Inhibition of binding of Helicobacter pylori to the glycolipid receptors by probiotic Lactobacillus reuteri

Abstract

We examined the competition of binding of Lactobacillus reuteri and Helicobacter pylori to gangliotetraosylceramide (asialo-GM1) and sulfatide which are putative glycolipid receptor molecules of H. pylori, and identified a possible sulfatide-binding protein of the L. reuteri strain. Among nine L. reuteri strains, two (JCM1081 and TM105) were shown to bind to asialo-GM1 and sulfatide, and to inhibit binding of H. pylori to both glycolipids by a thin layer chromatogram-overlay assay using biotin-labeled bacterial cells. The extract from the bacterial cells of strain TM105 with several detergents, including octyl β-d-glucopyranoside, retained binding to both glycolipids and also inhibited H. pylori binding, suggesting that a binding inhibitor(s) is associated with the bacterial cell surface. When the cell extract was applied to the agarose gel immobilized galactose 3-sulfate corresponding to the structure of sugar moieties of sulfatide, an approximately 47-kDa protein was found to bind to the gel. This observation strongly suggested that inhibition by selected L. reuteri strains help to prevent infection in an early stage of colonization in H. pylori and proposed that L. reuteri strains sharing glycolipid specificity with H. pylori have a potential as probiotics.

Probiotic, Inhibition, Glycolipid receptor, Lactobacillus reuteri, Helicobacter pylori

1 Introduction

Helicobacter pylori is a Gram-negative, spiral shaped, microaerophilic bacterium isolated from human gastric mucosa. This bacterium causes chronic antral gastritis, peptic ulcers and is associated with stomach cancer. Many factors involved in H. pylori virulence have been studied in detail, including urease, the VacA cytotoxin, CagA, the neutrophil-activating protein NapA and lipopolysaccharide [1].

Bacterial adhesion to the mucosal surface is an initial important step for colonization and infection. It was widely accepted that H. pylori adheres to receptors in the gastric epithelium by means of specific adhesins. Sialyl glycoconjugates, phosphatidylethanolamine and extracellular matrix proteins have been reported as putative receptors for H. pylori adhesion [2]. It was also demonstrated that H. pylori binds to glycolipids including gangliotetraosylceramide (asialo-GM1) and sulfatide by a thin layer chromatogram (TLC)-overlay assay [3,4] and that adhesion to a human gastric epithelial cell line, MKN-45 cells, was inhibited by an anti-sulfatide antibody [5]. Moreover, the recent study demonstrated that a high level of sulfatide expression in the gastric mucosa of Mongolian gerbils allows abundant colonization by H. pylori, resulting in the development of gastric lesions in this animal model [6]. These findings imply that sulfatide is the main receptor molecule and that other glycoconjugates such as asialo-GM1 also partially participated in the H. pylori adhesion.

Some of the Lactobacillus species are members of intestinal microbiota from birds and mammals. Since several intestinal Lactobacillus species are believed to play beneficial biological roles to hosts, they are widely used as probiotics. The most important effects include the protection against pathogens as well as various physiological effects [7]. It was recently reported that some Lactobacillus strains could inhibit the growth or attachment of H. pylori to human gastric epithelial cells in vitro and exhibit antagonistic activities against H. pylori in vivo [8–10]. Such inhibitory effects seem to be caused by antimicrobial substances or metabolic end products such as lactic acid as well as the induction of an immune response of host.

Lactobacillus reuteri is also considered a candidate for probiotics because some L. reuteri strains have several beneficial effects to host organisms including protective effects against pathogens such as Salmonella[11], Cryptosporidium[12] and rotavirus [13]. Our previous studies demonstrated that L. reuteri JCM1081 has a hemagglutination activity relating to a binding ability to glycolipids including asialo-GM1 and sulfatide [14]. Interestingly, this strain also showed anti-adhesion activity against food-born pathogens on human enterocyte-like Caco-2 cells [15]. We presumed that steric hindrance or competition for receptor molecules participated in the anti-adhesion effect because the JCM1081 strain did not produce antimicrobial substances against tested pathogens.

The aim of this study is to examine whether L. reuteri competes for the same glycolipids as H. pylori strains. For this purpose, we first screened the L. reuteri strains having a binding ability to asialo-GM1 and sulfatide using the TLC-overlay assay. Further, we identified a putative anti-adhesion protein from L. reuteri.

2 Materials and methods

2.1 Bacterial strains and culture conditions

L. reuteri strains (JCM1112^T, JCM1081, JCM1084 JCM2762, JCM2763 and JCM2764) were obtained from the Japan Collection of Microorganisms (Wako, Japan). Strains TM105, TM128 and TM135 were isolated from intestinal contents of chickens in our laboratory. The species of these isolates were identified by the DNA–DNA hybridization method as described by Ezaki et al. [16] using JCM1112^T as a type strain. All strains showed a high homology (more than 80%) with the type strain. Bacteria were propagated aerobically in MRS broth (Difco, Detroit, MI, USA) at 37°C for 18 h. H. pylori (ATCC43504) was obtained from the American Type Culture Collection (Rockville, MD, USA). The bacterium was grown on Brucella agar containing 10% fetal calf serum under microaerophilic conditions (5% O₂, 10% CO₂, 85% N₂) at 37°C. The bacteria were stored at −85°C until use in the study.

2.2 Biotin labeling of bacterial cells

Biotin-labeled bacterial cells were prepared essentially as described by Turner et al. [17]. After bacterial cells were washed three times with phosphate buffered saline (PBS, pH 7.2), they were resuspended in 1 ml of PBS (approximately 1×10¹⁰ cells). Sulfo-N-hydrosuccinimide (NHS)-biotin (Pierce, Rockford, IL, USA) was then added (4 µl of a 25 mg ml⁻¹ solution), and the cells were incubated at room temperature for 2 min. Unbound sulfo-NHS-biotin was removed by washing four times with PBS.

2.3 TLC-overlay assay

Asialo-GM1 and sulfatide were obtained from Sigma Chemical Co. (St. Louis, MO, USA). TLC-overlay assays were carried out as described previously with some modifications [14]. TLC was performed on silica gel-precoated plates (Polygram, Sil G plastic plate; Macherey-Nagel Co., Düran, Germany). The plates were developed in chloroform–methanol–0.25% KCl (50:40:10), and then dried. For visualization of glycolipids, the plate was sprayed with orcinol reagent (500 mg orcinol in 3 mol l⁻¹ sulfuric solution) and incubated at 100°C for color development. For binding assay, the plate was blocked with PBS containing 1% (w/v) gelatin and 0.02% (w/v) sodium azide for 2 h. After being washed three times with 1% (w/v) bovine serum albumin (BSA)-PBS with gentle shaking (50 rpm) for 10 min, the plate was overlaid with biotin-labeled bacterial cells (5×10⁸ cells ml⁻) in BSA–PBS with gentle shaking for 2 h at 37°C. The suspension was removed by aspiration and the plate was washed three times with BSA–PBS. The plate was then dipped into BSA–PBS for 1 h, washed three times with BSA–PBS and then incubated for 2 h with alkaline phosphatase-conjugated streptavidin (Roche Diagnostics GmbH, Mannheim, Germany) diluted to 1/2000 with BSA–PBS. Thorough washing with BSA–PBS followed by PBS preceded incubation with the enzyme substrate solution kit (Bio-Rad Laboratories, Hercules, CA, USA). The binding intensity was estimated by densitometric analysis of stained TLC plates using an imagescanner and a computer program (NIH Image).

For the inhibition assay, after developing glycolipids on the plates, the plates were incubated with the suspension containing non-labeled bacteria or inhibitors at the indicated concentration for 1 h, and then washed three times with PBS before the addition of labeled bacteria. Experiments were performed in duplicate in three separate experiments.

2.4 Extraction of glycolipid-binding proteins

Extraction of glycolipid-binding proteins was made from 200-ml cultures. The cells were harvested and a part of each cell was used to label with sulfo-NHS-biotin as described above. Then, intact and biotin-labeled cells were suspended in 10 ml of 0.5% (w/v) octyl β-d-glucopyranoside (OGP) (Sigma), 5 mmol l⁻¹ Nonidet P40, 2 mol l⁻¹ guanidine hydrochloride, 3 mol l⁻¹ lithium chloride, or distilled water and gently stirred at 4°C for 2 h. After centrifugation (10 000×g, 10 min), the supernatant was recovered and dialyzed against 20 mmol l⁻¹ phosphate buffer (pH 7.0) containing 150 mmol l⁻¹ NaCl. To check the binding ability of the extracted samples containing biotin-labeled components to glycolipids, TLC-overlay assays were performed as described above. The specific binding activity of each sample was defined as the binding intensity per mg protein. Protein was measured using the BCA protein assay kit (Pierce).

2.5 Identification of proteins binding to sugar moieties of sulfatide

Biotinylated galactose 3-sulfate probe (Seikagaku Industries Co., Ltd., Tokyo, Japan) in 100 µl of PBS was mixed with 50 µl of streptavidin agarose gels (ImmunoPure Immobilized Streptavidin, Pierce) and then incubated for 2 h at 4°C. After centrifugation, the gel was washed three times with PBS followed by washing with 20 mmol l⁻¹ phosphate buffer containing 150 mmol l⁻¹ NaCl and 0.01% Triton X-100 (binding buffer). The non-labeled OGP extracts (100 µl) from intact cells were added to the gel immobilized biotinylated galactose 3-sulfate probe and the mixture was incubated for 2 h at 4°C under a rotating head over tail. The agarose gel was harvested by centrifugation (10 000×g, 2 min), washed three times with the binding buffer, and once with PBS. The binding protein was eluted by heating the agarose gel immobilized biotinylated galactose 3-sulfate probe in 50 µl of 2× concentrated reducing SDS sample buffer (100 mM Tris–HCl, pH 6.8, 4% SDS, 10% glycerol, 0.02% bromophenol blue, 200 mmol l⁻¹ dithiothreitol) at 95°C for 5 min. After centrifugation, the supernatant was subjected to SDS–PAGE analysis using 10% acrylamide gel [18]. For N-terminal sequence analysis, the proteins on the gel were transferred to a polyvinylidene difluoride membrane followed by staining with Coomassie brilliant blue. The target band was excised and subjected to an Applied Biosystems 491 gas phase sequencer (PE Applied Biosystems, CA, USA).

3 Results

3.1 Binding of L. reuteri to glycolipids

We examined the binding ability of several L. reuteri strains to asialo-GM1 and sulfatide with biotin-labeled bacterial cells (5×10⁸ cells ml⁻¹). As shown in Table 1, binding to the glycolipids was strain specific. Strains JCM1081 and TM105 bound to both asialo-GM1 and sulfatide, strains JCM1112^T and TM128 bound only to asialo-GM1, while no binding was observed with strains JCM1084, JCM2762, JCM2763, JCM2764 and TM131. The relative binding intensity of strains JCM1081 and TM105 to asialo-GM1 and sulfatide, respectively, was evaluated by densitometric analysis using NIH image. The binding intensity of both cells to asialo-GM1 was almost at a similar level while binding of TM105 to sulfatide was slightly higher than that of JCM1081, i.e. the binding of JCM1081 and TM105 to sulfatide was 12% and 28%, respectively, compared with that to asialo-GM1 (data not shown).

Table 1

Binding of L. reuteri strains to asialo-GM1 and sulfatide

Strains	Glycolipid binding		Source
	asialo-GM1	sulfatide
JCM1112T	+	−	Human feces
JCM1081	+	+	Chicken intestine
JCM1084	−	−	Chicken intestine
JCM2762	−	−	Fermented molasses
JCM2763	−	−	Fermented molasses
JCM2764	−	−	Fermented molasses
TM105	+	+	Chicken intestine
TM128	+	−	Chicken intestine
TM131	−	−	Chicken intestine

+ indicates binding, − no binding even at a level of 4 µg of each glycolipid.

Binding of each strain to glycolipids was estimated by TLC-overlay assay.

3.2 Inhibitory effect of L. reuteri cells on binding of H. pylori to glycolipids

We next examined the effect of strains JCM1081 and TM105 on the binding of biotin-labeled H. pylori to asialo-GM1 and sulfatide. As shown in Fig. 1, a complete reduction in binding to both glycolipids was obtained on addition of JCM1081 and TM105 cells (5×10⁸ cells ml⁻¹). Addition of 5×10⁷ cells per ml TM105 cells caused approximately 85% and 78% reduction in binding to asialo-GM1 and sulfatide, respectively (data not shown). On the contrary, no inhibitory effect was observed on the addition of strain JCM2764 cells at a concentration of 5×10⁸ cells ml⁻¹ (Fig. 1).

Figure 1

Effect of L. reuteri cells on binding of H. pylori to asialo-GM1 and sulfatide. After developing the glycolipids, the TLC plates were preincubated with (B) or without (A) L. reuteri cells (5×10⁸ cells ml⁻¹) followed by washing, and then biotin-labeled H. pylori cells (5×10⁸ cells ml⁻¹) were added to the plates. Color development was performed as described in the text.

3.3 Inhibitory effect of cell surface extract on the binding of H. pylori to glycolipids

To extract glycolipid-binding proteins, TM105 cells were treated with various detergents and then the extracts were subjected to TLC-overlay assays. The yields were estimated as the specific activity (spec. act.) of the binding intensity to either asialo-GM1 or sulfatide per mg protein. The highest yield of the binding proteins was obtained using 0.5% OGP (spec. act. 453 for asialo-GM1 and 212 for sulfatide), indicating that the cell extract with OGP contains binding proteins for both glycolipids. In other extraction buffers including 5 mmol l⁻¹ Nonidet P40, 2 mol l⁻¹ guanidine hydrochloride, 3 mol l⁻¹ lithium chloride and distilled water, lower yields of binding proteins were obtained. The extract with all detergents used from JCM2764 did not show binding to sulfatide (data not shown).

We then examined the inhibitory effect of the cell extract with OGP on H. pylori binding. As shown in Fig. 2, a significant reduction was obtained in H. pylori binding following the addition of the extract. Complete inhibition was observed in binding to both glycolipids on the addition of extract at a concentration of 400 µg total protein ml⁻¹.

Figure 2

Effect of the cell extracts from L. reuteri TM105 cells on binding of H. pylori to asialo-GM1 (closed bars) and sulfatide (open bars). After developing the glycolipids, the TLC plates were preincubated with cell extracts (0, 40, 400 µg ml⁻¹) followed by washing, and then biotin-labeled H. pylori cells (5×10⁸ cells ml⁻¹) were added to the plates. Means and standard deviations of binding intensity determined by densitometric analysis of stained TLC plates in duplicate in three separate experiments are shown.

3.4 Identification and N-terminal amino acid sequence of the binding protein to sugar moieties of sulfatide in L. reuteri

To identify sulfatide-binding proteins of L. reuteri, cell extracts with OGP from JCM1081, TM105 or JCM2764 were applied to the agarose gel immobilized galactose 3-sulfate corresponding to the structure of sugar moieties of sulfatide. A peptide with a molecular mass of 47 kDa was eluted from the gels with the cell extracts of both JCM1081 and TM105, while no band was detected from the cell extract of JCM2764 (Fig. 3).

Figure 3

SDS–PAGE analysis of binding protein of L. reuteri to sugar moieties of sulfatide. The cell extract with octyl β-d-glucopyranoside was applied to the agarose gel immobilized galactose 3-sulfate. Binding proteins were eluted from the gel with reducing SDS sample buffer, and then subjected to SDS–PAGE. Lanes: 1, JCM1081; 2, TM105; 3, JCM2764.

The eluted proteins of JCM1081 and TM105 were separated by SDS–PAGE, transferred to a polyvinylidene membrane, and subjected to amino-terminal amino acid sequence analysis. Both sequences obtained were AEKEHYWETKPVNW.

4 Discussion

We demonstrated the binding of some L. reuteri strains to asialo-GM1 and/or sulfatide, but the binding specificities were different among the strains tested. It is of interest that the strains from either intestine or feces but not from fermented molasses exhibited binding to the glycolipids. Asialo-GM1 has been suggested to be important for the colonizing capacity of many pathogenic bacteria and non-pathogenic bacteria [19]. Also, sulfate-containing glycoconjugates have been found in the basolateral surfaces of cells in the small and large intestines and gastric epithelial cells [20]. Thus, the binding ability of L. reuteri strains to glycolipids may be responsible for the adhesion to gastrointestinal mucosa, although the biological function of binding to glycolipids is unknown. Similar glycolipid-binding abilities have been found in a L. casei strain [21] and several Bifidobacterium strains [22] among lactic acid bacteria.

Some probiotic strains including Lactobacillus and Bifidobacterium have been reported to inhibit the adhesion of enteropathogenic bacteria to intestinal cells [23]. One of the mechanisms of the anti-adhesion activity can be explained via (an) antimicrobial substance(s) in spent culture supernatant secreted by those bacteria. The other inhibitory mechanisms by probiotics are proposed through steric hindrance or through the effects of competitive inhibition for attachment sites on the targeted cells. It has been shown that L. acidophilus La1 inhibits adhesion of the enteropathogens to a human enterocyte Caco-2 cell by secreting (an) antimicrobial substance(s) [24,25]. Recently, this strain was also found to share carbohydrate-binding specificities with the enteropathogens, suggesting the possibility of competition for adhesion of the pathogens to the carbohydrate receptor molecules of the cells [26]. From these findings, it may be, at least partly, requisite that Lactobacillus strains for competitive inhibition share ‘receptor molecules’ with targeted pathogens. On the basis of such speculation, we demonstrated here that the selected L. reuteri strains inhibit the binding of H. pylori to both asialo-GM1 and sulfatide of the putative receptors. Fujiwara et al. [27] also found the Bifidobacterium strain secreting an asialo-GM1-binding protein in the culture supernatant which inhibits the binding of enterotoxigenic Escherichia coli to the same glycolipid.

The present results demonstrated that the fraction containing glycolipid-binding proteins could be extracted from L. reuteri cells with OGP as well as several detergents. We confirmed microscopically that detergents such as OGP and guanidine hydrochloride do not burst the cells (data not shown). The fraction was also slightly extracted even with water. In addition, sulfo-NHS-biotin, used in this study as a labeling reagent, does not pass through the bacterial cell membrane. Thus, it is strongly suggested that the glycolipid-binding protein is associated with the cell surface. A peptide with a molecular mass of 47 kDa was found from the cell extracts of both JCM1081 and TM105 strains as a putative sulfatide-binding protein (Fig. 3), and the amino-terminal amino acid sequence obtained was AEKEHYWETKPVNW. Roos and co-workers have demonstrated two molecular species of cell surface proteins of L. reuteri strains; one is a collagen-binding protein with a molecular mass of 29 kDa showing sequence similarities to the solute-binding component of the family of ABC transporters [28,29], and the other is a putative DEAD-box helicase with a molecular mass of 56 kDa mediating autoaggregation [30]. The putative sulfatide-binding protein found in this study was, however, different from both proteins in its amino-terminal amino acid sequence and in its molecular size.

This is the first study to show that L. reuteri possesses the cell surface protein that inhibits the binding of H. pylori to receptor glycolipids in vitro. H. pylori infection afflicts about 50% of the population in the developed countries and about 80% of the population in the developing world [31], and effective protection has not been proposed to date. A recent review has described the possibility of anti-adhesion drugs using receptor mimics such as carbohydrate analogues [32]. We here propose that probiotics including L. reuteri strains, which share glycolipid-binding specificity with H. pylori, might be an effective competitor to pathogens at the receptor site.

Acknowledgements

This study was partially supported by a Grant-in-Aid for Scientific Research (09876037 and 11760186) from the Ministry of Education, Science, and Culture of Japan, and by a Kitasato University Research Grant for Young Researchers (H10-12).