Evolution and Engineering of RNA-based Macromolecular Machines

Abstract

The “RNA World Hypothesis” posits that extant life was likely preceded by a simpler primordial biology lacking DNA and proteins, relying instead on RNA for both heredity and catalysis. An essential component of an RNA world scenario would be an RNA “replicase” – a ribozyme capable of self-replication as well as copying other RNA sequences. While such a replicase has not been found in nature, past work in our laboratory identified artificial ribozymes that are promising candidates for further engineering towards self-replication. One such example is the t5+1 RNA polymerase ribozyme, which polymerises triplets to produce the complementary strand of a given RNA template; while incapable of self-synthesis in its entirety, t5+1 can copy the sequence of its catalytic subunit if provided in 5 separate segments. In Chapter 2 of this thesis, I describe in vitro selection methods to further evolve the t5+1 RNA polymerase ribozyme, identifying mutations that improve its catalytic activity, while also mapping the comprehensive fitness landscape of the ribozyme. In Chapter 3, I explore the potential of appending novel domains to the ribozyme in order to increase its contacts upstream of the triplet ligation junction. Using directed evolution, I identify new 5’ domains which when grafted onto t5+1 alter its triplet substrate preference, leading to improved polymerisation activity on some RNA templates that were previously difficult to copy.

The invention of translation paved the transition away from the RNA world, providing a method for the coded-synthesis of proteins. In modern cells, a key player in translation is the ribosome, a large macromolecular machine that reads messenger RNAs in a triplet register, converting each codon into the correct amino acid. In Escherichia coli, RNA comprises two-thirds of the total mass of the ribosome, in the form of the 16S, 23S and 5S ribosomal RNAs (rRNAs). Despite the importance of the bacterial ribosome in biotechnological applications and as a target for antibiotics, most rRNA mutations have not been characterised for their effects on translation. In Chapter 4, I develop a novel system for the high-throughput quantification of mutational effects in the 23S rRNA of E. coli; I apply this system to map the higher-order fitness landscape of a short rRNA segment that can be read by short-read Next Generation Sequencing (NGS). In Chapter 5, I describe a new cloning method that enables the identification of mutations anywhere in the 23S rRNA through sequencing of a short 20nt barcode. Using this method, I map the fitness of all 23S rRNA point mutants.