Evolutionary rescue in populations of Pseudomonas fluorescens across an antibiotic gradient

Abstract

Environmental change represents a major threat to species persistence. When change is rapid, a population's only means of persisting may be to evolve resistance. Understanding such 'evolutionary rescues' is important for conservation in the face of global change, but also in the agricultural and medical sciences, where the objective is rather population control or eradication. Theory predicts that evolutionary rescue is fostered by large populations and genetic variation, but this has yet to be tested. We replicated hundreds of populations of the bacterium Pseudomonas fluorescens SBW25 submitted to a range of doses of the antibiotic streptomycin. Consistent with theory, population size, and initial genetic diversity influenced population persistence and the evolution of antibiotic resistance. Although all treated populations suffered initial declines, those experiencing the smallest decreases were most likely to be evolutionarily rescued. Our results contribute to our understanding of how evolution may or may not save populations and species from extinction.

Keywords: Pseudomonas fluorescens; Resistance; antibiotic; evolutionary rescue; pharmacology.

Figure 1

Population dynamics of rescued (solid lines and circles) and extinct (dashed lines and squares) bacterial populations by level of streptomycin (μg/mL) and by relative genetic diversity (clonal versus diversified) in experiment 1. Each line represents a single microcosm. Lines are nonlinear interpolations intended for illustration.

Figure 2

(A) Survival of clonal, low genetic-diversity populations and diversified, high genetic-diversity populations after 53 h in experiment 1, as a function of streptomycin concentration. (B) Mean population decline after 22 h (difference in log population size between 0 h and 22 h) of future-extinct and future-rescued populations as a function of population type and streptomycin concentration. (C) Mean population size (log-transformed) of rescued clonal and diversified populations after 53 h, as a function of streptomycin concentration. Error bars represent standard error (in (A) calculated from the binomial distribution; n = 10 populations).

Figure 3

(A) Survival of small populations (0.2-mL microcosms) and large populations (1.5-mL microcosms) after 86 h in experiment 2 as a function of streptomycin concentration. (B) Mean population decline during day 2 (difference in log population size between 23 h and 48 h) of future-extinct and future-rescued populations in 1.5-mL microcosms, as a function of streptomycin concentration. (C) Mean population size (log-transformed) of rescued populations after 86 h in 1.5-mL microcosms, as a function of streptomycin concentration. Error bars represent standard error (in (A) calculated from the binomial distribution; n = 12 populations).

Figure 4

Rescue dynamics in large (1.5-mL) bacterial populations in experiment 2 as a function of streptomycin concentration. Changes in population size (left axis) are shown by solid circles. Changes in the frequency of antibiotic-resistant cells (right axis) are shown in open circles. Dynamics in nonrescued (extinct) populations are shown in Fig. S2; antibiotic-resistant cells were not observed in these populations. Error bars represent standard errors.