Of the three major genomes that contribute to our health - the nuclear genome, the mitochondrial genome and the bacterial genomes of our microbiome – only the mitochondrial genome has yet to be programmed. The highly conserved mitochondrial DNA (mtDNA) genome consists of 37 genes essential for electron transport chain function, and mutations in mtDNA can cause disease. Much is not known about these diseases because of a lack of sufficient models, and there currently is no long-term treatment for these patients. The ability to edit the mtDNA genome would enable both the interrogation of gene function as well as facilitate the study of mtDNA-based genetic diseases.
Here, we describe to our knowledge the first evidence of mtDNA programming using site-specific mitochondrial DNA enzymes. We use zebrafish as our model system to test these tools. Zebrafish mitochondrial DNA are similar in gene number and order to mammalian models and are readily microinjected to deliver editing tools. We initially deleted a 5kb fragment of DNA between mt-nd5 and mt-atp8 by using novel custom enzyme tools. This corresponding deletion is found in humans to cause the disease Kearns-Sayre Syndrome (KSS). Approximately 35% of injected animals carried a detectable deletion, with no deletion detectable in uninjected animals. The resulting deletion was sequence-verified to be mtDNA. To develop mitochondrial DNA carrying a single gene deletion, we then demonstrated that a single 2kb region encoding mt-nd4 could also be deleted using the same strategy.
We next asked whether we could improve the efficiency of editing by selecting for deleted mtDNA using a site-specific nuclease that targets only the non-deleted genomes. Double-strand breaks are not repaired efficiently in mitochondria, and enzymes that cause DSBs degrade their target rather than inducing NHEJ repair. This represents a major technical challenge in editing mtDNA using standard approaches deployed in the nucleus. By coinjecting the mt-nd5 and mt-atp8 custom editors along with a nuclease targeting mt-nd4, we were able to detect deletions in over 90% of injected animals, a 2.5-fold improvement. The respiration capacity of these animals were measured, and those injected with custom mitochondrial editors and a nuclease were 4 times lower than uninjected animals and 3 times lower than nuclease-alone injected animals.
This method of using custom mitochondrial DNA editors combined with nucleases to program the mtDNA genome enable more precise mtDNA disease modeling and single gene function studies.