Plasmids facilitate the propagation of resistance genes and antibiotic treatment selects for increased prevalence of these mobile genetic elements. However, plasmid prevalence often does not decrease when they are no longer selected for. To address this puzzle, I will briefly discuss the existence conditions of plasmids from the classic Stewart & Levin (1977) paper to more recent work on the so-called plasmid paradox, leading on to our work on the effect of the usually neglected host range of plasmids.

Plasmids with a narrow host range (NHR) and a broad host-range (BHR) co-exist in nature, suggesting both have fitness advantages and disadvantages. Since a broader host-range makes it more likely that the plasmid has not yet adapted to a particular host or vice versa, it is to be expected that broad host-range plasmids impose a greater fitness cost on the host. Given this disadvantage, what are the advantages of having a broad host-range? We developed an individual-based mathematical model to explore the dynamics of BHR and NHR plasmids in well-mixed chemostats and spatially structured biofilms and a mass-action model for chemostats for comparison. The NHR plasmid is assumed to infect only one of two species, while the BHR plasmid can infect both otherwise identical host species. We found that a costly NHR plasmid cannot persist in a chemostat unless it is helped by a costlier BHR plasmid that can infect the competing species. The NHR plasmid had to be incompatible to persist in a chemostat. Higher transfer rates of the BHR generated oscillations in abundance leading to bottlenecks with a chance of extinction of both plasmids in the stochastic individual-based model. In biofilms, having a broad host-range was particularly advantageous. However, incompatibility and lower transfer proficiency reduced the advantage of BHR. Overall, broad host range was the better strategy in both biofilms and chemostats. The results suggest that the coexistence of BHR and NHR plasmids can be explained by a trade-off between fitness cost, favoring vertical transmission of NHR plasmids, and broad host-range, favoring horizontal transmission, especially in spatially structured communities. Further, incompatibility may have evolved as a defense mechanism increasing the fitness of narrow host-range plasmids.

In addition to this 2 plasmids 2 hosts model about host range, I will present some results from a 1 plasmid in a more complex environment model – the activated sludge stage of a wastewater treatment plant where plasmids are continuously entering the system with the sewage, hosted by Enteric bacteria, and the residence time of bacteria exceeds the residence time of the solutes. I will show the effects of non-hosts, selection by an antimicrobial, transfer to the indigenous activated sludge bacteria and boost of transfer rate in transconjugants versus donors.

This is joint work with Sonia Martins, Robert J Clegg, Qian Zhang, Akvile Zemgulyte, Chris M Thomas, Helen Wilson, Roberto de la Cruz, David Graham, Barth F Smets.

We are grateful for support from the Darwin Trust, Edinburgh, UK