The disulfide bonding system suppresses CsgD-independent cellulose production in Escherichia coli

Abstract

The bacterial extracellular matrix encases cells and protects them from host-related and environmental insults. The Escherichia coli master biofilm regulator CsgD is required for the production of the matrix components curli and cellulose. CsgD activates the diguanylate cyclase AdrA, which in turn stimulates cellulose production through cyclic di-GMP (c-di-GMP). Here, we identified and characterized a CsgD- and AdrA-independent cellulose production pathway that was maximally active when cultures were grown under reducing conditions or when the disulfide bonding system (DSB) was compromised. The CsgD-independent cellulose activation pathway was dependent on a second diguanylate cyclase, called YfiN. c-di-GMP production by YfiN was repressed by the periplasmic protein YfiR, and deletion of yfiR promoted CsgD-independent cellulose production. Conversely, when YfiR was overexpressed, cellulose production was decreased. Finally, we found that YfiR was oxidized by DsbA and that intraprotein YfiR disulfide bonds stabilized YfiR in the periplasm. Altogether, we showed that reducing conditions and mutations in the DSB system caused hyperactivation of YfiN and subsequent CsgD-independent cellulose production.

FIG 1

Δdsb colonies had increased spreading but curli levels and adrA transcription similar to those of the WT. (A) UTI89 colonies were grown on YESCA-CR plates for 48 h at 26°C. ΔdsbB and ΔdsbA colonies formed hyperspreading colonies that had a lack of wrinkling in the center of the colony. Bars, 0.25 cm. (B) ΔdsbB and ΔdsbA strains had increased colony diameters compared to that of the WT. Colony diameters were measured in biological triplicates, and error bars represent standard deviations of data from biological triplicates. Significance was determined by Student's two-tailed t test (*, P value of >0.01). (C) CsgA Western blot showing that WT, ΔdsbA, and ΔdsbB strains had similar levels of CsgA when grown at 26°C for 48 h. Hexafluoroisopropanol (HFIP) is a strong denaturant that was added to monomerize the aggregated curli fibers. CsgD Western blot analysis was performed on colonies that were grown at 26°C for 24 h. (D) β-Galactosidase assays were performed on WT and ΔdsbB strains transformed with pRJ800, pRJ800-adrA, or pRJ800-16S (rrsA). The averages of biological triplicate readings from pRJ800 were subtracted from readings from strains containing pRJ800-adrA (adrA promoter upstream of lacZ) and pRJ800-16S (16S promoter upstream of lacZ). adrA and 16S transcription levels were similar between the WT and ΔdsbB strains. Error bars are the standard deviations of Miller units from biological triplicates. Significance was determined by using Student's two-tailed t test. NS, not significant. (E) UTI89 derivatives were grown statically in 24-well plates after 2 μl of a culture grown overnight was inoculated into 2 ml YESCA broth that contained 1.67 μg/ml of CR. Pellicles formed at the air-liquid interface with WT UTI89 and DSB mutants. (F) Glass beads were added to WT and ΔdsbB pellicles every 5 s, until the pellicle could no longer hold the glass beads. The image at the top shows a pellicle holding glass beads, while the image at the bottom shows a collapsed pellicle. (G) WT pellicles held ∼5 glass beads in YESCA-CR medium at 48 h, while ΔdsbB pellicles held ∼13 glass beads. Mutation of the cellulose synthase BcsA in WT and ΔdsbB strains yielded cultures that could not produce pellicles or hold glass beads. Error bars represent standard deviations of data from six biological replicates. P values were calculated by using Student's two-tailed t test (P value of <0.01).

FIG 2

Δdsb colonies wrinkle due to csgD- and adrA-independent cellulose production. (A) Colonies were grown in YESCA medium for 48 h on YESCA-CR plates at 26°C. Bar, 0.25 cm. (B) CsgA protein levels were assessed via Western blot analysis. Samples were treated with HFIP to solubilize CsgA fibers.

FIG 3

UTI89 Δdsb colonies wrinkle at 37°C, on dextrose, and on LB plates. (A) Strains were grown on YESCA medium buffered with 15 mM MES to pH 6.6 at 26°C for 48 h. WT, ΔdsbB, and ΔcsgD ΔdsbB colony morphotypes were not affected by buffering of the plates. In the middle row, colonies were grown on YESCA medium with 0.4% dextrose buffered with 15 mM MES to pH 6.6 at 26°C for 48 h. In the bottom row, strains were grown on LB-CR (1:200) plates at 26°C for 48 h. WT colonies were white and nonspreading or wrinkling on LB, while ΔdsbB and ΔcsgD ΔdsbB colonies both wrinkled on LB. (B) Colonies were grown for 24 h at 37°C on YESCA-CR plates. Bars, 0.25 cm.

FIG 4

Cellulose production in the ΔcsgD ΔdsbB strain is dependent on the diguanylate cyclase YfiN. (A) Strains were grown on YESCA-CR medium. The top row shows strains that were grown at 26°C for 48 h, and the bottom row shows strains that were grown at 37°C for 24 h. (B) Strains were grown on YESCA-CR medium supplemented with 15 μM IPTG for 48 h at 26°C or for 24 h at 37°C. Bars, 0.25 cm.

FIG 5

yfiR overexpression inhibits alternate cellulose expression. (A) WT, ΔdsbB, and ΔcsgD ΔdsbB strains were transformed with pCKR101-EV (pEV) or pCKR101-yfiR (pyfiR). Colonies were grown on YESCA-CR plates with 10 μM IPTG. The colonies on the left were grown at 37°C for 24 h, and the plates on the right were grown at 26°C for 48 h. (B) Colony diameter measurements were taken from biological triplicates of colonies that were grown on YESCA medium at 26°C for 48 h. Error bars show the standard deviations of the measured diameters. P values represent values determined by Student's two-tailed t test (*, P value of <0.04; **, P value of <0.01). (C) UTI89 WT and ΔdsbA strains were transformed with pEV and pyfiR^His (pyfiR^H) and grown for 24 h at 37°C on YESCA-IPTG plates. Periplasms were isolated from the cells and were incubated with nonreducing SDS sample buffer with or without 5 mM DTT for 10 min. Anti-His Western blot analysis revealed that YfiR was oxidized in WT UTI89, as DTT treatment led to reduced YfiR levels with decreased migration through the gel. Without DTT, YfiR ran further on the gel (**); however, there was a spur on the left side of the band (*). Spurs develop from reducing agent diffusion from adjacent lanes causing a reduction of the protein on one side of the protein band and are indicative of proteins containing disulfide bonds (56). MBP levels were assayed to ensure that similar periplasmic protein concentrations were loaded for each sample. Underneath the Western blot are pictures of WT pEV and pCKR101-yfiR^His and ΔdsbA pEV and pCKR101-yfiR^His colonies grown on YESCA-CR plates for 48 h at 26°C. (D) Colonies were grown at 37°C for 24 h 1.5 cm away from a filter paper disk containing 20 μl of 500 mM DTT suspended in 200 mM Tris buffer (pH 8.6) or buffer alone. Bars, 0.25 cm.

FIG 6

Updated model of cellulose production in UTI89. YfiN is inhibited by YfiR, which requires the DSB system to properly fold. YfiN produces small amounts of c-di-GMP (yellow stars) that are required for normal CsgD protein levels. CsgD positively regulates adrA, which encodes the diguanylate cyclase that produces c-di-GMP (pink stars) to bind to PilZ of BcsA and activate the cellulose synthase complex. UDP–d-glucose molecules are the monomers of cellulose production. In the absence of the DSB system or YfiR, YfiR is not present, and YfiN dimerizes and leads to uninhibited production of c-di-GMP. YfiN produces c-di-GMP that leads to BcsA activation, even in the absence of CsgD or AdrA.