Evolution of the Quorum Sensing Regulon in Cooperating Populations of Pseudomonas aeruginosa

Abstract

In the opportunistic pathogenic bacterium Pseudomonas aeruginosa acyl-homoserine lactone quorum sensing (QS) can activate expression of dozens to hundreds of genes depending on the strain under investigation. Many QS-activated genes code for extracellular products. P. aeruginosa has become a model for studies of cell-cell communication and coordination of cooperative activities, which result from production of extracellular products. We hypothesized that strain variation in the size of the QS regulon might reflect the environmental history of an isolate. We tested the hypothesis by performing long-term growth experiments with the well-studied strain PAO1, which has a relatively large QS regulon, under conditions where only limited QS-controlled functions are required. We grew P. aeruginosa for about 1000 generations in a condition where expression of QS-activated genes was required, and emergence of QS mutants was constrained and compared the QS regulons of populations after 35 generations to those after about 1000 generations in two independent lineages by using quorum quenching and RNA-seq technology. In one lineage the number of QS-activated genes identified was reduced by over 60% and in the other by about 30% in 1000-generation populations compared to 35-generation populations. Our results provide insight about the variations in the number of QS-activated genes reported for different P. aeruginosa environmental and clinical isolates and, about how environmental conditions might influence social evolution. IMPORTANCE Pseudomonas aeruginosa uses quorum sensing (QS) to activate expression of dozens of genes (the QS regulon). Because there is strain-to-strain variation in the size and content of the QS regulon, we asked how the regulon might evolve during long-term P. aeruginosa growth when cells require some but not all the functions activated by QS. We demonstrate that the P. aeruginosa QS-regulon can undergo a reductive adaptation in response to continuous QS-dependent growth. Our results provide insights into why there is strain-to-strain variability in the size and content of the P. aeruginosa QS regulon.

Keywords: acyl-homoserine lactone; adaptive evolution; metatranscriptomics; quorum quenching.

FIG 1

Long-term evolution of P. aeruginosa PAO1 serially passaged in a medium (CAB) that requires quorum sensing for growth. (A) Experimental design. (B) Abundance of protease-negative (neg) cheaters in each lineage (A through E) at the indicated days. (C) Growth of lineages D (blue) or E (orange) in adenosine-only broth. Inocula were cells from 5-day (circles) or 50-day (triangles) CAB cultures. Growth of the parent strain PAO1 (black diamonds) is included for comparison. Data are the means of two biological replicates, error bars are the range. (D) Growth yields (colony-forming units [CFU] per mL) of indicated populations grown in CAB for 18 h. Black lines are the geometric mean of three or four biological replicates for each population; error bars are the geometric standard deviation.

FIG 2

Many QS-activated genes in day-5 populations do not show QS activation in day-160 populations. (A) Venn diagrams showing the relationship between QS-activated genes at day-5 versus day-160 populations for lineages D (blue) and E (peach). (B) Venn diagram showing the overlap of genes that remain under QS-control (Retained) in both populations after 160 days of serial passage in CAB. (C) Venn diagram showing the overlap of genes that are no longer under QS-control (Lost) in both populations. Numbers in the Venn diagrams were determined using Venny (46) and area calculated using the area-proportional Venn diagram plotter and editor found at http://apps.bioinforx.com/bxaf7c/app/venn/index.php. The numbers of genes in each category are indicated. Lists of genes shared in each category are in Table 2.

FIG 3

Normalized RNA-seq reads of select genes that are lost from QS control during serial passage in CAB for 160 days. (A) The only two lost genes with increased expression, nuh and PA0144, after 160 days of CAB passage. (B) Two lost genes, pqsA and hsiG2, with low expression levels in 160-day populations even in the presence of added AHLs. White bars are AiiA-treated and black bars are plus AHL signals. Data are the mean normalized transcript counts of the two biological replicates for populations (D or E) passaged for 5 or 160 days in CAB; error bars represent the range.

FIG 4

There are diverse routes to eliminate pqsA from the QS regulon. GFP fluorescence in isolates containing a P_pqsA-gfp reporter plasmid, grown in 96-well plates in CAB (top) and LB (bottom). The strains include isolates evolved for 160 days in CAB for population D (D1, D2) or E (E1, E2), the parent strain (PAO1), and a PqsR-variant mutant strain (S36N). Data are the mean relative fluorescence units over time of three technical replicates for two biological replicates; error bars represent standard deviations of means.