Engineering and evolution of precision genome editing agents

Abstract

Genome editing agents have had a transformative effect on the study of biological systems, the engineering of cells and organisms with novel properties, and even the treatment of genetic diseases in the clinic. This revolution was first enabled through the discovery of programmable nucleases, including zinc-finger nucleases (ZFNs)1-5, transcription activator-like effector nucleases (TALENs)6-11, and CRISPR-associated (Cas) nucleases12-15. These nucleases induce targeted double-stranded DNA breaks (DSBs), which stimulate cellular repair processes that ultimately result in genome modification. While effective for some desired genome editing applications, such as targeted gene disruption, the highly stochastic nature of DSB repair results in a variety of outcomes, typically small insertions or deletions (indels), making targeted modification and correction of a gene difficult and imprecise. To address this limitation, several precision genome editing technologies building upon CRISPR-Cas nucleases have been developed to enable precise installation of desired modifications without DSBs. These tools include base editing (BE), which enables the installation of any of the four nucleotide transitions (A-to-G, C-to-T, T-to-C, or G-to-A) and some transversions (C-to-G), prime editing (PE), which enables the programmable modification of a small region (generally 50 bp) of DNA, and CRISPR-associated transposons or PE+recombinases, which have the potential to enable the installation of gene-sized fragments. While in theory these tools greatly broaden the scope of genome modifications available to researchers, in practice their initial application was partially constrained by three major types of limitations: 1) efficiency and generalizability, 2) specificity, and 3) deliverability. As such, recent developments have focused on minimizing these limitations through a variety of natural discovery, rational engineering, or evolution approaches. In this thesis, I will describe efforts to improve the efficiency, generalizability, and specificity of precision genome editing technologies using engineering and directed evolution, with a focus on base editing. In chapter two, I describe the successful incorporation of natural and engineered variants of the canonically used SpCas9 to expand the targeting range of adenine base editing. In chapter three, we show that a novel, sequence-agnostic functional selection is better able to evolve high-activity variants of non-SpCas9 orthologs capable of efficient base editing. Lastly, in chapter four, I describe early efforts to incorporate a negative counterselection into the sequence-agnostic selection, with the goal of developing a generalizable method for generating tailor-made, genome site-specific Cas variants for use in precision genome editing. Together, the work described herein include novel tools that enhance the generalizability of precision genome editing while also providing a set of methods that will enable the further development of improved variants.