During recent years, our understanding of TGFbeta signalling through serine/threonine kinase receptors and Smads has increased enormously. Activation of R-Smads by receptor induced phosphorylation is followed by complex formation with co-Smads and translocation to the nucleus, where the transcription of specific genes is affected and ultimately results in changes in cell behaviour. Experimental analysis primarily of epithelial cells in culture has revealed that a number of members of the TGFbeta family are interchangeable in the effect they have on growth and differentiation. On the other hand, different ligands of the TGFbeta superfamily can result in different responses because of cell type specific expression of other components of the signalling pathway. The relative expression levels of receptors and Smads within the cell is an important determinant of TGFbeta induced responses. Functional analysis of genes in the TGFbeta superfamily signal transduction cascade in vivo in mice either lacking entire genes, or expressing dominant negative forms of particular proteins, are providing profound new insights into the signalling cascades, their interaction and their specificity (Table 3). For example, by phenotypical comparison and intercrossing different heterozygous mutants, it has become clear that nodal, until recently an orphan protein without receptor/signal complex, probably signals through the activin type II receptor, ALK-4 and Smad2 (Nomura and Li, 1998; Song et al., 1999). Many of the genes of this cascade that have been targeted in the mouse result in early embryonic lethal phenotypes, demonstrating an important function for the BMP and TGFbeta/activin-activated pathways in mesoderm formation and differentiation, but masking a possible role in later events. For example mutations in BMP2 and 4 are lethal at or soon after gastrulation so that their putative role in skeletogenesis cannot be studied in mice lacking these genes. The difference in severity of the phenotypes between ligand, receptor and Smad deficient mice suggest that other receptors and ligands may partially compensate for the loss of one protein. Chimeric analysis provides one tool for analysing later developmental functions. By rescuing the early defects it was demonstrated that TGFbeta family members have an important function in anterior development and left/right asymmetry. Temporal and spatial specific gene targeting will be a powerful tool for analysing the function of TGFbeta family members in for example, bone formation, angiogenesis and carcinogenesis. Isolation of cells from the different gene targeted mice provides a unique source of material to gain more insight in the biochemical mechanisms of specific pathways. For example, use of cells deficient in Smad2 for biochemical and cell biological assays could give a better view of the function of Smad3. Smad3 deficient mice already demonstrate that there is a clear difference between Smad2 and Smad3 during development. Full descriptions of the remaining gene ablation studies of this signal transduction cascade, namely those for ALK-5, BMPR-II and Smad1 and -7 are eagerly awaited to complete the puzzle. As more of these superfamily of ligands and their signalling pathways have been functionally dissected, it has become evident that this superfamily of growth factors plays a pivotal role in epiblast formation and gastrulation, signalling from both the epiblast as well as the extraembryonic tissues. Furthermore, it becomes clear that TGFbeta is indeed important for proper vessel formation and that it might use endoglin, as well as ALK-1, ALK-5 and Smad5 to mediate this function. Further analyses of these mice should provide a clearer understanding of the mechanism of TGFbeta action in vascular development and remodelling.