Phenotypic and integrated analysis of a comprehensive Pseudomonas aeruginosa PAO1 library of mutants lacking cyclic-di-GMP-related genes

Abstract

Pseudomonas aeruginosa is a Gram-negative bacterium that is able to survive and adapt in a multitude of niches as well as thrive within many different hosts. This versatility lies within its large genome of ca. 6 Mbp and a tight control in the expression of thousands of genes. Among the regulatory mechanisms widespread in bacteria, cyclic-di-GMP signaling is one which influences all levels of control. c-di-GMP is made by diguanylate cyclases and degraded by phosphodiesterases, while the intracellular level of this molecule drives phenotypic responses. Signaling involves the modification of enzymes' or proteins' function upon c-di-GMP binding, including modifying the activity of regulators which in turn will impact the transcriptome. In P. aeruginosa, there are ca. 40 genes encoding putative DGCs or PDEs. The combined activity of those enzymes should reflect the overall c-di-GMP concentration, while specific phenotypic outputs could be correlated to a given set of dgc/pde. This notion of specificity has been addressed in several studies and different strains of P. aeruginosa. Here, we engineered a mutant library for the 41 individual dgc/pde genes in P. aeruginosa PAO1. In most cases, we observed a significant to slight variation in the global c-di-GMP pool of cells grown planktonically, while several mutants display a phenotypic impact on biofilm including initial attachment and maturation. If this observation of minor changes in c-di-GMP level correlating with significant phenotypic impact appears to be true, it further supports the idea of a local vs global c-di-GMP pool. In contrast, there was little to no effect on motility, which differs from previous studies. Our RNA-seq analysis indicated that all PAO1 dgc/pde genes were expressed in both planktonic and biofilm growth conditions and our work suggests that c-di-GMP networks need to be reconstructed for each strain separately and cannot be extrapolated from one to another.

Keywords: Pseudomonas; biofilm; c-di-GMP; diguanylate cyclase; phosphodiesterase.

Figure 1

Pseudomonas aeruginosa genes encoding putative diguanylate cyclases (DGCs) or phosphodiesterases (PDEs). Genes encoding for GGDEF (A), GGDEF/EAL dual domain (B) or EAL/HD-GYP (C) motifs in P. aeruginosa PAO1. GGDEF domains are shown in light green, EAL domain in dark blue, and HD-GYP in orange. Accessory domains are indicated in pink (REC domain), yellow (MYHT domain), silver (HAMP domain), gray (MASE1), purple (GAF), brown (CHASE4), dark green boxes (PAS domain), light blue (PAC domain), magenta (PBPb domain), and black segments (transmembrane domain). PDE (phosphodiesterase), DGC (diguanylate cyclase) or not yet determined (ND) activity as well as experimentally assessed function (bold and colored) are described and corresponding studies listed.

Figure 2

Deletion of genes encoding c-di-GMP metabolizing proteins results in altered attachment profiles. Attachment properties of Pseudomonas aeruginosa PAO1 WT and deletion mutants of the 41 genes encoding for GGDEF domains, GGDEF/EAL dual domains as well as EAL or HD-GYP domains. Cells were grown statically at 37°C in LB in 96 wells plates for 6 h, liquid culture was removed and attached cells were stained with crystal violet. Absorbance of crystal violet was measured and is indicative of attached cells. Attachment was assessed in biological triplicates and normalized to PAO1 WT attachment. (Student’s t-test, ^*p < 0.05; and ^****p < 0.0001).

Figure 3

Impact of the deletion of genes encoding GGDEF proteins in Pseudomonas aeruginosa on biofilm structure characteristics. P. aeruginosa PAO1 WT and strains carrying a deletion of indicated c-di-GMP-metabolizing genes encoding GGDEF domain-containing enzymes were assessed for formation of biofilms by confocal scanning laser microscopy. GFP-tagged PAO1 WT and mutant strains were grown for 72 h in biofilm flow cell chambers under continuous feeding of ABTG medium. Confocal images were taken (three channels; seven images per channel). A representative image for each strain is shown.

Figure 4

Impact of the deletion of genes encoding GGDEF/EAL dual proteins in Pseudomonas aeruginosa on biofilm structure characteristics. P. aeruginosa PAO1 WT and strains carrying a deletion of indicated c-di-GMP-metabolizing genes encoding GGDEF/EAL dual domain-containing enzymes were assessed for formation of biofilms by confocal scanning laser microscopy. GFP-tagged PAO1 WT and mutant strains were grown for 72 h in biofilm flow cell chambers under continuous feeding of ABTG medium. Confocal images were taken (three channels; seven images per channel). A representative image for each strain is shown.

Figure 5

Impact of deletion of genes encoding EAL or HD-GYP proteins in Pseudomonas aeruginosa on biofilm structure characteristics. P. aeruginosa PAO1 WT and strains carrying a deletion of indicated c-di-GMP-metabolizing genes encoding EAL or HD-GYP domain-containing enzymes were assessed for formation of biofilms by confocal scanning laser microscopy. GFP-tagged PAO1 WT and mutant strains were grown for 72 h in biofilm flow cell chambers under continuous feeding of ABTG medium. Confocal images were taken (three channels; seven images per channel). A representative image for each strain is shown.

Figure 6

COMSTAT2 analysis of biofilm parameters of the c-di-GMP mutant library. Biofilm structure parameters, biomass (A), biofilm thickness (B) and roughness coefficient - a measure of how much biofilm thickness varies (C) were analysed using the biofilm quantification algorithm COMSTAT2. Analysis was performed in biological triplicates. Statistical analysis was performed using One-way ANOVA, Tukey’s multiple comparison test (Student’s t-test, ^*p < 0.05; ^**p < 0.01; ^***p < 0.001; ^****p < 0.0001).

Figure 7

Impairment of c-di-GMP metabolizing genes produced almost no differences in bacterial motility. (A) Swimming motility of PAO1 WT and deletion mutant strains with indicated genes knocked out was assessed. Strains were grown overnight in LB broth at 37°C and injected straight from overnight cultures (0.5 μl) into the middle of swimming agar plates (0.2% agar) and incubated at 37°C for 16 h. Swimming diameter was measured. Assay was performed in biological triplicates and in 3 technical replicates and data normalized to PAO1 WT swimming diameter. (Student’s t-test, ^****p < 0.0001). (B) Twitching motility was assessed for PAO1 WT and the indicated deletion mutants. Strains were grown on plate and picked up with the tip of a 200 μl pipette and inoculated on twitching plates (1% agar) for 48 h at 37°. Twitching diameter was measured and plotted. Assay was performed in three biological and technical replicates and data normalized to PAO1 WT twitching diameter. (Student’s t-test, ^****p < 0.0001).

Figure 8

Assessment of intracellular c-di-GMP levels upon deletion of individual genes encoding DGCs and PDEs. LC–MS/MS quantification of c-di-GMP levels in Pseudomonas aeruginosa PAO1 WT and indicated mutant strains containing deletions of GGDEF (A) or dual GGDEF/EAL, or individual EAL or HD-GYP encoding genes (B). Data are depicted as pmol c-di-GMP/mg of total protein and were normalized to PAO1 WT c-di-GMP levels. Each value is the average of two different biological replicates. (Student’s t-test, ^*p < 0.05; ^**p < 0.01; ^****p < 0.0001).

Figure 9

Expression level of dgc/pde genes in planktonic and biofilm growth. RNA-seq analysis was performed on total RNA extracted from Pseudomonas aeruginosa PAO1 WT grown either in planktonic conditions for 5 h at 37°C, under shaking (200 rpm) in ABTG medium or grown for 72 h in biofilm flow cell chambers under continuous feeding of ABTG medium. (A) Principal component analysis plot depicted the transcriptomes of P. aeruginosa PAO1 WT grown in either planktonic (triangle) or biofilm (circle) conditions. (B) RNA samples from planktonic (triangle) and biofilm (circle) conditions were prepared in triplicate from three independent biological samples and transcript abundance was analyzed and normalized using DESeq2 (Love et al., 2014) rlog transformation. (C) Log₂Fold change of transcript abundance of genes encoding c-di-GMP modulating enzymes in biofilm relative to planktonic cells.

Figure 10

Heatmap overview comparing phenotypic screens using different Pseudomonas aeruginosa mutant libraries in c-di-GMP networks. The phenotypes assessed and corresponding study are indicated above and below, while the deleted genes are listed on the left. For each phenotype (biofilm, swimming, twitching, swarming, c-di-GMP level or EPS) an increase in activity/level is indicated in blue, a decrease in red, no change in phenotypic output in white and not tested or the absence of a mutant in gray. The study published by Kulasekara et al. screened a set of mutants of P. aeruginosa PA14 with transposon inserted in c-di-GMP-related genes (except from HD-GYP domain encoding genes) for their ability to form biofilm attachment and the enzymatic activity of encoded proteins (characterized by overexpression, *). The screen conducted by Ha et al. assessed all c-di-GMP genes in PA14 for swimming, twitching, swarming and EPS production. The study published by Bhasme et al. screens only genes encoding for a GGDEF domain and assessed phenotypes such as biofilm, swimming, swarming, EPS production and cellular c-di-GMP levels.