Conjugation is a fundamentally important mechanism of horizontal DNA transfer between bacteria, bacteria x archea, and bacteria x eukaryotes. This work has concentrated on conjugation between bacteria x eukaryotes, specifically Escherichia coli x Saccharomyces cerevisiae. Four hypotheses were tested, investigating the barriers to this particular form of DNA transfer. The first investigated if a mutation that altered the cell-surface of the recipient S. cerevisiae could inhibit DNA transfer. The final three utilised a recombination-dependent-conjugation assay to investigate the barrier to DNA transmission through recombination. The hypotheses tested if the frequency of recombination, in this recombination-dependent-conjugation assay, differed when using similar or diverged DNA substrates, if a mismatch repair mutation within the recipient could affect the frequencies of recombination observed, and if the position on the plasmid of the gene of interest affected the frequency of transmission. Transmission of the Ura3 DNA sequence in the recipient S. cerevisiae was used to test all four hypotheses. The cell wall mutants mnn9, knr4, fks1 and kre6 were utilised to investigate if the cell-surface of the recipient could affect the frequency of transmission. The similar and diverged substrates utilised in the investigation of the affect of sequence similarity on recombination were the DNA sequences of ura3 from S. cerevisiae and Saccharomyces carlsbergensis, respectively and the MMR mutants utilised were msh2, pms1 and pol30-52. Cell wall mutants were not found to limit the frequency of transfer once donor-recipient contact was induced through the solid surface mating procedure. Sequence similarity, MMR and the relative position of the ura3 DNA sequence on the conjugative plasmids were shown to have little effect on the frequency of transmission in S. cerevisiae. This suggests that any DNA that enters the nucleus of S. cerevisiae (eukaryotes) can recombine with the chromosome and alter it to the same extent. However, trends within the data also suggest that DNA is transferred into the recipient and then transported to the nucleus to recombine with the chromosome as a single-stranded DNA molecule.

The genome of B. phytofirmans OLGA172 has been sequenced by Next Generation sequencing methods. Over 42 kbp of its genome surrounding its 3CBA degradative genes, tfdCIDIEIFI, was assembled and annotated. The most important method used was the synteny method, which implies homology between the genes, and descent from a common ancestor (Guttman, 2008). The conserved gene order between B. phytofirmans PsJN, B. xenovorans LB400, and OLGA172 was used as a confirmation of annotation through BLASTn, enabled closing of the gaps in the sequencing data, and allowed prediction of genes further downstream. Though the whole genome is not yet assembled, a very significant region carrying a concentrated area of mobile genetic elements (MGE) has been found to surround the degradative genes in OLGA172. This thesis details the sequence evidence that, upon examination of closely related strains, OLGA172 and its related strain from pristine soils may be the ancestral chlorobenzoate degraders.

Vibrio vulnificus is an opportunistic human and animal pathogen with the highest death rate of any foodborne disease agent. The capsular polysaccharide (CPS) is essential for virulence. Over 100 CPS types (carbotypes) have been identified among natural isolates, yet little is known about the genetic mechanisms that drive such diversity. Chitin, the second most abundant polysaccharide in nature, induces competence in Vibrio species. We found that transformation frequency varies by strain and (GlcNAc)2 was the shortest chitin-derived polymer capable of inducing competence. We confirmed that V. vulnificus can undergo chitin-dependent carbotype conversion following the uptake and recombination of complete cps loci from exogenous genomic DNA. The acquisition of a partial locus was also demonstrated when internal regions of homology between the endogenous and exogenous loci existed. Thus, the same mechanism governing the transfer of complete cps loci also contributes to their evolution by generating novel combinations of CPS biosynthesis genes.

The genome of B. phytofirmans OLGA172 has been sequenced by Next Generation sequencing methods. Over 42 kbp of its genome surrounding its 3CBA degradative genes, tfdCIDIEIFI, was assembled and annotated. The most important method used was the synteny method, which implies homology between the genes, and descent from a common ancestor (Guttman, 2008). The conserved gene order between B. phytofirmans PsJN, B. xenovorans LB400, and OLGA172 was used as a confirmation of annotation through BLASTn, enabled closing of the gaps in the sequencing data, and allowed prediction of genes further downstream. Though the whole genome is not yet assembled, a very significant region carrying a concentrated area of mobile genetic elements (MGE) has been found to surround the degradative genes in OLGA172. This thesis details the sequence evidence that, upon examination of closely related strains, OLGA172 and its related strain from pristine soils may be the ancestral chlorobenzoate degraders.

Vibrio vulnificus is an opportunistic human and animal pathogen with the highest death rate of any foodborne disease agent. The capsular polysaccharide (CPS) is essential for virulence. Over 100 CPS types (carbotypes) have been identified among natural isolates, yet little is known about the genetic mechanisms that drive such diversity. Chitin, the second most abundant polysaccharide in nature, induces competence in Vibrio species. We found that transformation frequency varies by strain and (GlcNAc)2 was the shortest chitin-derived polymer capable of inducing competence. We confirmed that V. vulnificus can undergo chitin-dependent carbotype conversion following the uptake and recombination of complete cps loci from exogenous genomic DNA. The acquisition of a partial locus was also demonstrated when internal regions of homology between the endogenous and exogenous loci existed. Thus, the same mechanism governing the transfer of complete cps loci also contributes to their evolution by generating novel combinations of CPS biosynthesis genes.

Conjugation is a fundamentally important mechanism of horizontal DNA transfer between bacteria, bacteria x archea, and bacteria x eukaryotes. This work has concentrated on conjugation between bacteria x eukaryotes, specifically Escherichia coli x Saccharomyces cerevisiae. Four hypotheses were tested, investigating the barriers to this particular form of DNA transfer. The first investigated if a mutation that altered the cell-surface of the recipient S. cerevisiae could inhibit DNA transfer. The final three utilised a recombination-dependent-conjugation assay to investigate the barrier to DNA transmission through recombination. The hypotheses tested if the frequency of recombination, in this recombination-dependent-conjugation assay, differed when using similar or diverged DNA substrates, if a mismatch repair mutation within the recipient could affect the frequencies of recombination observed, and if the position on the plasmid of the gene of interest affected the frequency of transmission. Transmission of the Ura3 DNA sequence in the recipient S. cerevisiae was used to test all four hypotheses. The cell wall mutants mnn9, knr4, fks1 and kre6 were utilised to investigate if the cell-surface of the recipient could affect the frequency of transmission. The similar and diverged substrates utilised in the investigation of the affect of sequence similarity on recombination were the DNA sequences of ura3 from S. cerevisiae and Saccharomyces carlsbergensis, respectively and the MMR mutants utilised were msh2, pms1 and pol30-52. Cell wall mutants were not found to limit the frequency of transfer once donor-recipient contact was induced through the solid surface mating procedure. Sequence similarity, MMR and the relative position of the ura3 DNA sequence on the conjugative plasmids were shown to have little effect on the frequency of transmission in S. cerevisiae. This suggests that any DNA that enters the nucleus of S. cerevisiae (eukaryotes) can recombine with the chromosome and alter it to the same extent. However, trends within the data also suggest that DNA is transferred into the recipient and then transported to the nucleus to recombine with the chromosome as a single-stranded DNA molecule.

The PhD thesis is composed of three publications on genomic, metabolic, and cellular integration between the host and its symbionts in the tripartite nested mealybug system. The articles revealed a path to an intimate endosymbiosis that can be compared to what we think happened before (and to some extent after) bacterial ancestors of key eukaryotic organelles, mitochondria and plastids, became highly integrated into their host cells. I argue that these much younger symbioses may tell us something about how the mitochondria and plastids came to be, at the very least by revealing what types of evolutionary events are possible as stable intracellular relationships proceed along the path of integration.

Horizontal gene transfer (HGT), a mechanism that facilitates exchange of genetic material between organisms from different lineages, has a profound impact on prokaryotic evolution. To infer HGT, we first developed a comparative genomics-based tool, APP, which can perform phyletic pattern analysis using completely sequenced genomes to identify genes are unique to a genome or have sporadic distribution in its close relatives. Performance assessment against currently available tools on a manually created 18-genome dataset and 2 benchmarking datasets revealed the superior accuracy of APP over other methods. We then utilized a parametric method to construct a gene exchange network. The composition-based method, Jenson-Shannon Codon Bias (JS-CB), groups genes into clusters based on similar codon usage bias. These clusters were analyzed using APP and examined for the enrichment HGT associated marker genes, then annotated as of native or alien origin based on these multiple lines of evidence. Intergenome clustering enabled identification of genes mobilized across alien components of the genomes (alien-alien transfer) and from native components of donor genomes to the recipient genomes (native-alien transfer). Functional classification of alien gene clusters revealed that metabolism associated genes are most frequently mobilized, in concurrence with previous reports, and additionally, a large number of genes with yet unknown functions were found to have been horizontally transferred, a important finding that needs to be further investigated.

La découverte de virus géants avec une taille de génome géante et des caractéristiques génomiques surprenantes soulève différentes questions sur leur origine et leur évolution. De nombreuses études phylogénétiques ont souligné le rôle décisif des HGT et des échanges génétiques sur l'évolution des MV, mais la plupart d'entre eux sont basés sur des familles MV étroitement liées. Pour enquêter sur les événements HGT, nous avons déterminé la distribution des gènes et les phylogénies de gènes pour les 86 ORFomes MV complets classés dans 6 familles définies et 4 putatives, dans le cadre de leurs homologues d'autres domaines de la vie. À l'aide d'un flux de travail phylogénétique automatisé MimiLook, 4577 OG ont été détectés, dont 91% des OG ont été jugés spécifiques à la famille, alors que 9% sont représentés par des protéines de 2 familles MV ou plus. 414 OG ont été détectés comme événement HGT. Nous avons appliqué une procédure similaire aux 7 898 protéines non orthologues pour détecter les événements de transfert et identifié 259 HGT à partir de protéines non orthologues. Les cas de HGT révèlent la spécificité des donneurs. En conclusion, une distinction claire peut être observée dans le mosaïque du génome des familles de Megavirale éloignées, où elles ont évolué par spécificité génomique et acquisitions de gènes spécifiques à la famille de leur créneau écologique respectif. Notre recherche systématique d'événements HGT d'origine non-mégavirale fournit la première estimation de la contribution totale de HGT dans le mosaïque du génome spécifique à la famille des Megavirales éloignés. / Discovery of giant viruses with giant genome size and surprising genomic features raises different question about their origin and evolution. Many phylogenetic studies have pointed out decisive role of HGTs and genetic exchanges on evolution of MVs, but, majority of them are based on closely related MV families. To investigate HGT events, we have determined gene distributions and gene phylogenies for the 86 complete MV ORFomes classified in 6 defined and 4 putative families, in context of their homologs from other domains of life. Using an automated phylogenetic workflow MimiLook, 4577 OGs were detected, out of which, 91% of OGs were found to be family specific, whereas, 9% are represented by proteins from 2 or more MV families. 414 OGs were detected as HGT event. We applied a similar procedure to the 7,898 non-orthologous proteins to detect transfer events and identified 259 HGTs from non-orthologous proteins. Instances of HGT were found to be depicting donor specificity, as viruses of vertebrates/invertebrates acquired genes from donors like Euteleostomii, Eutheria, Baculoviridae and proteobacteria; algal viruses and protozoan viruses were found to be acquiring genes from donors like Dictyostellium, Mammeillales, Firmicutes, Clostridiales. In conclusion, clear distinction can be seen in the genome mosaicism of distantly related Megavirale families, where they evolved via genome specificity and family specific gene acquisitions from their respective ecological niche. Our systematic search for HGT events of non-megavirale origin provides the first estimate of the total contribution of HGT in family specific genome mosaicism of distantly related Megavirales.

Genes are able to replicate horizontally and vertically- a given gene may be more successful on horizontally mobile elements than others. This includes genes that exhibit a post-segregational killing (PSK) phenotype. PSK is generated by expression of a toxin and antitoxin from a mobile element, such that if a bacterium loses the element the toxin becomes active in the cell and the cell dies. All PSKs described to date involve a toxin and an antitoxin function, though within a given group of toxin and antitoxin gene pairs only some are likely to exhibit this phenotype. Here, I investigate what differentiates genes that induce PSK from biochemically similar genes that do not. One group of genes of which some are known to induce PSK is toxinantitoxin (TA) systems, composed of a stable toxin and an unstable antitoxin. I analyzed computational data on the distribution of type I TA systems (RNA antitoxin), which appear to be less mobile than type II TA systems (protein toxin). Data on validated TAs suggests a correlation between distribution, mobility and the PSK phenotype. Differences in phylogeny could be due to differences in tendency to exhibit PSK in different environments. This connection between distribution and PSK was explored by experimentally testing a computationally described operon, plasmid_Toxin-ptaRNA1, that exhibited structural and distributional similarities to a mobile type I TA system. Despite this, expression of the predicted toxin ORFs did not reduce growth (as measured by saturation density) in E. coli, and the operon did not induce PSK. The conditions of PSK were further tested with the toxin (barnase) and antitoxin (barstar), which are not known to have the phenotype. A number of heterologous expression systems were developed with these genes in E. coli to test their ability to exhibit PSK in a manner akin to both type II TA systems, with a cytoplasmic toxin, and bacteriocins,which have a secreted toxin. I used equations of logarithmic decay to model the necessary expression of the proteins in the cell and their rate of decay after plasmid loss to enable PSK. My results suggest there is likely to be an evolutionary trend toward TA systems with high expression levels of very unstable antitoxins. Secreted barnase was also tested experimentally for its ability to induce PSK similar to bacteriocins, which exhibit a PSK-like phenotype in monoculture by driving maintenance of the immunity encoding plasmid. Barnase did not induce PSK, possibly due to its inability to cause antibiosis in our test system. Structural similarities and biochemical similarities are not sufficient to determine whether a given system will act as a PSK because numerous contextual factors have an effect on whether the genes are addictive. A given set of genes may have the phenotype in one species but not another, under one set of environmental conditions but not another, or on one replicon but not another. This is consistent with the competition hypothesis, which states that genes will be selected for on mobile elements due to their ability to increase horizontal reproductive success, depending on the environmental conditions.