Imaging the cell-cycle dependence of prime editing and adapting PE7 for pooled screens

Abstract

Prime editing enables precise insertions, deletions, and all twelve base substitutions without double-strand breaks, but its efficiency varies across cell types and edit types, and the cellular processes that resolve the heteroduplex DNA intermediate generated during editing remain poorly understood. Alongside these mechanistic questions, prime editing has become a promising platform for pooled genetic screens, enabling the interrogation of diverse genetic variants, though newer editors such as PE7 have not yet been characterized in this context. This thesis aims to address both of these areas. To test whether prime edit incorporation is linked to S phase, I developed a live-cell imaging platform pairing a strand-specific prime editing reporter with the PIP-FUCCI cell cycle reporter, along with a computational pipeline for single-cell segmentation, tracking, and fluorescence quantification. Using 4 bp substitution edits in U2OS cells, we found that coding-strand reporter activation, which requires transfer of the edit to the template strand, was significantly delayed relative to template-strand activation and strongly depleted in G1 phase. These results are consistent with a model in which DNA replication facilitates heteroduplex resolution for substitution edits, with direct implications for prime editing in post-mitotic cells. To evaluate PE7 in a pooled screening context, I designed and cloned two matched sub-libraries of 100 pegRNAs and epegRNAs targeting 42 essential genes selected as top dropout hits from a previous screen. A redesigned HiFi assembly workflow yielded uniform guide representation, producing screen-ready libraries for direct comparison of PE7 and PEmax with and without MLH1 knockout. Together, this work provides an imaging-based platform for investigating the cell-cycle dependence of prime editing and lays the groundwork for evaluating PE7 in pooled dropout screens for the investigation of genetic variants.