Search

Bacteria often evolve resistance to phage through the loss or modification of cell surface receptors. In Escherichia coli and phage λ, such resistance can catalyze a coevolutionary arms race focused on host and phage structures that interact at the outer membrane. Here, we analyse another facet of this arms race involving interactions at the inner membrane, whereby E. coli evolves mutations in mannose permease-encoding genes manY and manZ that impair λ's ability to eject its DNA into the cytoplasm. We show that these man mutants arose concurrently with the arms race at the outer membrane. We tested the hypothesis that λ evolved an additional counter-defence that allowed them to infect bacteria with deleted man genes. The deletions severely impaired the ancestral λ, but some evolved phage grew well on the deletion mutants, indicating that they regained infectivity by evolving the ability to infect hosts independently of the mannose permease. This coevolutionary arms race fulfils the model of an inverse gene-for-gene infection network. Taken together, the interactions at both the outer and inner membranes reveal that coevolutionary arms races can be richer and more complex than is often appreciated., Microbiology, 167 (5), ISSN:1350-0872, ISSN:1465-2080

While the major function of transfer RNA is conserved across the tree of life, organisms differ in the types and numbers of tRNA genes that they carry. The evolutionary mechanisms behind the emergence of different tRNA gene sets remain largely obscure. Here, we report the rapid and repeated evolution of a tRNA gene set in laboratory populations of the bacterium Pseudomonas fluorescens SBW25. Deletion of the non-essential, single-copy tRNA gene serCGA from SBW25 results in a sub-optimal tRNA gene set. Compensation occurs within 35 generations via large (45-290 kb), direct, tandem duplications in the chromosome. Each duplication contains a serTGA gene, and is accompanied by a two-fold increase in tRNA-Ser(UGA) in the mature tRNA pool. This work demonstrates that the composition of tRNA gene sets – and mature tRNA pools – can readily evolve by duplication of existing tRNA genes, a phenomenon that could explain the presence of multiple identical tRNA gene copies within genomes.

Phenotype switching is commonly observed in nature. This prevalence has allowed the elucidation of a number of underlying molecular mechanisms. However, little is known about how phenotypic switches arise and function in their early evolutionary stages. The first opportunity to provide empirical insight was delivered by an experiment in which populations of the bacterium Pseudomonas fluorescens SBW25 evolved, de novo, the ability to switch between two colony phenotypes. Here we unravel the molecular mechanism behind colony switching, revealing how a single nucleotide change in a gene enmeshed in central metabolism (carB) generates such a striking phenotype. We show that colony switching is underpinned by ON/OFF expression of capsules consisting of a colanic acid-like polymer. We use molecular genetics, biochemical analyses, and experimental evolution to establish that capsule switching results from perturbation of the pyrimidine biosynthetic pathway. Of central importance is a bifurcation point at which uracil triphosphate is partitioned towards either nucleotide metabolism or polymer production. This bifurcation marks a cell-fate decision point whereby cells with relatively high pyrimidine levels favour nucleotide metabolism (capsule OFF), while cells with lower pyrimidine levels divert resources towards polymer biosynthesis (capsule ON). This decision point is present and functional in the wild-type strain. Finally, we present a simple mathematical model demonstrating that the molecular components of the decision point are capable of producing switching. Despite its simple mutational cause, the connection between genotype and phenotype is complex and multidimensional, offering a rare glimpse of how noise in regulatory networks can provide opportunity for evolution., PLoS Biology, 13 (3), ISSN:1544-9173, ISSN:1545-7885

Transfer RNAs play a central role in protein synthesis. Organisms contain sets of genes that code for different tRNA isotypes, and these sets differ among species and strains. An interesting question is how and why these differences come to exist. This study is focused on answering this question by applying the Pseudomonas fluorescens as a model system. First, in the introduction chapter, a review of the functions and life cycle of a tRNA molecule, and the genetic organization of genes encoding tRNAs in all domains of life, is provided. Next, the tRNA gene sets characteristics in several strains of Pseudomonas fluorescens, with particular focus on three strains with differing degrees of evolutionary divergence (SBW25, A506 and Pf0-1), are discussed for further analyses. By comparing the tRNA gene sets of these strains, the tRNAs present in all three genomes (core tRNAs) and those present in only a subset (accessory tRNAs) are identified. In the next section, the fitness effects of removing accessory tRNAs from individual genomes are investigated. Unexpectedly, the results indicate no significant deleterious effects of removing accessory tRNAs from A506 and Pf0-1. Considering the general tolerance of A506 and Pf0-1 to the deletion of one copy of multi-copy tRNA genes, I conclude that there is a degree of redundancy in tRNA gene sets. This redundancy provides a certain level of robustness to tRNA mutations. I next hypothesized that there is a threshold for this redundancy. By deleting more than one copy of the redundant tRNA-GluUUC and tRNA-GlyGCC genes from the SBW25 background, the threshold was passed, and phenotyping showed a significant fitness defect. This strain was then used as a founder for five independent lineages of an evolution experiment along with five independent wild type control lineages. 1% of each population was passaged daily into fresh medium, for 21 days. The fitness defect was recovered after ~210 generations. To unravel the genetic bases of this fast recovery, the genomic DNA of a single isolate from day 20 of each lineage was subjected to whole genome sequencing. Based on the sequencing analyses, genomic rearrangements, which resulted in duplications as large as 1Mb, are the main molecular mechanisms behind the rapid adaptation. Interestingly, this duplicated region includes a copy of the tRNA-GlyGCC gene in every case. Thus, the duplications have increased the tRNA-GlyGCC gene dosage, likely resulting in a compensatory increase in mature tRNA-GlyGCC. This study has provided a real time example of gene duplication as a mechanism in the evolution of tRNA gene sets.

The genetic code is degenerate – 20 proteinogenic amino acids are coded for by 61 functional codons. One might then expect all the codons for an amino acid, called synonymous codons, to be used equally. However, across all domains of life, preferential use of synonymous codons has been observed. This phenomenon is called codon bias. Synonymous substitutions, nucleotide level changes that do not modify the amino acid sequence, were widely viewed to be selectively neutral. Mounting evidence, however, indicates otherwise – synonymous substitutions can induce non-neutral and measurable fitness effects. Experimental studies with randomised codon usage in genes from both prokaryotes and eukaryotes reveal large-scale fitness effects, ranging from altered growth, global changes in both transcription and translation and protein output and function. The effects of synonymous substitutions become particularly prominent in bacteria, where growth rate is often limited by the speed of translation. However, these observations mainly arise from selectively enriching certain codons, a pattern that is rare in naturally evolved genomes. To investigate the functional implications of naturally diverged patterns of codon usage we have identified an essential and highly expressed gene that exclusively exhibits synonymous differences across three genomes of Pseudomonas fluorescens bacteria. These variants were swapped between genomes to test for effects on gene expression and fitness. To swap the synonymous alleles between genomes, I have optimised a scar-free genetic engineering technique that is routinely used for SBW25, to A506 and Pf0-1. The resulting mutants varied from the corresponding wild type strains at only one locus – the locus of the gene that was swapped. I proceeded to test for the effects synonymous substitutions on gene expression and fitness and demonstrated that changing established codon usage patterns of a gene in one strain to that of another strain has considerable effects on gene expression (transcription). Having optimised a technique for manipulating genomes of P. fluorescens strains (besides SBW25), I proceeded to examine the genetic basis of one of the most studied and extensively characterised phenotype of P. fluorescens – the ability to form mats at the air-liquid interface. First, I compared mat formation phenotype in A506, Pf0-1 and PICF7 to the model strain SBW25; while, SBW25, A506 and PICF7 formed mats, Pf0-1 did not. To identify the genetic routes to mat formation in PICF7, I developed tools for transposon suppressor analysis, which revealed that unlike SBW25, which utilises the cellulose biosynthetic machinery encoded by the wss genes to form mats, PICF7 makes use of the pel operon to synthesise the Pel exopolysaccharide. Pel is an exopolymer known to form biofilms (mats) in P. aeruginosa. Using genetic engineering techniques I identified that A506 does not use wss genes either, but the pga operon to produce mats, a pathway seldom utilised in SBW25. The lack of mat forming ability in liquid medium by Pf0-1 is intriguing as it possesses some of the structural genes known from P. aeruginosa and is reported to form mats on solid surfaces. While different structural genes are exploited by the three mat forming strains (SBW25, A506 and PICF7), the central regulatory pathways that fine-tune the expression of the operons have remained the same. This is suggestive of a modular mechanism, wherein multiple structural genes can substitute for one another across genomes. So far, our understanding of P. fluorescens genome evolution has been limited to SBW25. With the development of genetic tools within the scope of this thesis, we now have the opportunity to expand on the comparative study of P. fluorescens and the efficacy of these tools for the same has been amply demonstrated through the examination of the effects of changing naturally evolved synonymous codons as well as the variation in molecular routes exploited to colonise the air-liquid interface. Der genetische Code ist degeneriert – das heisst die 20 proteinogenen Aminosäuren werden von 61 funktionellen Codons codiert. Man könnte daher erwarten, dass alle Codons für eine Aminosäure, auch synonyme Codons genannt, gleichermaßen verwendet werden. Es wird jedoch in allen biologischen Domänen die bevorzugte Verwendung von bestimmten synonymen Codons beobachtet. Dieses Phänomen nennt man verzerrte Codonverwendung oder codon bias. Solche synonymen Substitutionen, also Veränderungen des Nukleotidspiegels, die die Aminosäuresequenz nicht verändern, wurden weithin als selektiv neutral angesehen. Eine wachsende Zahl an Studien deuten jedoch auf etwas anderes hin - auch Substitutionen können nicht-neutrale und messbare Fitnesseffekte hervorrufen. Experimentelle Studien mit randomisierter Codonverwendung in Genen von Prokaryonten und Eukaryonten zeigen große Fitnesseffekte, die von verändertem Wachstum über globale Veränderungen der Transkription und Translation bis hin zu unterschiedlicher Proteinproduktion und -funktion reichen. Die Auswirkungen synonymer Substitutionen treten besonders bei Bakterien auf, bei denen die Wachstumsrate häufig durch die Geschwindigkeit der Translation begrenzt wird. Diese Beobachtungen ergeben sich jedoch hauptsächlich aus der selektiven Anreicherung bestimmter Codons, ein Muster, das in natürlich entwickelten Genomen selten vorkommt. Um die funktionellen Implikationen von natürlich divergierenden Mustern der Codonverwendung zu untersuchen, haben wir ein essentielles und hoch exprimiertes Gen identifiziert, das ausschließlich synonyme Unterschiede zwischen drei Genomen des Bakteriums Pseudomonas fluorescens aufweist. Diese Varianten wurden zwischen diesen Genomen ausgetauscht, um die Auswirkungen auf die Genexpression und Fitness zu testen. Um die synonymen Allele zwischen den Genomen auszutauschen, habe ich eine narbenfreie Gentechnik optimiert, die routinemäßig für Pseudomonas fluorescens SBW25, A506 und Pf0-1 angewendet wird. Die resultierenden Mutanten unterschieden sich von den entsprechenden Wildtyp-Stämmen nur an einem Ort - dem Ort des Gens, das ausgetauscht wurde. Ich habe dann die Auswirkungen von Synonym-Substitutionen auf die Genexpression und Fitness getestet und konnte zeigen, dass der Austausch der etablierten Codon-Verwendungsmuster eines Gens von einem Stamm zu einem anderen erhebliche Auswirkungen auf die Genexpression (Transkription) hat. Nachdem ich eine Technik zur Manipulation des Genoms von P. fluorescens-Stämmen (zusätzlich zu SBW25) optimiert hatte, untersuchte ich die genetische Basis eines der am besten untersuchten und am ausführlichsten charakterisierten Phänotypen von P. fluorescens - die Fähigkeit, Matten an der Luft-Flüssigkeits-Grenzfläche zu bilden . Zunächst verglich ich den Phänotyp der Mattenbildung in A506, Pf0-1 und PICF7 mit dem Modellstamm SBW25; während SBW25, A506 und PICF7 Matten bildeten, tat Pf0-1 dies nicht. Um die genetischen Wege zur Mattenbildung in PICF7 zu identifizieren, entwickelte ich Werkzeuge für die Transposon-Suppressor-Analyse. Diese zeigten, dass PICF7 im Gegensatz zu SBW25, bei dem die von den wss-Genen kodierten Zellulosebiosynthesemaschinen zur Bildung von Matten verwendet werden, das Pel-Operon zur Synthese des Pel-Exopolysaccharids verwendet. Pel ist ein Exopolymer, von dem bekannt ist, dass es Biofilme (Matten) in P. aeruginosa bildet. Unter Verwendung genetischer Techniken stellte ich fest, dass A506 auch keine wss-Gene verwendet um Matten herzustellen, sondern das pga-Operon, welches in SBW25 selten dazu genutzt wird. Das Fehlen der Fähigkeit zur Mattenbildung in flüssigem Medium durch Pf0-1 ist faszinierend, da es einige der von P. aeruginosa bekannten Strukturgene besitzt und Berichten zufolge auf festen Oberflächen Matten bildet. Während die drei mattenbildenden Stämme (SBW25, A506 und PICF7) unterschiedliche Strukturgene ausnutzen, sind die zentralen Regulationswege, welche die Expression der jeweiligen Operons optimieren, gleich geblieben. Dies lässt auf einen modularen Mechanismus schließen, bei dem mehrere Strukturgene sich über die Genome hinweg gegenseitig ersetzen können. Bisher beschränkte sich unser Verständnis der Genomentwicklung von P. fluorescens auf den Stamm SBW25. Mit der Entwicklung genetischer Werkzeuge im Rahmen dieser Arbeit haben wir nun die Möglichkeit, die vergleichenden Studien von P. fluorescens zu erweitern. Die Wirksamkeit dieser Werkzeuge wurde durch die Untersuchung natürlicher Veränderungen synonymer Codons sowie die Variation der molekularen Wege, die zur Besiedlung der Luft-Flüssigkeits-Grenzfläche genutzt werden, ausführlich demonstriert.