DNA bendability regulates transcription factor binding to nucleosomes [MNase-seq]

Cell fates are controlled by ‘pioneers’, sequence-specific transcription factors (TFs) that bind recognition motifs on nucleosomes (‘pioneer binding’). Pioneers occupy a minority of their recognition sequences in the genome, suggesting that the sequence context regulates their binding. Here, we developed PIONEAR-seq, a high-throughput biochemical assay to characterize pioneer binding to nucleosomes. We used PIONEAR-seq to assay 11 human TFs for binding to nucleosomes based on Widom 601 versus genomic sequences. We found that pioneer binding, while mediated primarily by TFs' recognition motifs, senses the broader nucleosome sequence context and that TFs previously found to be dyad or periodic binders on nucleosomes assembled on synthetic sequences exhibit exclusively end binding to nucleosomes based on genomic sequences. We propose a model where the nucleosome exploits the local bendability of the DNA sequence to position pioneer binding, revealing another cis-regulatory layer in eukaryotes. MNase digestion followed by sequencing for the DNA libraries used for PIONEAR-seq. These libraries are based on 4 DNA sequences of 147 bases (ALBN1, CX3CR1, NRCAM, W601) known to position nucleosome either in vivo or in vitro. W601, also known as Widom 601, is a 147-bp DNA sequence of synthetic origin with the highest reported affinity to the histone octamer (PMID: 9514715 ). ALBN1 is a 147-bp DNA sequence based on the genomic sequence of the nucleosome N1 within the ALB enhancer (PMID: 31303471). CX3CR1 is a 147-bp sequence from the CX3 chemokine receptor 1 locus that have been proposed to be a candidate nucleosome target for pioneer factors in hematopoietic differentation (PMID: 31303471). NRCAM is a 147-bp genomic nucleosomal sequence from the Neuron-Glia-CAM- related cell adhesion molecule locus that gets digestable by MNase upon neuronal reprogramming (PMID: 31303471). These sequences were tiled with random 20-mers at several position. The resulting DNA libraries were pooled together, assembled on nucleosome core particles (NCPs), digested with MNase and sequenced (Libraries: D1-WANC-1-Mnase, H1-WANC-2-Mnase). As controls, the libraries were sequenced also without the digestion of MNase(Libraries: C1-WANC-1-Input, G1-WANC-2-Input). In addition, MNase-seq was also perofmed on NCP samples based on single 147-bp sequences based on the NRCAM template (Libraries: F3-203-1-MNase, G3-203-2-MNase), eventually modified to include CEBPB specific binding sites (Libraries: E2-337-2-MNase, F2-338-1-MNase).