Proximity epigenomics for context-specific analysis of complex chromatin features

Project Summary Chromatin context defines all DNA transactions, from gene regulation to genome maintenance. Changes in chromatin composition are hallmarks of cancer cells that contribute to malignant transformation and have been studied extensively using a wide range of Next Generation Sequencing (NGS) approaches. However,

conventional genome-wide mapping efforts are limited to the detection of individual targets of interest, impeding the study of their often diverse biological roles. Multi-subunit chromatin-regulatory complexes as well as protein interactions with histones or noncanonical nucleic acid structures are largely defined through correlative analyses

of separate mapping efforts, which are unable to determine physical interaction, proximity on the same DNA fragment or even presence in the same cell. These limitations underline an urgent need for improved, context- dependent chromatin mapping and characterization efforts. In this application, we will establish a versatile and

broadly applicable technique that allows for the genome-scale analysis of two-component interactions on chromatin. The proposal builds on our extensive experience in the study of genome-wide chromatin responses to DNA damage, as well as the expertise of co-Investigator Dr. Michael Seidman in the analysis and visualization

of close-proximity molecular interactions in the context of DNA replication stress. Aim 1 will combine this complementary expertise to develop Proximity-based Chromatin Immunoprecipitation (ProxiChIP), a tool to characterize functional subsets of a given chromatin feature based on its interacting partners, shifting the current

ChIP paradigm towards combinatorial feature mapping. Using well-characterized, interacting chromatin binding proteins as a proof of principle, we will convert a method currently restricted to the imaging-based detection of protein interactions (proximity ligation assay, PLA) to a broadly applicable biochemistry tool suitable for

immunoprecipitation and a diverse set of downstream applications including NGS. In Aim 2, we will apply this methodology to advance the epigenomic characterization of pathological RNA:DNA hybrids, a complex and poorly understood feature of many cancer genomes thought to contribute to DNA replication stress, cancer

genome instability and therapy response. RNA:DNA hybrids have been mapped genome-wide using DNA:RNA immunoprecipitation (DRIP) and related methods. However, existing approaches fail to distinguish between physiological and pathological RNA:DNA hybrid subsets. ProxiChIP-based mapping of RNA:DNA hybrids in the

context of replication stress is expected to define the genomic features that underly pathological R loop formation, which presents an essential step towards understanding their impact on genome integrity and malignant transformation.