570. Transient Reconstitution of Cas9 Protein for Safer Gene Editing In Vivo - A Ferry Tale Starring rAAV Capsid

Abstract

Transforming the gene editing technology into therapeutic uses encounters several obstacles including the concern over safety. These gene editing platforms, such as the Cas9/sgRNA system, have been shown to induce off-target DNA double-stranded breaks (DSBs) throughout genomes, which is associated with toxicity. Such off-target effects not only stem from the intrinsic ambiguity of DNA sequence recognition by nucleases, but also attribute to the prolonged presence of an active gene editing system in a given cell. As a result, off-target DSBs accumulate over time, and ultimately lead to genotoxicity. To mitigate the potential toxicity due to prolonged expression of a gene editing system in vivo, we sought to transiently deliver the key component, i.e. the endonuclease protein, which will fulfill the task of inducing permanent gene editing followed by natural degradation. Specifically, we use the VP2 protein of AAV capsid as a protein delivery vehicle to ferry the Cas9 protein in vivo. We first constructed a sensitive gene editing reporter plasmid, and demonstrated that co-transfection of the reporter plasmid and a plasmid expressing the SpCas9-VP2 fusion protein induced gene editing in HEK293 cells. We then modified our rAAV packaging system to include a plasmid expressing VP1 and VP3, and another plasmid expressing either the SpCas9-VP2 fusion protein or the EGFP-VP2 fusion protein. We successfully produced EGFP-AAV2 (EGFP protein grafted on the AAV2 capsid). However, we were not able to produce rAAV particles carrying SpCas9 protein, likely because the large size of SpCas9 protein interfered with the AAV packaging process. Therefore, we split SpCas9 into halves aiming to utilize split intein-mediated protein trans-splicing (PTS) to transiently reconstitute the full-length SpCas9 (Figure)(Figure). When the two parts of a split intein (termed IntN and IntC, respectively) fused with two proteins, the split intein is able to mediate PTS, resulting in the generation of a fusion protein with the intein being spliced out. We generated plasmids expressing the fusion proteins SpCas9N-IntN and IntC-SpCas9C-VP2, respectively, and demonstrated productive intein-mediated reconstitution of SpCas9-VP2 protein in HEK293 cells by co-transfection. Importantly, co-transfection of plasmids expressing SpCas9N-IntN and IntC-SpCas9C-VP2 in HEK293 cells led to gene editing based on the reporter assay. Guided by structural analysis and prediction, we strategically screened and identified the SpCas9 split sites close to the C-term of SpCas9. Therefore, the IntC-SpCas9C protein to be grafted on VP2 is equal or smaller than EGFP. Because we were able to produce EGFP-AAV2, we reasoned that the IntC-SpCas9C-VP2 will likely be amenable to the rAAV packaging process, which is underway. If successful, our novel off-target mitigating strategy described here will greatly facilitate the translational application of gene editing technology in vivo. In addition, this transient in vivo protein reconstitution approach is applicable to the delivery of therapeutic but highly toxic proteins for the treatment of cancer and other diseases.FigureReconstituting SpCas9 by intein-mediated protein trans-splicing (PTS). In the first AAV vector, the AAV genome encodes the N-terminal portion of SpCas9 (SpCas9N) fused with IntN. The second AAV vector carries Intc and the C-terminal portion of SpCas9 fused to VP2. In vivo transduction of the first AAV vector produces the fusion protein SpCas9N-lntN, which is followed by delivery of the second vector. PTS occurs to reconstitute the full-length SpCas9 protein.View Large Image | Download PowerPoint Slide