Adenosine-to-inosine RNA editing catalyzed by the family of enzymes Adenosine Deaminases Acting on RNA (ADARs) plays several important roles in biological homeostasis of metazoans. Hydrolytic deamination of adenosine by ADARs results in the formation of inosine, which is read by cellular machinery as guanosine. This A-to-G information change can impact miRNA/siRNA biogenesis, splicing, mRNA recoding, and the innate immune response. Humans encode three ADAR genes, with ADAR1 and ADAR2 being catalytically active enzymes with distinct substrate preferences. Structural biology has been shown to be an important tool for understanding the basic mechanism and substrate specificity of ADARs. Information gleaned from structural studies can be used to optimize therapeutic site-directed RNA editing platforms, develop new genome editing technologies, and design ADAR inhibitor drugs for the treatment of cancer. Chapter 1 of this dissertation begins with an overview of the various modifications that occur in RNA, with an emphasis on adenosine-to-inosine RNA editing by ADARs. Next, I discuss the mechanism and substrate specificity of ADARs, the diverse biological implications of ADAR editing, and how aberrant editing is implicated in several human diseases. I go on to discuss how ADARs have been used to develop therapeutic site-directed RNA editing technologies. Lastly, I finish with a discussion of the structural information of ADARs to date and how these structures have provided important mechanistic and substrate specificity information.
Chapters 2 and 3 describe highly collaborative studies to optimize guide oligonucleotide design for the site-directed RNA editing with endogenous ADARs. Chapter 2 examines modified nucleotides at the orphan base position to enhance on-target editing by endogenous ADARs. Benner’s base dZ was shown to enhance ADAR base-flipping, catalysis, and on-target editing by wild type ADAR. A high-resolution crystal structure of the deaminase domain of ADAR2 bound to a duplex with dZ at the orphan base position shows close contact between the dZ Watson-Crick-Franklin face and the glutamate flipping residue with bifurcated hydrogen bonding. Chapter 3 examines ways to enable robust editing at difficult to edit sites containing a guanosine 5’ to the target adenosine. Previous structural studies suggest a clash between the backbone of G489 and the exocyclic amine of the 5’ G. These studies show a G: G adjacent to the target site enables robust editing. A 2.7 Å resolution structure shows that the 5’ guanosine adopts the syn conformation presenting its Hoogsteen face to the Watson-Crick-Franklin face of the guanosine pair on the guide strand. The adoption of the syn conformation flips the offending exocyclic amine out of the minor groove and ameliorates the clash with G489.
Chapter 4 deviates away from human ADARs to characterize the first active ADAR ortholog form Hydra vulgaris, a freshwater poly that is a model organism for aging and regeneration. Traditional genome editing technologies including CRISPR/Cas9 are challenging to implement in Hydra vulgaris, making gene knockout studies problematic. Previous studies have shown that ADARs can edit the DNA strand of DNA/RNA hybrids, making it possible that endogenous hydra ADAR could be used for loss-of-function studies. This chapter shows that hyADAR has a short 5’ binding loop correlating with a glutamine flipping residue, a motif present in medusozoan cnidarians while anthozoan cnidarians more closely resemble human ADARs. Furthermore, this chapter details the 2.0 Å crystal structure of the deaminase domain of hyADAR and its unique dimeric form not observed in human ADARs.
Chapter 5 returns back to human ADARs, detailing experiments to optimize the purification of ADAR1 for structural studies The recent implication of the cytoplasmic ADAR1 p150 in cancer evasion has made structural studies for drug development imperative, but historically ADAR1 has been challenging to work with in vitro. This chapter shows a new intein-based protein purification results in highly pure and highly active protein. Furthermore, the addition of equimolar L-arginine and L-glutamate to purification conditions allows the protein to be concentrated to a level to begin structural studies.
In Chapter 6, I examine the highly conserved hydrogen bonding network that links the buried inositol hexakisphosphate molecule to the ADAR active site. Protein expression and activity studies show that mutation of this network is highly detrimental to protein expression and/or folding further supporting the structural role of IP6.
Lastly Chapter 7 details preliminary studies towards the development of an ADAR genome editing tool. I detail the development of a fluorescence-based assay for the detection of DNA editing by ADARs in human cells. Next, I show that Next Generation Sequencing detected minimal editing of a pre-formed DNA/RNA hybrid in HEK293T cells in a deaminase activity dependent manner. Lastly, I begin to explore the propensity for DNA strand invasion by guide oligonucleotides containing locked nucleic acids.