Developing a polymerase-driven method to incorporate base modifications & Structurally characterizing m6A- and colibactin- modified nucleic acids

Abstract

Living cells need to adapt their processes to environmental stimuli more quickly and more precisely than the slow changes in DNA genomes, or even the fast turnover of messenger RNAs. To this end, chemical modifications on the bases of DNA and RNA constitute a layer of epigenetic and epitranscriptomic regulation on top of the underlying nucleic acid sequences. These function either as detectable markers forming a binding surface for specific interactions or as structural switches interfering with base pairs and thereby changing the underlying DNA or RNA structure. Accordingly, a given modification's function is largely determined by its location on the base, relative to the canonical base pairs and by the type of modification: small modifications on the Watson-Crick edge (e.g. N6-methyladenosine) or the Hoogsteen and sugar edge (e.g. 5-methylcytosine) alter the secondary structure and the binding surface, respectively. However, if the N6-methyl moiety is pointing away from the base pairing surface, or the 5- methylcytosine is engaged in non-canonical (i.e. Hoogsteen) base pairs, their respective functional consequences may be inverted. Meanwhile, large adducts (e.g. bacterial genotoxins) form structures that cause a steric hindrance disrupting the original function of the base altogether. This interplay between such local effects and the biological processes it affects and regulates further depends on the sequence environment around the modification, as well as the available binding partners. Thus, to understand the mechanism of the downstream effects of base modifications, we need to know their structural effects in their appropriate context. Current recognition of the prevalence and chemical diversity of base modifications in all domains of life highlight both the complexity and importance of their effects on the structure and function of nucleic acids. As such, methods to detect, quantify, and construct modified DNA and RNA samples for characterization are necessary. In this dissertation, we first define a novel method for the segmental and site-specific incorporation of modified bases in RNA and then use it to characterize an N6-methyladenosine (m6A) mediated switch on the hepatitis B virus (HBV) pre-genomic RNA. In addition, we also describe a structure of the bacterial genotoxin colibactin crosslinking the strands of a DNA helix. The available tools to introduce site-specific modifications are complex and laborious due to the challenging nature of RNA chemistry. Current methods rely on chemical synthesis with strict limits on the product length and enzymatic synthesis suffering from the functional limitations of the available enzymes. Here we used a DNA polymerase mutant from the bacterium Thermococcus gorgonarius (Tgo) and developed and optimized a protocol to present SegModTeX ('Segmental labeling and site-specific Modifications by Template-directed eXtension'), a versatile, one-pot, copy- and-paste approach method for site-specific and segmental labeling of RNA. By precise, stepwise construction of a diverse set of RNA molecules for solution NMR, we demonstrated the technique to be superior to RNA polymerase driven and ligation methods owing to its substantially high yield, fidelity, and selectivity. We also showed the technique to be useful for incorporating some fluorescent and a wide range of other probes, as well as for the synthesis of RNAs starting with any base. This significantly extends the toolbox of RNA biology in general. The high yields in each step also allow for the stepwise construction of long RNAs carrying site-specific modifications, with possible future applications beyond structural RNA biology including the construction of mRNAs. Next, using this method, we investigated the mechanism of an m6A modification in the pre-genomic RNA (pgRNA) of HBV. HBV, a reverse transcribing virus with a small 3.2 kb dsDNA genome, causes chronic infections in humans leading to an estimated one-million deaths per year. The circular extrachromosomal genome is the product of packaging and reverse transcription of the pgRNA by the viral reverse transcriptase. Recently, an m6A located at the end of a stem-loop termed epsilon, was shown to be required for pgRNA packaging by its translation product called core protein. Thereby a single m6A constitutes a functional switch between the pgRNA's mRNA and replication template functions in cellulo. Here we performed sequence and base pair conservation analysis and found regions beyond epsilon, but proximal to the m6A site, to form a hitherto unreported bulge structure, which we confirmed by NMR. The bulge structure exists in a tertiary conformation known as an arginine sandwich motif, a protein binding surface known for its specific interaction with arginines and thus a likely candidate for the hitherto unknown core binding site. Interestingly, the motif's structure is independent of the presence of the adjacent m6A modification, indicating it modulates a pre-existing binding surface. These results highlight a role for m6A in the tertiary structure of a functional switch in the HBV life cycle, which may have implications for core binding during pre-genome packaging. Lastly, colibactin is a genotoxic natural product of clb+ E. coli strains, only recently shown to give rise to increased tumor formation in mouse models and detected more frequently in colorectal cancer patients. These observations combined with colibactin induced mutational patterns resembling those of double strand DNA breaks, led to the hypothesis that colibactin forms DNA interstrand crosslinks (ICLs). Here, we elucidate the structure of the colibactin-DNA ICL and describe an unexpected α-iminiumketone present in the central region of colibactin that serves as a key base recognition element, similar to arginine interactions with DNA in minor grooves. Combined, our structural data provides a strong foundation for understanding colibactin’s structure and its crosslinking of DNA and explains its specificity for the 5′-AAWWTT- 3′ motif (W = A or T). The tightly bound molecule almost entirely concealed in the minor groove also indicates how colibactin evades detection by the cellular repair enzymes continuously scanning the DNA. Altogether, this dissertation emphasizes the importance of base modifications for the function of underlying RNAs and DNAs, as well as the urgency of adequate methods for biology.