Reprogramming Human Transcription Factors for Epigenetic Editing

Programmable genome and epigenetic editing platforms have changed how we do research today while providing enormous therapeutic potential. The domains that were initially used for these purposes, Cys2His2 zinc fingers and homing endonucleases, were difficult to engineer and offered inconsistent efficacy. Not surprisingly, these platforms were quickly replaced by TALE and CRISPR systems due to their high activity and extreme ease of use. However, a zinc finger design code would provide therapeutic advantages such as smaller size for routine AAV delivery and reduced immunogenic risk, while providing a better approximation of natural transcription factor function for academic applications. Therefore, to overcome the limitations of zinc finger design we have spent nearly a decade evolving zinc fingers in different adjacent finger contexts to provide a detailed understanding of both specificity and domain compatibility. We used the screen of 28 libraries and over 114 billion protein-DNA interactions to provide the data necessary to train an AI-based model of zinc finger specificity and array design. We have applied this model to simply design zinc finger arrays that can be used as nucleases, activators, and repressors with similar activity to alternative platforms. This model represents a solution to a 30-year design challenge for the routine production of highly active zinc finger arrays!

Zinc fingers are the most common DNA-binding domain utilized by eukaryotic transcription factors, representing roughly 50% of the human transcription factors. Therefore, we reasoned that our design model could be used to commandeer the natural functions of these transcription factors. We have applied our model to a series of human transcription factors and seamlessly reprogrammed their DNA-binding domains while keeping the rest of the protein intact. In all cases this approach has allowed us to direct the transcription factor's function to novel sequences across the genome. This transcription factor reprogramming enables the targeting of activation, repression, or both to any sequences in the genome. As these factors are small enough that several are able to be packed into a single AAV cassette, reprogrammed transcription factors provide robust tools for multiplexed epigenetic editing with proteins completely composed of human components.