Advancing precision genome and transcriptome engineering

Abstract

The cells within all living organisms contain genomic DNA that encodes the information for their structure, function, and capabilities. According to the central dogma of biology, this DNA is transcribed to RNA and this RNA is translated to proteins that perform these cellular functions. Each level within this gene expression cascade is tightly regulated to ensure that every cell produces the correct quantities of each RNA and protein for its designed role. Thus, even though most cells within an organism harbor the same genomic sequence, different gene expression programs can give rise to diverse cell types, such as neurons, T cells, and pancreatic β islet cells. In humans, dysregulation at any stage within the gene expression cascade—for example from a DNA mutation or from defective splicing of an RNA transcript—can often lead to cellular dysfunction and disease. The ability to precisely manipulate the genome and transcriptome therefore enables the study and potential correction of these pathogenic defects. In this thesis, I describe the development and advancement of programmable tools for editing the DNA and RNA within living cells.

In Chapter 2, I describe efforts to elucidate the mechanism of prime editing, a recent genome editing technology that affords high precision and versatility. Using a chimeric Cas9 nickase–reverse transcriptase fusion protein complexed with a guide RNA, prime editing makes a nick on a targeted strand of genomic DNA and synthesize new DNA from this nick. The resulting 3′ DNA flap containing the desired gene editing product is then processed by endogenous DNA repair pathways to either incorporate the edit into the genome or reject the edit. To identify the cellular determinants of prime editing, we performed pooled CRISPR interference screens to assess the effects of hundreds of genetic perturbations on prime editing outcomes. These screens revealed DNA repair pathways that strongly antagonizes the efficiency and fidelity of traditional prime editing approaches. In addition, we also explored the genetic modulators of other prime editing systems, revealing the involvement of diverse cellular processes. Collectively, these efforts have deepened our understanding of prime editing and informed new strategies for improving its capabilities.

Next, in Chapter 3, I describe the development of next-generation prime editing technologies. Despite the versatility and precision of prime editing, its widespread adoption has been limited by low and inconsistent efficiency. We pursued three orthogonal approaches to enhance prime editing efficiency. First, we manipulated DNA repair to bias the productive resolution of prime editing intermediates. Second, we engineered prime editing guide RNAs with greater stability and resistance to exonucleolytic degradation. Finally, we improved the cellular expression, localization, and activity of the prime editor protein component. Together, these three approaches synergistically improve prime editing efficiency and substantially expand the applicability of prime editing.

In Chapter 4, I build upon our work on prime editing mechanisms by systematically characterizing and learning the determinants of prime editing efficiency. We assessed the performance of prime editing across many diverse edits and used these data to train a deep learning model that predicts prime editing efficiency and optimal guide RNA design. These efforts not only reveal new insights about prime editing, but also provide a general approach for maximizing prime editing efficacy.

Finally, in Chapter 5, I describe a novel tool for manipulating cellular transcriptomes in a precise and programmable manner. Among the biomolecules in the central dogma, RNA is the most dynamically regulated. One such form of regulation that has only been recently appreciated is chemical modifications that decorate RNA nucleotides within transcripts. N6-methyladenosine, or m6A, is the most abundant non-cap messenger RNA modification in humans and has been shown to play a role in differentiation, immunity, neurogenesis, as well as cancer and psychiatric disorders. Despite the significance of m6A, current research on this modification is largely correlative due to a lack of methods for selectively perturbing specific m6A states. To address this need, we developed programmable tools for site-specific installation of m6A in RNA, enabling its functional interrogation on transcripts of interest.