c-di-GMP-related phenotypes are modulated by the interaction between a diguanylate cyclase and a polar hub protein

Abstract

c-di-GMP is a major player in the switch between biofilm and motile lifestyles. Several bacteria exhibit a large number of c-di-GMP metabolizing proteins, thus a fine-tuning of this nucleotide levels may occur. It is hypothesized that some c-di-GMP metabolizing proteins would provide the global c-di-GMP levels inside the cell whereas others would maintain a localized pool, with the resulting c-di-GMP acting at the vicinity of its production. Although attractive, this hypothesis has yet to be demonstrated in Pseudomonas aeruginosa. We found that the diguanylate cyclase DgcP interacts with the cytosolic region of FimV, a polar peptidoglycan-binding protein involved in type IV pilus assembly. Moreover, DgcP is located at the cell poles in wild type cells but scattered in the cytoplasm of cells lacking FimV. Overexpression of dgcP leads to the classical phenotypes of high c-di-GMP levels (increased biofilm and impaired motilities) in the wild-type strain, but not in a ΔfimV background. Therefore, our findings suggest that DgcP activity is regulated by FimV. The polar localization of DgcP might contribute to a local c-di-GMP pool that can be sensed by other proteins at the cell pole, bringing to light a specialized function for a specific diguanylate cyclase.

Figure 1

DgcP interacts with FimV. Schematic diagrams of DgcP (A), its C-terminus cloned in the T25-DgcP_473–671 construct (B) and FimV (C). The GGDEF domain of DgcP is shown, as well as the transmembrane region (TM), TPR motifs and the LysM domain of FimV. The red box in FimV corresponds to the prey fragment and lines show the regions cloned to confirm the interaction. The E. coli host strain BTH101 was co-transformed with pKT25_DgcP (full length) and pUT18_FimV constructs, with the T18 tag in the C-terminal; the interactions were measured using β-galactosidase activity as a reporter (C). Data are the means ± SD from three replicates. ***p < 0.001.

Figure 2

The GGDEF domain, but not its activity, is needed for DgcP localization at the cell poles. msfGFP was fusioned to wild type DgcP (1–670), to a DgcP where the conserved GGEEF motif was mutated to GGAAF and to a truncated version without the GGDEF domain (1–560) and the fusions were produced from a plasmid in PA14. Only the wild-type and GGAAF fusions localize at the cell poles. White bars represent 4 μm.

Figure 3

DgcP localizes at the cell poles in a FimV-dependent manner. msfGFP was fusioned to wild-type DgcP full-length and expressed in ΔdcgP and ΔfimV. (A) Cells were observed by bright field (left panels) and fluorescence microscopy (middle panels), and a merged image was obtained (right panels). (B) The intensity of the GFP fluorescence was measured in 300 cells of each strain and a heat map of DgcP_msfGFP localization was obtained with MicrobeJ, as described in the Methods section (C).

Figure 4

Mutation in dgcP affects biofilm formation. PA14 and the ∆dgcP strains were inoculated at OD₆₀₀ = 0.05 in 48 well polystyrene plates with the referred media and kept at 30 °C for 16 h without agitation. The medium was discarded, and adherent cells were washed and stained with 1% crystal violet, washed and measured at OD₅₉₅. (A) The same procedure was carried out for the strains overproducing DgcP in LB with 0.2% arabinose. (B) 3D pictures resulting from CLSM after 16 h (C) and 72 h (D) of biofilm formation in LB at 30 °C in an 8-well Lab-Tek chambered coverglass system. Data in (A,B) are the means ± SD from five replicates. ***p < 0.001. Bars in C represent 20 μm; in D, top panels, 40 μm; in lower left panel in D, 12 μm and 3.5 μm in right lower panel in D.

Figure 5

∆dgcP mutant presents defects related to surface behaviors. Cells were stabbed into the bottom of an agar plate by using a toothpick and incubated upright at 37 °C overnight, followed by 48 h of incubation at room temperature. The medium was discarded, and adherent cells stained with crystal violet. (A) Diameter of the twitching colonies was measured in triplicates. Data are the means ± SD. (B) Light microscopy images of PA14 and ΔdgcP twitching colonies. Interstitial biofilms formed at the interface between a microscope slide coated in solidified nutrient media (Gelzan Pad) after four hours of colony expansion. The ΔfimV mutant was used as a negative control of twitching. (C) A silicon slide was placed upright in a culture tube and after the different time points cells at the air-liquid interface were washed, fixed and the spread of cells during the initial stages of biofilm formation was observed by FESEM. (D) *p < 0.05; **p < 0.01; ***p < 0.001.

Figure 6

DgcP activity is FimV dependent. Full-length wild type DgcP (pDgcP) or DgcP mutated in its GGDEF domain (pDgcP-GGAAF), only DgcP C-terminal wild-type region (pDgcP-Cterm) and WspR (pWspR) were overproduced from the pJN105 vector in PA14, ΔdgcP and ΔfimV backgrounds. Biofilm (A), swimming motility (B), c-di-GMP quantitation (C) and cdrA mRNA relative levels (D) were assayed. FimV, DgcP or both were expressed in E. coli and the EPS production was assessed in Congo red plates. (E) Data are the means ± SD from three replicates. *p < 0.05; **p < 0.01; ***p < 0.001.

Figure 7

Model of FimV-dependent localization and activity of DgcP. In the PA14 wild type strain (left side of the cartoons), DgcP in its active form (green spheres) is located at the pole due to interaction with FimV (blue box) and may contribute to a local (upper panel) or global (lower panel) c-di-GMP (red dots) pool. In a ∆fimV background, DgcP is scattered in the cytoplasm in a less active form (gray spheres) and may have a small contribution to the c-di-GMP pool (right). Nucleoid, blue dashed line; T4P, orange wavy lines and flagella, dark green lines.