Fig. 2. The Xvent2-BRE contains a motif that mediates Bmp responsiveness in Drosophila. (A) The Drosophila element contains Mad and Med sites (boxed) separated by a five-nucleotide spacer that is required for recruitment of Shn. The sequence of Smad sites and their relative spacing are maintained in the Xvent2-BRE with one mismatch from the consensus (GT) at position 4. Point mutations and deletions in the Xvent2-BRE variants are marked in green. (B) Mutant BREs were tested for their response to Bmp signaling in Xenopus animal cap assays. Xvent2-BRE-luciferase reporters (top) were microinjected into two- to four-cell stage embryos (bottom), animal cap explants were dissected at stage 8-9 and cultured until siblings reached early gastrula stages, then processed for luciferase assays. (C, left) Mutant Xvent2-BREs respond identically in Xenopus and Drosophila. Both the wild-type BRE and Xvent2-sub2, which contains two transversions within the five-nucleotide spacer (see A), are stimulated 10- to 11-fold in response to CABR, compared with reporter alone. By contrast, Xvent2-del2 bearing a two-nucleotide spacer deletion (see A) fails to respond. In the experiment shown, 100 pg of CABR mRNA/embryo was used to stimulate expression; however, the Xvent2-del2 reporter fails to respond even at 2 ng. All results are presented as fold activation relative to wild type in the absence of CABR. We consistently observed higher luciferase counts for Xvent2-sub2 relative to wild type. (C, right) Wild type and Xvent2-sub2 respond to endogenous Bmp signaling, whereas the Xvent2-del2 mutant does not. Transgenic embryos containing BRE multimers driving GFP are at tailbud stage 31/32. In situ hybridization shows that wild type and Xvent2-sub2 mutants direct expression in a pattern similar to the endogenous Xvent2 mRNA and Xvent2 transgenes described previously. Transgenic Xvent2-del2 embryos had variable expression patterns that are likely to reflect position effects due to random integration sites.