| dc.description.abstracteng | Posttranscriptional and posttranslational modifications are key regulatory mechanisms to expand biological properties of proteins and nucleic acids in living cells. A tremendous number of chemical modifications is found on all RNA species, with the highest diversity in composition and density found in transfer RNA (tRNA). Occurring all over the tRNA body, modifications increase tRNA stability, induce proper folding, and modulate translational fidelity. Non-canonical nucleobases include simple modifications as methylations and alkylations to highly complex extensions as seen in wybutosine and queuosine (Q), of which the latter requires a complete base exchange. Because mammalian tRNA molecules can harbour on average 13 modification at the same time, studying single modifications can be challenging. Especially since modifications exhibit a certain degree of “cross-talk” by influencing each other as seen for Q. Recent studies revealed a link between the hypermodified 7-deaza guanine derivative Q occurring at the wobble base position 34 (Q34) of tRNAAsp and the methylation of cytosine at position 38 (m5C38) introduced by Dnmt2. Here, deposition of m5C38 is stimulated upon prior Q34 modification with an enigmatic underlying biochemical mechanism. Thus, the aim of this study was to gain structural and biochemical insights into RNA modifying enzymes introducing (i) a methyl group to guanine at position 7 (m7G) by TrmB, (ii) methylation of cytosine at position 4 (m5C) by Dnmt2, and (iii) ribosyl transfer and isomerisation to preQ1 by QueA.
m7G is found not only as the mRNA-cap structure to protect mRNA from degradation, but also in tRNAs at position 46 in the variable loop introduced by the TrmB/Trm8 enzyme family. Prior to this thesis, several crystal structures of TrmB/Trm8 enzymes have been determined, revealing differences in their biological assembly. In this thesis, the first crystal structure of the homodimeric B. subtilis TrmB in complex with the methyl group donor S-Adenosylmethionine (SAM) and post catalytic product SAH is reported. Analysis of the SAM/SAH crystal structures revealed conserved ligand binding across TrmB/Trm8 enzymes. Structural, biochemical, and computational approaches revealed a 2:2 binding stoichiometry of tRNA to protein, and resulted in the TrmB-tRNAPhe complex model in which two tRNA molecules bind to the homodimeric TrmB. Interestingly, biochemical analysis of TrmB activity at physiological SAM conditions showed a half-of-the sites reactivity, even though each monomer of TrmB is capable of tRNA and ligand binding. Subsequently, the presented biochemical and structural data give valuable insights into TrmB activity and substrate binding.
The second part of the thesis focusses on the m5C writer enzyme Dnmt2. Dnmt2 substrate specificity was long enigmatic, as Dnmt2 was first identified as DNA methyltransferase (MTase), however, exhibiting weak methyltransferase activity on DNA. Since the discovery of highly specific MTase activity of Dnmt2 on tRNAAsp, a few more substrates have been identified, including tRNAGly and tRNAVal. However, only tRNAAsp harbours Q at position 34 which was identified to increase Dnmt2 methylation activity. Of the four Q34 tRNAs tRNAAsp, tRNAAsn, tRNATyr, and tRNAHis only the latter harbours a cytosine at position 38, rendering tRNAHis a putative Dnmt2 substrate. Even though m5C38 in tRNAHis could not be observed so far, human Dnmt2 was identified to modify human tRNAHis during the course of this thesis. In contrast to prior work on S. pombe Dnmt2, stimulating effects upon Q modification could not be observed in the human context, implying different functions of Q34 in S. pombe and humans. Furthermore, Dnmt2 activity on tRNAHis was found to be highly pH-dependent rendering tRNAHis from cognate to non-cognate tRNA by shifting the pH from 8.0 to 7.4. The identification of tRNAHis as Dnmt2 substrate gives more insight into Dnmt2 substrate specificity.
The S-adenosylmethionine:tRNA ribosyltransferase isomerase (QueA) was the study focus of the last part of this thesis. QueA is of special interest as (i) QueA catalyses the ribosyl-transfer in the penultimate step of Q-biosynthesis, (ii) Q is a major driving force in the virulence of Shigella bacteria, and (iii) QueA inhibition represents and interesting starting point in treating Shigella infection. Prior to this thesis incomplete QueA crystal structures have been determined. During this work, a full length QueA structure model on basis of a QueA crystal structure determined in this thesis was proposed. Investigation of this model enabled the identification of a putative SAM-binding pocket and amino acids possibly involved during catalysis. Furthermore, molecular docking experiments gave rise to a putative QueA-tRNA complex model. The proposed complex model gives valuable insight into QueA activity and provides a model for initial computer-based fragment screening in order to perform structure-based drug design to inhibit Shigella bacterial infection and treat shigellosis.
Overall, this thesis provides insights into protein-RNA complexation and substrate specificity by studying complex formation by means of biochemical, structural, and computational methods. Even though, both subunits of the homodimeric TrmB are capable of binding SAM and tRNA, TrmB exhibits at physiological SAM concentration a half-of-the sites reactivity. Furthermore, a connection of enzyme activity to stress response was identified, as human Dnmt2 shows altered enzyme activity on tRNAHis depending on the used pH value. Lastly, the identification of the putative SAM-binding pocket and proposition of a QueA-tRNA complex model represents a valuable starting model for computational and experimental methods in structure-based drug design. | de |