The evolution of new genes and functions is considered to be a major contributor to biological diversity in organisms. Through de novo origination, “duplication and divergence”, and horizontal gene transfer, organisms can acquire new genetic material that can evolve to perform novel functions. In this thesis, we investigate how functional trade-offs, “gene duplication and amplification”, and neutral divergence contribute to the emergence of a new function from a preexisting gene.

 In Paper i, we investigated the ability of Salmonella enterica to compensate for the loss of peptide release factor 1 (RFI) and the potential of peptide release factor 2 (RF2) to gain a new function to replace RFI. The amplification of RF2 and accumulated mutations within RF2 were the main evolutionary routes by which the fitness cost was restored. However, further characterization of the evolved RF2 showed a toxic effect to the cell due to the termination on tryptophan codon (UGG). This evolutionary trade-off - which we named “collateral toxicity” - might present a serious barrier for evolving an efficient RF2 to replace RF1.

In Paper ii, we determined whether we could evolve a generalist enzyme with two functions (HisA + TrpF) from the specialist enzyme HisA, which can only synthesize histidine. In a previous study, we showed that HisA evolved a TrpF activity through strong trade-off trajectories. Here, we developed a selection scheme in which we constantly selected for keeping the original function (HisA), while intermittently selecting for the new function (TrpF). Our results showed that all evolved lineages shared the same “stepping stone” mutations in the hisA gene, which enabled them to grow well in the absence of both histidine and tryptophan. Additional accumulated mutations in the hisA gene gave the strains an increased ability to grow without both amino acids, indicating that the HisA enzyme evolved to be an efficient generalist.  

In Paper iii, we explored how differences between diverged orthologs influence evolvability. We generated artificial orthologs using a random mutagenesis approach. First, we screened for orthologs with a lower HisA activity and then selected for orthologs with a higher HisA activity; these steps were repeated in alternating rounds. We then tested the ability of each ortholog to evolve  TrpF activity. As expected, the orthologs showed varying abilities to evolve the new function. In particular, orthologs with higher HisA activity levels showed both a higher potential to evolve the new function and a higher TrpF activity when they acquired the new function. 

Translation termination in bacteria involves precise reading of stop codons (UAA, UAG, UGA) and coordinated peptidyl-tRNA hydrolysis by the class-I release factors (RFs) on the ribosome (70S). This thesis investigates the evolutionary and post-translational modification mechanisms of these RFs and the concurrent effect on bacterial translation termination.

Unlike eukaryotes with a single RF, bacteria have two RFs. Release factor 1 (RF1) reads UAA and UAG; release factor 2 (RF2) reads UAA and UGA codons. So, why do bacteria host two RFs? To answer this, we performed an in vivo evolution experiment to explore how RF2 evolution compensates for the loss of RF1. Characterization of the evolved RF2 mutants, specifically E167K RF2, using both in vivo and in vitro peptide release assay reveals its ability to read the RF1-specific UAG codon and also the tryptophan (UGG) codon, displaying a functional trade-off termed “collateral toxicity”. Further, fast-kinetics-based peptide release assay shows that E167K RF2 is generally efficient in peptide release on UAA and UGA but significantly more efficient on UAG and UGG than WT RF2. This increased efficiency is primarily due to the higher affinity of E167K RF2 to the 70S. Our 2.8 Å cryo-EM structures demonstrated K167 to be engaged in hydrogen bond interactions with the rRNA, that are absent in WT RF2 having E167. Further, the mutant displays somewhat destabilized conformation when unbound, bypassing the conformational change check-point and facilitating the reading of near-cognate UAG and UGG codons. 

Post-translational methylation on the conserved GGQ motif of RF1/2 increases the efficiency of translation termination, but its role in termination accuracy was unknown. We compared the methylated and unmethylated variants of RF1/2 for cognate and near-cognate codon recognition. The unmethylated RFs exhibited lower termination accuracy, likely caused by the loss of conformational stability in the absence of GGQ methylation.

In summary, these studies reveal the compensatory evolution of E167K RF2 as a tighter binder of the ribosome with the destabilized compact conformation that enhances UAG reading at the expense of UGG reading. Additionally, our study shows that GGQ methylation maintains the conformational stability of RF2 and facilitates accurate stop codon recognition.