The Section on Vertebrate Development (SVD) is interested in mechanisms that regulate the differentiation of the ectoderm into its primary derivatives, i.e., epidermis, central nervous system, and neural crest. Our primary animal model system is the gastrula-stage embryo of the frog Xenopus laevis, but we are also working with the related frog Xenopus tropicalis, which is more suitable for genetic and transgenic approaches. In addition, the laboratory has experience with transgenic mouse and zebrafish models. Our general approach is to use molecular techniques to identify transcription factors that play roles in regulating ectoderm differentiation and then to exploit the biological advantages of the appropriate model organism to test for such functions and to establish relationships among individual regulatory factors. Since its inception, SVD's program has produced several interesting findings. Two recent discoveries involve, first, the mechanism by which Dlx and Msx homeobox genes serve as interpreters of dorsoventral and mediolateral spatial information and, second, the role of the transcription factor AP-2 in the integration of BMP and Wnt signals that control ectodermal cell fate.

Detection of Morphogenetic Gradients by Homeobox Genes
Luo, Matsuo-Takasaki, Lim, Sargent
Msx1, Dlx3, Dlx5, and Dlx6 are members of the Msx and Dlx family of homeobox genes. Results from our and other laboratories have shown that these genes are expressed very early in Xenopus development, with Msx1 active in mesoderm and ectoderm while the Dlx genes are transcribed only in ectoderm. All are ventral-specific, but the boundaries between dorsal and ventral expression domains differ in a systematic way: Msx1 is the most "ventral," followed by Dlx3, and Dlx5, and Dlx6 expression extends the farthest toward the midline. SVD investigators found that these different boundaries reflect different sensitivity of the Msx and Dlx genes to inhibition of BMP signaling by antagonists originating in the Spemann organizer and that there are different downstream target genes for the various homeobox factors. Based on these results, we have proposed a model for the mechanism by which boundaries between epidermis-neural crest and neural crest-neural plate are established during gastrulation (see articles by Luo et al. below). One current project is testing this model by using transgenic frogs: the regulatory elements from the Dlx5 gene will be used to drive ectopic expression of Dlx3 and Msx1 proteins. If the model is correct, the resulting embryos should have disrupted cranial neural crest and forebrain development.

Modulation of Wnt and BMP Signals by AP-2
Luo, Sargent
Several years ago, we determined that the transcription factor AP-2 was playing an important role in the regulation of at least one of the major keratin genes expressed in presumptive epidermis. More recently, we used ectopic expression and loss-of-function experiments in Xenopus to extend our earlier findings and have confirmed that AP-2 is a direct positive regulator of keratin gene expression, representing at least part of the mechanism by which epidermis is specified by BMP signaling. Manipulating AP-2 expression in the embryo results in reactivation of epidermal genes in neural tissue but does not block neural gene expression. Thus, we conclude that AP-2 functions downstream of the neural-epidermal "switch." Other interesting results demonstrate that the high levels of AP-2 expression in the presumptive neural crest is induced by Wnt signaling and that AP-2 function in that tissue is necessary for activation of early neural crest regulatory factors such as Xenopus Twist and Slug. We are trying to learn how these two important embryological signals control the pattern of AP-2 expression and what target genes lie downstream from AP-2 in epidermis and neural crest. One approach to the latter investigation relies on DNA microarrays to identify genes whose expression responds to changes in AP-2 levels. Glass chip microarray of about 1,000 genes from an ectodermal subtracted cDNA library has been generated and is being probed with cDNAs from experimentally manipulated embryonic ectoderm.