This proposal aims to manipulate DNA deaminase enzymes to generate hyperactive and controllable base editors that can be targeted for precise gene editing. Base editing of the immunoglobulin locus by AID, the ancestral member of the AID/APOBEC family of cytosine deaminase enzymes, normally initiates maturation of antibody responses in B-cells, while APOBECs provide protection against retroviruses. Out of their physiological context, when DNA deaminases are directed by catalytically-impaired CRIPSR/Cas proteins, their base editing activity can be used to introduce targeted mutations at a desired genomic locus. While this system offers a potentially powerful means to edit the genome for biological or therapeutic purposes, base editors have two barriers that limit their broader application in basic and translational research. First, DNA deaminases have naturally evolved to be constrained enzymes with low overall catalytic activity, as hyperactivation is associated with increased oncogenic mutations. Second, when dysregulated, AID/APOBECs are known to act outside of their targets, promoting cancer mutagenesis, chromosomal translocations, and resistance to chemotherapy. Given that natural regulatory constraints on DNA deaminases are lost in base editor complexes, these constructs pose similar risks to the genome. In this proposal, we harness our extensive knowledge of the mechanism, structure and function of deaminase enzymes in order to overcome these challenges. For one, hyperactive deaminases have been generated to overcome the naturally attenuated activity, and we will exploit these variants to evaluate the hypothesis that increasing the deamination rate can improve the efficiency of the base editing reaction, while simultaneously improving precision. Second, we have devised split deaminases that can only be reconstituted at the targeted locus under the control of a small molecule. This strategy newly offers spatiotemporal control, a critical requirement that will facilitate the use of base editors in the lab and is essential to therapeutic applications in patients. Given the wide range of potential uses for base editors, we will demonstrate the importance of efficiency and control broadly across diverse genomic sites, and then specifically by generating enhanced chimeric antigen receptor expressing T (CAR-T) cells as a model system. The tools developed here will globally advance deaminases as base editors and will readily translate to other innovations in CRISPR/Cas proteins, and to genome engineering more generally.