Genomic sequence context and histone marks regulate pioneer binding to nucleosomes
Date:
Luca Mariani1, Xiao Liu1,2, Kwangwoon Lee1, Stephen S Gisselbrecht1, Philip A Cole1,3, Martha L Bulyk1,4
1Brigham and Women’s Hospital and Harvard Medical School, Div. Genetics, Dept. Medicine, Boston, MA, 2Harvard Medical School, Masters Program in Biomedical Informatics, Boston, MA, 3Harvard Medical School, Dept. Biological Chemistry and Molecular Pharmacology, Blavatnik Institute, Boston, MA, 4 Brigham and Women’s Hospital and Harvard Medical School, Dept. Pathology, Boston, MA
Cell fate decisions are controlled by sequence-specific transcription factors (TFs), referred to as ‘pioneers’, that invade nucleosomal chromatin and render it accessible to other TFs. Pioneers are typically defined as binding DNA within nucleosomes (‘pioneer binding’). However, they bind just a minority of their recognition sequences in the genome and their binding sites differ between cell types, suggesting that local sequence context features and epigenetic features may regulate pioneer binding. To study these features in vitro, we developed PIONEAR-seq, a highly parallel sequencing-based assay for high-throughput characterization of TF binding to nucleosomes tiled with random 20-bp sequences. Using PIONEAR-seq, we surveyed 7 putative human pioneers. Analysis of binding to nucleosomes based on the synthetic Widom 601 model sequence versus three genomic sequences (ALB1, CX3, NRCAM) revealed that the sequence context drives nucleosomal binding modes. In cases of TF binding to the ends of nucleosomes, our data suggest that the asymmetric unwrapping of nucleosomal DNA ends facilitates TF pioneer binding by exposing the less bendable half. To validate this result, we quantitatively evaluated the DNA bendability in published NCAP-SELEX data for the binding of 149 TFs to nucleosomes that contain random 99-bp sequences. We confirmed that most TFs prefer binding at nucleosome ends adjacent to the more rigid ends, which was not the case for TFs that bind more internally to nucleosomes (e.g., dyad, periodic and gyre-spanning binding). We propose a model for how genomic sequence composition ~100-bp away from TF DNA binding sites on nucleosomes can influence TF binding through nucleosome dynamics. Furthermore, our PIONEAR-seq data revealed that the histone marks H3K4me1, H3K27ac, H3K9me3, and H3K27me3 modulate pioneers’ DNA binding site specificities and binding modes. Altogether, our results revealed that rather than the nucleosomal binding preferences being intrinsic to a pioneer, the composition of nucleosomes also determines pioneers’ binding properties, and suggest that sequence features and chromatin marks regulate pioneer binding in cells.