Biofilm formation by and multicellular behavior of Escherichia coli O55:H7, an atypical enteropathogenic strain

Abstract

Enteropathogenic Escherichia coli (EPEC) is an important causal agent of diarrheal illness throughout the world. Nevertheless, researchers have only recently begun to explore its capacity to form biofilms. Strain O55:H7 (DMS9) is a clinical isolate belonging to the atypical EPEC (aEPEC) group, which displays a high degree of genetic relatedness to enterohemorrhagic E. coli. Strain DMS9 formed a robust biofilm on an abiotic surface at 26 degrees C, but not at 37 degrees C. It also formed a dense pellicle at the air-liquid interface and developed a red, rough, and dry (RDAR) morphotype on Congo red agar. Unlike a previously described E. coli O157:H7 strain, the aEPEC strain seems to express cellulose. Transposon mutagenesis was used to identify biofilm-deficient mutants. One of the mutants was inactivated in the csgFG genes, required for assembly and secretion of curli fimbriae, while a second mutant had a mutation in crl, a thermosensitive global regulator that modulates sigma(S) activity and downstream expression of curli and cellulose. The two mutants were deficient in their biofilm formation capabilities and did not form a pellicle at the air-liquid interface. Unlike in Salmonella, the csgFG mutant in aEPEC completely lost the RDAR phenotype, while the crl mutant displayed a unique RDAR "pizza"-like morphotype. Genetic complementation of the two mutants resulted in restoration of the wild-type phenotype. This report is the first to describe and analyze a multicellular behavior in aEPEC and support a major role for curli and the crl regulator in biofilm development at low temperatures corresponding to the nonmammalian host environment.

FIG. 1.

Quantification of biofilms formed on a 96-well polystyrene plate, using the CV method. Strains were incubated statically in LB broth for 3 days at 26°C. Data are from a representative experiment with three replicates. 1, E. coli TG-1; 2, O55:H10 (EPEC); 3, O55:H7 (EPEC DMS9); 4, O7:H4 (enterotoxigenic E. coli); 5, O153:H− (enterotoxigenic E. coli); 6, O126:H27 (enteroaggregative E. coli); 7, O26:H11 (EHEC); 8, O157:H7 (EHEC) 85/170; 9, O127:H6 e2348/69 (EPEC). The error bars represent the standard errors of the means.

FIG. 2.

EPEC strain DMS9 displays a characteristic RDAR morphotype. Strain DMS9 was inoculated on the center of Congo red plates and incubated for 4 days at 26°C (A and B) or 37°C (C). Panel A presents a section of the peripheral region of a single representative colony, and panel B shows its central region. The bar represents 1 mm.

FIG. 3.

Pellicle formation at the air-liquid interface. Strain DMS9 incubated statically for 3 days in LBNS broth forms a dense pellicle at 26°C (A) but not at 37°C (B). The band at the top seen at 37°C corresponds to light reflection. The robustness and elasticity of the pellicle are illustrated by the inability of a plastic quad-loop inoculation device, pushed gently against the pellicle, to penetrate through (C) and by its capacity to hold a liquid column of 4 ml broth when the tube was held upside down (D).

FIG. 4.

Screening of the transposon library for biofilm-deficient mutants. Strains were incubated statically in LBNS broth for 3 days at 26°C, and the biofilm was quantified by the CV method. Data represent the OD₅₉₅ values of the different strains relative to that of the wt strain, which was set as 100%. Average data of three independent experiments are presented. The biofilm masses (CV values) of all mutants were significantly smaller than that of the wt (P < 0.0001). The error bars represent the standard errors of the means.

FIG. 5.

Schematic representation of the ORFs located near the mini-Tn10 insertion in two selected mutants: A59 (A) and C24 (B). Chromosomal localization of the insertion sites and coordinates are based on the genome map for E. coli O157:H7 Sakai (http://genome.gen-info.osaka-u.ac.jp/bacteria/o157/index.html).

FIG. 6.

Biofilm quantification on 96-well polystyrene plates. Bacteria were grown in static LBNS cultures for 3 days at 26°C. The biofilm mass was determined using the CV method. The average OD₅₉₅ values from three separate experiments are presented. The wt strain tested was DMS9S. The biofilm masses of the A59 and C24 mutants were significantly smaller than that of the wt (P < 0.0001). The error bars represent the standard errors of the means.

FIG. 7.

Biofilm formation on a glass surface. Cells were incubated for 3 days at 26°C in chamber slides, washed, and then stained with AO. Confocal microscopy image stacks were collected using FluoView 500 software, and three-dimensional reconstitution was performed using Imaris software. The pictures were transformed into black-and-white images and processed using Microsoft Office Picture Manager software to enhance their clarity.

FIG. 8.

Pellicle formation at the air-liquid interface. Bacteria were incubated in LBNS for 3 days at 26°C in glass tubes, and their pellicles were visualized by staining with CV (A). The wt pellicle was removed from the tube and stained with AO (B). AO-stained pellicles of the wt and complemented mutants were examined by confocal microscopy (C). Stacks were collected and processed using FluoView 500 software. The z-section values of representative areas are denoted. Pictures were processed using Microsoft Office Picture Manager software to enhance their clarity.

FIG. 9.

Curli expression during pellicle formation. Bacteria were grown in LBNS broth for 3 days at 26°C. (A) Western blots of the wt (lanes 1 and 2), the A59 crl⁺ strain (lanes 3 and 4), the C24 csgFG⁺ strain (lanes 5 and 6), and the A59 crl mutant (lanes 7 and 8). Since the C24 csgFG mutant did not form a pellicle, planktonic bacteria from the liquid culture and aggregates at the bottom were collected for Western analysis (lanes 11 and 12). A mixture of pellicle, planktonic cells, and aggregates of the wt strain was used as the control (lanes 9 and 10). Samples were pretreated with (+) or without (−) HFIP before being loaded onto an SDS-polyacrylamide gel and probed with CsgA antibody. (B) Transmission electron microscopy of the wt, A59 crl⁺, C24 csgFG⁺, and A59 crl strains collected from pellicles. Scale bars are equal to 500 nm.

FIG. 10.

Comparison of the RDAR morphotypes of different strains. Strains were seeded in the center of LBNS agar plates supplemented with Congo red, and the plates were incubated for 5 days at 26°C. Photographs were taken through a binocular microscope with a digital camera. Each bar represents 1 mm.