DNA and RNA Polymerases with Expanded Substrate Scope : Synthesis of Modified Nucleic Acids Using Engineered Polymerases Generated by Directed Evolution

Abstract

Modified RNA is used in a series of nucleic acid-based, cutting-edge technologies such as antisense oligonucleotide, small interfering RNA (siRNA) or as aptamer. Furthermore, modifications within RNA are used for studying RNA structure and dynamics. The 2’-position of the sugar moiety is a desirable position for introducing modifications within the RNA target because modifications attached to this position provide the RNA with desired properties such as increased stability. Additionally, the 2’-position is important because it can be advantageous for introducing selenium into RNA for use in X-ray crystallographic structure determination. 2’-Methylseleno (2’-SeCH3)-modified RNA is accessible by traditional chemical synthesis, and thus limited in its available length due to the inherent limitations of the synthesis on solid support. In the present work, the enzymatic synthesis of 2’-SeCH3-modified RNA has been elaborated by using engineered variants of T7 RNA polymerase. This approach provides access to long selenium modified RNA, which cannot be accomplished by chemical synthesis. Furthermore, this research shows the effectiveness of an established fluorescence readout-based screening assay that was used to identify a variant of T7 RNA polymerase with an increased acceptance of 2’-SeCH3-NTPs by directed evolution.First, the enzymatic synthesis of 2’-SeCH3-modified RNA was investigated by using two T7 RNA polymerase mutants known to have a high tolerance for 2’-modified NTPs. These investigations were conducted in collaboration with Prof. Dr. Ronald Micura from Innsbruck University in Austria, where his research team assisted in synthesizing the four 2’-SeCH3-NTPs. In vitro transcription reactions were performed in the presence of 2’-SeCH3-UTP or 2’-SeCH3-CTP involving the T7 RNA polymerase mutants previously reported by Sousa (Y639F, H784A)[1] and Ellington (E593G, Y639V, V685A, H784G)[2]. It was discovered that both mutants can incorporate the 2’-SeCH3-modified nucleotides into an 89-mer transcript at various positions, whereas the wildtype T7 RNA polymerase fails to do so. Both mutants possess amino acid substitutions at position Y639 and H784, two residues known to be involved in nucleotide discrimination.[3,4] Encouraged by these results, saturation mutagenesis of amino acid positions Y639 and H784 was applied to construct a library comprising of 3200 variants. This library was screened for active T7 RNA polymerase variants by using a screening assay based on fluorescence readout after transcription by using the RNA specific stain SYBR Green II. The assay was established in 384-well plates to screen mutant libraries of T7 RNA polymerase in high-throughput directly with diluted E. coli lysates containing the overexpressed variants. The screening procedure was validated by screening the first library and then used to screen a second mutant library composed of 1600 variants. The second mutant library was constructed by random mutagenesis based on the Ellington mutant (E593G, Y639V, V685A, H784G) using error-prone PCR. Active variants from both libraries were purified and tested in 32P-based transcription assays for increased incorporation efficiencies of the 2’-SeCH3-modified nucleotides. In doing so, a T7 RNA polymerase variant from the second library was identified, referred to as 2P16, that exhibits an increased acceptance for the tested 2’-SeCH3-NTPs compared to the parental mutant from which it derived. DNA sequencing revealed that this 2P16 has seven mutations in addition to the four initial ones, totalling 11 mutations that surprisingly did not affect its ability to polymerize natural NTPs. These results highlight the benefit of directed protein evolution and show that the established screening system is well-suited to identify T7 RNA polymerase variants with altered properties like the ability to incorporate 2’-modified nucleotides. Furthermore, this approach resulted in the identification of a T7 RNA polymerase mutant that can further be used for the efficient enzymatic synthesis of 2’-SeCH3-modified RNA for application in X-ray crystallography.The second part of this work focuses on the enzymatic synthesis of four unnatural nucleic acid polymers comprising of CeNA, HNA, ANA and dXNA (cyclohexenyl nucleic acid, 1,5-anhydrohexitol nucleic acid, arabino nucleic acid and deoxyxylo nucleic acid). The investigation used a selection of DNA polymerases and mutants of eukaryotic, prokaryotic, and archaic origin representing different polymerase families. Over the past two decades, the scientific community has intensively studied the chemical and enzymatic synthesis, the reverse transcription and structural characteristics of synthetic genetic polymers, also referred to as xeno-nucleic acids (XNA). This research was originally driven by the question of how life evolved on earth and why DNA and RNA were naturally selected over other possible information storage systems. In this context, it was investigated in the present work whether DNA polymerase mutants with broadened substrate spectra, which had either been generated by site-directed mutagenesis or by directed protein evolution, are able to synthesize the unnatural nucleic acids in a DNA-dependent manner. Homopolymer formation of the four different sugar-modified XNAs was examined in 32P-based primer extension assays in order to identify DNA polymerases that show potential to polymerize the respective modified nucleotides. In addition to testing 2’-deoxyxyloadenosine 5’-triphosphate (dXATP), four sugar modified thymidine 5’-triphosphates were also tested: hTTP, araTTP, ceTTP and dXTTP. The investigations revealed that family-B polymerases and mutants were able to efficiently incorporate arabino, cyclohexenyl and hexitol nucleotides into a growing XNA strand, whereas family-A polymerases had significant problems with the sugar-modified nucleotides. Furthermore, it was observed that the tested DNA polymerase mutants showed increased incorporation behaviours of the nucleotide analogues compared to their respective wildtype polymerase. These results are promising and indicate that family-B polymerases might be well-suited to evolve into DNA-dependent XNA polymerases. The investigations on the DNA-dependent enzymatic synthesis of deoxyxylose nucleic acids (dXNA) have revealed that dXATP and dXTTP are poor substrates for all of the tested DNA polymerases irrespective of their original polymerase family. These observations are consistent with the proposed orthogonality of dXNA. Nevertheless, a mutant of human DNA polymerase beta was observed to efficiently incorporate and elongate the tested deoxyxylose nucleotides and might therefore be an interesting candidate to develop into a DNA-dependent dXNA polymerase using directed protein evolution.