Edward M. De Robertis - US grants

We have recently identified a novel family of developmentally controlled genes in the frog Xenopus laevis. We demonstrated for the first time that, using a conserved region of Drosophila development-controlling genes called the homeobox, it was possible to identify and isolate vertebrate genes that resemble the Drosophila genes in several respects. Since then, similar genes have been isolated by others from a number of vertebrates, including cDNA clones from human transformed cells. We have by now isolated one of the largest collections of homeobox-containing cDNA clones available in any vertebrate, which are expressed at the gastrula stage of development. Some of these genes are expressed maternally in the oocyte and will be the focus of our studies, because they could potentially be examples of the elusive "cytoplasmic determinants" thought to be laid down in egg cytoplasm for the control of the initial steps of embryonic differentiation. The objective of this proposal is to ascertain the developmental function of Xenopus homeobox-containing genes. A variety of techniques will be utilized towards this goal, and many of them take advantage of the ready availability of large amounts of Xenopus embryos during the earliest developmental stages, during which equivalent experiments with mammalian embryos would be much more difficult.

In amphibia the egg is radially symmetric, but fertilization triggers a rotation of the egg cortex that determines the aids of bilateral symmetry. On the dorsal side the large yolk platelets move upwards, coming in intimate contact with animal pole cytoplasm. Our working hypothesis is that the dorsal yolk could release small hydrophobic molecules that activate transcription factors of the nuclear receptor family. Eight nuclear receptors of the retinoic acid (RA)/thyroid hormone family have been isolated from a cDNA library prepared from unfertilized Xenopus eggs. Using functional assays we have shown that two of them are retinoic acid receptors (RARs) of the classical type and two others are members of the novel retinoid "X"receptor family (RXRs). An antibody against one of the latter proteins shows that it becomes localized to a small subset of blastula nuclei that gives rise to mesoderm, particularly on the dorsal side. The discovery of maternal retinoid receptors, which become asymmetrically distributed during development, may lead to a better understanding of the potent teratogenic effects of RA in early vertebrate development. The other receptors correspond to so-called orphan receptors, for which the ligand is not yet known. The sequence of their DNA binding regions strongly suggests that they may bind to an overlapping subset of the gene network activated by retinoid receptors. The identification of ligands for the maternal orphan receptors could result in the identification of new morphogens and teratogens that cause congenital malformation in vertebrates. The objective of this proposal is to ascertain the developmental function of maternal retinoid and orphan receptors, including the identification of the endogenous ligand molecules involved, the study of their teratogenic effects, their mode of action, the localization of receptors and their ligands in the developing egg, and the molecular mechanisms of their segregation. A number of molecular techniques will be used in this analysis, exploiting the accessibility of amphibian embryos to experimental manipulation at the early stages of development, and of transgenic mice at later stages.

Homeobox genes encode DNA-binding proteins that control key steps in animal development. Spemann and Mangold (1924) identified the dorsal lip of the blastopore, the organizer, as a region of the embryo able to induce twinning, or a secondary body axis, when transplanted to a host embryo. Recent experiments indicate that homeobox genes, such as goosecoid and Xnot-2, play an important role in the organizer phenomenon. Microinjection of goosecoid mRNA can mimic Spemann's experiment, recruiting neighboring cells into twinned axes in Xenopus. Because homeobox genes encode nuclear proteins, their effects on neighboring cells must be mediated by secreted molecules. We have recently identified the two major downstream targets activated by organizer homeobox genes: chordin and cerberus. Both are novel secreted proteins. In addition, goosecoid represses the expression of BMP-4, a ventralizing signal that is antagonized by the chordin protein. By studying organizer-specific homeobox genes and their secreted targets chordin, cerberus and BMP-4 we hope to better understand the molecular mechanisms that pattern vertebrate development. In particular, we will ask how these genes pattern the three germ layers, the mesoderm, the ectoderm and the endoderm, and how they control the sequential deployment of Hox genes in the antero-posterior body axis. Because the molecular mechanisms of development are very conserved, these studies in Xenopus should provide insights into the molecules that control development in all vertebrates, including humans.

Description (provided by applicant): The early embryo of the frog Xenopus provides a favorable system for discovering the molecular mechanisms that control cell signaling and differentiation. When an embryo is cut in half, it can self-regulate to form a well-proportioned organism. Recent experiments have elucidated that self-regulation is mediated by a novel extracellular biochemical pathway. Signaling by the BMP family of growth factors is antagonized by a dorsal protein called Chordin. Chordin is degraded by a ventrally-expressed zinc metalloproteinase of the Tolloid family. In turn, this protease is inhibited by a ventrally secreted Frizzled-Related Protein (sFRP). Here we propose to investigate cell-cell signaling by exploring three broad questions: 1) How do cells in the dorsal and ventral poles of the embryo communicate with each other across long distances? Our hypothesis is that secreted molecules of similar biochemical activity expressed at opposite poles of the embryo, but under opposite transcriptional control, provide the key to understanding how developmental fields of differentiating cells self-regulate. 2) Are other sFRPs inhibitors of zinc metalloproteinases? The hypothesis to be tested is that Frizzled domains, which are present in many extracellular proteins, might provide a module that regulates the proteolysis of extracellular molecules, perhaps under the control of Wnt growth factors. 3) How are multiple growth factor signaling inputs integrated intracellularly? We will test whether the BMP, Wnt, and Receptor Tyrosine Kinase signaling pathways converge on the phosphorylation state of the transcription factor Smadl, regulating its activity. Investigating these three specific aims will help provide an integrated view of how cells communicate with each other, not only in embryos but also in adult tissues and in human embryonic stem cells. Genes in the signaling pathways to be studied are involved in many processes of medical relevance, such as central nervous system induction, cell and tissue differentiation, the formation of identical and conjoined twins, neural induction, fibrotic and connective tissue disease, and cancer.