CRISPR-Cas systems, which consist of protein and RNA components, originally developed as a natural defense mechanism of bacteria to fend off intruding viruses. Over the last decade, re-engineering these so-called "gene scissors" has revolutionized genetic engineering in science and medicine. The tools can be programmed to find a specific location in our DNA and edit the genetic information in a precise manner. For example, a disease-causing mutation in the DNA can be reverted to its healthy state.
It was recently discovered that Cas proteins evolved from much smaller proteins, with TnpB being the progenitor of Cas12. Since the large size of Cas proteins creates challenges when trying to deliver them to the right cells in the body, recent studies tried to use their smaller evolutionary progenitors as a genome editing tool. The problem with these small alternatives is that they function less efficiently.
This hurdle has now been tackled by a research team headed by Gerald Schwank from the Institute of Pharmacology and Toxicology at the University of Zurich (UZH) together with colleagues from the ETH Zurich. "By engineering the small but powerful protein TnpB, we were able to design a variant that shows a 4.4-fold increase in efficiency of modifying DNA - making it more effective as a gene editing tool," says Schwank.
TnpB proteins are found in a variety of bacteria and archaea. The TnpB studied by the researchers comes from the bacterium Deinococcus radiodurans. This microbe survives cold, dehydration, vacuum and acid, and is one of the most radiation-resistant organisms known to humans. The compact TnpB protein has previously been shown to work for genome editing in human cells, albeit with low efficiency and limited targeting ability due to its recognition requirements when binding DNA.
Better binding ability and broader range of DNA target sequences
Therefore, the researchers optimized TnpB so that it edits the DNA of mammalian cells more efficiently than the original protein. "The trick was to modify the tool in two ways: first, so that it more efficiently goes to the nucleus where the genomic DNA is located, and second, so that it also targets alternative genome sequences," says Kim Marquart, PhD candidate in Gerald Schwank’s lab and first author of the study.To identify which features in the DNA sequences of the target sites determine the genome editing efficiency, the researchers tested TnpB at 10,211 different target sites. In collaboration with the team of Michael Krauthammer, also professor at UZH, they developed a new artificial intelligence model capable of predicting TnpB editing efficiencies at any given target site. "Our model can predict how well TnpB will work in different scenarios, making it easier and faster to design successful gene editing experiments. Using these predictions, we achieved up to 75.3% efficiency in mouse livers and 65.9% in mouse brains," Marquart adds.
"For the animal experiments, we were able to use clinically viable Adeno-associated viral vectors to efficiently transport the tools into mouse cells. Due to its small size, the TnpB gene editing system can be packaged into a single virus particle," Marquart says. In contrast, the CRISPR-Cas9 components have to be packaged into multiple virus particles, which means that higher vector doses need to be applied.
In the current project, the researchers studied whether the TnpB tool could be employed to treat patients with familial hypercholesterolemia. This genetic disease leads to lifelong severely elevated high cholesterol affecting approximately 31 million people globally. The disease increases the risk of early-onset atherosclerotic cardiovascular disease. "We were able to edit a gene that regulates cholesterol levels, thereby reducing the cholesterol in treated mice by nearly 80%. The goal is to develop similar gene editing strategies in humans in order to treat patients suffering from hypercholesterolemia," says Gerald Schwank.
Literature
Kim Marquart et al. Effective Genome Editing with ISDra2 TnpB and Deep Learning-Predicted ’RNAs. Nature Methods. 23 September 2024. DOI: 10.1038/s41592’024 -02418-z