Lactobacillus reuteri is a heterofermentative lactic acid bacteria that belongs to the authochtonous microbiota of humans and animals 1. Animal studies have shown that certain Lb. reuteri strains used as probiotics can provide protection against detrimental effects of certain microbiological, chemical and physical stressors; lower cholesterol; modulate immune responses; and improve the development of ileal tissue 234. Lb. reuteri SD2112 (ATCC 55730) has been shown to reduce the incidence and severity of diarrhoea of different origins in children and adults 567, and to survive the passage through the stomach and upper intestine and transiently colonize the human gastrointestinal tract 6. Indeed, this strain has been used for more than ten years as probiotic and/or starter culture in food and health care products 4.

In the presence of glycerol Lb. reuteri can synthesize 3-hydroxypropionaldehyde (3-HPA). 3-HPA is excreted into the medium where it forms together with the hydrate and the dimer (reuterin), a dynamic multi-component equilibrium (HPA, HPA system: named reuterin in most publications) 4. Reuterin is a potent antimicrobial agent active against Gram positive and Gram negative bacteria, as well as yeasts, moulds and protozoa 8. Reuterin is synthesized in vitro under pH, temperature and anaerobic conditions similar to those of the gastrointestinal tract 9. In vivo, active reuterin synthesis could occur in the colon via Lb. reuteri metabolism, if sufficient amounts of glycerol become available as a product of luminal microbial fermentations, digestion of luminal fats, sloughed mucus and desquamated epithelial cells, and intestinal clearing of endogenous plasma glycerol 2. To our knowledge, there are presently no data concerning the amount of glycerol available for bacterial conversion to reuterin in the human intestine.

Reuterin is water soluble, effective in a wide range of pH, resistant to proteolytic and lipolytic enzymes and has been therefore studied as a food preservative or auxiliary therapeutic agent 4810. Reuterin is also believed to play a role in the probiotic effects of Lb. reuteri SD2112, but its mode of action on microbial growth has not been elucidated due to the high complexity of the HPA-system chemistry 4. It has been postulated that reuterin might inhibit the activity of bacterial ribonucleotide reductase, an enzyme catalyzing the first step in DNA synthesis, by competition (HPA-dimer) with ribonucleotides for binding sites or by reaction (3-HPA) with unstable sulfhydryl groups of ribonucleotide reductase or with thioredoxin which is required for enzymatic activity 11. The inhibition of the conversion of ribonucleotides to deoxyribonucleotides would explain the broad-spectrum activity of reuterin 11. It has also been suggested that the mechanism of action of reuterin might be directed towards sulfhydryl enzymes, and that the tolerance of bacteria to reuterin might be associated with intracellular levels of LMW-SH such as glutathione (unpublishd data).

Reduced glutathione (GSH), is a cysteine-containing tripeptide, mostly found in Gram negative bacteria such as purple and cyanobacteria. The highest GSH concentrations are observed in aerobes or facultative anaerobes, whereas strict anaerobes are generally deprived of GSH but possess other soluble thiol compounds 12. However, some faculative anaerobic Gram positive bacteria (eg. Lactococcus lactis subsp. lactis, Streptococcus agalactiae ATCC 12927 or Enterococcus faecalis ATCC 29212) can synthesize significant amount of GSH, whereas others (e.g. Lactococcus lactis subsp. cremoris Z8 or Streptococcus mutans ATCC 33402) can import GSH from the growth medium, but this faculty is strain dependent 1314. GSH plays important functions in bacteria including control of redox potential, protection against oxidative stress, detoxification of endogenously and exogenously derived toxins, protein folding and storage and transport of organic sulfur 15. This potent antioxidant is also believed to play a key role in detoxification of reactive aldehydes, such as acrolein and protects cells from oxidative damage such as produced by H₂O₂ 16. Recently, glutathione reductase has been shown to contribute to Lactobacillus sanfranciscensis oxygen tolerance 17. Various studies have reported the MIC of reuterin against pathogens and lactic acid bacteria 91118. However, the activity of pure reuterin against intestinal bacteria has never been reported, even though this data is important for understanding the probiotic effects of Lb. reuteri in vivo when administered to humans and animals. The aim of this study was therefore to determine the in vitro antibacterial activity of pure reuterin on a representative panel of human intestinal bacteria. The intracellular concentration of LMW-SH and the sensitivity of intestinal bacteria to hydrogen peroxide were also recorded to tentatively correlate reuterin sensitivity to intracellular LMW-SH such as GSH and oxygen tolerance.

In this study, the MIC and MBC of pure reuterin for different intestinal bacteria were determined using a serial dilution microtiter plate assay. With the exception of Dobrogosz et al. 19, who express the inhibitory activity of reuterin with arbitrary units defined, after analysis with ¹⁴C-labeled glycerol, as 1 U equal 8 μg reuterin/ml, most of the authors have used poorly defined arbitrary units to express the inhibitory concentration of reuterin 1020. Therefore, a direct comparison of published data for antimicrobial activities of reuterin against different bacteria is difficult. Another problem of these studies is that the molar concentration of reuterin was calculated using the molecular mass of the dimer form 9112122, although reuterin occurs in its monomeric form at the low concentrations used 4. Moreover arbitrary units are changed by many experimental factors as shown for bacteriocins, such as variation of sensitivity among indicator strains of the same species, medium and incubation conditions for the activity tests and concentrations of antimicrobial compounds in tested preparations 2324. For data comparison, test conditions must be standardized and MIC data expressed in μg ml^-1 or mM. In this study the molar concentration of reuterin was determined by a method originally developed for acrolein quantification, using the acid-catalyzed formation of a colored complex between acrolein/HPA and tryptophan 25, which has been adopted by several other authors to correctly quantify reuterin 4262728. Furthermore, all bacteria were tested under controlled test conditions and the same medium for the activity test.

Our data showed that reuterin is an effective inhibitor for most tested intestinal bacteria, with the exception of lactobacilli. Among the 20 species non belonging to the Lactobacillus genus, 14 were inhibited and/or killed by reuterin concentrations equal or below 3.8 mM. They exhibited higher sensitivity than E. coli, the usual sensitive indicator used to test reuterin 4. Dobrogosz et al. 11 and Chung et al. 9 reported, using a two-fold dilution test, MIC for E. coli K12 of 4–5 U/ml (0.9–1.1 mM, after recalculation based on 3-HPA monomer and 1 U = 8 μg reuterin/ml), 32 U/ml (6.9 mM) for Lb. reuteri and ranging from 4 to 42 U/ml (0.9–9.1 mM) for different lactobacilli. Our data show the same trends; MIC values for lactobacilli were at least 2 times higher than for E. coli (Table 1). However, under our test conditions, MIC measured for lactobacilli were approximately 10 times higher than reported in previous studies. Dobrogosz et al. 11 found that Listeria monocytogenes showed a similar low sensitivity to reuterin as for lactobacilli, with MIC of 4–42 U/ml (0.9–9.1 mM), which was different from our data for L. ivanovii and L. innocua, which were much more sensitive to reuterin than lactobacilli. These differences for MIC measured in different studies can be due to various factors of the arbitrary activity tests, including pH, culture media, incubation temperature, concentration and physiological state of the tested bacteria which affect both the growth 23 of the tested strains and the activity of reuterin 27. As well, a strict criteria for MIC calculation (lowest reuterin concentration resulting in complete inhibition of the tested strain) was used in our study. The aldehyde function of 3-HPA is highly reactive due to its β-hydroxy moiety and therefore the molecule spontaneously reacts with available amino- and sulfhydryl-functional groups present in compounds of the growth medium and bacteria 4. According to Lüthi-Peng et al. 27, only 30 % of reuterin remains free after 24 h of incubation at 37°C in a complex medium. The lactobacilli were also tested in the LAMVAB growth medium 29, which is used as selective medium to enumerate lactobacilli in fecal samples, at pH 5.0. With this medium MIC were 2–5 times lower than the MIC obtained with supplemented BHI (data not shown). This result clearly shows that the comparison of data from different studies on reuterin activity is difficult, but comparison within one study is much more valid. This limitation should also be taken into account when sensitivity data to antimicrobial compounds such as reuterin obtained with in vitro tests, are extrapolated to other in vitro systems or in vivo conditions such as in the gastrointestinal tract.

To tentatively explain the different behavior of intestinal bacteria toward reuterin activity, we analyzed their intracellular concentration of LMW-SH including GSH, which has been shown to be involved in aldehyde detoxification mechanisms 30. With this intention, we used the Ellmann's reagent that specifically reacts with thiol groups. To make the distinction between LMW-SH, and thiol groups of proteins, the concentration of intracellular LMW-SH was determined in the cell-free supernatant of bacterial crude extracts after protein elimination by precipitation with sulfosalicylcylic acid. Although, all LMW-SH and not only GSH were measured in the present study, this test gives a good estimation of GSH concentration because other LMW-SH occur only in low concentration 31. LMW-SH were present in all tested bacteria. The highest intracellular concentrations (> 20 μmol/10¹² cells) were measured in two Gram positive strict anaerobic species (C. aerofaciens and Cl. difficile) and one facultative anaerobic bacteria (Lb. salivarius). Although glutathione appears to be mostly present in cyanobacteria and purple bacteria 12, high intracellular glutathione concentrations have already been measured in Gram positive Closdriales, including strict anaerobes such as Clostridium perfringens 32 and Lactobacillus spp. 3334 and Bifidobacterium spp. 35. It is usually suggested that Gram positive organisms do not produce GSH but accumulate it from the growth medium. However, other evidences indicate that some Gram positive bacteria (e.g. Enterococcus faecalis ATCC 29212 or Streptococcus agalactiae ATCC 12927) can synthetize GSH 13. Moreover, Fernandes and Steele 33 showed that intracellular GSH levels in lactic acid bacteria varied widely between organisms and were very dependent on strains tested as well as growth conditions, including growth medium composition (presence or absence of GSH or precusors) and aeration, as well as bacterial growth phase. Very few data are reported on the capacity of intestinal bacteria to synthesize or harvest GSH from the growth medium. It has been shown that some probiotic lactobacilli strains can produce and even release high amount of GSH 36. In our study, the growth medium was deprived of cystein, which is a major precursor for GSH synthesis. However, GSH from yeast extract was likely available for bacterial uptake in the growth medium 12. We recently observed that external addition of GSH could conteract the toxicity of reuterin (3-HPA) toward E. coli K12, which suggests a correlation between induced toxicity of 3-HPA on E. coli K12 and depletion of GSH (unpublished data). However, in this study, the lack of correlation between intracellular GSH concentration and reuterin sensitivity suggests that other detoxification mechanisms prevent 3-HPA toxicity, such as 1,3-propanediol-dehydrogenase, which converts 3-HPA to the unreactive 1,3-propanediol. No inhibitory activity of 1,3-propanediol against intestinal bacteria is known until now.

GSH is considered to play a key role in protecting cells against oxidative stress. It is usually the most abundant intracellular, low-molecular-weight thiol in E. coli 32 and might contribute to protect E. coli cells 30 as well as Gram positive bacteria 34 from the lethal effects of oxidative damage. It is known that GSH can reduce hydrogen peroxide in exchange for oxidation of GSH to glutathione disulfide (GSSG) 30. The tolerance of intestinal bacteria to hydrogen peroxide was then determined to assess a possible correlation between their sensitivity to reuterin and oxygen detoxification. Oxygen toxicity is considered to result from the effects of activated compounds including superoxide, hydrogen peroxide and hydroxyl radicals 37. Facultative anaerobic bacteria decompose and detoxify these activated oxygen metabolites enzymatically with superoxide dismutase and catalase 37. Most of the intestinal bacteria, classified as strict anaerobic organisms, are devoid of these enzymes and therefore their growth can be inhibited when H₂O₂ is present 37. Indeed, the most oxygen sensitive species tested, which belongs to the orders Bacteroidales (Bacteroides spp.) and Clostridiales (Clostridium spp., Ruminococcus sp., Eubacterium spp. and Collinsella sp.) were completely inhibited with 10 g l^-1 hydrogen peroxide. Five bifidobacteria among the six tested, were not inhibited by 10 g l^-1 hydrogen peroxide although they are classified as a strict anaerobic bacteria. Using the same technique, Shimamura et al. 37 reported variations in H₂O₂ sensitivity of Bifidobacterium spp., and found similar results with B. infantis being the least sensitive to H₂O₂, followed by B. breve, B. adolescentis and B. longum. The facultative anaerobic bacteria tested were found to largely differ in their sensitivities to hydrogen peroxide. Among lactobacilli, Lb. acidophilus and Lb. reuteri DSM 20016 were the most tolerant.Lb. acidophilus and Bifidobacterium spp. have NADH oxidase and NADH peroxidase to scavenge hydrogen peroxide and prevent cell death in the presence of oxygen 3738. Moreover, it is known that some Lb. acidophilus strains produce hydrogen peroxide and peroxidase 39. In addition, certain lactobacilli produce superoxide dismutase 40 and a non-heme pseudocatalase 41.

Surprinsigly, E. coli and Listeria spp. also classified as facultative anaerobes and which possess a catalase were inhibited by 10 g l^-1 hydrogen peroxide. But it has also been shown that the high H₂O₂ sensitivity of E. coli can be related to DNA damage caused by hydrogen peroxide 42. In this study, we did not establish any correlation between bacterial hydrogen peroxide sensitivity and intracellular LMW-SH concentration or reuterin sensitivity. However, the wide range of organisms tested an the multiplicity of oxygen detoxification mechanisms as well as the wide functions of LMW-SH and GSH in microorganisms 15 suggest that strain specific mechanisms are likely to be involved rather than a general mechanism applicable to all bacteria.

This study reports for the first time the antimicrobial activity of reuterin on various intestinal bacteria. Large differences among species were observed. With the exception of Lactobacillus strains which were more resistant, most intestinal bacteria were inhibited by reuterin concentration below 7.5 mM. Moreover, the MBC (7.5–15 mM) recorded for the two Listeria spp. tested were higher than for the main intestinal commensals. The most sensitive species were inhibited by reuterin concentration around 2 mM, which suggests that low concentration of reuterin production in situ (if sufficient amount are present either via in situ production or ingested as encapulated form with a probiotioc bacteria) seems to be sufficient to alter the intestinal bacterial growth. However caution should be applied when extrapolating these results obtained with in vitro activity tests to specific in vivo conditions, where other important factors such as nutrient competition and antagonistic and/or synergistic relationships between the intestinal microbiota can largely influence the inhibition effects of antimicrobials such as reuterin. Complementary experiments are therefore needed in more complex in vitro systems, such as in vitro intestinal fermentation models, as in vivo experimentation is very difficult. No correlation was found between bacterial sensivity to reuterin and intracellular LMW-SH and/or hydrogen peroxide sensitivity. Our data suggest that protection via LMW-SH such as GSH is not a general protection mechanism toward reuterin toxicity.

GLB, CL and VC designed the experiments. SV performed the production and purification of reuterin. VC and MD performed the activity test experiments. VC performed the determination of LMW-SH and the cell survival to hydrogen peroxide. GLB, CL and SV revised the manuscript. All authors read and approved the final manuscript.