Enhancing precision genome editing agents through directed evolution and protein engineering

Abstract

Over the past decade, advances in genome editing with CRISPR-Cas systems have transformed basic research and have enabled the development of targeted therapeutic approaches to treat life-threatening diseases. Due to the growing interest and application of these technologies in basic and translational science, we seek to address limitations in the safety, programmability, and efficacy of emerging genome editing tools. Here, I describe the use of protein engineering and directed evolution, a method that employs Darwinian principles to evolve biomolecules with unique properties, to enhance the efficiencies of two genome editing platforms.

In the first portion of my thesis, I describe efforts to improve prime editing (PE), a programmable genome editing technology capable of installing all point mutations, small insertions, or small deletions into the genome of living cells. We employed a multipronged approach to engineer and evolve the prime editor protein which comprises of a Cas9 nickase and a reverse transcriptase (RT) enzyme. Our strategies included: 1) surveying RTs from diverse phylogenetic origins for activity, 2) rationally engineering RTs and the Cas9 nickase for improved performance by leveraging sequence and structure information, and 3) evolving RTs and the Cas9 nickase using a phage-assisted directed evolution platform tailored for PE. Collectively, these efforts resulted in next-generation prime editor variants PE6a-i that exhibit enhanced editing across different sites in various cell types including T cells and patient-derived fibroblasts. When taken in vivo, the improvements observed with PE6 were even more substantial; remarkably, PE6 installed a 42-base pair (bp) sequence with 40% efficiency, representing a 23-fold improvement compared to the previous prime editor used for in vivo applications. Collectively, we demonstrate that PE6 prime editors enhance prime editing in vitro and enable new classes of edits to be precisely installed in vivo.

A long-standing goal in gene therapy and genome engineering has been to efficiently and precisely integrate large DNA sequences at any target genomic locus. Classical methods for large gene integration, including programmable nucleases followed by either random end- joining or homology-directed repair rely on the generation of deleterious free double-stranded DNA breaks, whereas prime editing can only install sequences up to 800 bp at low efficiencies. In the second portion of my thesis, I describe efforts to enhance prime editing-assisted site- specific integrase gene editing (PASSIGE), a platform that couples the programmability of PE with the specificity of large serine recombinases (LSR) to integrate large DNA sequences into the genome. Although PASSIGE can precisely integrate multi-kilobase (kB) DNA sequences (up to 5.6 kB) into the genome in a targeted manner, it’s modest integration efficiencies (up to 6.8%), largely attributed to the suboptimal activity of the LSR, limit its broad applicability. To address this limitation, we established a recombinase phage-assisted directed evolution platform and employed it to evolve Bxb1, one of the most well-characterized LSR. Through directed evolution and rational engineering, we generated highly active recombinase variants, which, when used in PASSIGE, consistently achieved targeted gene integration efficiencies exceeding 30% across several genomic loci. These improvements establish PASSIGE as one of the most versatile and efficient methods for targeted gene integration, significantly expanding its potential applications across various fields including research, medicine, biotechnology, and agriculture. Finally, I conclude by discussing the implications of my thesis findings and outline possible directions for further investigation.