Project Summary / Abstract The current gold standard to deliver gene therapies, adeno-associated virus (AAV), has shown safe and stable transgene expression in many applications. However, the combination of AAV with CRISPR introduces unique risks of genotoxic side effects from long-term nuclease expression and integration of viral DNA into sites of DNA breaks in the host genome. Furthermore, the ssDNA genome of AAV limits its packaging capacity so that two AAVs must be used to deliver Cas9 from Streptococcus pyogenes (SpyCas9) and its guide RNA, limiting efficiency and increasing costs. Interest has switched to using a smaller Cas9 orthologue derived from Staphylococcus aureus (SauCas9), although several enzyme features have been shown to be significantly different. Furthermore, pre-existing immunity to AAV capsids, as well as SpyCas9 and SauCas9, has been identified in humans, potentially limiting the therapeutic use of these molecules. While the host response to the two orthologues is dependent on previous exposure, SauCas9 was found to elicit a stronger immune response than SpyCas9 in human subjects when measured by immunoblot, ELISA, and ELISpot assays. Therefore, although SauCas9 is a smaller nuclease that enables delivery by AAV, there are several questions about safety that must be addressed. Previous work in the Doudna laboratory has shown that the SpyCas9 endonuclease can be engineered with cationic residues to make the ribonucleoprotein (RNP) inherently cell-penetrating in neural precursor cells in vitro and in neurons in vivo by non-viral delivery. I have now engineered SauCas9 to also act as a cell-penetrating RNP. The purpose of this research proposal is to evaluate the outcomes of engineered Cas9 orthologues delivered as RNPs or AAVs in the mammalian brain. Aim 1 employs an unbiased protein engineering and screening strategy to make small deletions across Cas9 that improve the cellular host immune response to the protein. Aim 2 will test these variants as ribonucleoproteins or as adeno-associated viruses delivered by stereotaxic injection into the striatum of a fluorescent reporter animal model to assess genome editing outcomes and the host immune response; and Aim 3 will apply these findings to genetically correcting SOD1 mutations in an animal model of amyotrophic lateral sclerosis. Taken together, developing a cell-penetrating and immune-stealthy Cas9 RNP for transient and local genome editing would improve the safety of CRISPR therapies and accelerate the pace of clinical trials that could immediately benefit patients.