Bistable expression of CsgD in biofilm development of Salmonella enterica serovar typhimurium

Abstract

Bacterial persistence in the environment and in the infected host is often aided by the formation of exopolymer-enclosed communities known as biofilms. Heterogeneous gene expression takes place in microcompartments formed within the complex biofilm structure. This study describes cell differentiation within an isogenic bacterial cell population based on the example of biofilm formation by Salmonella enterica serovar Typhimurium. We analyzed the expression of the major biofilm regulator CsgD at the single-cell level with a chromosomal CsgD-green fluorescent protein (GFP) translational fusion. In individual cells, CsgD-GFP expression is mostly found in the cytoplasm. Quantitative expression analysis and results from three different models of S. Typhimurium biofilms demonstrated that CsgD is expressed in a bistable manner during biofilm development. CsgD expression is, however, monomodal when CsgD is expressed in larger amounts due to a promoter mutation or elevated levels of the secondary signaling molecule c-di-GMP. High levels of CsgD-GFP are associated with cellular aggregation in all three biofilm models. Furthermore, the subpopulation of cells expressing large amounts of CsgD is engaged in cellulose production during red, dry, and rough (rdar) morphotype development and in microcolony formation under conditions of continuous flow. Consequently, bistability at the level of CsgD expression leads to a corresponding pattern of task distribution in S. Typhimurium biofilms.

FIG. 1.

Regulatory network of CsgD-mediated biofilm formation in S. Typhimurium UMR1. CsgD expression requires the stress sigma factor RpoS and the response regulator OmpR (34). CsgD is required for the expression of the csgBA operon, encoding structural subunits of curli fimbriae (19), and adrA, encoding a diguanylate cyclase required for the activation of cellulose biosynthesis, on agar plates (55). CsgD expression can be positively modulated by c-di-GMP, the steady-state level of which is determined by the activity of c-di-GMP-specific phosphodiesterases, such as STM1703 and diguanylate cyclases (23, 43).

FIG. 2.

Functionality and expression of the CsgD-GFP fusion protein. (A) Colony appearance of the strains MAE851 (csgD-gfp), UMR1 (wild-type; csgD), and MAE50 (ΔcsgD) after 48 h of growth on LB agar without salt supplemented with Congo Red. Bar, 5 mm. (B) Expression of csgD, adrA, and csgB in S. Typhimurium UMRI (wild type), MAE851 (csgD-gfp), and MAE50 (ΔcsgD). Quantitative real-time RT-PCR was performed to compare the expression of csgD and the CsgD-regulated genes csgB and adrA. RNA was isolated from cells grown for 24 h at 28°C in liquid LB medium without NaCl. One representative experiment is shown. (C) Detection of CsgD by Western blotting in cells from strains MAE851 (csgD-gfp) and UMR1 (wild type; csgD) grown on LB agar without salt for 24 h. (D) Stability of CsgD and CsgD-GFP. Shown is detection of CsgD by Western blot analysis from cells treated with chloramphenicol to inhibit protein synthesis. Cells were removed from the culture for analysis after the indicated time points. (E) Growth phase-dependent expression of CsgD in strain MAE851 (csgD-gfp) as analyzed by Western blotting in cells grown on LB agar without salt. Cell lysates were made at the indicated time points.

FIG. 3.

Visualization of CsgD-GFP in cells grown on agar plates. (A) Fluorescence micrograph (confocal) of CsgD-GFP-expressing strain MAE851 after 22 h of growth on LB agar without salt. Bar, 2 μm. (B) The CsgD-GFP-expressing strain MAE851 was grown as a giant “spot” colony on LB agar without salt for 72 h. Bar, 3 mm. (C and D) GFP fluorescence micrographs of bacteria that were scraped off the agar surface either from the inner (i.e., aging) parts (C) or from the outermost (i.e., young) parts (D) of the same “spot” colony. (Inset in C) Membrane localization of CsgD. Green, CsgD-GFP; red, membrane stain FM4-64. Bar, 2 μm.

FIG. 4.

Visualization of CsgD-GFP during rdar morphotype development. Strain MAE851 (csgD-gfp) was grown on LB agar without salt at 28°C for 8 h (A), 12 h (B), 24 h (C), or 41 h (D). Bacteria were carefully scraped off the agar surface and mounted on agarose-coated slides in LB without salt. (B and C) Two distinct cell populations characterized by high or low intensity of GFP fluorescence are evident. Red, membrane stain FM4-64. Bar, 2 μm.

FIG. 5.

Cellulose production in the subpopulation of cells with high levels of CsgD. Detection of cellulose production by calcofluor staining was performed with the following strains: MAE851 (csgD-gfp) (A and B; red, GFP fluorescence; green, calcofluor), MAE50 (ΔcsgD) (C; red, membrane staining with FM4-64; blue, calcofluor), and MAE222 (ΔbcsA) (D [calcofluor] and E [corresponding phase-contrast image]). Bar, 2 μm.

FIG. 6.

Role and expression of CsgD in flow cell biofilms. (A and B) Confocal micrographs of 48-h-old biofilms of wild-type UMR1 (A) and isogenic strain MAE50 (ΔcsgD) (B). The large panels represent top-down views of the biofilms, and the side panels show orthogonal projections (XZ and YZ) taken along the thick and the thin lines, respectively. The cells were stained with Live/Dead stain. Bars, 20 μm. (C) The double mutant MAE265 (ΔcsgD ΔbcsA) did not form biofilm. The cells were stained with Live/Dead stain. Bar, 20 μm. (D) Confocal micrograph of 24-h-old biofilm of strain MAE851 (CsgD-GFP). Two subpopulations of cells are evident, one, with high GFP fluorescence, involved in building up the biofilm structure, and one subpopulation of individual cells with low GFP signal. Size bar, 2 μm.

FIG. 7.

(A) CsgD and CsgD-GFP expression of S. Typhimurium UMR1 and derivatives. UMR1 showed detectable CsgD expression at 24 h, but not at 8 h. MAE52 displayed high CsgD expression at 8 h and 24 h. UMR1 ΔbcsA ΔcsgBA and MAE52 ΔbcsA ΔcsgBA showed upregulation of CsgD expression at all time points compared to the respective parental strains, UMR1 and MAE52. Strain UMR1 Δ1703 displayed upregulation of CsgD expression compared to wild-type UMR1, although the upregulation in the MAE52 strain was more pronounced. The relative expression of CsgD-GFP was similar to the relative expression of CsgD for all strains. CsgD and CsgD-GFP expression was not observed in UMR1 ΔompR. UMR1 ΔcsgD served as a negative control. (B) Fluorescence microscopy studies of CsgD-GFP expression in S. Typhimurium UMR1 and mutants with altered CsgD expression. Strain UMR1 ΔbcsA ΔcsgBA csgD-gfp, deficient for the production of the extracellular matrix components cellulose and curli fimbriae, exhibits a time-dependent bistable expression pattern of CsgD-GFP comparable to that of wild-type strain UMR1 csgD-gfp (Fig. 4). In strain MAE52 ΔbcsA ΔcsgBA csgD-gfp, the majority of cells expressed high levels of CsgD-GFP at 8 and 24 h. Strain UMR1 Δ1703 csgD-gfp expressed high levels of CsgD-GFP. Extensive cell aggregation and production of extracellular matrix components can be observed in strain UMR1 Δ1703 csgD-gfp. The microscopic observations are consistent with the Western blot analysis shown in panel A. Green, CsgD-GFP fluorescence; red, fluorescence of membrane stain FM4-64. Bars, 2 μm.

FIG. 8.

FACS analysis of CsgD-GFP fluorescence detected in strains UMR1 ΔbcsA ΔcsgBA csgD-gfp (A) and MAE52 ΔbcsA ΔcsgBA csgD-gfp (B) after 8 h, 12 h, 24 h, and 41 h of growth (black shading). CsgD-GFP was virtually unexpressed in UMR1 ΔbcsA ΔcsgBA csgD-gfp at 8 and 12 h of growth. At 24 and 41 h of growth, a bistable expression pattern was evident in strain UMR1 ΔbcsA ΔcsgBA csgD-gfp. Due to the elevated CsgD-GFP expression of strain MAE52 ΔbcsA ΔcsgBA csgD-gfp, the majority of cells were positive for CsgD-GFP expression at all time points. A bistable expression pattern was not observed. Dark gray, the control strain UMR1 ΔompR csgD-gfp, expressing no CsgD-GFP; light gray, the positive control (strain MAE119). The x axis shows arbitrary units (AU) of fluorescence in a logarithmic scale. The y axis shows the cell count in the cell population-representing gate, where total numbers ranged between a minimum of 58,000 and a maximum of 89,500 events.