Secreted Compounds of the Probiotic Bacillus clausii Strain O/C Inhibit the Cytotoxic Effects Induced by Clostridium difficile and Bacillus cereus Toxins

Abstract

Although the use of probiotics based on Bacillus strains to fight off intestinal pathogens and antibiotic-associated diarrhea is widespread, the mechanisms involved in producing their beneficial effects remain unclear. Here, we studied the ability of compounds secreted by the probiotic Bacillus clausii strain O/C to counteract the cytotoxic effects induced by toxins of two pathogens, Clostridium difficile and Bacillus cereus, by evaluating eukaryotic cell viability and expression of selected genes. Coincubation of C. difficile and B. cereus toxic culture supernatants with the B. clausii supernatant completely prevented the damage induced by toxins in Vero and Caco-2 cells. The hemolytic effect of B. cereus was also avoided by the probiotic supernatant. Moreover, in these cells, the expression of rhoB, encoding a Rho GTPase target for C. difficile toxins, was normalized when C. difficile supernatant was pretreated using the B. clausii supernatant. All of the beneficial effects observed with the probiotic were abolished by the serine protease inhibitor phenylmethylsulfonyl fluoride (PMSF). Suspecting the involvement of a secreted protease in this protective effect, a protease was purified from the B. clausii supernatant and identified as a serine protease (M-protease; GenBank accession number Q99405). Experiments on Vero cells demonstrated the antitoxic activity of the purified protease against pathogen supernatants. This is the first report showing the capacity of a protease secreted by probiotic bacteria to inhibit the cytotoxic effects of toxinogenic C. difficile and B. cereus strains. This extracellular compound could be responsible, at least in part, for the protective effects observed for this human probiotic in antibiotic-associated diarrhea.

FIG 1

Effect of C. difficile VPI 10463 supernatant (Cd-SN) on Vero cell viability in the presence or absence of B. clausii O/C supernatant (OC-SN). Bars represent the percentage of cell attachment (A) or mitochondrial dehydrogenase activity (B) after incubation of Vero cells with the different treatments assayed (control, Cd-SN, OC-SN, and Cd-SN/OC-SN, with or without the protease inhibitors PMSF and EDTA). Vero cells were also pretreated with OC-SN before incubation with Cd-SN (OC-SN/cells – Cd-SN). The percentage of attachment or remaining mitochondrial dehydrogenase activity was calculated as 100 × (A/A_c), where A is the absorbance of treated cells and A_c is the absorbance of untreated control cells. Results were the averages ± standard deviations (SD) from 3 independent assays. **, significant differences from the corresponding control (P < 0.01); ϕ, significant differences from the Cd-SN condition (P < 0.01).

FIG 2

Effect of C. difficile VPI 10463 supernatant on the relative mRNA expression of Caco-2 cells in the presence or absence of B. clausii O/C supernatant. Cd-SN and OC-SN were preincubated (Cd-SN/OC-SN), with or without protease inhibitor (PMSF), before contact with cells. Bars represent the relative expression of genes (in the three independent assays performed) encoding RhoB (A), ZO-1 (B), and ROCK2 (C). **, significant differences from the corresponding control (P < 0.01).

FIG 3

Effect of B. cereus ATCC 14579 supernatant (Bc-SN) on Vero cell viability in the presence or absence of B. clausii O/C supernatant (OC-SN). Different times of coincubation between Bc-SN and OC-SN were tested before incubation with Vero cells (15 min, 30 min, and 2 h). Bars represent the percentage of cell attachment (A) or mitochondrial dehydrogenase activity (B) of cells coincubated with Bc-SN, OC-SN, or Bc-SN/OC-SN with or without protease inhibitors (PMSF and EDTA). Vero cells were also pretreated with OC-SN before incubation with Bc-SN (OC-SN/cells – Bc-SN). The percentage of cell attachment or remaining mitochondrial dehydrogenase activity was calculated as 100 × (A/A_c), where A is the absorbance of treated cells and A_c is the absorbance of untreated control cells. Results were the averages ± SD from 3 independent assays. *, significant differences from the corresponding control (P < 0.05); **, significant differences from the corresponding control (P < 0.01); ϕ, significant differences from the Bc-SN condition (P < 0.01).

FIG 4

Effect of B. cereus ATCC 14579 supernatant (Bc-SN) on human red blood cells in the presence or absence of B. clausii O/C supernatant (OC-SN). Bars represent the concentration of released hemoglobin when red blood cells were coincubated with Bc-SN, OC-SN, or Bc-SN/OC-SN with or without protease inhibitors (PMSF and EDTA). Results are averages ± SD from 6 values. **, significant differences from the corresponding control (P < 0.01).

FIG 5

SDS-PAGE of proteins and purified protease secreted by B. clausii O/C. Lane 1, molecular mass marker proteins (Precision Plus Protein standard; Bio-Rad, France); lanes 2 and 4, size exclusion chromatography-purified protease; lane 3, B. clausii O/C spore culture supernatant (crude extract).

FIG 6

Effect of C. difficile VPI 10463 supernatant (Cd-SN) on Vero cells viability in the presence or absence of B. clausii O/C purified protease (OC-protease). Bars represent the percentage of cell attachment of cells coincubated with Cd-SN, OC-protease, or Cd-SN/OC-protease with or without protease inhibitors (PMSF and EDTA). The percentage of cell attachment was calculated as 100 × (A/A_c), where A is the absorbance of treated cells and A_c is the absorbance of untreated control cells. Results were averages ± SD from 4 independent assays. **, significant differences from the corresponding control (P < 0.01); ϕ, significant differences from the Cd-SN condition (P < 0.01).

FIG 7

Effect of B. cereus ATCC 14579 supernatant (Bc-SN) on Vero cell viability in the presence or absence of B. clausii O/C purified protease (OC-protease). Bars represent the percentage of cell attachment of cells coincubated with Bc-SN, OC-protease, or Bc-SN/OC-protease with or without protease inhibitors (PMSF and EDTA). The percentage of cell attachment was calculated as 100 × (A/A_c), where A is the absorbance of treated cells and A_c is the absorbance of untreated control cells. Results were averages ± SD from 4 independent assays. **, significant differences from the corresponding control (P < 0.01); ϕ, significant differences from the Bc-SN condition (P < 0.01).