Transgenic frogs and FGF signalling in early development

Abstract

Time was, vertebrate developmental biologists worked on a particular species and stuck to the techniques appropriate to that species. If you worked on frogs you injected I~,tA into the eggs and did some dissections and grafts; if you worked on zehrafish you looked for mutations and followed the fates of transplanted cells; if you worked on cbickens you did grafts and followed the fates of cells; and if you .studied gastmlatinn in mice 1 think you must have been crazy, because the embryos are so small and inaccessible. With the realization that developmemal mechanisms are highly conserved, people began switching between species. It is much easier, for example, to overexpress a gene in a frog than in a mouse embryo, so why not do itt? And if you believe that regulatory elements are conserved, why not rake advantage of the small genome of Fugu and analyse promotets from thts species in transgenic mice2? This general approach has undoubtedly yielded important results hut it has the inevitable disadvantage of produdng a homogeneous view of vertebrate development in which all species are the same, and one is left wondering why we don't all look like a hybrid frog/monse/fish/chicken (or a fromtxshken, as Hazel Sive would put it3). Attempts are being made to overcome this problem. For example, Rossaot's and Sive's groups have cartied out frog-like in ~*t~o tissue recombinations in mouse'LS and zebrafish 6, and Beddington 7 and Shill and Fraser s,9, have earned out 'organizer grafts' in mouse and zebrafish, respectively. Indeed, Fl~¢er's work has revealed surprising differences between zebrafish and Xe~opus I0 I=.oth in the normal fate of the organizer and in its fate after transplantation. The latest advance reverses this trend of applying microsurgical techniques to species more usually respected for their contributions to genetics, for Amaya and Kroll have now introduced transgenesis to Xenoplt# 1.12. In the mouse, the ability to make transgenic animals has allowed detailed analysis of promoter and enhancer sequences 13, and, once suitable elements have been identified, it has also allowed investigators to overexpress a wild-type or mutant gene at the desired place and at the desired time 1~. Such experiments have proved very valuable in our understanding of the comrol of gene expression and of cell interactions in development. These approaches have been denied to the Xenopus embryologist, however, because injected DNA does not integrate into chromosomes during the early cell cycles of this species. As a result, the embryo expresses genes in a highly mosaic fashion and in the absence of integration it is likely that accurate temporal and tissue-specific control is lost. Although the cla.~sical Xenopusahemative, of injecting RNA, suffers only slightly from mosaieism, and one can inject into blasromeres wire known fates, there is no tempoml control of expression: translation occurs as soon as the RNA is iniected 15. It is possible to obtain effective temporal control by means of tricks, such as fusing one's protein of interest to the hormone-binding domain of a steroid receptor 16, but this only works for some genes. The importance of the correct timing of gene expression is illustrated by Xwm-8, where over-expression of this gene product by KNA injection causes dorsalization tT,rs, while more appropriate expression after the mid-blastula transition reveals that it is more likely to be involved in the specification of ventral mesoderm 19,