DNA targeting by compact Cas9d and its resurrected ancestor

Abstract

The type II-A CRISPR effector SpCas9 has gained widespread popularity as an efficient and programmable genome editing tool. However, much remains to be known about novel compact variants that may overcome some limitations of current systems1,2. Recently, alternative CRISPR-Cas systems with highly compact nucleases capable of genome editing in mammalian cells have been discovered through metagenomic analysis of uncultivated microbes, including Cas9d (a type II-D CRISPR-Cas effector)3. Here, we report the cryo-EM structures of a Cas9d nuclease (747 amino acids in length) in multiple functional states, revealing a stepwise process of DNA targeting involving a conformational switch in a REC2 domain insertion. Our structures provide insights into the intricately folded guide RNA which acts as a structural scaffold to anchor small, flexible protein domains and facilitate DNA target recognition. We find that the sgRNA can be truncated by up to [~]25% yet still retain activity in vivo. We also show that despite preferentially targeting an NGG PAM, Cas9d exhibits a unique mechanism for PAM recognition. Finally, we identify the first Cas9d smaller than 800 amino acids exhibiting robust nuclease activity in mammalian cells. Using ancestral sequence reconstruction, we demonstrate that it is possible to generate compact nucleases capable of efficient genome editing by expanding the diversity of Cas9d families. Collectively, our results provide mechanistic insights into the evolution and DNA targeting of diverse type II CRISPR-Cas systems, providing a molecular blueprint for future rational re-engineering of minimal RNA-guided DNA nucleases.