Correspondence: Borja Sánchez, Departamento de Microbiología y Bioquímica de Productos Lácteos, Instituto de Productos Lácteos de Asturias, Consejo Superior de Investigaciones Científicas (IPLA-CSIC), Ctra. Infiesto s/n, 33300, Villaviciosa, Asturias, Spain. Tel.: +34 985 892 131; fax: +34 985 892 233; e-mail: email@example.com
Bacillus cereus CH is a probiotic strain used in human nutrition whose adhesion to mucin is dependent on its surface-associated flagellin. Flagellins from the surface of several probiotic Bacillus strains were efficiently extracted with 5 M LiCl and identified by peptide fingerprinting. Based on the proteomic analysis, cloning of the gene coding for the flagellin of B. cereus CH was performed in the lactococcal vector pNZ8110 under the control of a nisin-inducible promoter. The resulting strain, Lactococcus lactis CH, produced a surface-associated flagellin after 6 h of induction with nisin. The recombinant Lactococcus strain adhered strongly to mucin-coated polystyrene plates, whilst inhibiting competitively the adhesion of the pathogens Escherichia coli LMG2092 and Salmonella enterica ssp. enterica LMG15860 to the same molecule. Strain CH could be used in further experimentation for the characterization of the molecular mechanism of action of this probiotic B. cereus CH flagellin.
Flagellins are the major constituents of bacterial flagella, long and narrow filaments present on the surface of certain bacterial groups; they rotate rhythmically, allowing cells to move (Kuwajima et al., 1986; Nuijten et al., 1990). In addition, flagella have a basal body and a hook, both responsible for up to 2% of the final flagellar mass (LaVallie & Stahl, 1989). Together, basal body and hook form a type III-like secretion system, by which flagellin monomers are specifically exported to the bacterial surface, where they auto-assemble and give the flagella its typical helicoid shape (Hueck, 1998). Flagellin is formed by four domains: D0, D1, D2 and D3. D0 and D1 are the N-terminal and C-terminal domains of the flagellin, respectively, being highly conserved among species. D2 and D3 are globular domains, very variable in terms of amino acid sequence, which present differences of up to 1000 residues, depending on the microorganism (Beatson et al., 2006). Whereas D0 and D1 domains are buried in the flagellar filament, D2 and D3 domains are surface exposed and represent the targets of antibody responses.
Both D0 and D1 domains, as highly conserved zones, represent special molecular patterns that are recognized by the human innate immune system through Toll-like receptor 5 (TLR5) and the ICE protease-activating factor (IPAF) (Gewirtz, 2006; Zamboni et al., 2006). Because of their differential subcellular locations in human epithelial cells, TLR5 respond to extracellular flagellin, whereas IPAF detects cytosolic flagellin (Miao et al., 2007). Flagellin signalization through TLR5 involves the secretion of proinflammatory cytokines such as interleukin-8 (IL-8) and tumour necrosis factor-α, always by means of nuclear factor-κB translocation (Means et al., 2003). In contrast, flagellin signalization through IPAF triggers a caspase-1 response, inducing IL-1β and IL-18 secretion, the latter leading respectively to local inflammation and natural-killer cell activation (Takeda et al., 1998; Harrison et al., 2008; Khan et al., 2008; Massis et al., 2008; Kinnebrew et al., 2010). Interestingly, recent data support the hypothesis that IPAF may be involved in the recognition of other bacterial molecules (Abdelaziz et al., 2010). The interaction of TLR5 and IPAF signalizations might thus detect the presence of cellular invasion by flagellated microorganisms. Although still unclear, some scientific evidence supports the potential involvement of other receptors such as Naip5 in flagellin recognition (Miao et al., 2007).
In a previous study, we have shown that a flagellin present on the surface of the probiotic strain Bacillus cereus CH was able to adhere to mucin and fibronectin (Sánchez et al., 2009a). It is known that flagellins are responsible for the adhesion to mucosal cells, their absence being related to a deficient binding of the flagellated microorganism (Ramarao & Lereclus, 2006). In the present work, the gene coding for the flagellin was cloned, and a recombinant Lactococcus lactis strain expressing the B. cereus CH flagellin obtained. Induced cultures of this strain were able to compete with Escherichia coli LMG2092 and Salmonella enterica ssp. enterica LMG15860 for the attachment to mucin.
Materials and methods
All the strains used in this study and their source of isolation or reference are listed in Table 1. Bacillus strains were routinely grown in Mueller–Hinton (MH) broth (Becton, Dickinson and Company, Le Pont de Claix, France) at 30 °C under constant agitation (150 r.p.m.) to avoid veil formation. Lactococcus lactis ssp. cremoris SMBI198, kindly provided by Bioneer A/S (Hørsholm, Denmark) and the recombinant strain L. lactis ssp. cremoris CH were grown at 30 °C in M17 medium (Becton, Dickinson and Company), supplemented with 1% w/v glucose and 5 μg mL−1 of chloramphenicol for strain selection when needed. Lactococcus lactis ssp. cremoris CH cultures were induced for flagellin expression by addition of 33 ng mL−1 nisin A (Sigma) when cultures reached an A600 nm of 0.3. Escherichia coli LMG2092 and S. enterica ssp. enterica LMG15860 were grown overnight from stocks stored at −80 °C in brain-heart infusion broth (Becton, Dickinson and Company) at 37 °C in an anaerobic cabinet (Bactron Anaerobic/Environmental Chamber, Sheldon Manufacturing Inc., Cornelius, OR) in an atmosphere of 5% CO2, 5% H2, 90% N2. These cultures were used to inoculate fresh media (1% v/v) and the pathogens were collected at stationary phase of growth.
Table 1. Strains used in this study and protein yield of 5 M LiCl extractions as measured with the BCA protein assay kit (mean ± SD of three independent extractions)
Source of isolation/reference
Yields represent the total amount of protein extracted from a pellet coming from 150 mL of culture. ATCC, American Type Culture Collection; BCCM/LMG, Belgium Coordinated Collection of Microorganisms.
Flagellins were extracted from the surface of all Bacillus strains by cell treatment with 5 M LiCl. First, overnight precultures were used to inoculate 150 mL of fresh MH broth. Cells were collected at early stationary phase (around 18 h of culture) by centrifugation (5000 g, 10 min, 4 °C), and resuspended in 5 mL of 5 M LiCl in phosphate-buffered saline (PBS) (final pH 7). Protease inhibitors EDTA (Sigma-Aldrich Chimie S.a.r.l., Saint-Quentin Fallavier, France) and phenylmethylsulphonyl fluoride (PMSF, Sigma-Aldrich) were added at final concentrations of 5 and 1 mM, respectively. Suspensions were kept at 37 °C for 30 min under gentle agitation, and cells were removed by centrifugation (5000 g, 10 min, 4 °C). Supernatants were recovered and filtered to avoid the presence of vegetative cells (cellulose acetate filters, 0.45-μm pore size, Sartorius AG, Goettingen, Germany) and extensively dialyzed against mQ water supplemented with 5 mM EDTA (dialysis tubing, cut-off=7000 Da, Medicell International Ltd, London, UK). Protein concentration was measured with the BCA Protein Assay kit (Pierce, Rockford, IL) according to the manufacturer's instructions. Aliquots of 40 μg of protein were loaded onto sodium dodecyl sulphate (SDS) polyacrylamide gels using a final polyacrylamide concentration of 10% or 12.5% w/v (Laemmli 1970). Proteins were resolved at a constant voltage of 170 V and visualized by GelCode Blue staining (Pierce). Selected bands were excised from gels and digested with trypsin using standard protocols, the resulting peptide mixture being analysed by tandem MS (MS/MS). Data were acquired using a MALDI Q-Tof Premier mass spectrometer (Waters, Manchester, UK), with α-cyano-4-hydroxy-cinnamic acid (Sigma-Aldrich) used as a matrix (3.6 mg mL−1 solution in 50% acetonitrile in 0.1% v/v aqueous trifluoroacetic acid TFA). Monoisotopic masses were corrected using the pseudomolecular ion of Glu-fibrinopeptide as a lock mass (1570.6774 Da).
Proteins were identified using the MS/MS search module from mascot software (http://www.matrixscience.com) against the nonredundant protein NCBI database, using the monoisotopic masses derived from trypsinolysis. The following parameters were used: peptide charge +1, peptide tolerance ± 0.1 Da, MS/MS tolerance ± 0.1 Da and one missed cleavage allowed for trypsin. Gels were repeated three times from independent cultures.
Cloning of B. cereus CH flagellin in L. lactis ssp. cremoris NZ9000
Flagellin was amplified from B. cereus CH total DNA, which was extracted using the DNeasy Blood and Tissue Kit following the manufacturer's instructions (Qiagen S.A., Courtaboeuf, France), with the primers FBC-Dir: 5′-GGGGCGCCGGCATGGATTTTTTCGCATATTAC-3′ and FBC-Rev: 5′-CGGGGGCCGGCCTATTGTAATAATTTAGAAAC-3′, in which NaeI sites are shown underlined. The relevant sequence was cloned into the blunt-end NaeI site of a pNZ8048-based lactococcal vector denominated pNZ8110, under the control of the nisin A-inducible promoter PnisA (de Ruyter et al., 1996). In addition, pNZ8110 carries just after PnisA an in-frame sequence coding from the signal peptide of the lactococcal protein Usp45, which allowed flagellin secretion (van Asseldonk et al., 1990). The resulting plasmid, denominated pNZ8110-CH, was used to transform L. lactis ssp. cremoris NZ9000 by electroporation. This strain carries chromosomal nisRK genes needed for the nisin-induced activation of PnisA. The plasmid insert was sequenced to be sure that no undesirable mutations were introduced (GenBank accession HQ262412). Because of instability of the recombinant flagellin, the plasmid was transferred to L. lactis ssp. cremoris SMBI198, which are derived from strain NZ9000 by a deletion in the chromosomal htrA gene (Poquet et al., 2000; Rigoulay et al., 2004). The resulting strains, L. lactis ssp. cremoris CH, produced exclusively a surface-associated recombinant flagellin.
Flagellin production and identification
Flagellin production was induced in L. lactis ssp. cremoris CH cultures at an A600 nm of 0.3 using a concentration of 33 ng mL−1 nisin. For flagellin subcellular localization, surface-associated and secreted proteins were obtained 6 h after induction as already described (Sánchez et al., 2009b, c). These extracts were analysed by SDS-polyacrylamide gel electrophoresis (PAGE) and differential bands, as deduced after comparison with uninduced cultures, and were excised from gels and identified by MS as described above.
Adhesion to mucin and competition assays
Adhesion of L. lactis ssp. cremoris CH, L. lactis SMBI-pNZ8110, E. coli LMG2092 and S. enterica ssp. enterica LMG15860 to mucin was performed in Immuno 96 MicroWell™ plates (Nunc, Roskilde, Denmark) as described before (Tallon et al., 2007). Bacteria from overnight cultures were collected by centrifugation (10 000 g for 5 min at 4 °C), washed twice and resuspended in PBS to an A600 nm of 0.7, being CFU (CFU mL−1) determined by plate count. Cellular suspensions containing 107 CFU mL−1 were incubated with 100 μmol L−1 carboxyfluorescein diacetate (CFDA) (Molecular Probes, OR), at 37 °C for 30 min as already described (Laparra & Sanz, 2009). Suspensions were washed twice and resuspended in the same volume of PBS. Volumes of 300 μL of CFDA-labelled suspensions were loaded onto mucin-coated 96-well plates and incubated at 37 °C for 1 h. After the incubation period, the media were aspired with a micropipette and wells were washed three times with 300 μL PBS. Then, 300 μL of a solution containing 1% w/v SDS in 0.1 N NaOH were added to wells and incubated at 37 °C for 1 h. Finally, the well contents were homogenized and transferred to black 96-well plates (Nunc), suitable for fluorescence scanning. The fluorescence was read in a Cary Eclipse Fluorescence Spectrophotometer (Varian, Palo Alto, CA) at λex 485 nm and λem 538 nm. Negative controls without bacteria were used to calculate the unspecific CFDA adsorption to the wells.
Adhesion was expressed as the percentage of fluorescence recovered after binding to mucin corrected by the fluorescence of the bacterial suspension added to the wells. Each assay was performed in duplicate, and conducted in three independent experiments.
For competition assays, 107 CFU mL−1 CFDA-labelled E. coli LMG2092 and S. enterica ssp. enterica LMG15860 were submitted to adhesion assays in the presence of 107 or 108 CFU mL−1 of nisin-induced L. lactis CH cultures.
Results and discussion
In a previous work, a flagellin produced by the probiotic B. cereus CH strain was shown to bind to mucin and fibronectin, two common attachment molecules of the human gastrointestinal surface (Sánchez et al., 2009a). In the present work, our aim was to characterize the phenotype of a recombinant L. lactis strain able to produce flagellin regarding its interaction with mucin, pathogens and eukaryotic cells. This was achieved by studying its ability to inhibit the adhesion of two well-known enteropathogens to mucin.
Bacillus flagellins are efficiently extracted from the cell surface by LiCl treatment
Five B. cereus and two B. subtilis strains were used in this study (Table 1). Five of the seven were isolated as the bacterial species identified on the labels of commercial probiotic or biocontrol products (Sánchez et al., 2009a). In addition, B. cereus ATCC 14579, the B. cereus type strain, and B. cereus KF1, a strain isolated from artisanal kefir produced in Vietnam, were included. For flagellin extraction, we set up a protocol involving the use of 5 M LiCl. LiCl is a chaotropic agent which destabilizes both hydrophobic/electrostatic interactions and hydrogen bonds, leading to the extraction of noncovalent surface-associated proteins (Sánchez et al., 2008). LiCl (5 M) solution was always supplemented with EDTA 5 mM and PMSF 1 mM, the absence of the latter leading to complete proteolysis of the flagellins (data not shown). As shown in Fig. 1a, incubations of 30 min at 4 °C led to the extraction of prominent bands with observed molecular masses of 28–60 kDa.
All bands were identified as Bacillus flagellins by MS, or by tandem MS (MS/MS) in cases that needed a more powerful analytic technique (Table 2). A single flagellin product was detected in all the cases except for B. cereus ATCC 14579 and B. cereusN, in which two bands identified as flagellins were observed. These flagellins could be extracted from SDS gels and renaturalized as described (Peant & LaPointe, 2004); they were able to bind mucin, as shown in a previous work (data not shown) (Sánchez et al., 2009a).
Table 2. MS and MS/MS data concerning flagellin identification
Lactococcus lactis CH produced a surface-associated and an extracellular flagellin
The size of the different flagellins varied from approximately 30 to 60 kDa (Fig. 1). In the B. cereus group, flagellins comprise a set of variable proteins in terms of sequence and molecular masses. This is due to the fact that the domains D2 and D3 are highly variable, producing potentially infinite variants (Beatson et al., 2006).
The gene coding from the B. cereus CH flagellin of 45 kDa was amplified using its total DNA as template. Primer sequences were deduced from protein sequence AAZ22698, to which B. cereus CH flagellin showed the higher degree of identity. This gene was cloned into the blunt-end NaeI site of pNZ8110, the resulting ligation being electroporated in L. lactis NZ9000. In this way we isolated a Lactococcus clone producing the recombinant flagellin, which was denominated L. lactis ssp. cremoris NZ9000-CH. Surface-associated protein and secreted protein profiles were obtained and, surprisingly, two bands of around 30 and 25 kDa were identified in the supernatant fraction of the induced cultures (data not shown). After MS analysis, these bands were properly identified as the B. cereus CH flagellin (Table 1). These results suggested that flagellin was proteolyzed on the Lactococcus surface, an aberrant form of the flagellin being released into the bacterial environment. It is known that the housekeeping protease HtrA from L. lactis, is targeted to the cell surface, where it degrades abnormal proteins, somehow monitoring the quality of surface proteins (Lyon & Caparon, 2004). We thought of the possibility that the recombinant protein was recognized as aberrant, being degraded by the action of this enzyme (or other surface-associated proteases) and released to the surrounding media. For that reason, we transferred the plasmid containing the CH flagellin gene, to L. lactis ssp. cremoris SMBI198, a strain derived from NZ9000, knocked out in the chromosomal htrA gene (Poquet et al., 2000; Rigoulay et al., 2004). The resulting strain produced only a surface-associated form of the recombinant flagellin (Fig. 1b). Interestingly, two bands showed homology with B. cereus flagellin, one of around 45 and the other one of around 63 kDa. It is known that, in certain cases, protein aggregates are difficult to disassociate, producing this kind of artefact in SDS-PAGE (Kankainen et al., 2009). This tendency to aggregation may lead to bacterial autoaggregation, and the physical–chemical dynamics of this process are currently under investigation. This could reflect the trend to auto-assembly that flagellins display in vivo (Hueck, 1998). In addition, it reinforces the role of HtrA as the major housekeeping protease on the L. lactis surface as, in the absence of it, the aggregated flagellin cannot be proteolyzed and thus shed into the bacterial surroundings.
Lactococcus lactis CH, expressing B. cereus CH flagellin, competitively inhibits the adhesion of E. coli and S. enterica to mucin
Lactococcuslactis ssp. cremoris CH showed a better ability to adhere to mucin when flagellin production was induced with nisin (Fig. 2). Adhesion of both L. lactis ssp. cremoris SMBI198 and L. lactis ssp. cremoris SMBI198 (pNZ8110) strains was similar to uninduced L. lactis ssp. cremoris CH cultures (data not shown). After gene induction, the adhesion was increased by a factor of 4.7. Nisin-induced L. lactis CH cultures inhibited the adhesion to mucin of the two enteropathogens used in this study in a dose-dependent manner (Fig. 2). A lower inhibition was also observed when uninduced L. lactis ssp. cremoris CH cultures were used. This is not surprising as L. lactis is also able to bind to mucin; an interference with enteropathogen adhesion to mucin is thus expected.
The adhesion data, corrected by the inhibitory effect observed for the uninduced L. lactis CH strain cultures, showed that L. lactis ssp. cremoris CH expressing the Bacillus flagellin was able to inhibit the adhesion of E. coli 4.4 times (1.8 times in the case of uninduced cultures), and 3.9 times in the case of S. enterica (1.6 times in the case of uninduced cultures), when the E. coli/L. lactis ratio was 1 : 10. In our previous work, we showed that adhesion of B. cereus CH to mucin could be explained, to a large extent, by the presence of a flagellin on its surface (Sánchez et al., 2009a).
In addition, preliminary results concerning the interaction of surface-associated protein extracts of the CH strain with peripheral blood mononuclear cells have suggested that the response driven by this flagellin may be different in terms of cytokine production, mainly by an increase of IL-6 and IL-1β (A. Suárez, pers. commun.). However, this statement deserves further experimentation. In this sense, it is known that Caco-2 cell monolayers have an atypical response to the flagellin from E. coli Nissle 1917, a probiotic strain, involving increases in the production of IL-8 (Schlee et al., 2007).
In conclusion, we have characterized a recombinant L. lactis strain expressing the B. cereus CH flagellin gene. This strain was able to inhibit the adhesion of two enteropathogens to mucin. Lactococcus lactis ssp. cremoris CH may be used as reference model for further studies addressed to the study of the molecular mechanism of action of this probiotic flagellin.
B.S. was the recipient of a Juan de la Cierva postdoctoral contract from the Spanish Ministerio de Ciencia e Innovación, and P.L. is the recipient of a postdoctoral contract from the project AGL2007-61805. Research in our group is supported by grant AGL2007-61805 from the Spanish Ministerio de Ciencia e Innovación.