Investigation of RNA-Protein Interactions in PRC2 Function

Chromatin regulation contributes to control of gene expression and what identity a cell will adopt. In the last decade the role that RNA plays in chromatin regulation has become increasingly clear. RNA mediates protein recruitment and eviction from chromatin, forms nuclear condensates with proteins and DNA, and contributes to proper chromatin organization. Yet our knowledge of the mechanisms that govern RNA activity on chromatin lags significantly and limits our ability to understand nuclear function. To effectively answer some of the questions of RNA function in the nucleus we need a comprehensive atlas of RNA-protein interactions, which would enable generation of protein mutants defective in RNA-binding. The goal of my thesis was to develop an unbiased method to profile RNA-binding proteins in the nucleus and apply it to Polycomb repressive complex 2 (PRC2). PRC2 is an epigenetic regulatory complex that deposits mono, di- and tri- methyl lysine onto histone H3 (H3K27me3) and maintains gene silencing during development. PRC2 shows extensive contacts with RNA but their function remains unclear. In the first chapter, we present a novel method, dubbed RBR-ID, for the identification of RNA-protein interactions, which uses UV-crosslinking of photosensitive nucleotide analogs to proteins followed by high resolution mass spectrometry (LC-MS/MS). We identified over 800 RNA-binding proteins, of which 427 were novel and enriched for chromatin-related functions. In the second chapter we adapted RBR-ID to study PRC2, identifying RNA-binding-regions (RBRs) on every subunit of the complex. An RBR identified on EED fell near the regulatory center of PRC2, and we showed that RNA-mediated inhibition of PRC2 can be reversed by stimulatory peptides that bind in the regulatory center, reflecting the antagonistic relationship between RNA and PRC2. In the final chapter we present a testing method we developed for the SARS-CoV-2 virus. Our method, COV-ID, uses reverse transcription and loop-mediated isothermal amplification (RT-LAMP) from patient saliva paired with high-throughput sequencing. Using this method we can detect as little as 5-10 SARS-CoV-2 virions/μL, and we successfully replicate classification of saliva samples (10/10) from clinical COVID-19 patients. We show that COV-ID can be multiplexed to detect influenza as well as SARS-CoV-2. Finally we demonstrate that COV-ID can process saliva samples collected on filter paper with sensitivity as low as 50 virions/μL