Expanding the Capabilities of Genome Editing

Abstract

The ability to precisely and efficiently edit DNA sequences within the genome of living cells has been a major goal of the life sciences since the first demonstration of restriction cloning. Recently, the use of RNA-programmable nucleases from CRISPR systems has been transformative for researchers seeking to knock out genes to interrogate their function. Nucleases stimulate gene knockout by generation of a double-stranded DNA break (DSB) in the target gene, a process that leads to incorporation of mutations – mostly in the form of insertions or deletions (indels) - around the cleavage site. The inherent stochasticity of this process means that DSB-based methods do not enable precise DNA changes with single nucleotide resolution. When such precise changes are required, researchers can supply a donor DNA that contains the desired mutation and homology arms surrounding the cleavage site. Through the cellular process of homology-directed repair (HDR), the desired mutation may be incorporated into the genome. However, because it requires a DSB, this process produces a great excess of undesired indels and is low in efficiency. Previous work in the Liu lab established that a catalytically disabled Cas9 nucleases (a nickase, capable of cleaving only one of the two DNA strands) can be fused to a cytosine deaminase enzyme and a uracil DNA glycosylase inhibitor protein to generate a cytosine base editor (CBE), a construct capable of converting a cytosine to a uracil in genomic DNA. Upon DNA replication or repair, the uracil is repaired to a thymine, so CBEs convert a cytosine to a thymine (C-to-T). This tool was transformative thanks to its high efficiency in generation of C-to-T mutations and low prevalence of undesired additional mutations. In this thesis, I begin by addressing two of the challenges associated with the original CBE - DNA target specificity and delivery. First, we engineer and characterize a high-fidelity version of the CBE that reduces undesired editing at off-target loci in the genome. Second, we establish a method to deliver the cytosine base editor as a ribonucleoprotein complex, and demonstrate its utility for editing DNA in the mouse inner ear. Next, I present our characterization of a new class of base editor – the adenine base editor (ABE) – an editor that is capable of converting adenosine to inosine in the context of genomic DNA. Inosine is converted to guanine by DNA replication or repair. We characterize the DNA and RNA specificity of ABE, demonstrating that the ABE is surprisingly specific to its target site in DNA but can edit RNA in a non-targeted fashion at a low level. We directly compare the ABE to a recently developed method for HDR. The adenine base editor is capable of incorporating the tested point mutations with an >1000-fold improvement in the ratio of desired product: undesired insertion than HDR. Finally, I developed and characterized a new method for performing HDR that does not rely on generation of DSBs. Instead, we develop a fusion construct between hRad51 and Cas9(D10A) nickase that can mediate efficient HDR at a nick site. This improves the HDR:indel ratio by up to 52-fold and maintains the efficiency of Cas9 DSB-based HDR. I discover that mutants of hRad51 unable to bind to its native binding partners (BRCA2 or hRad51) further improve the ratio of desired product: undesired insertion or overall efficiency of the tool.