Background
The communication of bacteria with each other is termed “quorum sensing” (QS)[
1]. It is an important global gene regulatory mechanism, which is used by gram-negative as well as gram-positive bacteria, enabling individual bacteria to communicate and coordinate their behavior in populations. In general terms, it is often defined as cell density-dependent regulation of gene expression via extracellular signals. Bacteria produce small, diffusible signals, termed “autoinducers”. When these signals reach a critical threshold concentration, the targeted QS genes are activated or repressed. Acyl-homoserine lactone signal-mediated QS systems are the primary QS system discovered in gram-negative bacteria. AHL were originally identified in marine bacteria, where they play a pivotal role in the regulation of bioluminescence in
Vibrio fischeri[
2‐
4]
. Bacteria may produce one or more different AHL, which regulate diverse phenotypes, such as biofilm formation, swarming, production of proteases, antibiotics, siderophores or bioluminescence, conjugation, the modulation of the immune system and the induction of apoptosis[
5‐
11]. While AHL are responsible for intraspecies communication, a second Autoinducer (AI-2) was discovered, which is responsible for the interspecies communication; furanosyl borate diester is synthezised by the gene
lux S, which is found in the genome of many gram-negative as well as gram-positive bacteria. It is an integral component of the activated methyl cycle, which detoxifies S-adenosyl-L-Methionine (SAM)[
12‐
14].
Probiotics are, according to the FAO/WHO “live bacteria which when administered in adequate amounts confer to a health benefit to the host”. The numerous mechanisms of the probiotic acting microorganisms include induced expression of certain cytokines, as well as increased secretion of immunoglobulin A and mucin, activation of lymphocytes and macrophages and inhibition of the adhesion and invasion of epithelial cells[
15].
Escherichia coli Nissle 1917 (Mutaflor) is one of the most extensively studied probiotic bacterium. It was isolated in 1917 by Prof. Nissle from feces from a soldier during the First World War, who did not suffer from diarrhoea as his comrades did. Prof. Nissle realized the potential health benefits early on[
16]. In addition, several important discoveries were made during recent years:
E. coli Nissle is inducing human ß-defensin 2 (hBD-2) expression in the cell culture in a time and density dependent manner[
17]. In another study it has been shown that
E. coli Nissle is decreasing TNF-α secretion[
18]. Also,
E. coli Nissle is outcompeting several nonpathogenic and pathogenic bacteria including ETEC and EPEC strains in biofilm formation[
19]. Maintaining remission in ulcerative colitis using
E. coli Nissle have shown equivalent efficacy to the gold standard mesalazine[
20‐
22]. However, the exact mechanisms by which
E. coli Nissle is exerting its beneficial effects are still not completely understood[
23].
If E. coli Nissle is using QS to communicate with itself or with other species has not been elucidate so far. In this report we want to elucidate whether or not this bacterium is using QS. For the first time we were able to show that E. coli Nissle is producing AI-2 in a density dependent manner. We silenced the lux S gene by intron insertion and studied the effect of QS in the DSS mouse model of acute colitis. We observed differential expression of different cytokines and mBD-1, suggesting that indeed QS is influencing the probiotic properties of E. coli Nissle.
Discussion
E. coli Nissle 1917 is a well established probiotic bacterium. Since its first isolation and description 95 years ago, the interest in this bacterium has increased steadily. Especially during the last few years, the understanding on the mechanisms this bacterium is employing for its beneficial effects, has increased substantially[
23]. However, the exact mechanism of action still remains to be elucidated to completely “understand” this bacterium. One important gene regulation mechanism, QS, has, so far, not been looked at in
E. coli Nissle. It is a density dependent genetic regulatory mechanism which allows single bacterial cells to measure the concentration of certain bacterial signal molecules. In case that
E. coli Nissle is using QS, we would like to study its influence on the probiotic properties of the bacterium.
Not surprisingly, this bacterium, as all other
E. coli strains having been studied so far, is not synthesizing homoserine lactones (AI-1 molecules)[
13]. However, we showed for the first time, that AI-2 is produced by
E. coli Nissle in a density dependent manner. During the logarithmic and late logarithmic growth phase the amount of AI-2 being produced is the largest. This finding suggests, that
E. coli Nissle is using interspecies communication and is “talking” with other bacterial species. In order to study the effect of AI-2 in
E. coli Nissle, we silenced the corresponding gene via intron insertion. We performed control experiments and were able to show, that no AI-2 was produced by the mutant any longer.
We studied the effect of
lux S silencing of
E. coli Nissle in the DSS mouse model of acute colitis. The wild type
E. coli Nissle and the
E. coli Nissle::
lux S mutant behaved similar in respect to colonisation and stability, which is essential in animal experiments. Thus, AI-2 in
E. coli Nissle is not necessary for survival of the bacterium, otherwise we would not have been able to isolate the mutant bacterium after days in the mouse intestine. Measuring the body weight of the mice during the course of the experiments showed that the mice which were fed with the wild type
E. coli Nissle lost 15% of their initial weight. The other two groups of mice (
E. coli Nissle::
lux S; DSS control) lost on average only 5% of their weight. In addition, the group of mice fed with the
E. coli Nissle started to lose weight about two days early than the other two groups. In accordance with these data, these mice looked sicker: their fur was scrubby and their movement around the cages was rather erratic. We can speculate, that
E. coli Nissle is fitter than its mutant, which makes the bacterium more prone for potential translocation: while under healthy conditions there is minimal translocation of intestinal bacteria in mesenterial lymph nodes, under inflammatory conditions the epithelial barrier is broken down and intestinal bacteria and also
E. coli Nissle are translocated through the Peyer patches and the MLN[
33]. The length of the colon is proportional to the level of its inflammation: increasing levels of inflammation results in shortening of the colon. The colon of these mice, which received only DSS, were the shortest, while the colon of the mice, which were inoculated with
E. coli Nissle or its corresponding mutant were longer. To elucidate the effect of the oral infection of mice with
E. coli Nissle wild type or its corresponding
lux S mutant on the expression of cytokines, we choose four well characterized pro- and anti-inflammatory cytokines and one defensine. In mice, which were fed with the wild type
E. coli Nissle, the pro-inflammatory cytokine IFN-γ, was suppressed in the colon mucosa, while the anti-inflammatory cytokine IL-10 was suppressed by
E. coli Nissle::
lux S. The expression of the pro-inflammatory cytokines Il-6 and TNF-α was around 8 fold higher in mice infected with the
lux S mutant, than in mice of the other two groups. On the other hand the expression of mBD-1 was suppressed in the mice infected with the mutant, in comparison to the mice which were infected with
E. coli Nissle or were not infected at all. Taken together, we observed significant differences in the expression of cytokines between the different groups of mice. Eventhough the results were partly only moderately significant (p<0.1), we have generated evidence that AI-2 from
E. coli Nissle is indeed influencing the expression of cytokines and defensins and thus may influence the probiotic properties. It is well known that
E. coli Nissle is capable of inducing anti-inflammatory cytokines. Early on it was shown, that
E. coli Nissle is stimulating the epithelial defense in Caco-2 cells[
17]. The same authors showed that patients with Crohns disease, who have a mutation in NOD2 have a low level of expression of the defensins HD-5 and HD-6[
17]. The normal colon mucosa is producing human beta-defensin-1 (hBD-1; the homolog to mBD-1)[
34]. The functional importance of defensins was shown in elegant experiments with HD-5 expressing transgenic mice. These mice became resistant towards an infection with salmonellae[
35]. While pathogenic bacteria seem to suppress the production of defensins probably for self-defense[
36],
E. coli Nissle is protecting its host by defensin induction[
17,
34].
It has been shown in a number of studies that homoserine lactones (AI-1) regulate the expression of cytokines and virulence factors, for example in
Vibrio cholerae,
Pseudomonas aeruginosa or EHEC[
37‐
39]. However, the influence of AI-2 on the cytokine expression was discussed only in two reports so far[
40,
41]. Using microarrays the group showed the differential regulation of a number of genes involved in the complement pathway, regulation of cytokine expression and antigen presentation when infecting RAW264.7 macrophages with the wild type
Vibrio vulnificus or with the corresponding
lux S mutant.
Is QS also used by other probiotic bacteria? Studies were published only with bacteria of the genus
Lactobacillus. Early on it was shown that
Lactobacillus rhamnosum GG is communicating via AI-2 molecules[
42]. It was shown that the
lux S gene has a clear role in the acidic stress response; AI-2 activity increased by lowering the pH in a dose dependent manner[
43]. A second probiotic bacterium,
L. acidophilum strain La-5, showed that its supernatant is influencing the AI-2 concentration and the expression of virulence genes of the enterohemorrhagic
E. coli (EHEC) 0157:H17 in the gut, inhibiting its colonization. In addition it was shown, that the supernatant is reducing the attaching and effacing lesions in HeLa cells. Also a significant inhibition of bacterial adhesion in Hep-2 cells was observed. In elegantly designed mouse experiments, the authors used slow-scan CCD cameras to show reduced fecal shedding of luminescent EHEC constructs when
L. acidophilus was fed additionally[
44,
45]. Another
Lactobacillus species,
L. plantarum was used successfully to inhibit the pathogenic activity of
Pseudomonas aeruginosa. The AHL production, as well as the production of elastase and biofilm was inhibited by
L plantarum cultures and filtrates, but not by isolated, washed cells[
46]. In another set of experiments with
L. plantarum, it was shown that its LamBDCA quorum sensing system is responsible for the modulation of cytokine response in human PBMC[
47]. These studies showed that indeed QS is influencing the probiotic properties of the studied
Lactobacillus species. In the future, QS in other probiotic species should be studied, to hopefully confirm that QS is indeed influencing the probiotic properties.
Examining the histologic sections of the colon of the different groups of mice, we did not observe any significant difference between the groups. In contrast to the histopathology of a chronic inflammation,
E. coli Nissle has no influence in case of an acute DSS induced colitis. Here one can observe histologic patchy mucosal damage with the loss of crypts, followed by the acute transmural infiltration of inflammatory cells. No T- or B-cells are necessary[
48].
In this report we showed for the first time that the probiotic bacterium E. coli Nissle is producing AI-2 molecules. Here we show that AI-2 is affecting the regulation of cytokine expression in the DSS mouse model of acute colitis. In comparison to the E. coli Nissle lux S mutant, the mice which were infected with the wild type lost more weight, looked sicker regarding their fur and their lack of movement. On the other hand, mice, which were infected with the E. coli Nissle::lux S mutant showed a higher expression of pro inflammatory cytokines, but a reduced expression of the anti-inflammatory cytokine IL-10 or the mBD-1. Thus, it remains to be seen if AI-2 is influencing the probiotic properties of this important bacterium.
Competing interest
The authors declare that they have no competing interests.
Authors’contribution
CAJ designed and performed most of the experiments. He supervised the experiments which were not performed by him. He has written the manuscript. The work of SG focused on the in vitro experiments while C-JH performed parts of the RT PCR and gave valuable advice. JSF introduced CAJ into the DSS mouse model and was of great help concerning the mouse experiments. PA was examining the histological sections. GL, IBA, MG and PM gave valuable advice concerning the experiments and for the preparation of the manuscript. All authors read and approved the manuscript.