Using both fluctuation and MA assays, it was determined that the genome-wide spontaneous mutation rate is significantly greater in an anaerobic environment than in an aerobic one. The average mutation rate obtained for aerobically grown E. coli cells via fluctuation assays was 2.6 × 10-10 mutations per nucleotide per generation, which is
comparable to the genome-wide spontaneous mutation rate of 2.5 × 10-10 mutations per nucleotide per generation obtained via MA assays. The aerobic mutation rates obtained
in this study are within the range of recent estimates obtained for a strain of E. coli K-
12, though GCRs were not included in the analysis of that study. For anaerobically grown E. coli cells, an average mutation rate of 1.3 × 10-9 mutations per nucleotide was
obtained via fluctuation assays. However, when using MA assays, a three-fold lower
rate of 4.1 × 10-10 mutations per nucleotide per generation was obtained, revealing discrepancies that can be obtained between fluctuation assay based estimates and those based on the more comprehensive, genome-wide mutation detection.
While this study was a comprehensive analysis of the mutation rate and spectra, the mutation rates were estimated from only a small number of genomes (24 in each
172 environment). More robust estimates could be obtained from sequencing larger sample sizes. Additionally, the mutation rates and spectra obtained from this study may not precisely match the mutation rates or spectra that are found in natural microbial populations, where populations may experience other stresses such as extreme temperatures or starvation. However, the reductionist approach undertaken within this thesis provides valuable foundation information regarding differences in mutation rate under aerobic and anaerobic conditions that would contribute to rates displayed by natural populations. Finally, it is important to note that the reported genome-wide mutation rates are still underestimations of the true spontaneous mutation rates under both aerobic and anaerobic environments. While there is negligible selection in this study (section 4.2.1.1.3), such that the substitution rate can be equated with the spontaneous mutation rate, any lethal and/or highly deleterious mutations would not have accumulated in the MA lineages and so, would not have been included in the calculations.
Another notable finding of this study was that, depending on how the mutation rate is expressed, with regard to generation or absolute time, different trends among mutation classes can be obtained. Due to the slower growth rate of anaerobically grown cells, mutation rates expressed per unit time were also calculated. Per day mutation rates for BPSs, indels and GCRs were greater in an aerobic environment, as compared to an anaerobic one, with BPSs and indels exhibiting significantly two-fold greater rates. Mutation rates calculated per generation displayed a different trend, with rates for BPSs, indels and GCRs being 1.3-, 1.5- and 2.6- fold greater, respectively, during growth in an anaerobic environment, as compared to aerobic. These differences in the per generation and the per day mutation rates can be attributed to the aerobically and anaerobically grown cells spending differing proportions of time in the different stages of the cell cycle, and indicate that some mutation types may arise independently of genome replication during cell division.
The types of mutations that prevailed in the two environments were also found to differ. In the aerobic environment, there were biases towards G Æ T transversions, which was expected as a result of ROS-induced DNA damage. In addition, IS186 transposition
occurred at significantly 4.6-fold greater rates in aerobically grown cells. In contrast, in the anaerobic environment, there was an unexpected propensity for C Æ A, T Æ G and
173 A Æ C transversions. These results were not consistent with our understanding of oxidative stress-induced DNA damage and repair, suggesting that other exogenous agents or cellular mechanisms present in anaerobically grown cells may be responsible for these mutations. In addition, while the general GCR mutation rate was significantly higher in anaerobically grown cells, IS element insertions displayed the greatest mutation rates. In particular, IS150 transposition was significantly greater under
anaerobic growth conditions, as compared to aerobic conditions. While IS elements are known to respond to starvation and other cellular stresses (60), the reasons for increased IS element activity under anaerobic conditions are not immediately clear.
Overall, these findings highlight the need for further studies of the mutagenic and physiological pressures associated with aerobic and anaerobic growth. In particular, determining and characterizing the agents of mutation behind the mutation biases observed in the anaerobic environment and determining the extent to which these sources contribute to the spontaneous mutation rate. Determining conditions that promote particular mutation classes (e.g. GCRs) would also greatly aid in understanding the mutational processes under both environments. Furthermore, in their study, Lee et al. (2012) demonstrated that methylated bases serve as mutational hotspots, and contributed to the prevalence of G Æ A and C Æ T transitions in aerobically grown
E. coli (13). By analysing the sequence context of the BPSs, it would be possible to
determine if anaerobically grown E. coli exhibit a similar trend or not. Moreover, even
in the case of aerobically grown E. coli, there is conflicting information regarding the
relationship between gene expression and mutation rates (13, 307, 308). So to determine whether the higher mutation rates can in part be explained by mutations in genes with low expression, it would be worthwhile to investigate these associations in anaerobically grown E. coli. Studies have also indicated that the leading and lagging
strand of the E. coli chromosome are replicated with differential fidelity; such that the
two strands differ in their susceptibility to mutations (258, 309). Further studies could involve elucidating any differences in the mutation rates of the leading and lagging strands, and to determine whether they can help explain the observed mutation spectra. The rates at which mutations occur in populations are the combined result of the mutational pressures experienced by the DNA, accurate DNA replication and the efficiency of the pathways that find and repair DNA damage. Therefore, to determine
174 how genome integrity is maintained during growth under aerobic and anaerobic conditions, the activities of the many DNA repair pathways in E. coli were investigated.
Overall, expression of genes involved in repair and replication was greater under anaerobic growth conditions, than aerobic conditions. The greatest differential activity was observed for genes involved in GCR repair, consistent with findings that GCRs were more prevalent in anaerobic MA lineages. While there were some difficulties in relating gene expression of repair genes under aerobic and anaerobic conditions to the observed mutational spectra in the MA study, this study is one of few to provide insight about how genome fidelity is maintained under anaerobic conditions. As very little is known about how anaerobically grown E. coli maintain genome fidelity, better
understanding of how the DNA repair and replication pathways function under anaerobic growth conditions will provide valuable information on how anaerobes maintain genome integrity and to establish which DNA repair and replication genes are responsible for the accumulation of certain mutations. Thus, analysing gene expression of more biological replications at different stages of the growth cycle would be beneficial. Finally, by conducting MA assays of repair-deficient strains, further insight on the mode of action of anaerobic repair genes could be gained.
Many repair and stress response genes, including those involved in acid resistance, displayed greater expression under anaerobic conditions, than under aerobic conditions. Hence, it is possible that these systems were induced by the intracellular pH generated within cells as they underwent anaerobic fermentation. In future, monitoring intracellular pH or for gas production in anaerobically grown cells will aid in better understanding the physiological conditions experienced by the cells.
We sought to understand how GCRs contribute to and impact adaptive evolution by using sbcC, a gene involved in maintaining genome fidelity, and experimental evolution
techniques. After E. coli strains containing a disrupted sbcC had been evolved for 1,000
generations, sbcC mutants exhibited higher rates of mutations and GCRs, though these
differences were not statistically significant. In addition, the relative population fitness of selected sbcC mutant populations had increased, though the genetic bases behind
these increases were not able to be resolved. Moreover, there were no clear correlations between the number of GCRs detected in sequenced clones and the relative fitness of evolved populations. In fact, these results suggested that GCRs can play a significant
175 role in adaptation under anaerobic environments. In the presence of functional SbcC, GCRs are well regulated and that an increase in GCRs can lead to an increase in fitness. On the other hand, in the absence of functional SbcC, perhaps genome instability has drastically increased, such that population fitness does not increase in large increments due to the occurrence of too many lethal or highly deleterious GCRs. Additionally, after 1,000 generations, some of the populations also displayed mixed colony morphology.