Regulatory RNAs and the HptB/RetS signalling pathways fine-tune Pseudomonas aeruginosa pathogenesis

Abstract

Bacterial pathogenesis often depends on regulatory networks, two-component systems and small RNAs (sRNAs). In Pseudomonas aeruginosa, the RetS sensor pathway downregulates expression of two sRNAs, rsmY and rsmZ. Consequently, biofilm and the Type Six Secretion System (T6SS) are repressed, whereas the Type III Secretion System (T3SS) is activated. We show that the HptB signalling pathway controls biofilm and T3SS, and fine-tunes P. aeruginosa pathogenesis. We demonstrate that RetS and HptB intersect at the GacA response regulator, which directly controls sRNAs production. Importantly, RetS controls both sRNAs, whereas HptB exclusively regulates rsmY expression. We reveal that HptB signalling is a complex regulatory cascade. This cascade involves a response regulator, with an output domain belonging to the phosphatase 2C family, and likely an anti-anti-sigma factor. This reveals that the initial input in the Gac system comes from several signalling pathways, and the final output is adjusted by a differential control on rsmY and rsmZ. This is exemplified by the RetS-dependent but HptB-independent control on T6SS. We also demonstrate a redundant action of the two sRNAs on T3SS gene expression, while the impact on pel gene expression is additive. These features underpin a novel mechanism in the fine-tuned regulation of gene expression.

Fig. 2

Influence of hptB overexpression in PAK, PAKΔhptB or PAKΔretS strains, on biofilm formation and exopolysaccharide production. A. Bacterial colony staining on Congo red-containing agar plates (upper row) and glass tube assay showing biofilm formation (lower row). The name of the tested strains is indicated above each panel. B. Quantification of the adherence ring formed in the glass tube. Each experiment was repeated three times. The error bars indicate standard deviations. The name of the strains used is indicated under each bar. Filled bars correspond to strains carrying pUCPhptB whereas open bars correspond to strains carrying pUCP18. The pUCPhptB allowed overexpression of the hptB gene cloned into the pUCP18 vector.

Fig. 1

Comparison between PAKΔhptB and PAKΔretS mutants for biofilm formation and exopolysaccharide production. A. Glass tube assay showing biofilm formation (upper part). Quantification of the crystal violet-stained adherence ring formed in the glass tube (lower part). Each experiment was repeated three times. The error bars indicate standard deviations. The name of the tested strain is indicated above each panel. B. Bacterial colony staining on Congo red-containing agar plates. The name of strains used is indicated under each panel.

Fig. 3

Expression of lacZ transcriptional fusion in PAK, PAKΔhptB or PAKΔretS strains. Activity was recorded at different growth stages. A. Activity of the pelA–lacZ transcriptional fusion. B. Activity of the exoS–lacZ transcriptional fusion (carried on pBS307). Open signs correspond to strains carrying pUCPhptB whereas filled signs correspond to strain carrying pUCP18. β-Galactosidase activities are expressed in Miller units. Values are averages of at least three independent experiments.

Fig. 4

Effect of the gacS and gacA mutations on biofilm formation. The gacS or gacA mutation was introduced in PAK, PAKΔhptB and PAKΔretS. A. Glass tube assay showing biofilm formation. The name of the strain is indicated above each panel. B. Quantification of the crystal violet-stained adherence ring formed in the glass tube. Each experiment was repeated three times. The error bars indicate standard deviations. The name of strains used is indicated under each bar.

Fig. 6

Influence of the rsmY and rsmZ mutations in PAK, PAKΔhptB and PAKΔretS strains for biofilm formation and exopolysaccharide production. For each row the name of the corresponding strain is indicated on the left. For each strain the upper row corresponds to the bacterial colony staining on Congo red-containing agar plates and the lower row to the glass tube assay showing biofilm formation. In each strain additional mutations in rsmY, rsmZ or rsmY/rsmZ have been introduced as indicated at the top of each column.

Fig. 5

Expression of the rsmY and rsmZ genes in various P. aeruginosa strains. A. Activity of the rsmY–lacZ (open signs) and rsmZ–lacZ (filled signs) transcriptional fusion in PAK, PAKΔhptB or PAKΔretS strains was recorded at different growth stages. B. Activity of the rsmY–lacZ transcriptional fusions in PAK, PAKΔhptB or PAKΔretS strains, carrying pUCP18 (filled signs) or pUCPhptB (open signs), was recorded at different growth stages. β-Galactosidase activities are expressed in Miller units. Values are averages of at least three independent experiments.

Fig. 7

Expression of lacZ transcriptional fusion in PAK, PAKΔhptB or PAKΔretS strains. Activity was recorded after 4 h growth. A. Activity of the pelA–lacZ transcriptional fusion. B. Activity of the exoS–lacZ transcriptional fusion (carried on pSB307). White bars correspond to PAK, grey bars to PAKΔhptB and black bars to PAKΔretS. Each additional mutation, i.e. gacA, rsmY, rsmZ or rsmYZ, which was introduced in each of these strains, is indicated under the corresponding bar. β-Galactosidase activities are expressed in Miller units. Values are averages of at least three independent experiments.

Fig. 10

Effect of the overexpression of PA3346 (pBBR3346) and PA3347 (pBBR3347) on biofilm formation. The cloning vextor (pBBRMCS4) and the appropriate recombinant plasmids were introduced in PAK. A. Glass tube assay showing biofilm formation. The name of the strain is indicated above each panel. B. Quantification of the crystal violet-stained adherence ring formed in the glass tube. Each experiment was repeated three times. The error bars indicate standard deviations. The name of strains used is indicated under each bar.

Fig. 9

Influence of the PA3346 and PA3347 mutations in PAK, PAKΔhptB and PAKΔretS strains for biofilm formation and exopolysaccharide production (Congo red assay). For each column the name of the corresponding strain is indicated. For each strain the upper row corresponds to the bacterial colony staining on Congo red-containing agar plates, the middle row to the glass tube assay showing biofilm formation and the bottom row to the quantification of crystal violet staining as seen in the middle row.

Fig. 8

Interaction between HptB, PA3346 andPA3347. A. PA3345 encodes Hpt and is clustered with PA3346 and PA3347 genes. B. Two-hybrid experiment showing interaction between HptB and the receiver domain of PA3346 and between the PP2C domain of PA3346 and PA3347 (shown with the insert and the red-stained colonies on Mc Conkey agar plates). Interaction between HptB and the PP2C domain of PA3346 or between the receiver domain of PA3346 and PA3347 are negative (shown with the insert and the white colonies on Mc Conkey agar plates). The N- and C-termini for PA3346 have been indicated.

Fig. 11

Model for the HptB regulatory network. HptB has a negative impact on PA3346 activity. In the absence of HptB, PA3346 dephosphorylates the putative anti-anti-σ factor PA3347 (yellow) through the activity of its PP2C domain (in blue). Dephosphorylated PA3347 could bind a putative anti-σ factor (purple), which allows the release of a yet uncharacterized σ factor (green). This σ factor may have a specific impact on rsmY gene expression, but not on rsmZ gene expression. The controlled expression of rsmY through the HptB/PA3346/PA3347 cascade is still GacA-dependent, suggesting GacA synergistically acts on the rsmY promoter together with the unknown σ factor. Overproduction of RsmY alone (through the PA3346/PA3347 pathway) results in overexpression of pel genes and repression of T3SS genes but not in overexpression of the T6SS genes, which are specifically controlled through the RetS pathway. The rest of the model integrates previous published data. The activity of RsmY and RsmZ is through RsmA titration, which is not represented in the figure. The RetS control through interference with the GacS activity was previously published (Goodman et al., 2009). The potential role of three hybrid sensors, PA2824, PA1611 or PA1976 on HptB activation and the retro-transfer of phosphate from HptB onto RetS comes also from previously published data (Hsu et al., 2008).