The mechanism underlying the killing activity of Lactobacillus strains against bacterial pathogens appears to be multifactorial. Here, we investigate the respective contributions of hydrogen peroxide and lactic acid in killing bacterial pathogens associated with the human vagina, urinary tract or intestine by two hydrogen peroxide-producing strains. In co-culture, the human intestinal strain Lactobacillus johnsonii NCC933 and human vaginal strain Lactobacillus gasseri KS120.1 strains killed enteric Salmonella enterica serovar Typhimurium SL1344, vaginal Gardnerella vaginalis DSM 4944 and urinary tract Escherichia coli CFT073 pathogens. The cell-free culture supernatants (CFCSs) produced the same reduction in SL1344, DSM 4944 and CFT073 viability, whereas isolated bacteria had no effect. The killing activity of CFCSs was heat-stable. In the presence of Dulbecco's modified Eagle's minimum essential medium inhibiting the lactic acid-dependent killing activity, CFCSs were less effective at killing of the pathogens. Catalase-treated CFCSs displayed a strong decreased activity. Tested alone, hydrogen peroxide triggered a concentration-dependent killing activity against all three pathogens. Lactic acid alone developed a killing activity only at concentrations higher than that present in CFCSs. In the presence of lactic acid at a concentration present in Lactobacillus CFCSs, hydrogen peroxide displayed enhanced killing activity. Collectively, these results demonstrate that for hydrogen peroxide-producing Lactobacillus strains, the main metabolites of Lactobacillus, lactic acid and hydrogen peroxide, act co-operatively to kill enteric, vaginosis-associated and uropathogenic pathogens.
The mechanism underlying the killing activity of Lactobacillus strains against bacterial pathogens appears to be multifactorial and to include the production of hydrogen peroxide, lactic acid, bacteriocin-like molecules and unknown heat-stable, nonlactic acid molecules (Servin, 2004). The vaginal microbial flora plays a role in maintaining human health (Pybus & Onderdonk, 1999; Aroutcheva et al., 2001a, b), and within this flora, resident Lactobacillus species exercise antibacterial activity by producing metabolites, including hydrogen peroxide, lactic acid and other antibacterial molecules (Eschenbach et al., 1989; Klebanoff et al., 1991; Hillier et al., 1992; Saunders et al., 2007). Hydrogen peroxide-producing lactobacilli that colonize the vagina have been reported to reduce the prevalence of bacterial vaginosis (Wilks et al., 2004; Antonio et al., 2005). For women with recurrent urinary tract infections (UTIs), who often display persistent vaginal colonization by Escherichia coli (Johnson & Russo, 2005), the absence of hydrogen peroxide-producing strains of Lactobacillus appears to be important in the pathogenesis of recurrent UTI by facilitating colonization by E. coli (Gupta et al., 1998). In the intestine, the role of hydrogen peroxide-producing strains in killing enteric pathogens has been poorly documented. Recently, Pridmore (2008) reported for the first time that the human intestinal isolate Lactobacillus johnsonii NCC533, which exhibits antimicrobial properties against Salmonella typhimurium (Bernet et al., 1994; Bernet-Camard et al., 1997; Fayol-Messaoudi et al., 2005; Makras et al., 2006) and Helicobacter pylori (Michetti et al., 1999; Felley et al., 2001; Cruchet et al., 2003; Gotteland & Cruchet, 2003; Sgouras et al., 2005), produced hydrogen peroxide that was effective in killing S. typhimurium.
Here, we investigate the respective contributions of hydrogen peroxide and lactic acid in killing bacterial pathogens associated with the human vagina, urinary tract or intestine by two hydrogen peroxide-producing strains: enteric L. johnsonii NCC933 (Pridmore et al., 2008) and vaginal Lactobacillus gasseri KS120.1 (Atassi et al., 2006b). The human bacterial pathogens we used were Gardnerella vaginalis strain DSM 4944, uropathogenic E. coli (UPEC) strain CFT073 (UPEC CFT073) and Salmonella enterica serovar Typhimurium strain SL1344 (S. typhimurium SL1344). Gardnerella vaginalis is a heavily pilated, gram-negative bacterium (Boustouller et al., 1987) that produces cytolysin (Cauci et al., 1993) and attaches to epithelial cells (Scott et al., 1987). It is of particular importance in the etiology of bacterial vaginosis (Mikamo et al., 2000; Aroutcheva et al., 2001a). Strain CFT073 is the prototype UPEC strain involved in inducing recurrent UTI (Johnson & Russo, 2005). It displays various virulence factors (Marrs et al., 2005) such as toxins, and type 1 pili that help to form an intracellular reservoir of the pathogen by invading uroepithelial cells (Mysorekar & Hultgren, 2006), as well as flagella that enables them to ascend to the upper urinary tract and to disseminate throughout the host (Lane et al., 2007). Salmonella typhimurium is the etiological agent of gastroenteritis in humans (Grassl & Finlay, 2008), and produces a large panel of virulence factors, including flagella (Chevance & Hughes, 2008) and type-III secretion system-associated bacterial effectors (Patel & Galan, 2005; Galan & Wolf-Watz, 2006). It has the ability to interact with and invade host cells, and then to live within these cells (Ly & Casanova, 2007). We report that the two Lactobacillus strains display killing activity against G. vaginalis, UPEC and S. typhimurium by substances present in the cell-free culture supernatants (CFCSs). Moreover, our results show that the main metabolic product of Lactobacillus, lactic acid, displays no killing activity at the concentration present in Lactobacillus cultures, whereas hydrogen peroxide dose-dependently killed these pathogens. We also provide evidence that at the concentration present in Lactobacillus cultures, lactic acid considerably enhances the killing activity of hydrogen peroxide.
Materials and methods
Bacterial strains and culture conditions
The prototype UPEC strain CFT073 (Mobley et al., 1990) and S. typhimurium SL1344 (Finlay & Falkow, 1990) were used. Bacteria were cultured in Luria–Bertani (LB) agar (Difco Laboratories, Detroit, MI) and incubated at 37 °C for 24 h. Gardnerella vaginalis DSM 4944 was grown on Gardnerella agar plates purchased from BioMerieux (Lyon, France), as described previously (Atassi et al., 2006a, b). Bacteria were suspended in pH 7.0 buffered sodium chloride-peptone solution at about 106 CFU mL−1. Five hundred microliters of the prepared suspension was spread on the agar plate. The inoculated plates were dried under a sterile laminar airflow. The agar plates were then incubated under anaerobic conditions in a sealed anaerobic jar (Becton Dickinson) at 37 °C for up to 36 h. Before being used, G. vaginalis was subcultured in brain–heart infusion supplemented with yeast extract (1%), maltose (0.1%), glucose (0.1%) and horse serum (10%) under anaerobic conditions in a sealed anaerobic jar at 37 °C for up to 36 h. For each experiment, bacteria were subcultured for the exponential phase in appropriate media.
Lactobacillus johnsonii strain NCC533 was from the Nestec Research Center at Vers-chez-les-Blanc (Switzerland). The L. gasseri KS120.1 strain isolated from the vaginal flora of a healthy woman (Department of Obstetrics and Gynecology, Zurich University Hospital, Switzerland) was from Medinova (Zurich, Switzerland) (Atassi et al., 2006a, b). All the Lactobacillus strains were grown in De Man, Rogosa, Sharpe (MRS) broth (Biokar Diagnostic, Beauvais, France) for 24 h at 37 °C. The Lactobacillus culture was adjusted to pH 4.5 by adding HCl or NaOH to ensure standardized conditions.
Preparation of bacteria and CFCSs
Cultures of the Lactobacillus strains (24 h) were centrifuged at 10 000 g for 30 min at 4 °C. Bacteria were collected and washed three times with sterile phosphate-buffered saline (Coconnier et al., 1997, 2000). Supernatants of the centrifuged cultures were collected and passed through a sterile 0.22-μm filter unit Millex GS (Millipore, Molsheim, France). The filtered CFCSs were tested to confirm the absence of bacterial colonies by plating on tryptic soy agar (TSA).
Lactic acid determination
A commercial d- and l-lactic acid determination kit was used (Test-Combination d-lactic acid/l-lactic acid UV-method, Boehringer Mannheim GmbH, Germany) to determine the concentration of lactic acid in the Lactobacillus cultures.
Killing activity in co-culture condition
The killing activities of Lactobacillus cultures and isolated Lactobacillus bacteria were examined under co-culture conditions as described previously (Atassi et al., 2006a, b). Briefly, an exponential culture of bacterial pathogen in an appropriate culture medium (108 CFU mL−1, 500 μL) was incubated with or without Lactobacillus culture (500 μL of a 24-h culture) at 37 °C for 4 h. In a separate experiment, an exponential culture of bacterial pathogen in an appropriate culture medium (108 CFU mL−1, 500 μL) was incubated with or without Lactobacillus bacteria (108 CFU mL−1, 500 μL) or Lactobacillus CFCS (500 μL) isolated from a 24-h culture at 37 °C for 4 h.
Characterization of the killing activity of Lactobacillus CFCSs
The Lactobacillus CFCSs were heated to 110 °C for 1 h (Coconnier et al., 1997). To test their sensitivity to protease, the Lactobacillus CFCSs were incubated at 37 °C for 1 h with and without pronase (200 μg mL−1), trypsin (200 μg mL−1), proteinase K (100 μg mL−1) or pepsin (200 μg mL−1) (Sigma-Aldrich Chimie SARL, L'Isle d'Abeau Chesnes, France) (Coconnier et al., 1997). To determine the killing effect attributable to hydrogen peroxide, the CFCSs were treated at 37 °C for 1 h with catalase (from bovine liver, Sigma-Aldrich Chimie SARL) at a final concentration of 5 μg mL−1 (Atassi et al., 2006a, b). Hydrogen peroxide solution was used to control the activity of catalase and bovine serum albumin to control that of proteolytic enzymes. To determine whether lactic acid was involved in the killing activity, the experimental conditions used were as described previously (Fayol-Messaoudi et al., 2005). Briefly, an exponential culture of bacterial pathogen in an appropriate culture medium (108 CFU mL−1, 500 μL) was incubated with Lactobacillus CFCS (500 μL of a 24-h culture) with or without Dulbecco's modified Eagle's minimum essential medium (DMEM) (500 μL) (Life Technologies, Cergy, France) at 37 °C for 4 h. To eliminate low–molecular-weight factors, the Lactobacillus CFCSs were passed through a Microcon SCX-filter (cut-off 3 kDa) (Millipore) (De Keersmaecker et al., 2006). Aliquots of the co-culture medium were removed, serially diluted and then plated on appropriate media as described above to determine the bacterial colony counts of the pathogen. The bacterial colony counts of the pathogen were determined as described above.
Killing activity of lactic acid and hydrogen peroxide
An exponential culture of bacterial pathogen (108 CFU mL−1, 500 μL) was incubated with or without increasing concentrations of dl-lactic acid or hydrogen peroxide (Sigma-Aldrich Chimie SARL) at 37 °C for 4 h. To test the effect of lactic acid on the killing activity of hydrogen peroxide, pathogens (108 CFU mL−1, 500 μL) were incubated in the presence of increasing concentrations of hydrogen peroxide and 65 mM of dl-lactic acid. The bacterial colony counts of the pathogen were performed as described above.
Results are expressed as the mean±SEM. Statistical comparisons and Student's t-test were performed, with P <0.01 considered statistically significant.
Results and discussion
Killing activity of L. johnsonii NCC533 and L. gasseri KS120.1
Cultures of L. johnsonii NCC533 and L. gasseri KS120.1 were tested for their killing activity against the UPEC CFT073, G. vaginalis DSM 4944 and S. typhimurium SL1344. The results reported in Fig. 1 show that the 24-h cultures of both the Lactobacillus strains reduced the viability of the pathogens, but with different efficacies. Our results show that the killing activity of Lactobacillus cultures results from substances present in CFCS, and that isolated bacteria display no killing activity.
Characteristics of the killing activity of L. johnsonii NCC533 and L. gasseri KS120.1
We next investigated the characteristics of the killing activity of L. johnsonii NCC533 and L. gasseri KS120.1 CFCSs. Part of the killing activity was attributable to heat-stable components (Fig. 2), which were not sensitive to protease treatment (not shown). The killing activity was not attributable to a pH effect, because MRS at pH 4.5 shows no activity (Fig. 2). Fayol-Messaoudi (2005) have previously demonstrated that the killing activity of lactic acid against S. Typhimurium was inhibited after adding DMEM to the LB culture medium. Here, when DMEM was added to each appropriate pathogen culture medium, the killing activity of Lactobacillus CFCSs was slightly decreased compared with that without DMEM (Fig. 2). In order to investigate the role played by hydrogen peroxide in the killing activity of Lactobacillus CFCS against the three pathogens, the CFCSs were exposed to catalase treatment. As shown in Fig. 3, the killing activity of the CFCS of L. johnsonii NCC533 and of L. gasseri KS120.1 was considerably reduced after catalase treatment. Collectively, these findings indicate that the killing activity of L. johnsonii NCC533 and L. gasseri KS120.1 strains against S. typhimurium SL1344, UPEC CFT073 and G. vaginalis DSM 4944 was mainly attributable to heat-stable secreted molecules and hydrogen peroxide.
Lactic acid kills the pathogens at high concentrations
Lactic acid was present in culture media of the hydrogen peroxide-producing strain (L. johnsonii NCC533: 61±16 mM, and L. gasseri KS120.1: 63±12 mM). Makras (2006) and De Keersmaecker (2006) concluded that the antibacterial effect of probiotic L. johnsonii NCC533 and Lactobacillus rhamnosus GG strains against serovar Typhimurium results from the accumulation of their main metabolite, lactic acid. The above results (Fig. 2) show that the killing activity of L. johnsonii NCC533 and L. gasseri KS120.1 CFCSs in the presence of DMEM is slightly decreased, which prompted us to investigate the concentration-dependent killing activity of MRS at pH 4.5 containing increasing concentrations of lactic acid. We found that lactic acid alone displayed significant killing activity at concentrations from 100 mM that was greater against G. vaginalis DSM 4499 and S. typhimurium SL1344 than against UPEC CFT073 (Fig. 4). Like Fayol-Messaoudi (2005), we found that the killing activity of lactic acid against G. vaginalis DSM 4499, S. typhimurium SL1344 and UPEC CFT073 was totally abolished in the presence of DMEM (Fig. 4). These results indicate that lactic acid did indeed kill these pathogens at concentrations higher than those present in the 24-h cultures of the Lactobacillus strains tested.
Lactic acid increases the killing activity of hydrogen peroxide
We tested the killing activity of hydrogen peroxide alone and in the presence of lactic acid. The data in Fig. 5 show that MRS at pH 4.5 containing hydrogen peroxide alone was able to kill G. vaginalis DSM 4944, S. typhimurium SL1344 and UPEC CFT073 strains in a concentration-dependent manner. In the presence of lactic acid at the concentration present in CFCSs (65 mM), we found that the killing activity of hydrogen peroxide against G. vaginalis DSM 4944 and S. typhimurium SL1344 was significantly greater than that of hydrogen peroxide alone, whereas that against UPEC CFT073 was increased to a lesser extent. Collectively, these data show that lactic acid, which is present in the CFCSs of L. johnsonii NCC533 and L. gasseri KS120.1, co-operates with hydrogen peroxide to kill G. vaginalis DSM 4499, S. typhimurium SL1344 and UPEC CFT07 more efficiently.
The data reported here show that upon co-culture, the enteric and vaginal strains L. johnsonii NCC533 and L. gasseri KS120.1 reduced the viability of G. vaginalis, S. typhimurium and UPEC strains with differing levels of efficacy. We found that hydrogen peroxide, which dose-dependently kills the pathogens, displays enhanced killing activity against G. vaginalis and S. typhimurium and to a lesser extent against UPEC, in the presence of lactic acid at the concentration present in the Lactobacillus cultures. The role of acidity in antipathogenicity is controversial. It has been established that pathogens have developed sophisticated adaptive systems. For example, E. coli O157:H7 possesses adaptive systems that protect it against acid stress (Foster, 2004). In G. vaginalis, the adaptive system(s) that provide protection against acid stress have not been identified, but a recent report indicates that increased tolerance to hydrogen peroxide and lactic acid can result from its ability to form a biofilm (Patterson et al., 2007). In S. Typhimurium, the PhoP/PhoQ two-component system that controls several physiological and virulence functions is activated by low Mg2+, acidic pH and antimicrobial peptides (Kato & Groisman, 2008). Moreover, gene products including RpoS, an alternative σ factor involved in stationary-phase physiology and stress responses, and Fur, a regulator of iron metabolism, have been shown to be involved in the adaptive response of Salmonella to acid stress (Audia et al., 2001). These regulators are able to protect against two types of acid stress, with RpoS and Fur protecting against organic acid stress and PhoP and RpoS protecting against inorganic acid stress (Baik et al., 1996; Bearson et al., 1998). In addition, Bearson (2006) have recently identified the phoP, rpoS, fur and pnp genes as being involved in protecting serovar Typhimurium against exposure to lactic acid.
Our group has previously reported that, at the concentration present in Lactobacillus CFCSs, lactic acid plays no role in the anti-Salmonella activities of L. johnsonii NCC533, L. rhamnosus GG, Lactobacillus casei Shirota YT9029, L. casei DN-114 001, L. rhamnosus GR1 or Lactobacillus acidophilus LB strains (Bernet et al., 1994; Coconnier et al., 1997, 2000; Hudault et al., 1997; Lievin-Le Moal et al., 2002; Fayol-Messaoudi et al., 2005). Here, we found that at the concentration present in Lactobacillus CFCS lactic acid alone plays no role in the killing effect of L. johnsonii NCC533 or vaginal L. gasseri KS120.1 against two other pathogens: UPEC CFT073 and G. vaginalis DSM 4944 strains. The observation that the killing activity of lactic acid develops at high concentrations is consistent with Makras (2006), who have shown that activities of lactic acid started at 100 mM. In contrast, based on the fact that the activity of L. rhamnosus GG CFCS against the growth and survival of serovar Typhimurium disappears after dialysis eliminating lactic acid, whereas it is still present after dialysis against a lactic acid solution, De Keersmaecker (2006) have concluded that lactic acid is responsible for the activity of L. rhamnosus GG. However, eliminating lactic acid could have an effect on some other molecule(s) secreted by Lactobacillus that kill pathogens in co-operation with lactic acid. Consistent with this hypothesis, Niku-Paavola (1999) have proposed that compounds secreted by Lactobacillus plantarum act synergistically with lactic acid, and Makras (2006) observed that L. johnsonii NCC533 CFCS was effective against serovar Typhimurium by unknown inhibitory substance(s) that are only active in the presence of lactic acid. These nonlactic acid, heat-resistant anti-Salmonella molecule(s) present in the CFCSs of probiotic Lactobacillus strains have not yet been identified (McGroarty & Reid, 1988; Bernet-Camard et al., 1997; Coconnier et al., 1997; Hudault et al., 1997; Ocana et al., 1999; Aroutcheva et al., 2001b; van de Guchte et al., 2001; Sgouras et al., 2004, 2005; Fayol-Messaoudi et al., 2005, 2007; Atassi et al., 2006a). It has already been suggested that pyroglutamic acid may be responsible for the antimicrobial activity of L. rhamnosus GG and L. casei strains LC-10 and LB1931 (Silva et al., 1987; Huttunen et al., 1995; Yang et al., 1997), but it has been found to be intrinsically present in MRS medium and it does not increase during bacterial growth (De Keersmaecker et al., 2006). Adding increasing concentrations of acetic or formic acid to MRS medium has no effect on the viability of serovar Typhimurium (De Keersmaecker et al., 2006).
Vaginal Lactobacillus strains, by producing hydrogen peroxide, displayed antibacterial activity against vaginosis-associated pathogens including G. vaginosis and Prevotella bivia (Aroutcheva et al., 2001a). Consistent with this, Lactobacillus strains have been isolated from the human vaginal microbiota for probiotic use against vaginosis-associated pathogens (Reid & Burton, 2002; Reid et al., 2003) on the basis of their ability to produce high levels of hydrogen peroxide (Klebanoff et al., 1991; Hillier et al., 1992, 1993). Moreover, Pridmore (2008) first reported for an intestinal Lactobacillus that hydrogen peroxide contributes to the killing activity of L. johnsonii NCC533 against serovar Typhimurium. Consistent with these reports, here, we observed that hydrogen peroxide concentration-dependently kills serovar Typhimurium, G. vaginosis and UPEC strains. Moreover, we report that lactic acid acts synergistically with hydrogen peroxide to kill G. vaginalis, S. typhimurium and UPEC more efficiently. The mechanism underlying the stimulatory effect of lactic acid observed could be related to the observation by Greenacre (2006), who have reported that the lactic acid-induced acid tolerance response causes hydrogen peroxide sensitivity in serovar Typhimurium via the downregulation of the OxyR regulon. A second mechanism could also be proposed, resulting from the permeabilizing effect of lactic acid on the gram-negative bacterial outer membrane (Alakomi et al., 2000), thus facilitating the passage of molecules across the membrane, and in turn increasing the killing effects of antimicrobial compounds (Niku-Paavola et al., 1999; Alakomi et al., 2000).