Polymorphism and selection of rpoS in pathogenic Escherichia coli

Abstract

Background: Though RpoS is important for survival of pathogenic Escherichia coli in natural environments, polymorphism in the rpoS gene is common. However, the causes of this polymorphism and consequential physiological effects on gene expression in pathogenic strains are not fully understood.

Results: In this study, we found that growth on non-preferred carbon sources can efficiently select for loss of RpoS in seven of ten representative verocytotoxin-producing E. coli (VTEC) strains. Mutants (Suc++) forming large colonies on succinate were isolated at a frequency of 10-8 mutants per cell plated. Strain O157:H7 EDL933 yielded mainly mutants (about 90%) that were impaired in catalase expression, suggesting the loss of RpoS function. As expected, inactivating mutations in rpoS sequence were identified in these mutants. Expression of two pathogenicity-related phenotypes, cell adherence and RDAR (red dry and rough) morphotype, were also attenuated, indicating positive control by RpoS. For the other Suc++ mutants (10%) that were catalase positive, no mutation in rpoS was detected.

Conclusion: The selection for loss of RpoS on poor carbon sources is also operant in most pathogenic strains, and thus is likely responsible for the occurrence of rpoS polymorphisms among E. coli isolates.

Figure 1

Catalase activity and RpoS expression in representative Suc⁺⁺ mutants of VTEC strains EDL933, CL106 and EC3-377. (A) Samples were separated by native PAGE and stained for catalase activity. Catalase HPI (KatG) and HPII (KatE) are indicated. (B) Expression of RpoS and RpoS-regulated AppA by Western analysis. Mutations in rpoS were identified in these tested Suc⁺⁺ mutants by sequencing, and sequences are provided in Supplemental material Figure S1 and Figure S2. To confirm equal protein loading, identical gels were run in parallel and stained by Coomassie Blue R-250 [14,71].

Figure 2

Growth of EDL933 and derivative Suc⁺⁺ mutants on M9 glucose (Glu) and succinate (Suc) media. Colony size (diameter) was determined under a light microscope at 40× magnification.

Figure 3

Virulence-related traits, RDAR and cell adherence. (A) Development of RDAR morphotype is impaired in Suc⁺⁺ mutants. Cells were replica-plated on CR (Congo Red) plates and incubated at 25°C for 48 h. (B) Cell adherence to epithelial cells. The adherence was expressed as the percentage of cells surviving the washing process. rpoS designates the constructed rpoS null-deletion mutant.

Figure 4

Growth of EDL933 and derivative mutants on different carbon sources. "ND": not detected. Cells were grown in LB media to OD₆₀₀ 0.6, washed and inoculated to fresh media to a starting OD₆₀₀of 0.05. Cultures were then grown at 37°C with vigorous shaking (200 rpm) and sampled every hour for 10 hours to monitor growth. D-glucuronic acid, threonine, glutamine or proline were added to M9 minimal media as the sole carbon source to a final concentration of 20 mM.

Figure 5

Dynamic view of RpoS status and metabolic fitness in natural E. coli populations. It is postulated that the ancestral E. coli strain possesses a functional rpoS allele (RpoS⁺) but cannot grow well on poor nutrient sources (Suc^--). Upon exposure to nutrient limitation, mutants (Suc⁺⁺) exhibiting enhanced metabolic activity can be selected and become dominant among the population. These mutants consist of two groups, RpoS⁺ and RpoS^-. Under stress conditions, however, the proportion of RpoS^- mutants decreases because of the loss of protection by RpoS-controlled functions, and the abundance of strains with functional RpoS increases. Because cells likely are constantly facing selection between nutrient limitation and stress in nature, the population of E. coli isolates is in a dynamic status in terms of RpoS function and metabolic fitness.