Understanding how base editing tools work at the molecular level

You may have seen it in the news recently: a baby in Pennsylvania with a rare genetic disorder was healed with a personalized treatment that repaired his specific genetic mutation. The treatment was created using a form of gene editing called base editing -a method created by Alexis Komor when she was a postdoctoral scholar in molecular biologist David Liu's group at Harvard University.

Since that work was published in 2016, Komor, who is now an associate professor of chemistry and biochemistry at the University of California San Diego, has continued to study base-editing tools to better understand and further develop their capabilities. Her latest research, published in Nature Communications, outlines the way certain DNA repair proteins can be manipulated to produce desired outcomes.

Our genomic DNA is comprised of four bases - cytosine (C), thymine (T), guanine (G) and adenosine (A). These bases join together into approximately 3 billion different base pairs, arranged in a double-helix structure.

Humans are 99.9% identical in their genetic makeup, while the remaining 0.1% accounts for any difference between one person and another. Where one person has a C base, another person might have a T base. There are millions of genetic variations possible between any two people, and although many are harmless, others can lead to debilitating or terminal genetic diseases.

For many people with genetic diseases, gene editing is their only hope of a cure.

Gene editing is traditionally done using CRISPR-Cas9 to make a physical change in the DNA. A guide RNA directs the Cas9 protein to a specific DNA location, where Cas9 completely severs the DNA - called a double stranded break. There are many proteins within the cell that can detect DNA damage and then fix it through a process called a repair pathway.

Normally these pathways take the two broken ends of the DNA and fuse them back together, called a ligation. With CRISPR-Cas9, as the number of breaks and repairs increases, so do unwanted insertions and deletions, called indels. When this happens, the DNA strand no longer matches the original and the editing process ends.

As a postdoc, Komor found a way to achieve gene-editing with higher efficiencies and a lower incidence of indels by avoiding double-stranded breaks. She called this new class of tools "base editing" because it chemically changes a DNA base one letter at a time.

With base editing, not only do we achieve a better outcome, but the steps leading to the outcome are also improved. Double-stranded breaks can be toxic and can cause cell death. They can also cause larger-scale genomic rearrangements because you're physically cutting up the DNA. Base editors avoid that."

Alexis Komor, associate professor of chemistry and biochemistry, University of California San Diego

Komor developed two tools, an adenine base editor (ABE), which converts an A base to G base, and a cytosine base editor (CBE), which converts a C base to a T base. Base editors make conversions through an intermediary. In the case of CBEs, the cytosine is first converted to uracil, a nucleic acid found in RNA. During repair, the DNA reads the uracil as thymine.

Although there's no double-stranded break, base editors do create a nick in one strand. An enzyme is attached to the Cas9 and chemically changes the base. CBEs can have a 90-95% conversion rate with minimal unwanted byproducts.

We know the base editor works, but how? That's the main question Komor's group wanted to answer. They wondered how the uracil was being handled by the cell. What role does the nick play? How do all of the different proteins in the cell affect the editing outcomes?

One particular protein called uracil N-glycosylase (UNG) finds and deletes uracils. When that protein is present, the incidence of unwanted outcomes rose. When UNG was inhibited, CBE efficiency increased. But Komor's group didn't have a full understanding of the process.

To uncover the answer, the team used a technique called gene knockdown, which turns down a gene's ability to express itself. They did this for every single different DNA repair or processing protein in the human genome – 2,015 proteins in total.

Then they used green fluorescent protein markers to identify cells containing the desired edits, either C to T mutations or the UNG-affected C to G edits, while discarding the UNG-affected edits they didn't want.

They did this work using a CRISPRi screen, which takes a guide RNA to a gene of interest and reduces its expression. Komor's CRISPRi cells had two guide RNAs - one to activate the fluorescent protein maker with the CBE and one directed to knock down the specific repair protein. All told, Komor had a library of around 12,000 different guide RNA combinations for the CRISPRi screen.

Once they collected all the cells that had fluoresced green, they sequenced the guide RNAs to see which genes were knocked down (and therefore their expression was detrimental to base editing).

They discovered that a ligase called Lig3 inhibited base editing. Ligases attached the ends of broken DNA strands back together and when Lig3 was present, CBE editing was lower.

"We think that the Lig3 can sneak in and seal that nick back together. Then we don't have a chance for that uracil to be turned into a thymine anymore," stated Komor. "It's like the Lig3 is working against us."

They also found a repair pathway called mismatch repair that actually helped cytosine base editing. The MutS-alpha protein complex, a component of mismatch repair, is composed of two proteins and is expressed throughout a cell's lifetime. When it was knocked down, conversions of C to T dropped. Komor thinks it's because MutS-alpha recognizes the uracil intermediate and helps convert it to thymine.

Why all the fuss? If the tool works, does it matter why? Yes, says Komor.

"These base editing tools are good, but they're not perfect," she stated. "If we can better understand how they're functioning in the cell, that will help us understand how we can improve their efficiency or even create new types of base editors."

Komor also notes that it's important to understand how these tools work for safety reasons. Does base editing activate an unforeseen cellular response? Is there DNA damage or cell death? The more we know about the mechanisms of the base editors and how cells respond, the better positioned we are to observe and stop unintended consequences.

"That's one of the reasons this work was supported by the National Science Foundation. This is basic science - understanding why things are the way they are and how they function in certain environments," Komor said.

Often it takes decades to see the real-world impact of basic science research, but Komor's base editing techniques were being deployed in hospitals in much shorter time. She notes that, to date, at least 10 people have been saved using CBE. Additionally, there are several clinical trials underway, meaning it's likely that number will increase greatly in the coming years.

This research was supported, in part, by the National Science Foundation (MCB-2048207), the National Institutes of Health (T32GM146648) and the Research Corporation for Science Advancement through the Cottrell Fellowship (27975).