DEFINING SITES OF REPLICATION FORK COLLAPSE CAUSED BY ATR INHIBITION Nishita K. Shastri Eric J. Brown Replication stress, characterized by stalling of DNA replication and the accumulation of abnormal replication intermediates, has been linked to the genomic instability observed in cancer. Previous studies have defined specific genomic sequences that are difficult to replicate to be more vulnerable to replication-associated breaks and rearrangements. However, many of these sequences have been identified through indirect and potentially biased approaches. To identify DNA sequences that contribute to replication-associated genomic instability, I will describe genome-wide screens I have performed to determine the location, sequence, and frequency of replication perturbations within the mammalian genome upon replication stress. Ataxia telangiectasia and Rad3-related protein (ATR) is a checkpoint kinase that is a key upstream regulator of the response pathway to replication fork stalling during replication stress that prevents fork collapse. Through inhibition of this response pathway in mouse embryonic fibroblasts, my aims are to 1) characterize regions that lead to frequent replication fork stalling and collapse, and 2) further define genomic regions that become processed into double-strand breaks. Since replication protein A (RPA) binds to single-stranded DNA that becomes exposed when replication forks stall, RPA ChIP-Seq has been performed to map sites of frequently collapsed replication forks; however, not all stalled replication forks result in breaks. To differentiate a replication fork that has simply stalled from a fork that has become sensitized to double-strand break formation, I developed and applied a novel and specific break-detection assay, BrITL. With these complementary approaches to map replication-problematic loci, subsequent bioinformatics methods have been utilized to characterize features of the identified genomic regions that make it prone to fork collapse and detrimental DNA break formation when cells experience replication stress. While well-established difficult-to-replicate sequences (e.g. triplet and telomere repeats) exhibited enhanced fork collapse in RPA ChIP’d cells exposed to replication stress, these sequences were overshadowed by sites composed of previously uncharacterized simple tandem repeats. Circular dichroism and thermal difference absorption spectra indicate that the most commonly observed simple repeat at RPA-enriched sites (CAGAGG) folds into a stable intramolecular secondary structure and is sufficient to stall DNA replication in vitro and in vivo. BrITL analysis confirmed that these repetitive regions of RPA accumulation are also sites of DNA breakage. Interestingly, a majority of break sites identified by BrITL do not associate with RPA accumulation, but rather tend to locate around inverted retroelements that are predicted to form highly stable intrastrand stem-loop structures. Due to the lack of available ssDNA at these potential hairpin-forming sites, RPA accumulation would be limited. Overall, my studies represent the first unbiased identification of mammalian genomic sites that are vulnerable to replication stress and rely on ATR for stability.