Precision Editing of Nuclear and Mitochondrial Genomes

Abstract

The ability to convert a target nucleotide sequence into any desired nucleotide sequence has been a longstanding goal in genome editing. RNA-guided CRISPR-Cas systems have transformed this field because genome editing agents could now be directed to almost any target sequence by simply varying the choice of a guide RNA. CRISPR-Cas systems have since been engineered extensively by our lab and others to perform a myriad of precise DNA modifications, including introduction of a single base pair change using base editing, and performing targeted insertion, deletion and multiple base pair replacement using prime editing. Apart from the nuclear genome, the mitochondrion contains its own genome that encodes for proteins and RNAs critical for energy production. Pathogenic point mutations in the mitochondrial genome (mtDNA) have been identified for mitochondrial disorders including MELAS, LHON, MERRF and Leigh’s disease. Thousands of somatic mtDNA mutations remain uncharacterized for their association with human diseases and ageing. Given the importance of the mtDNA in human health, there is a critical need to develop tools that enable precise mtDNA manipulation. While base editors and prime editors have been shown to edit the nuclear DNA in living cells with high efficiencies, the challenge of RNA delivery to the mitochondrial matrix have precluded the use of CRISPR-based systems for mtDNA engineering. This dissertation seeks to address the challenge of precision mtDNA editing. We developed a CRISPR-free mitochondrial base editor (DdCBE) that enables the first precise C•G to T•A base pair conversion within the mtDNA. DdCBE contains a bacterial deaminase toxin, DddA, that exhibits unprecedented double-stranded DNA cytidine deaminase activity. We engineered non-toxic split halves of DddA, then fused them to programmable DNA-binding TALE array proteins to reassemble active DddA at target DNA site, resulting in efficient and sequence-specific base editing in both human mtDNA and nuclear DNA. Next, we used laboratory evolution to generate DdCBE variants that result in improved activity and expanded targeting scope. The canonical DdCBE showed modest editing efficiencies at selected mtDNA sites and was limited to editing cytosines in a TC context. Using phage-assisted continuous evolution (PACE), we evolved DdCBEs for higher editing activity at TC and non-TC targets. Compared to canonical DdCBEs containing wild-type DddA, those with DddA6 improved mtDNA base editing efficiencies at TC by an average of 3.3-fold across nine tested loci. DdCBEs containing DddA11 offered substantially broadened HC (H = A, C, or T) target sequence context compatibility for both mitochondrial and nuclear base editing. We next used these evolved DdCBEs to efficiently install disease-associated mtDNA mutations in human cells at non-TC target sites, resulting in cells with impaired oxidative phosphorylation and reduced respiration rates. Finally, we observed a modest increase in average mitochondrial genome off-target editing associated with DddA6 and DddA11 for specific TALE designs, but overall ratios of on-target:off-target editing efficiencies of evolved DddA variants were comparable to those of canonical DdCBE. DdCBE enables the installation of disease-associated mtDNA mutations in human cells lines and animal models to accelerate preclinical research. Its potential as a future therapeutic for debilitating mitochondrial disorders may be realized with further developments and innovations.