Mesoderm Induction A Postmodern View

and that the active form is present at undetectable levels in vivo. In support of a role for Vg1, a recent study using a dominant-negative mutant indicates that Vg1 is David Kimelman and Kevin J. P. Griffin Department of Biochemistry University of Washington Seattle, Washington 98195-7350 required for the development of dorsal mesoderm and dorsal endoderm (Joseph and Melton, 1998). However, the dominant-negative Vg1 had no effect on ventral and lateral mesoderm (ventral and lateral endoderm were The origin of all tissues in the adult animal can be traced back to one of three primary germ layers: endoderm not examined), arguing that a different TGF-b family member is involved in mesoderm, and possibly endo(gut), mesoderm (muscle, bone, and connective tissues), and ectoderm (epidermis and neural tissue). Germ layer derm, induction. Several authors have presented evidence opposing and favoring a role for activin in early formation is one of the first subdivisions that occurs in embryonic development, and its regulation has engaged Xenopus development, and it still remains possible that the early embryo contains a yet undiscovered member developmental biologists for over a century. The seminal work of Nieuwkoop (1969) demonstrated that a signal of the TGF-b family. And the Mesoderm-Inducing Signal Is ... released by the most vegetal cells, the prospective endoderm, converts the overlying prospective ectoderm a Transcription Factor? The work of Zhang et al. (1998) in this issue of Cell toward a mesodermal fate, creating the three germ layers of the amphibian embryo (Figure 1A; reviewed by seriously challenges the orthodox view of mesoderm induction. Their work shows that a crucial component Harland and Gerhart, 1997). Later experiments showed that the endogenous mesoderm-inducing signal is presof the vegetal maternal mesoderm-inducing signal is not a secreted factor, but a member of the intriguing T-box ent as early as the 32-cell stage (Jones and Woodland, 1987), many hours before transcription occurs in the transcription factor family. Since transcription factors cannot act until the start of zygotic transcription at the embryo, and is mimicked in vitro by members of the TGF-b and FGF families of secreted growth factors. It mid-blastula (4000-cell) stage, a clear implication of this work is that mesoderm (and endoderm) induction occurs is now clear that mesoderm induction requires a TGF-b signal operating in concert with an FGF signal. (TGF-b much later than originally supposed. Over the last two years, four groups have described is used throughout this review to connote any member of the TGF-b superfamily.) Endoderm development also a novel Xenopus transcription factor containing a T-box DNA-binding motif and gave it a variety of names includrequires TGF-b signaling, suggesting that mesoderm and endoderm may be induced by a common pathway ing VegT, Xombi, Antipodean, and Brat (citations can be found in Zhang et al., 1998). VegT is first observed (Henry et al., 1996). However, it is not at all clear how mesoderm and endoderm induction are spatially sepaas a maternal transcript localized to the vegetal hemisphere of eggs and embryos, which corresponds prirated in the embryo. The widely accepted synthesis of these data has been marily to the prospective endoderm and possibly some of the mesoderm (Figure 1B), in a pattern quite similar that the endogenous mesoderm-inducing signal must be present as a maternal mRNA or protein encoding a to that of Vg1 (Weeks and Melton, 1987). Just before gastrulation, zygotic VegT transcripts are found throughsecreted factor, most likely a TGF-b family member, that is localized to the vegetal cytoplasm during oogenesis. out the mesoderm (Figure 1C). Ectopic expression experiments by all of the groups indicated an important Fortunately, Melton and colleagues identified a maternal, vegetally localized mRNA encoding the TGF-b family role for VegT in regulating mesoderm and endoderm specification and morphogenesis, but it was not possimember Vg1 (Weeks and Melton, 1987). The active form of Vg1 protein, however, has never been detected in ble to distinguish between the maternal and zygotic roles of VegT. vivo, and ectopically expressed wild-type Vg1 does not induce mesoderm or endoderm. Advocates for Vg1 Zhang et al. (1998) used antisense oligonucleotides to specifically deplete the maternal VegT mRNA and speculate that its processing must be tightly regulated,