Genetic Alternatives for Experimental Adaptation to Colistin in Three Pseudomonas aeruginosa Lineages

Abstract

Pseudomonas aeruginosa is characterized by a high adaptive potential, developing resistance in response to antimicrobial pressure. We employed a spatiotemporal evolution model to disclose the pathways of adaptation to colistin, a last-resort polymyxin antimicrobial, among three unrelated P. aeruginosa lineages. The P. aeruginosa ATCC-27833 reference strain (Pa_ATCC), an environmental P. aeruginosa isolate (Pa_Environment), and a clinical isolate with multiple drug resistance (Pa_MDR) were grown over an increasing 5-step colistin concentration gradient from 0 to 400 mg/L. Pa_Environment demonstrated the highest growth pace, achieving the 400 mg/L band in 15 days, whereas it took 37 and 60 days for Pa_MDR and Pa_ATCC, respectively. To identify the genome changes that occurred during adaptation to colistin, the isolates selected during the growth of the bacteria (n = 185) were subjected to whole genome sequencing. In total, 17 mutation variants in eight lipopolysaccharide-synthesis-associated genes were detected. phoQ and lpxL/PA0011 were affected in all three lineages, whereas changes in pmrB were found in Pa_Environment and Pa_MDR but not in Pa_ATCC. In addition, mutations were detected in 34 general metabolism genes, and each lineage developed mutations in a unique set of such genes. Thus, the three examined distinct P. aeruginosa strains demonstrated different capabilities and genetic pathways of colistin adaptation.

Keywords: Pseudomonas aeruginosa; antibiotic resistance; colistin; experimental evolution; mutation.

Figure 1

Time-lapse images of fifty-day growth patterns and pace over the colistin concentration gradient for three experimental P. aeruginosa lineages. Three P. aeruginosa strains, including Pa_ATCC (section 1), Pa_Environment (section 2), and Pa_MDR (section 3), were inoculated in the outermost left bands of each compartment of the experimental plate containing no colistin (red dots) and allowed to grow for indicated time. Each compartment consisted of 5 bands containing an exponential gradient of colistin from 0 to 400 mg/L. The growing bacteria appeared as a white mass against the black ink background. An image was taken from above the plate every 24 h (a movie assembled over the entire experiment is provided in Supplement). Each panel represents an unedited image obtained on the indicated day of the experiment. Bar, 40 mm.

Figure 2

Graphical representation of the spreading dynamics over the colistin concentration gradient for three experimental P. aeruginosa lineages. Three experimental P. aeruginosa strains were grown, as described in Figure 1’s legend. The Y axis indicates the day of the experiment on which the corresponding colistin concentration band was reached. Note that smaller bars indicate faster growth, as the lineage takes less time to reach a certain colistin gradient band.

Figure 3

Genome mutation rate and individual gene alterations during P. aeruginosa’s adaptation to colistin. (A) Pa_ATCC, (B) Pa_Environment, (C) Pa_MDR. Bars show the median number with the Q3 boundary of the core genome mutations per isolate collected from the corresponding colistin concentration band, as indicated above the bars. The signs above the bars in (B,C) indicate statistically significant differences (p value of the Mann–Whitney U test < 0.05) in the mutation rate as follows: *, Pa_Environment vs. Pa_ATCC; #, Pa_Environment vs. Pa_MDR; +, Pa_MDR vs. Pa_ATCC. Individual affected genes are listed below the bars; red font indicates the LPS-synthesis-associated genes related to colistin resistance, blue font indicates general metabolism genes not directly related to colistin resistance. Alteration type is indicated by the rectangle fill color as follows: red, nonsense mutation/out-of-frame or large indel/gene disruption; yellow, in-frame indel; green, missense mutation. Asterisk indicates a large 262,402 bp deletion (239 ORF), which included the galU gene, involved in LPS synthesis.