004).

105. Rivera-Perez, J. A. & Magnuson, T. Primitive streak formation in mice is preceded by localized activation of Brachyury and Wnt3. Dev. Biol. 288, 363–371 (2005).

106. Liu, P. et al. Requirement for Wnt3 in vertebrate axis formation. Nature Genet. 22, 361–365 (1999).

107. Yamamoto, M. et al. Nodal antagonists regulate formation of the anteroposterior axis of the mouse embryo. Nature 428, 387–392 (2004).

108. Perea-Gomez, A. et al. Nodal antagonists in the anterior visceral endoderm prevent the formation of multiple primitive streaks. Dev. Cell 3, 745–756 (2002).

109. Kemp, C., Willems, E., Abdo, S., Lambiv, L. & Leyns, L. Expression of all Wnt genes and their secreted antagonists during mouse blastocyst and postimplantation development. Dev. Dyn. 233, 1064–1075 (2005).

110. Mishina, Y., Suzuki, A., Ueno, N. & Behringer, R. R. Bmpr encodes a type I bone morphogenetic protein receptor that is essential for gastrulation during mouse embryogenesis. Genes Dev. 9, 3027–3037 (1995).

111. Winnier, G., Blessing, M., Labosky, P. A. & Hogan, B. L. Bone morphogenetic protein-4 is required for mesoderm formation and patterning in the mouse. Genes Dev. 9, 2105–2116 (1995).

112. Taverner, N. V. et al. Microarray-based identification of VegT targets in Xenopus. Mech. Dev. 122, 333–354 (2005).

113. Chung, H. A. et al. Screening of FGF target genes in Xenopus by microarray: temporal dissection of the signalling pathway using a chemical inhibitor. Genes Cells 9, 749–761 (2004).

114. Okabayashi, K. & Asashima, M. Tissue generation from amphibian animal caps. Curr. Opin. Genet. Dev. 13, 502–507 (2003).

115. Fukui, Y. et al. Long-term culture of Xenopus presumptive ectoderm in a nutrient-supplemented culture medium. Dev. Growth Differ. 45, 499–506 (2003).

116. Amaya, E. Xenomics. Genome Res. 15, 1683–1691 (2005).

117. Chen, J. A., Voigt, J., Gilchrist, M., Papalopulu, N. & Amaya, E. Identification of novel genes affecting mesoderm formation and morphogenesis through an enhanced large scale functional screen in Xenopus. Mech. Dev. 122, 307–331 (2005).

118. Loose, M. & Patient, R. A genetic regulatory network for Xenopus mesendoderm formation. Dev. Biol. 271, 467–478 (2004).

119. Feng, X. H. & Derynck, R. Specificity and versatility in TGF-β signaling through Smads. Annu. Rev. Cell Dev. Biol. 21, 659–693 (2005).

120. van Es, J. H., Barker, N. & Clevers, H. You Wnt some, you lose some: oncogenes in the Wnt signaling pathway. Curr. Opin. Genet. Dev. 13, 28–33 (2003).

121. Thisse, B. & Thisse, C. Functions and regulations of fibroblast growth factor signaling during embryonic development. Dev. Biol. 287, 390–402 (2005).

122. Shen, M. M. & Schier, A. F. The EGF–CFC gene family in vertebrate development. Trends Genet. 16, 303–309 (2000).

123. Dorey, K. & Hill, C. S. A novel Cripto-related protein reveals an essential role for EGF–CFCs in Nodal signalling in Xenopus embryos. Dev. Biol. 20 February 2006 (doi:10.1016/j.ydbio.2006.01.006).

124. Cheng, S. K., Olale, F., Bennett, J. T., Brivanlou, A. H. & Schier, A. F. EGF–CFC proteins are essential coreceptors for the TGF-β signals Vg1 and GDF1. Genes Dev. 17, 31–36 (2003).

125. Lane, M. C., Davidson, L. & Sheets, M. D. BMP antagonism by Spemann’s organizer regulates rostral-caudal fate of mesoderm. Dev. Biol. 275, 356–374 (2004).

126. Lieschke, G. J. et al. Zebrafish SPI-1 (PU.1) marks a site of myeloid development independent of primitive erythropoiesis: implications for axial patterning. Dev. Biol. 246, 274–295 (2002).

AcknowledgementsI wish to thank A. Schier, E. Amaya, U. Pyati, and D. Szeto for critical comments on this manuscript; C. Stern, H. Isaacs, G. Lieschke, R. Behringer and A. Schier for providing valuable information; and S. Dougan and J. Heasman for communicat-ing unpublished results. D.K.’s work on mesoderm induction is supported by the US National Science Foundation and National Institutes of Health.

Competing interests statementThe author declares no competing financial interests.

DATABASESThe following terms in this article are linked online to:Entrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=genebmp2b | Bmp4 | cyc | Ntl | Spt | sqt | Xnr5 | Xnr6

FURTHER INFORMATIONAxeldb: http://www.dkfz-heidelberg.de/molecular_embryology/axeldb.htmXenopus Mesendoderm Network: http://www.nottingham.ac.uk/biology/Genetics/staff/rogerpatient/networks/mesendoderm/mesendoderm.htmZFIN: http://zfin.orgAccess to this links box is available online.

R E V I E W S



Mesoderm induction: from caps to chips

Documents