Plasmid Viability Depends on the Ecological Setting of Hosts within a Multiplasmid Community

Abstract

Plasmids are extrachromosomal genetic elements, some of which disperse horizontally between different strains and species of bacteria. They are a major factor in the dissemination of virulence factors and antibiotic resistance. Understanding the ecology of plasmids has a notable anthropocentric value, and therefore, the interactions between bacterial hosts and individual plasmids have been studied in detail. However, bacterial systems often carry multiple genetically distinct plasmids, but dynamics within these multiplasmid communities have remained unstudied. Here, we set to investigate the survival of 11 mobilizable or conjugative plasmids under five different conditions where the hosts had a differing ecological status in comparison to other bacteria in the system. The key incentive was to determine whether plasmid dynamics are reproducible and whether there are tradeoffs in plasmid fitness that stem from the ecological situation of their initial hosts. Growth rates and maximum population densities increased in all communities and treatments over the 42-day evolution experiment, although plasmid contents at the end varied notably. Large multiresistance-conferring plasmids were unfit when the community also contained smaller plasmids with fewer resistance genes. This suggests that restraining the use of a few antibiotics can make bacterial communities sensitive to others. In general, the presence or absence of antibiotic selection and plasmid-free hosts (of various fitnesses) has a notable influence on which plasmids survive. These tradeoffs in different settings can help explain, for example, why some resistance plasmids have an advantage during a rapid proliferation of antibiotic-sensitive pathogens whereas others dominate in alternative situations. IMPORTANCE Conjugative and mobilizable plasmids are ubiquitous in bacterial systems. Several different plasmids can compete within a single bacterial community. We here show that the ecological setting of the host bacteria has a notable effect on the survival of individual plasmids. Selection for opportunistic genes such as antibiotic resistance genes and the presence of plasmid-free hosts can determine which plasmids survive in the system. Host bacteria appear to adapt specifically to a situation where there are multiple plasmids present instead of alleviating the plasmid-associated fitness costs of individual plasmids. Plasmids providing antibiotic resistance survived under all conditions even if there was a constant migration of higher-fitness plasmid-free hosts and no selection via antibiotics. This study is one of the first to observe the behavior of multiple genetically different plasmids as a part of a single system.

Keywords: antibiotic resistance; multiresistance; plasmid ecology; plasmid evolution; plasmid stability; plasmid-mediated resistance.

FIG 1

The experimental design of the serial culture experiment. Ampicillin was used for beta-lactam and kanamycin for aminoglycoside. The cultures were refreshed 41 times (n = 4/treatment).

FIG 2

Maximum growth rates and maximal optical density (600 nm) of cultures of each individual strain (A and C) and communities under different treatments at the start and end of the experiment (B and D). Maximum growth rate indicates a maximum change in optical density in 1 h during the 24-h measurement. C1 and C42 refer to the serial culture experiment after day 1 and day 42, respectively. Statistical difference of growth rates and optical density was determined using two-way ANOVA (P < 0.05). The growth rate for the CMK communities at the starting point was the highest (P < 0.01, 2-way ANOVA; post hoc test, Tukey HSD), whereas no difference in the growth rate was shown between C, CA, and CMKA (P = 0.067, 2-way ANOVA; post hoc test, Tukey HSD). At the endpoint, significant difference was observed among all treatments (P < 0.01, 2-way ANOVA; post hoc test, Tukey HSD) (B). Optical densities of the communities at the endpoint were higher than at the starting point between all treatments (P < 0.01, 2-way ANOVA; post hoc test, Tukey HSD) (D). Different letters indicate statistically significant differences in results.

FIG 3

Prevalence of plasmids pEC3pl1, pEC3pl2, pEC13, pEC14pl1, pEC14pl3, and pEC14pl2+pEC15pl1 under different conditions as normalized with the gene for 16S rRNA. Plasmid prevalence was followed over 42 cycles of the serial culture experiment. Culture conditions are depicted in Fig. 1. Alternative visualizations of the results are available in Data Set S1, sheet D.

FIG 4

Prevalence of plasmids pEC14pl2+pEC15pl1, pEC15pl2, pEC16pl1, pEC16pl2, and RP4 under different conditions as normalized with the gene for 16S rRNA. Plasmid prevalence was followed over 42 cycles of the serial culture experiment. Culture conditions are depicted in Fig. 1. Alternative visualizations of the results are available in Data Set S1, sheet D.

FIG 5

Observed mutations in plasmids after 42 cycles in the serial culture experiment (∼320 generations). Only plasmids with mutations are shown. The black dots mark the mutation region and its frequency; the bigger the black dot, the higher the mutation frequency (for exact values and mutation types, see Data Set S1, sheet G).

FIG 6

Observed mutations in the host chromosomes obtained from the community after 42 cycles in the serial culture experiment (∼320 generations). Mutations in noncoding regions are not labeled. The black dots mark the mutation region and its frequency; the bigger the black dot, the higher the mutation frequency (for exact values and mutation types, see Data Set S1, sheet G).