The pathogenicity island encoded PvrSR/RcsCB regulatory network controls biofilm formation and dispersal in Pseudomonas aeruginosa PA14

Abstract

Pseudomonas aeruginosa biofilm formation is linked to persistent infections in humans. Biofilm formation is facilitated by extracellular appendages, some of which are assembled by the Chaperone Usher Pathway (Cup). The cupD gene cluster is located on the PAPI-1 pathogenicity island of strain PA14 and has probably been acquired together with four genes encoding two-component signal transduction proteins. We have previously showed that the RcsB response regulator activates expression of the cupD genes, which leads to the production of CupD fimbriae and increased attachment. Here we show that RcsB activity is tightly modulated by two sensors, RcsC and PvrS. While PvrS acts as a kinase that enhances RcsB activity, RcsC has a dual function, first as a phosphorelay, and second as a phosphatase. We found that, under certain growth conditions, overexpression of RcsB readily induces biofilm dispersal. Microarray analysis shows that RcsB positively controls expression of pvrR that encodes the phosphodiesterase required for this dispersal process. Finally, in addition to the PAPI-1 encoded cupD genes, RcsB controls several genes on the core genome, some of which encode orphan response regulators. We thus discovered that RcsB is central to a large regulatory network that fine-tunes the switch between biofilm formation and dispersal.

Figure 1

Protein–protein interaction between selected components of the Rcs/Pvr regulatory system investigated by bacterial two-hybrid analysis. Plasmids (pUT18c or pKT25 respectively) expressing relevant domains of potentially interacting proteins were co-transformed into E. coli DHM1 cells, which were grown on MacConkey agar. Interactions were quantified by β-galactosidase assays in biological triplicates. The TorR/TorS and RocA1/RocS1 two-component systems were included as positive controls. D indicates receiver domain, H indicates Hpt domain and DH indicates both. E indicates empty vector. Significant increases compared with the double empty vector control are indicated with asterisks (Student's t-test, **P ≤ 0.01; ***P ≤ 0.001). Inset: schematic diagram of proteins of interest with sensors inserted into the inner membrane (IM). Black boxes: transmembrane domains; shaded boxes: HisKA-ATPase domains; ovals: receiver domains; light grey boxes: histidine phosphotransfer (Hpt) domains; white hexagons: output domains, either helix–turn–helix (HTH) or phosphodiesterase (EAL).

Figure 2

The effect of rcsC and pvrS overexpression on cupD gene expression. β-Galactosidase assays of P_BAD-rcsB::DZ carrying empty vector or overexpressing either rcsC (A) or pvrS (B) as indicated.

Figure 3

Epistatic analysis of RcsC and PvrS. β-Galactosidase assays of the P_BAD-rcsB::DZ reference strain and isogenic deletion mutants carrying empty vector or overexpressing either rcsC (A) or pvrS (B) as indicated.

Figure 4

Functional analysis of PvrS and RcsC. β-Galactosidase activity of P_BAD-rcsB::DZΔpvrS (A) or P_BAD-rcsB::DZΔrcsC (B) complemented with constructs encoding either wild-type or mutant proteins of PvrS or RcsC respectively. The domain organization of PvrS and RcsC are shown in (A) and (B) respectively. TM is for transmembrane domain, HisKA-ATPase is the sensor transmitter domain carrying the kinase and ATPase activity, Rec is the receiver domain, Hpt is the histidine phosphotransfer domain, finally PAS is a domain found in several signalling proteins and belongs to the Pfam family PF00989.

Figure 5

Identification of putative RcsB binding sites.A. Alignment of target promoter sequences. Numbers indicate distance from +1. Black or grey shading indicates putative conserved or partly conserved nucleotides respectively. Asterisks indicate nucleotides in the cupD1 promoter that are mutated in (B) (GAA→TTC).B. β-Galactosidase assays of the P_BAD-rcsB reference strain carrying either wild-type or mutated cupD–lacZ promoter fusions in the att site.

Figure 6

Biofilms formation in microfermentors. PA14 or PA14ΔcupD carrying empty vector or overexpressing rcsB as indicated. Top: microfermentors after 4 days growth; bottom: biofilms formed on the glass spatulas.

Figure 7

RcsB-dependent dispersal of pre-formed biofilms requires pvrR. Biofilms were grown in 24-well plates in static conditions for 6 h, and arabinose was added to induce rcsB. Biofilm dispersal was then monitored using crystal violet (CV) staining.A. Time-course of dispersal after addition of arabinose at 45 min, 2 h, 3 h, 4 h, 5 h, 7 h and 18 h. Error bars represent standard deviation (n = 2).B. Images of CV stained biofilms in wells after 18 h rcsB induction with arabinose.This figure is available in colour online at wileyonlinelibrary.com.

Figure 8

Regulation model. The cupD gene cluster is inversely regulated by the response regulators RcsB and PvrR. RcsB is activated by the PvrS kinase, while RcsC acts as a phosphorelay and a phosphatase. In addition to its primary target, the cupD promoter, RcsB also activates the expression of other regulatory proteins encoded on the core genome. RcsB also positively influences expression of pvrR, which downregulates production of CupD fimbriae and reduces biofilm formation. Shaded boxes: HisKA-ATPase domains; ovals: receiver domains; light grey boxes: histidine phosphotransfer (Hpt) domains; white hexagons: output domains, either helix–turn–helix (HTH) or phosphodiesterase (EAL).