In a biological beating of swords into plowshares, researchers have converted a bacterial toxin into a genome editing tool that, for the first time, can make precise changes to DNA in mitochondria, the cell’s power plants. The tool, which worked in lab experiments with human cells, could open the door to new studies of—and one day therapies for—dozens of hard-to-treat diseases caused by mutations in mitochondrial DNA (mtDNA). These rare conditions, which include Leber hereditary optic neuropathy and lethal infantile cardiomyopathy, collectively affect about one in 4000 people. Until now, research on these illnesses has been stymied in part because there was no way of reproducing the mutations in strains of mice.
The new DNA editor is “quite innovative and pioneering,” says Joseph Hacia, a medical geneticist at the University of Southern California. “It’s highly likely that it will work in mice, and I’m hopeful that it will have therapeutic implications down the line.”
Thousands of mitochondria, which likely evolved from bacteria, exist in most every human cell, and each contains its own genes. Researchers have made little headway correcting the genetic defects that lead to mitochondrial diseases, many of which are caused by “point mutations.” In such mutations, a single DNA base—adenine, cytosine, thymine, or guanine—is replaced by one that disrupts a needed protein or otherwise impairs the power plant. One difficulty is that a key component of the most famous genome editor, CRISPR, is too large to enter mitochondria. And other genome editors that can reach mtDNA do not have the subtlety to correct point mutations.
To create the new tool, which combines features of CRISPR and an older technology called transcription activator-like effectors (TALEs), three teams joined forces. “Part of what made the project so fun to work on, and ultimately successful, is the fact that three labs came together organically because the science led us to each other,” says David Liu, a chemist at the Broad Institute and the last author of a paper that describes the work in Nature today.
The first step toward that collaboration was a finding by Marcos de Moraes, a postdoc working in a University of Washington, Seattle, lab run by microbiologist Joseph Mougous. The team studies how bacteria secrete toxins to kill off other bacteria when there are scarce resources. In 2018, de Moraes stumbled on a bacterial toxin that helps catalyze the conversion of cytosine into uracil. (This base is normal in RNA, but in DNA it naturally converts into thymine.) What’s more, the toxin creates this mutation on both strands of the DNA double helix, which had not been seen before.
Mougous, who like Liu is a Howard Hughes Medical Institute (HHMI) investigator but did not know the scientist, emailed him to ask whether he was interested in collaborating because Liu’s lab previously had developed cytosine and adenine base editors that used a similar catalyzing agent—an enzyme known as a deaminase—combined with two components of CRISPR technology. These deaminases only work on single-stranded DNA. The CRISPR duo includes a strand of RNA that helps untwist the double helix and shuttles the deaminase to precise targets on single strands. But this guide RNA (gRNA) cannot enter mitochondria.
Mougous says the two groups recognized from the outset that the new base editor had no obvious advantage over the ones Liu’s team had developed. “That pushed us to look for its niche,” he says, which proved to be altering mtDNA. Vamsi Mootha, a Broad Institute specialist in mitochondrial dysfunction and another HHMI investigator, also joined the collaboration. “I’ve been in the field for 25 years, and this is the first time ever that we’ve been able to [manipulate cells], and, voilà, a few days later, you have edits to the mitochondrial DNA.”
TALE and zinc-finger nucleases, another genome editor that predates CRISPR, can both sever the double-stranded DNA of mitochondria, destroying them. That has the protentional to treat some mitochondrial diseases, but it cannot correct mtDNA point mutations. To make a more refined tool, Beverly Mok, a graduate student in Liu’s lab, attached the toxin-derived deaminase from Mougous’s lab to a TALE, a protein that can enter mitochondria and, like gRNA, leads the complex to the target.
Because the deaminase is toxic to mitochondria, the researchers split it into two halves that come together only at the mtDNA target. “We had to tame the beast,” Liu says. In experiments with human cells, the conversion of cytosine to thymine occurred up to 50% of the time, the collaboration reports today in Nature. Importantly, they did not find a significant number of “off-target” edits, which potentially can cause serious harm.
Michio Hirano, who studies mitochondrial diseases at Columbia University and was not involved with the work, says this is a “very clever” strategy that “addresses a holy grail in the mitochondrial field.”
On top of trying to create cell and mouse models of human mitochondrial diseases, the researchers will look for other bacterial deaminases that can modify double-stranded DNA. They also hope to improve editing efficiency and reduce off-target edits so that mtDNA base editing can eventually be tested in humans. “We recognize that it’s a long road to get there,” Liu says. “I’m hopeful that the energy and the resourcefulness of the field will take these tools and continue to improve them now that we have the blueprints.”