Fig. 5. Model for differential gene expression in the neurogenic ectoderm. (A) Cross-section through a cellularizing embryo. The nuclear Dorsal gradient is shown with peak levels in ventral regions and lower levels in more lateral regions. The presumptive neurogenic ectoderm (NE) exhibits at least three distinct patterns of gene expression: sim and m8 are expressed only in the ventral-most line of cells in the NE, the mesectoderm; brk, vnd, rho and vn are expressed in the 5-6 cell wide ventral domain of the NE; and sog and CG12443 are expressed in broad lateral stripes throughout the NE. DE, dorsal ectoderm. (B) A stylized representation of the enhancers active in the NE. Enhancers active in the mesectoderm (e.g. sim) contain a large number of Su(H)-binding sites (red boxes), but few optimal dorsal sites (black boxes). By contrast, enhancers that direct broad expression throughout the NE (sog and CG12443) contain several optimal Dorsal sites, but no Su(H) sites. Enhancers that direct expression in an intermediate pattern, i.e. in ventral regions of the NE (rho, vnd, brk and vn), contain a mixture of high-affinity and low-affinity Dorsal sites, as well as a few Su(H) sites. Additionally, CA-Eboxes (CACATGT, blue boxes) and the CTGWCCY motif (not shown) are only found in the mesectodermal and ventral neurogenic ectodermal enhancers, and not in the enhancers driving broad expression in the NE. This implies that genes exhibiting overlapping expression patterns (e.g. sog and brk) are not activated solely by a gradient of nuclear Dorsal, but also by a variety of transcription factors, and also that they are activated in the same regions by different means.