Tadao Saito, Laboratory of Animal Products Chemistry, Graduate School of Agricultural Science, Tohoku University, Tsutsumidori-Amamiyamachi 1-1, Aoba-ku, Sendai, Miyagi 981-8555, Japan. E-mail: email@example.com
Aims: To identify and characterize a new adhesin-like protein of probiotics that show specific adhesion to human blood group A and B antigens.
Methods and Results: Using the BIACORE assay, the adhesion of cell surface components obtained from four lactobacilli strains that adhered to blood group A and B antigens was tested. Their components showed a significant adhesion to A and B antigens when compared to the bovine serum albumin (BSA) control. The 1 mol l−1 GHCl fraction extracted from Lactobacillus mucosae ME-340 contained a 29-kDa band (Lam29) using SDS–PAGE. The N-terminal amino acid sequence and homology analysis showed that Lam29 was 90% similar to the substrate-binding protein of the ATP-binding cassette (ABC) transporter from Lactobacillus fermentum IFO 3956. The complete nucleotide sequence (858 bp) of Lam29 was determined and encoded a protein of 285 amino acid residues. Phylogenetic analysis and multiple sequence alignments indicated this protein may be related to the cysteine-binding transporter.
Conclusions: The adhesion of ME-340 strain to blood group A and B antigens was mediated by Lam29 that is a putative component of ABC transporter as an adhesin-like protein.
Significance and Impact of the Study: Lactobacillus mucosae ME-340 expressing Lam29 may be useful for competitive exclusion of pathogens via blood group antigen receptors in the human gastrointestinal mucosa and in the development of new probiotic foods.
Lactobacilli are one of the indigenous micro-organisms living in the mammalian gastrointestinal tract and have the ability to adhere to mucosal surfaces (Lindgren et al. 1992; Bongaerts and Severijnen 2001; Roos and Jonsson 2002; De Leeuw et al. 2006). The adhesion of some lactobacilli to the gastrointestinal mucosa or mucus is thought to be of importance in the host to promote the modulation of the intestinal immune system and to exert inhibitory effects against pathogenic bacteria (Bernet et al. 1994; De Ambrosini et al. 1998). In lactobacilli, several studies show bacterial cell surface proteins are involved in host–bacteria interactions. The collagen-binding S-layer protein (CbsA) of Lactobacillus crispatus (Sillanpääet al. 2000), the mucus-binding protein (Mub) of Lactobacillus reuteri (Roos and Jonsson 2002) and the binding of cell surface–associated glyceraldehyde-3-phosphate dehydrogenase (GAPDH) of Lactobacillus plantarum to human colonic mucin (HCM) (Kinoshita et al. 2008a) are reported. A component of ATP-binding cassette (ABC) transporter is also an important adhesion factor to gastrointestinal tissues in host. A collagen-binding protein of Lact. reuteri (Roos et al. 1996), a mucus adhesion promoting protein (MapA) of Lact. reuteri (Havenith et al. 2002; Miyoshi et al. 2006) and mucus/mucin-binding protein (32-Mmubp) of Lactobacillus fermentum (Macías-Rodríguez et al. 2009) are identified as components of the ABC transporter.
We previously found several Lactobacillus species adhered specifically to each human ABO blood group antigen present on HCM using the BIACORE assay (Uchida et al. 2006a,b). It is well known that pathogens such as the Norwalk virus and Helicobacter pylori also recognize blood group antigens (Hutson et al. 2002; Aspholm-Hurtig et al. 2004). Therefore, blood group antigens expressed in mucin may be essential for initial colonization and later proliferation, especially with nonmotile lactic acid bacteria. However, the adhesion molecules and mechanisms of Lactobacillus strains with affinity to blood group antigens are still largely unknown. Here, we investigated cell surface proteins of Lactobacillus strains that showed specific adhesion to blood group A (bgA) and B (bgB) antigens and identified a new adhesin-like protein of Lactobacillus mucosae ME-340.
Materials and methods
Bacterial strains and growth media
Four strains of lactic acid bacteria, Lact. mucosae ME-340, Lactobacillus casei ME-341, Lact. crispatus ME-342 and Lactobacillus gasseri OLL 2877, were obtained from the culture collection of Meiji Dairies Corporation (Tokyo, Japan). They were isolated from human faeces and were identified using the API 50 CHL system (BioMerieux, Marcy I’Etoile, France). All strains were previously isolated as strains that adhere to the bgB antigen except for ME-341 (Uchida et al. 2006a,b). Lact. plantarum LA 318 was used as the positive control. We previously showed the cell surface GADPH from LA 318 isolated from human transverse colon bound to bgA and bgB antigens (Kinoshita et al. 2008b). Bacterial strains were cultured twice before the experiments at 37°C for 16–24 h in Man Rogosa Sharpe (MRS) broth (Difco Laboratories, Detroit, MI, USA) using 2% (v/v) inoculums.
Preparation of bacterial cells and cell surface proteins
After cultivation, all bacterial cells were washed three times with distilled water to remove media components and then lyophilized. To prepare cell surface proteins from each strain, bacterial cells were washed three times with phosphate-buffered saline (PBS, pH 7·4), and the first PBS wash fraction was collected and retained. After washing, the cells were resuspended in 1 mol l−1 guanidine hydrochloride (GHCl) and incubated at 37°C for 2 h. The cells were removed by centrifugation (8500 g, 30 min, 4°C), and the supernatant (the GHCl-treated fraction), the pellet (bacterial cells) and the first PBS wash fraction were dialysed against distilled water at 4°C for 2 days and then lyophilized. The lyophilized bacterial cells were used for the BIACORE assay as the analyte, and the lyophilized GHCl-treated fraction and the lyophilized first PBS wash fraction were used for SDS–PAGE and the BIACORE assay.
ABO blood group antigen probes
The following artificially prepared neoglycoproteins were used as blood group antigen probes: blood group A-trisaccharide bovine serum albumin [GalNAcα1-3 (Fucα1-2) Galβ-BSA: BSA-A] (Calbiochem, San Diego, CA, USA) and B-trisaccharide BSA [Galα1-3 (Fucα1-2) Galβ-BSA: BSA-B] (Calbiochem).
BIACORE assay using surface plasmon resonance (SPR)
The interaction between the analytes (bacterial cells or cell surface proteins) and the ligands (bgA or bgB antigen probes) was evaluated using a biosensor BIACORE 1000 (GE Healthcare Bio-Sciences K.K., Tokyo, Japan). This quantitative method is based on the principle of SPR (Uchida et al. 2004). The apparatus can detect ‘real-time’ specific interactions between the two molecules, analytes and ligands, without labelling (a colour coupler or a second antibody) that maintains an intact molecular structure. bgA or bgB antigen probes were immobilized on a CM5 research grade censor chip (GE Healthcare Bio-Sciences K.K.) using the amine coupling reaction according to the manufacture’s instructions. To determine nonspecific binding, BSA (Waco Pure Chemical Industries, Ltd, Osaka, Japan) was used. The analyte (0·1 mg ml−1) was suspended in HBS-EP buffer (HEPES-buffered saline with EDTA and polysorbate 20: 0·01 mol l−1 HEPES, pH 7·4, 0·15 mol l−1 NaCl, 3 mmol l−1 EDTA, 0·005% surfactant P20; GE Healthcare Bio-Sciences K.K.) and eluted at a flow rate of 3 μl min−1 at 25°C for 5 min. After washing with HBS-EP buffer to remove unbound analyte, the resonance units (RU) were measured for 180 s after the cessation of sample addition. The ligand immobilized on the sensor chip was regenerated by elution using 0·5–3 mol l−1 GHCl solution at a flow rate of 3 μl min−1 for 2 min. A response of 1 RU represents a weight change of 1 pg mm−2 analyte bound to the sensor chip surface.
Cell surface proteins of each bacterial strain were analysed using SDS–PAGE analysis according to the method of Laemmli (Cleveland et al. 1977). The polyacrylamide slab gel was prepared using a discontinuous buffer system with a 4·5% stacking gel and a 12·5% separation gel (10 × 12 cm). Molecular marker proteins and samples were dissolved in SDS buffer (62·5 mmol l−1 Tris, 2% SDS, 5%β-mercaptoethanol, 0·001% bromophenol blue and 10% glycerol, pH 6·8), followed by heating at 99°C for 10 min. SDS–PAGE was performed using an X Cell SureLock™ Electrophoresis Cell (Invitrogen, Carlsbad, CA, USA) at a constant voltage of 125 V with a Power PAC 3000 (Bio-Rad, Hercules, CA, USA) in running buffer containing 0·025 mol l−1 Tris, 0·2 mol l−1 glycine and 0·1% SDS. Proteins were visualized with Coomassie brilliant blue (CBB) (Rapid CBB; Kanto Chemical Co., Inc., Tokyo, Japan). Molecular weight markers from 14 to 79 kDa (middle range) were obtained from Waco Pure Chemical Industries, Ltd.
N-terminal sequence analysis
After the SDS–PAGE analysis, cell surface proteins were electrophoretically transferred to polyvinylidene difluoride (PVDF) membranes (Millipore, Billerica, MA, USA) using a semi-dry blotting apparatus (SEMI-PHOR™; Hoefer Scientific, CA). The 29-kDa protein (referred later as Lam29) was stained with CBB and excised from the PVDF membrane. The N-terminal sequence was determined using a Procise ABI 473A sequencer (Applied Biosystems, Tokyo, Japan).
Cloning and DNA sequencing of the lam29 gene
The lam29 gene fragment was amplified from ME-340 chromosomal DNA using the forward primer (5′-ACTCCTACCTTTACGCTTGG-3′) and the reverse primer (5′-GAAGTACGGCTTGAGTTCCT-3′) based on the complete genome sequence of Lact. fermentum IFO 3956 (INSD accession no. YP 001843639) (Morita et al. 2008). Amplification was performed using the Takara Ex Taq™ polymerase (Takara Bio Inc., Shiga, Japan) with an iCycler™ Thermal Cycler (Bio-Rad) with the following conditions: extension for 5 min at 96°C followed by 30 cycles: 30 s at 96°C, 30 s at 56°C, 60 s at 72°C and a final extension for 3 min at 72°C before holding at 4°C. The PCR cloning with the primers yielded a 200-bp fragment. DNA sequencing of the PCR products was performed using a BigDye Terminater ver. 1.1 Cycle Sequencing Kit (Applied Biosystems). The electrophoresis was performed using an ABI PRISM 3130xl Genetic Analyzer (Applied Biosystems). The analysis was performed using Sequencing Analysis Software ver. 3.7 (Applied Biosystems). Based on the partially determined genome sequence, the complete sequence of lam29 gene was determined using the primer-walking sequencing method with a BigDye Terminator ver. 3.1 Cycle Sequencing Kit (Applied Biosystems) (Kieleczawa et al. 1992).
Sequence analysis of Lam29
The complete nucleotide sequence of Lam29 obtained by PCR was translated using the GENETYX-SV/RC ver. 9 (GENETYX CO., Tokyo, Japan). Blast search for the N-terminal or complete amino acid sequence similarity to Lam29 was performed in the GenBank at the National Center for Biotechnology Information (NCBI) (http://blast.ncbi.nlm.nih.gov/Blast.cgi/). Calculation of the relative molecular mass and a theoretical isoelectric point were obtained using the public tool at Expasy (http://au.expasy.org/tools/protparam.html/). For prediction of cleavage sites, SignalP 3.0 (http://www.cbs.dtu.dk/services/SignalP/) was used. To find the conserved protein family, BLAST and PFAM (http://pfam.sanger.ac.uk/) were performed. For phylogenetic analysis, protein sequences from the putative or proven arginine-, histidine-, lysine–arginine–ornithine (LAO)-, glutamine-, cystine- and cysteine-binding ABC transporter proteins of the SBP bac 3 family (PF00497) were collected from the publicly available databases. Multiple sequence alignments were performed using the program Clustal W (Thompson et al. 1994). Evolutionary distances were calculated using Kimura’s two-parameter model (Kimura 1980). A phylogenetic tree was constructed using the neighbour-joining method (Saitou and Nei 1987) employing 1000 bootstrap trials. The tree was drawn using the program TreeView. Multiple sequence alignments with a representative set of sequences that are highly homologous (>50% identity) to Lam29 were performed using the GENETYX-SV/RC ver. 9. To identify the conserved amino acid residues of the cysteine-binding proteins, the sequence of CjaA from Campylobacter jejuni was used (Müller et al. 2005). CjaA is the cysteine-binding protein of the ABC transporter determined using crystal structure analysis (Müller et al. 2005). Lactobacillus fermentum IFO 3956 (accession no. YP 001843639) (Morita et al. 2008), Lactobacillus johnsonii NCC 533 (NP 965162) (Pridmore et al. 2004), Lactobacillus acidophilus ATCC 4796 (ZP 04021767) (unpublished in the literature), Lactobacillus casei ATCC 334 (YP 796444) (Makarova et al. 2006), Bacillus coagulans 36D1 (ZP 04431427) (unpublished in the literature), Leuconostoc mesenteroides ssp. mesenteroides ATCC 8293 (YP 817646) (Makarova et al. 2006), Bifidobacterium adolescentis L2-32 (ZP 02028369) (unpublished in the literature), Streptococcus thermophilus LMD-9 (YP 820887) (Makarova et al. 2006) and Campylobacter jejuni ssp. jejuni NCTC 11168 (CAL35100) (Parkhill et al. 2000) sequences were obtained from public databases.
All experiments were performed in triplicate and reported as the mean ± SD.
For adhesion test of cell surface proteins using the BIACORE assay, statistical analysis of the data was performed using one-way analysis of variance (anova test) followed by the post hoc Dunnet test when the F value showed significant differences at P < 0·05. For adhesion tests between untreated and GHCl-extracted bacterial cells using the BIACORE assay, the data were analysed using the Student’s t-test. In every statistics, significant differences were inferred whenever the P-value was <0·05.
Cell surface proteins involved in the adhesion of Lactobacillus mucosae ME-340 to BSA-A and BSA-B
We previously found the adhesin-like protein was removed in the fraction after washing bacterial cells with PBS buffer before lyophilization (Kinoshita et al. 2008a). Therefore, bacterial cell surface proteins in both fractions, washed with PBS and treated with 1 mol l−1 GHCl solution, were tested for adhesion to bgA or bgB antigen probes using the BIACORE assay. In almost all of the bacteria, both in the PBS wash fractions and the GHCl extract fractions bound BSA-B and also BSA-A equally compared to BSA, the control (Fig. 1).
Cell surface proteins in both fractions (PBS and GHCl) were analysed using SDS–PAGE analysis (Fig. 2). Several bands in ME-341, ME-342 and OLL 2877 were observed to be strain specific. However, a single prominent 29-kDa band in the GHCl-treated fraction from ME-340 was obtained, in contrast to no band in the PBS wash fraction. To identify the new protein component related to the adherence to bgA and bgB antigens, we further studied the 29-kDa protein (Lam29). Untreated ME-340 cells adhered to BSA-A and BSA-B at 382·0 ± 21·5 RU (the highest RU: 405·8) and 445·8 ± 24·9 RU (the highest RU: 463·9), respectively, whereas a significant reduction was shown with GHCl-treated whole cells to BSA-A and BSA-B at 18·2 ± 5·1 RU (the highest RU: 22·9) and 30·4 ± 7·1 RU (the highest RU: 38·6), respectively (P < 0·001) (Fig. 3). Untreated cells showed no adhesion to BSA as a control (4·6 ± 2·2 RU).
Identification of the new cell surface protein with specific binding activity to bgA and bgB antigens using gene sequencing
After the primary band of Lam29 was electrically transferred to a PVDF membrane, we sequenced the N-terminal amino acids. Ten amino acid residues were determined: NH2-ADNQSSVSAI. Comparison of the sequence to the database showed the closest identity (90%) to an ABC transporter substrate-binding component of Lact. fermentum IFO 3956 (INSD accession no. YP 001843639) (Morita et al. 2008). Based on the complete genome sequence of IFO 3956, we determined a predicted primary protein of 858 bp and 285 amino acids using PCR cloning experiments. The N-terminal 35 amino acid residues had the typical basic–hydrophobic–polar pattern of the signal peptide that is a potential substrate for signal peptidases I (van Roosmalen et al. 2004). Actually, the N-terminal amino acid of the mature Lam29 was alanine at position 36, which was in accordance with the prediction of the cleavage site. The calculated molecular mass of the mature was 27930·9 and was similar to that observed using SDS–PAGE. The nucleotide sequence of lam29 was deposited in the DDBJ with accession no. AB458523.
Using a BLAST search, the putative protein sequence of Lam29 was 38–77% identical to substrate- and solute-binding proteins of ABC transporters from both Gram-positive (Lactobacillus, Streptococcus, Bifidobacterium, Leuconostoc, Bacillus, Clostridium, Alkaliphilus and Desulfitobacterium) and Gram-negative bacteria (Campylobacter, Helicobacter, Pseudomonas, Acinetobacter, Aeromonas and Fusobacterium). Especially, the putative protein sequence of Lam29 showed the highest identity (77%) to this protein from Lact. fermentum IFO 3956.
A search of a putative conserved protein family performed using Blast and PFAM showed that Lam29 belongs to the PBPb superfamily. Lam29 is similar to the bacterial extracellular solute-binding proteins, family 3 (SBP bac 3 family). This family is primarily comprised of the amino acid-binding proteins, although the major ligand specificity has not been determined experimentally (Tam and Saier 1993). To speculate on the putative function of Lam29, a phylogenetic analysis was performed using a large number of proteins that are putative or proven arginine-, histidine-, LAO-, glutamine-, cystine- and cysteine-binding proteins of the ABC transporter systems of the SBP bac 3 family (Fig. 4). Six well-defined groups can be seen in the resulting tree, and Lam29 is clearly closely related to the cysteine-binding proteins from both Gram-positive and Gram-negative bacteria.
Figure 5 shows the data for the multiple sequence alignments of the representative sequences that are highly similar (>50% identity) to Lam29 with the sequence of CjaA, the cysteine-binding protein of the ABC transporter, in Camp. jejuni (Müller et al. 2005). Although the protein sequence similarity to Lam29 was different in various strains, these proteins including Lam29 had the cysteine-binding pocket of ABC transporter conserved in CjaA (boxed line in Fig. 5). This suggested Lam29 from Lact. mucosae ME-340 may be a putative cysteine-binding ABC transporter.
We previously found some Lactobacillus species specifically adhered to the bgB antigen present on HCM (blood group B) using the BIACORE assay (Uchida et al. 2006a,b). However, a re-examination showed that all four strains adhered not only to the bgB antigen but also to the bgA antigen (data not shown). Here, we show the adhesion of cell surface proteins from these strains to both the bgA and bgB antigens (Fig. 1), suggesting these bacterial cells adhere to the blood group antigens using surface protein interactions. The reason why the previously used strains did not show the adherence to bgA antigen may be related to the difference in culture conditions of the bacterial cells and/or the kind of bgA antigen probes (BSA-conjugated neoglycoprotein or biotinyl polymer probe).
We previously reported the surface layer protein A (SlpA) of Lactobacillus brevis OLL 2772 adhered to the bgA antigen (Uchida et al. 2006b), and a cell surface GAPDH from Lact. plantarum LA 318 strongly recognized bgA and bgB antigens (Kinoshita et al. 2008b). However, the adhesion molecules and the precise mechanisms for other Lactobacillus species adhering to blood group antigens are poorly understood. Here, we investigated a new surface protein causing the adhesion. Lam29 appeared primarily in the GHCl-extracted sample from Lact. mucosae ME-340 using SDS–PAGE analysis (Fig. 2) and was identified as the substrate-binding protein of the ABC transporter using N-terminal amino acid sequencing. The adhesion test strongly suggested Lam29 was the ME-340 adhesin and recognized bgA and bgB antigens at the same time because the adhesion of whole cells treated with the 1 mol l−1 GHCl solution to both bgA and bgB antigens was significantly decreased compared to untreated cells (Fig. 3).
The calculated isoelectric point of mature Lam29 was 9·87, indicating that it is strongly cationic at neutral and acidic pH conditions. Therefore, Lam29 may be anchored using electrostatic interaction with acidic groups on the cell surface as Turner et al. hypothesized in the solute-binding protein of ABC transporter BspA of Lact. fermentum (Turner et al. 1997). Lam29 could be removed with 0·05 mol l−1 glycine–HCl at pH 1·6 that would neutralize the negative charge (data not shown).
Extracellular substrate- and solute-binding proteins of the ABC transporters are generally known to play a central role in chemoreception and transmembrane transport (Tam and Saier 1993). Further, in addition to this fundamental function, it is suggested that some of the substrate-binding proteins are involved in host–bacteria interactions and act as adhesin-like proteins. The adhesion of Mn2+-binding protein of Streptococcus pneumoniae (pneumococcal surface adhesin A, PsaA) to type II pneumocytes (Berry and Paton 1996), the binding of the glycerol-3-phosphate-binding protein of Brucella species to HeLa cells (Castañeda-Roldán et al. 2006) and the binding of oligopeptide/dipeptide-binding protein (BopA) of Bifidobacterium bifidum to Caco-2 cells (Guglielmetti et al. 2008) are reported. In Lact. reuteri and Lact. fermentum, phylogenetically similar to Lact. mucosae, it is also shown that the substrate-binding proteins of the ABC transporters are involved in the adhesion to the gastrointestinal mucus (Miyoshi et al. 2006; Macías-Rodríguez et al. 2009). Therefore, many pathogenic and nonpathogenic bacteria possess the substrate-binding protein of the ABC transporter that shows binding to multiple tissues in the host.
Based on the Blast and PFAM analysis, a phylogenetic tree was constructed using the sequence of the amino acid-binding proteins of the ABC transporter systems of the SBP bac 3 family (Fig. 4). This family is involved in active transport of polar amino acids and opines across the cytoplasmic membrane. The phylogenetic analysis showed that Lam29 is part of the cysteine-binding proteins of the ABC transporters from both Gram-positive and Gram-negative bacteria. In comparison with the reported crystal structure of the cysteine-binding ABC transporter, CjaA, in Camp. jejuni (Müller et al. 2005), Lam29 also had conserved amino acid residues for the substrate-binding pocket of the ABC cysteine transporter (boxed line in Fig. 5). This suggests Lam29 may have bifunctional properties for the lectin-like interaction to blood group antigens and the substrate-binding for cysteine uptake. Miyoshi et al. reported that MapA, a putative cysteine-binding protein of the ABC transporter, of Lact. reuteri 104R has the binding properties to sheep erythrocytes, and this was inhibited by sugars (Havenith et al. 2002; Miyoshi et al. 2006). In their report, however, the binding of MapA to the carbohydrate antigens was not determined. Therefore, our report is the first showing Lam29 of Lact. mucosae ME-340 binds to blood group antigens through the lectin-like function. We are now undertaking further studies to determine whether Lam29 has recognition site(s) for carbohydrate antigens.
Some pathogens recognize the blood group antigen receptors. It has been reported that the adhesin BabA of H. pylori (Aspholm-Hurtig et al. 2004), the capsid protein of Norwalk virus (Hutson et al. 2002; Choi et al. 2008) and the heat-labile enterotoxin of enterotoxigenic Escherichia coli (Holmner et al. 2007) show specific binding to ABO blood group antigens. This indicates the blood group dependence of these strains and thus the adhesion of pathogenic bacteria to blood group antigens in the human gastrointestinal tract may be one of the important factors in establishing infections. Because Lact. mucosae ME-340 adheres to the same receptors, it is suggested that this strain could competitively inhibit infections with these pathogens.
In this study, we determined Lam29 of Lact. mucosae ME-340 was an adhesin with specific affinity to bgA and bgB antigens which are the nonreducing terminal sugar chains of the HCM. As some pathogens recognize the same blood group antigen receptors, Lact. mucosae ME-340 may be useful for the competitive exclusion of pathogens and in the development of new probiotic foods.
This study was in-part supported by a Grant-in-Aid for Scientific Research (B) (2) (No. 19380151) and for Exploratory Research (No. 18658030) from the Japan Society for the Promotion of Science (JSPS) to Prof T. Saito. We appreciate the support by Ryoshoku Kenkyukai (Odawara, Japan) and a grant from the National Institute of Genetics Cooperative Research Program (2009-A44) to Prof. T. Saito.