FlhA influences Bacillus thuringiensis PlcR-regulated gene transcription, protein production, and virulence

Abstract

Bacillus thuringiensis and Bacillus cereus are closely related. B. thuringiensis is well known for its entomopathogenic properties, principally due to the synthesis of plasmid-encoded crystal toxins. B. cereus appears to be an emerging opportunistic human pathogen. B. thuringiensis and B. cereus produce many putative virulence factors which are positively controlled by the pleiotropic transcriptional regulator PlcR. The inactivation of plcR decreases but does not abolish virulence, indicating that additional factors like flagella may contribute to pathogenicity. Therefore, we further analyzed a mutant (B. thuringiensis 407 Cry(-) DeltaflhA) previously described as being defective in flagellar apparatus assembly and in motility as well as in the production of hemolysin BL and phospholipases. A large picture of secreted proteins was obtained by two-dimensional electrophoresis analysis, which revealed that flagellar proteins are not secreted and that production of several virulence-associated factors is reduced in the flhA mutant. Moreover, we quantified the effect of FlhA on plcA and hblC gene transcription. The results show that the flhA mutation results in a significant reduction of plcA and hblC transcription. These results indicate that the transcription of several PlcR-regulated virulence factors is coordinated with the flagellar apparatus. Consistently, the flhA mutant also shows a strong decrease in cytotoxicity towards HeLa cells and in virulence against Galleria mellonella larvae following oral and intrahemocoelic inoculation. The decrease in virulence may be due to both a lack of flagella and a lower production of secreted factors. Hence, FlhA appears to be an essential virulence factor with a pleiotropic role.

FIG. 1.

Growth curves of B. thuringiensis 407 Cry⁻ [plcA′Z] (▪) and B. thuringiensis 407 Cry⁻[plcA′Z] ΔflhA (○) in LB medium at 37°C. Each point is the mean of five independent experiments. Vertical bars indicate standard errors. OD 600, optical density at 600 nm.

FIG. 2.

Comparison of two-dimensional gels from B. thuringiensis 407 [plcA′Z] (A) and B. thuringiensis 407 [plcA′Z] ΔflhA (B) strains. For each strain, proteins were extracted from the culture supernatant harvested at T₂ in stationary phase. Twenty micrograms of these proteins was loaded onto immobilin polyacrylamide gel strips in the linear pH range of 4 to 7 and were separated by two-dimensional electrophoresis and silver stained. Protein identification was determined by mass spectrometry or by comparison with previously published reference gels.

FIG. 3.

Effect of flhA mutation on plcA (A) and hblC (B) transcriptions. (A) β-Galactosidase activity of the B. thuringiensis 407 Cry⁻ [plcA′Z] (▪) and B. thuringiensis 407 Cry⁻ [plcA′Z] ΔflhA (○) strains. (B) β-Glucuronidase activity of the B. thuringiensis 407 Cry⁻ [plcA′Z] (▪) and B. thuringiensis 407 Cry⁻ [plcA′Z] ΔflhA (○) strains carrying pHT304-hbl′G. The cells were grown at 37°C in LB medium. Time zero indicates the onset of the stationary phase, and T_n is the number of hours before (−) or after time zero. Each point is the mean of three or four independent experiments. Vertical bars indicate standard errors.

FIG. 4.

Cytotoxicity to HeLa cells is FlhA dependent. (A) Cells were infected with the supernatant of B. thuringiensis 407 Cry⁻ [plcA′Z] (1) and B. thuringiensis 407 Cry⁻ [plcA′Z] ΔflhA (2) strains at a final dilution of 1:25. After 2 h, trypan blue dye was added to the cells. Untreated control cells are also shown (3). Nonpermeabilized cells remained unstained, whereas permeabilized cells allowed the dye to enter inside the cytoplasm, and cells were therefore stained blue. (B) At least 300 cells were visually counted, and the percentage of blue cells compared with that of unstained cells accounted for the percent cytotoxicity. Results are mean values of three independent experiments.

FIG. 5.

Effect of flhA mutation on virulence against the insect G. mellonella by force-feeding (A) and intrahemocoelic injection (B). (A) Last-instar larvae were force-fed with 2 μg Cry1C toxin and 5 × 10⁶ log-phase bacteria larva⁻¹. No mortality was observed with Cry1C toxin or vegetative bacteria alone. Results are mean values of four independent experiments. Verticals bars indicate the standard errors of the means. (B) Dose-mortality responses observed after injection into the hemocoel of B. thuringiensis 407 Cry⁻ [plcA′Z] (hatched line) and B. thuringiensis 407 Cry⁻ [plcA′Z] ΔflhA (solid line) vegetative bacteria. No mortality was observed in the control (NaPi buffer alone).