Author: Andrew Veenis

Publisher:

Published: 2022

Total Pages: 0

ISBN-13:

Noncovalent interactions play an integral role in RNA structure and function. RNA residues contain myriad donating and accepting groups that can form hydrogen bonds. Metal cations interact favorably with the highly negative RNA backbone and with various nucleobase atoms. The pi system of the aromatic nucleobases facilitate stacking interactions between the nucleobases and interactions with charged moieties. Hydrogen bonding and cation binding can directly assist in the function of small RNA enzymes (ribozymes) which use four general catalytic strategies (denoted [alpha], [beta], [gamma], and [delta]) to catalyze a self-cleavage reaction. Noncovalent interactions can also influence the properties of specific RNA moieties, such as the pKa of the N1 of guanine. In this work, cheminformatics and molecular dynamics (MD) simulations are employed to study how these interactions contribute to ribozyme catalysis and the properties of RNA moieties with comparison to past experimental or theoretical results. Noncovalent interactions, catalysis by small ribozymes, and the computational methods used herein are introduced in Chapter 1 along with the objectives of this thesis. In Chapter 2, we focus on the use of catalytic strategies by small ribozymes. These ribozymes catalyze a transphosphorylation at their scissile phosphate which involves the attack of the nucleophilic O2' on a adjacent phosphorus and release of the O5' leaving group, resulting in cleavage of the RNA phosphate backbone. The four general catalytic strategies entail alignment of the nucleophilic O2', phosphorus, and O5' leaving group for nucleophilic attack ([alpha]), deprotonation of the nucleophilic O2' ([gamma]), neutralization of the negative charge on the nonbridging phosphoryl oxygens (NPOs) of the scissile phosphate ([beta]), and stabilization of the O5' leaving group ([delta]). We compared how the catalytic strategies were used in various families of small ribozymes by analysis of structural data. This was achieved through the implementation of a computational pipeline where the use of catalytic strategies by small ribozymes could be inferred from the positioning of atoms within their active sites. Over 80 experimentally determined or computationally modeled structures of the glmS, hammerhead, twister, and hairpin ribozymes were initially considered followed by an extension of this analysis to four crystal structures of the pistol ribozyme. Evidence of all four strategies are observed for each of these small ribozymes. Numerous different moieties reside near the NPOs and O5' leaving groups, indicating a diversity in how the ribozymes employ the [beta] and [delta] strategies. Additionally, the active sites appear to be mostly devoid of uracil. Unlike the other three nucleobases, uracil lacks an amidine functionality which may be important in promoting catalysis. Moreover, we identified two new catalytic strategies, [gamma]' and [gamma]'', based on our findings in this study and prior experimental and theoretic results. We also compared scissile phosphates to nonscissile phosphates and found that scissile phosphates are unique in their use of the catalytic strategies. In Chapter 3, we investigated the pKa of the nucleophilic O2' in the hairpin ribozyme. Deprotonation of the O2' alcohol (the [gamma] strategy) results in a more nucleophilic O2' oxyanion. The pKa of this alcohol was hypothesized to be shifted down towards neutrality in the environment of the ribozyme active site, favoring its deprotonated state. Because the unperturbed pKa of an O2' is roughly 12-14, measuring the pKa of the nucleophilic O2' experimentally would be difficult since the ribozyme would become denatured and degraded under elevated pH conditions. Therefore, MD simulations were used to calculate this pKa. By simulating the dynamics of all of the atoms in the ribozyme, this approach also enabled the identification of factors that likely influence the O2' pKa. Contrary to our hypothesis, the pKa was estimated to be shifted away from neutrality to 18.5 ± 0.8. The O2' consistently resides near the pro-SP NPO in its protonated and deprotonated states. This results in strong O2'-to-pro-SP NPO hydrogen bonding in the O2'-protonated state and electrostatic repulsion between the two oxygens in the O2'-deprotonated state. Both of these interactions would contribute towards an increased pKa. Additionally, a decrease in hydration of the O2' upon deprotonation was observed and may also contribute to the pKa elevation. The O2' made an inner-sphere coordination with at least one Na+ ion throughout 5.2% of the O2'-protonated trajectory. In a separate calculation, we estimated that the O2' pKa drops to about 16.1 ± 1.0 when this interaction occurs. From our structural analysis of ribozymes in Chapter 2, we noticed that an exocyclic amine of guanine is often positioned near the nucleophilic O2' in small ribozymes and postulated that it may help depress the O2' pKa. In the hairpin ribozyme, the exocyclic amine of G8 resides near the O2'. When this group is replaced by a hydrogen, the O2' pKa drops to 16.4 ± 1.7. Because the pKa did not increase in the absence of the exocyclic amine, this result suggests that this group does not have a pKa lowering effect on the O2'. Overall, the high pKa calculated for the O2' supports a catalytic mechanism where deprotonation occurs during, rather than before, nucleophilic attack when the O2' pKa decreases. In Chapter 4, algorithms were written to search structural data for guanines that exist in three possible rare forms. The first is a guanine that has a stabilized zwitterionic resonance form due to dual hydrogen bond donation by its exocyclic amine. The other two are guanines that lack a proton at their N1. These forms of guanine could exist through deprotonation, yielding an anionic guanine, or through tautomerism, resulting in a neutral enol tautomer. Guanines candidates that are identified by the algorithms are awarded a score based on their potential interactions with their environment. The functionality of the algorithms was demonstrated by applying them to a crystal structure of two 70S ribosomes. In total, over a thousand candidates from both ribosomes were identified, most with relatively low scores. Six of these candidates are presented and help facilitate discussion on how the research can be advanced. Potential ways the algorithms can be improved to search for guanine candidates are considered. After they are refined and undergo further testing, the algorithms could be used to collect statistics. The incorporation of quantum mechanical (QM) calculations and experimental methods to further study guanine candidates are also discussed. The thesis is concluded in Chapter 5, and future directions are proposed based on the work presented in Chapters 2-4. In Chapter 2, a notable difference is observed in the use of the catalytic strategies between the scissile and nonscissile phosphates in small ribozymes. Thus, algorithms that search for the potential use of catalytic strategies could be combined with computational methods of RNA 3D structure prediction to find new families of small ribozymes. In Chapter 3, MD simulations revealed that a strong hydrogen bond forms between the nucleophilic O2' and a nearby pro-SP NPO and that a Na+ ion occasionally interacts with the O2'. These interactions can be interrogated experimentally using phosphorothioate substitutions and a series of alkali metal cations. In Chapter 4, algorithms were designed to search for rare forms of guanine and to score candidates based on their interactions. These algorithms can be expanded to search for rare forms of the other three nucleobases and to consider additional interactions when scoring the candidates.