Abstract
The regulation of gene expression is mediated at the transcriptional level by enhancer regions that are bound by sequence-specific transcription factors (TFs). Recent studies have shown that the in vivo binding sites of single TFs differ between developmental or cellular contexts. How this context-specific binding is encoded in the cis-regulatory DNA sequence has, however, remained unclear. We computationally dissect context-specific TF binding sites in Drosophila, Caenorhabditis elegans, mouse, and human and find distinct combinations of sequence motifs for partner factors, which are predictive and reveal specific motif requirements of individual binding sites. We predict that TF binding in the early Drosophila embryo depends on motifs for the early zygotic TFs Vielfaltig (also known as Zelda) and Tramtrack. We validate experimentally that the activity of Twist-bound enhancers and Twist binding itself depend on Vielfaltig motifs, suggesting that Vielfaltig is more generally important for early transcription. Our finding that the motif content can predict context-specific binding and that the predictions work across different Drosophila species suggests that characteristic motif combinations are shared between sites, revealing context-specific motif codes (cis-regulatory signatures), which appear to be conserved during evolution. Taken together, this study establishes a novel approach to derive predictive cis-regulatory motif requirements for individual TF binding sites and enhancers. Importantly, the method is generally applicable across different cell types and organisms to elucidate cis-regulatory sequence determinants and the corresponding trans-acting factors from the increasing number of tissue- and cell-type-specific TF binding studies.
