Skip to main content
Premium Trial:

Request an Annual Quote

Broad Institute Researchers Develop Faster, More Efficient Base Editor

NEW YORK – Researchers led by the Broad Institute's David Liu and University of California, Berkeley's Jennifer Doudna have developed a new adenine base editor (ABE) that is far faster and more efficient than its predecessor. It's also the first base editor that can be paired with a variety of Cas9 and Cas12 homologs. 


In a paper published on Monday in Nature Biotechnology, the researchers described their phage-assisted non-continuous and continuous evolution of the deaminase component of ABE7.10, which resulted in the new base editor ABE8e. Until now, applications of ABEs have been constrained by the limited compatibility of the deoxyadenosine deaminase component with Cas homologs other than SpCas9. But ABE8e is compatible with several Cas9 and Cas12 homologs, increasing its editing efficiency, the researchers said.  


It also contains eight additional mutations that increase activity 590-fold compared with ABE7.10. ABE8e is more processive than ABE7.10, which could be beneficial for screening, disrupting regulatory regions, and multiplex base editing applications. 


"We speculated that different Cas proteins might differ in how long they hold open the target DNA site for the evolved deaminase domain to convert the target A [base], and therefore evolving a deaminase domain that is as fast as possible might allow the resulting base editor to more efficiently install the target base conversion before the Cas protein part of the base editor lets go of the DNA," Liu explained in an email to GenomeWeb. "We used our phage-assisted continuous evolution (PACE) system to rapidly evolve our original ABE into a much faster variant by rewarding fast base editing with ABE gene replication, then gradually dialing up the speed required to survive this Darwinian selection, while allowing the ABE to acquire new mutations." 


To determine if their continuous evolution system resulted in improved deamination kinetics, the researchers compared the DNA adenine deamination kinetics of ABE7.10 and ABE8e in vitro by measuring A-to-I conversion. They found that deoxyadenosine deamination happened at a rate 590-fold faster for ABE8e than for ABE7.10, suggesting that it may be fast enough to yield efficient DNA adenine deamination even when coupled to non-SpCas9 Cas effectors that have decreased residence times on DNA substrates. 


The researchers also applied the PACE selection system in order to broaden the applicability of ABEs by improving their compatibility with a variety of Cas domains. When they transfected the base editor into HEK293T cells with a single-guide RNA targeting a site with a cognate PAM for SpCas9 (NGG), SaCas9 (NNGRRT), or LbCas12a (TTTV), the researchers found large improvements in A-T-to-G-C base editing efficiency — up to 9.4-, 12-, and 24-fold when tethered to SpCas9, SaCas9, and dLbCas12a, respectively, without any evident changes to the very low indel formation levels of ABE7.10. Most notably, they observed efficient A-T-to-G-C base editing activity with dLbCas12a for the first time, despite the lack of a suitable Cas12a nickase to nick the non-edited strand to enhance editing efficiencies. 


Further, they observed substantially enhanced editing with the ABE8e variant compared to the ABE7.10 variant with all Cas homologs. For example, editing levels at the second-most edited A base ranged from 1.7 percent to 20 percent with SpABE7.10, while editing levels ranged from 18 percent to 86 percent with SpABE8e.  


"Different Cas proteins each provide access to a different set of PAM sequences, and therefore a different set of disease-associated mutations" than the current base editor, Liu added. 


The researchers were concerned that the higher speed of ABE8e might also result in a higher number of off-target edits, particularly considering that ABE7.10 has been shown to have low off-target activity compared to most genome editing agents. While they did observe higher off-target editing by ABE8e compared with ABE7.10, these activities were greatly reduced when they installed an additional mutation into ABE8e called V106W, Liu said, adding, "Importantly, this additional mutation did not impair the average on-target editing efficiency of ABE8e, demonstrating how our understanding of some structure-function relationships within base editors is allowing tuning of important properties such as the reduction of off-target DNA and off-target RNA editing." 


Liu and his colleagues are now turning their attention to using ABE8e and the V106W variant to directly correct pathogenic mutations in animal models of human genetic diseases.  


"Thus far, ABE8e is working remarkably well," he said. "Which is not to say that there won't be other desired features that we or other labs might want to install in future base editors. But this study … and many additional innovations in base editing from other labs around the world demonstrates that the base editing field continues to tailor these molecular machines to offer increasingly powerful and therapeutically relevant properties."