A screen for developmental enhancers targeted by the Notch effector Su(H). Elizabeth K. Stroebele¹, Andrew Brittain¹, Xin Yuan¹, Seth Brown², Albert J. Erives¹. 1) Department of Biology, University of Iowa, Iowa City, IA; 2) Carver College of Medicine, University of Iowa, Iowa City, IA.

   Notch signaling regulates gene expression in cells undergoing asymmetric division and those forming borders between adjacent epithelial territories. Notch receptor-ligand interactions result in the nuclear translocation of Notch intracellular domain. A single transcription factor, Su(H), is thought to be a common transducer of NICD signaling in development. Su(H) binding sites are found in a set of neurogenic ectoderm enhancers (NEEs) characteristic of the vnd, brk, rho, and vn loci of diverse Drosophila species. These NEEs possess precisely-linked binding sites for Dorsal and Twist, and a separate, more loosely-linked binding site for Su(H). While the former spacing controls the dorsal extent of expression, the Su(H) site seems to control the strength of expression. To understand better how Su(H) works with different co-factors in different contexts, we wish to identify other functional enhancer families with Su(H) binding sites. We conducted a computational screen for conserved, non-coding sequences harboring Su(H) sites. We identified all such sites in 4 divergent Drosophila species. To detect conserved modules centered on these sites, we devised a reciprocal BLAST method optimized for regulatory regions (R-BLAST) using the NEEs as internal positive controls. R-BLAST is robust to the patterns of change seen at enhancer sequences such as frequent indels, elevated microsatellite content, and high binding site turnover that destroys serial sequence alignment. We find 1,128 regions that are conserved across the genus, and are likely to represent conserved developmental enhancers targeted by Notch signaling. We explored bioinformatic methods to determine the extent to which the NEE subset could be detected in principle given no knowledge of their binding sites other than those for Su(H). We cannot yet detect the NEEs de novo, but we identified and tested sequences from a subset of 346 modules containing an over-represented DNA-bending element.