Classical experiments with amphibians revealed that bisected embryos develop into properly proportioned tadpoles and that transplantation of a small group of dorsal cells called "Spemann’s organizer" into the ventral side of an early embryo leads to the formation of a properly proportioned second axis. Thus, the morphogen activation gradients that drive patterning in the early embryo appear to scale with size. Two theoretical models (an inhibition-based model and the shuttling-based model) have been proposed to explain how the bone morphogenetic protein (BMP) activity gradient that drives axis formation is established. Ben-Zvi et al. provide evidence for the shuttling-based model, which previously had only experimental support in fly embryos, in frog (Xenopus laevis) embryos and suggest that the additional BMP ligand antidorsalizing morphogenic protein (ADMP), which is produced in the dorsal pole and which is not present in Drosophila, allows the Xenopus embryos to scale patterning with size. First, they created a mathematical model for the core patterning network containing the BMP ligands BMP (which represented BMP2, 4, and 7) and ADMP, the BMP antagonist chordin, and the protease Xlr, which degrades chordin. In this model, chordin and ADMP were produced at the dorsal pole, and BMP signaling repressed ADMP. Systematic screening for parameters that allowed the activation gradient of BMP to be robust and to scale with embryo size resulted in the identification of 21 networks that scaled properly out of the 26,000 possible networks. In each of the networks that scaled, the ligands were concentrated at the ventral pole, suggesting that activation gradients were generated by shuttling of the ligands from their site of production to the ventral pole and not simply by localized production of an inhibitor. Further, the models were consistent with experimentally proven properties of the core patterning components: Properly scaling networks required that chordin binds BMP with higher affinity than it binds ADMP, that binding of chordin facilitates diffusion of BMP ligands, and that chordin is degraded primarily when complexed with ligand. A shuttling model that relied solely on BMP and chordin was insufficient to provide scaling; however, the addition of ADMP with its BMP-mediated repression was sufficient. Experiments in which Xenopus embryos were injected with a tagged BMP4 with or without coinjection of morpholino oligonucleotides (MOs) against chordin showed that BMP4 only concentrated at the ventral pole in the presence of chordin. Axis inversion, which was also consistent with a shuttling mechanism, occurred when chordin was depleted from the organizer at the dorsal side with MOs and a second organizer was induced on the ventral side. Induction of a second organizer on the ventral side in conjunction with MOs for ADMP resulted in the expression of a ventrally expressed gene between the original dorsal organizer and the ventral organizer, which is also consistent with shuttling of BMP ligands. Thus, as in fly embryos, vertebrates also appear to use a shuttling mechanism for robust patterning that scales with embryo size.
D. Ben-Zvi, B.-Z. Shilo, A. Fainsod, N. Barkai, Scaling of the BMP activation gradient in Xenopus embryos. Nature 453, 1205-1211 (2008). [PubMed]