The evolution and determination of plasmid transfer rate and subsequent effect on competition

Abstract

Horizontal gene transfer by conjugation is one of the processes that determines the persistence, prevalence and transmission of antibiotic resistance genes that can be found on bacterial plasmids. In order to appropriately tackle the spread of antibiotic resistance we must therefore understand how plasmid dynamics function in complex microbial communities. Various aspects of plasmid dynamics and how they contribute to the spread of antibiotic resistance are unclear and require attention. For example, plasmid transfer rates vary widely, but the ways in which environmental, plasmid and host factors explain this variation and the relative importance of each factor is unclear. In addition, the evolutionary forces that differentially affect plasmids and hosts to determine specific transfer rates have not been fully explored; in particular, the effects of host-plasmid conflicts in non-selective conditions and the impact of the relationship between plasmid cost on host growth and plasmid transfer rate. A theoretical understanding of transfer rates must then be placed within the context of the other parameters that affect plasmid dynamics (e.g. plasmid cost, loss etc.) to make assertions on plasmid persistence and prevalence, and theoretical results must be compared with experimental data in increasing microbial complexity. Experiments are rarely conducted using multiple species and the impacts and interactions of plasmid presence on a community have yet to be explored fully in the lab.

The first data chapter of this thesis (chapter 2) seeks to address the question of how transfer rate variation can be attributed to various environmental variables in addition to the effects of plasmid, donor and recipient identities. A meta-analysis of published transfer rates was therefore conducted and the variation assessed by applying series of multivariate linear models to the data. Over three quarters of the variation from the meta-analysis could be explained, with plasmid repression and media type explaining the most variation. The results also identify the recipient identity as an important variable that explains up to 34\% of the variation.

Given the variation in transfer rates, the next chapter (chapter 3) asks how the various selection pressures on host and plasmid may interact to determine specific rates of transfer. In particular, it asks how the costs of plasmid transfer impact transfer rates, and how host-plasmid conflicts in transfer rate may subsequently affect plasmid prevalence. Adaptive dynamics and invasion analyses were applied to simple conjugation models under selective and non-selective conditions, and using different plasmid transfer-cost relationships. The findings were then combined to model the effects that host-plasmid conflicts in non-selective conditions may have on transfer rates and plasmid prevalence. The results of separate analyses demonstrate the role of the recipient in controlling transfer rates, and show that plasmid-controlled transfer rate can be predicted with only three parameters (host growth rate, plasmid loss rate and the cost of plasmid transfer on growth). Low frequency genetic variation in transfer rate is predicted to accumulate, which can facilitate rapid adaptation to changing conditions. Further modelling showed that in order to substantially affect plasmid prevalence (and corresponding cumulative costs a plasmid has on a population in non-selective conditions) a host may need to decrease the transfer rate by several orders of magnitude, indicating that hosts must have strong control mechanisms to be valuable.

In the final data chapter (chapter 4) I ask if and how plasmid dynamics focusing on the interaction of plasmid presence and inter-species competition in simple microbial communities can be predicted using independently measured parameters. In particular, how does the rate of plasmid transfer impact species competitive advantage and the outcomes of competition? A series of experiments were conducted to estimate parameters for two plasmids and two bacterial species for use in a simplified two-species bacterial conjugation model to make predictions of competitive advantage. These predictions were then compared with a series of corresponding competition assays. The effects of the plasmid distribution on competition and the presence of multiple species on plasmid stability in the community were also noted and described. The model accurately predicted many of the experimental results, but deviated from those results where specific parameters were over or underestimated. The results emphasise the importance of appropriate parameter measurement. Plasmid presence reduced the competitive ability of each host and incurred higher costs from the plasmid with a higher transfer rate. These effects were limited or exacerbated dependent on whether the plasmid was able to successfully invade the other species where it incurred similar costs. These results demonstrate the complex effects of plasmid transfer, cost and host interactions on plasmid dynamics in a microbial community and the competitive dynamics of that community.

These results show that transfer rates are highly variable according to environmental conditions and that, while the majority of the variation can be assigned to some variables, additional work is required to evaluate the effects of particular variables, such as temperature and the effects of plasmid-host coevolution. While this work demonstrates how selective pressures act on transfer rates, more work is also required to link particular observed transfer rates to the conditions in which they evolve. The results highlight the importance of the variable and potentially conflicting selection pressures on host and plasmid that combine to determine the rate of transfer, emphasising the sometimes neglected role of the recipient. The relationship between plasmid cost and plasmid transfer rate is identified as a key part of transfer rate evolution and also requires future attention to describe this relationship in order to fully understand how plasmid transfer rates are constructed. These results increase our understanding of the factors that affect plasmid dynamics, have implications for the way we consider and handle the spread of antibiotic resistance, and provide direction for future research opportunities.