Appl Microbiol Biotechnol (2012) 96:431–441 DOI 10.1007/s00253-012-4031-2
APPLIED GENETICS AND MOLECULAR BIOTECHNOLOGY
Probiotic features of Lactobacillus plantarum mutant strains Pasquale Bove & Anna Gallone & Pasquale Russo & Vittorio Capozzi & Marzia Albenzio & Giuseppe Spano & Daniela Fiocco
Received: 12 January 2012 / Revised: 8 March 2012 / Accepted: 15 March 2012 / Published online: 11 May 2012 # Springer-Verlag 2012
Electronic supplementary material The online version of this article (doi:10.1007/s00253-012-4031-2) contains supplementary material, which is available to authorized users.
Abiotic stresses associated to small intestine poorly affected bacterial viability. All the bacterial strains significantly adhered to Caco-2 cells, with the ΔctsR mutant strain exhibiting the highest adhesion. Induction of immunerelated genes resulted higher upon incubation with heatinactivated bacteria rather than with live ones. For specific genes, a differential transcriptional pattern was observed upon stimulation with different L. plantarum strains, evidencing a possible role of the knocked out bacterial genes in the modulation of host cell response. In particular, cells from Δhsp18.55 and ΔftsH mutants strongly triggered immune defence genes. Our study highlights the relevance of microbial genetic background in host–probiotic interaction and might contribute to identify candidate bacterial genes and molecules involved in probiosis.
P. Bove : D. Fiocco Department of Biomedical Sciences, University of Foggia, Via L. Pinto 1, 71122 Foggia, Italy
Keywords Bacterial adhesion . Immune modulation . Lactobacillus plantarum mutants . Probiotic . Mucosal barrier
Abstract In this study, the probiotic potential of Lactobacillus plantarum wild-type and derivative mutant strains was investigated. Bacterial survival was evaluated in an in vitro system, simulating the transit along the human oro-gastro-intestinal tract. Interaction with human gut epithelial cells was studied by assessing bacterial adhesive ability to Caco-2 cells and induction of genes involved in innate immunity. L. plantarum strains were resistant to the combined stress at the various steps of the simulated gastrointestinal tract. Major decreases in the viability of L. plantarum cells were observed mainly under drastic acidic conditions (pH≤2.0) of the gastric compartment.
D. Fiocco e-mail:
[email protected] A. Gallone Department of Basic Medical Sciences, University of Bari “Aldo Moro”, Bari, Italy P. Russo : V. Capozzi : G. Spano (*) Department of Food Sciences, University of Foggia, Via Napoli 25, 71122 Foggia, Italy e-mail:
[email protected] M. Albenzio Department of Production Sciences and Innovation in Mediterranean Agriculture and Food Systems (PrIME), University of Foggia, Foggia, Italy
Introduction The human gut harbours a complex ecosystem composed of 1013–14 microbes, including both transient and normal inhabitants (Guarner and Malagelada, 2003; Eckburg et al. 2005). Thanks to its intense metabolic activity, this extremely composite microbiota contributes to the assimilation of nutrients and production of vitamins; moreover, it protects the host from potentially pathogenic microbes and stimulates the development of the immune system (Hooper and Gordon 2001; O’Hara and Shanahan 2006). The microbial communities that inhabit the human gut constitute a potential source of probiotics, defined as ‘live microorganisms which when administered in adequate
432
amounts confer a health benefit on the host’ (FAO/WHO 2002). Most probiotics belong to the genera of Lactobacillus and Bifidobacterium, but Lactococcus, Streptococcus and Enterococcus genera, as well as some non-pathogenic strains of Escherichia, and certain yeast strains are also used as probiotics (Ouwehand et al. 2002; de Vrese and Schrezenmeir, 2008; Ohland and MacNaughton, 2010). Probiotics-based therapies can cure or prevent the risk of gastro-intestinal disorders associated to intestinal microbiota imbalances and gut barrier dysfunctions including infections, diarrhoea, food allergies and inflammatory bowel diseases (Isolauri 2001; Ouwehand et al. 2002). Modulation of the gut immune function seems to be one of the main mechanisms through which probiotics provide beneficial effects to the host. Indeed probiotic and commensal bacteria may influence the production of immune factors secreted by the gut-associated lymphoid tissue as well as by the intestinal epithelium (Borchers et al. 2009). Because antimicrobial peptides, mucous components, microbicidal enzymes and cytokines play key roles in the barrier and immune function of the intestinal mucosa, the expression of genes encoding such molecules is frequently analysed when assessing the microbial probiotic potential (Mack et al., 1999; Morita et al., 2002; Wehkamp et al. 2004). Documented clinical efficacy and safety are essential attributes of probiotics; other desirable features include the ability to reach, survive and colonize the human gut (Ouwehand et al. 2002; Isolauri et al. 2004; Kalliomäki et al. 2008). Resistance to the extreme conditions of the gastro-intestinal (GI) tract is an essential criterion for the selection of orally delivered probiotics. In vivo studies on model organisms are often too demanding in the initial screenings for potential probiotics and for suitable food matrices. Therefore, implementing in vitro models, which simulate the different conditions of the human GI tract, is a prerequisite to subsequent in vivo experiments (FAO/WHO 2002) and several recent studies have addressed this issue (Mainville et al. 2005; Fernández de Palencia et al. 2008; Lo Curto et al. 2011). Adhesion to the intestinal mucosa is another desirable feature of probiotics as it prolongs persistence in the intestinal tract, which is usually required for their beneficial effect (Isolauri et al. 2004). Due to difficulty in performing in vivo studies, preliminary studies of adherent strains are mainly based on in vitro adhesion assays that use tissue cultures of human colon carcinoma cell lines, such as Caco-2 and HT-29. Although their tumoral nature can somehow limit the extrapolation of results, these cells offer a relevant tool for in vitro screening. Caco-2 cells also provide an excellent system for immunological studies (Ou et al. 2009). Lactic acid bacteria (LAB) are natural members of the human GI microflora. Several LAB strains are considered beneficial to the host and have been selected for probiotic applications (Ljungh and Wadström 2006). Lactobacillus
Appl Microbiol Biotechnol (2012) 96:431–441
plantarum is a versatile LAB which has been traditionally used as a starter in food fermentation processes; like other LAB, L. plantarum is a normal inhabitant of the human GI tract (Ahrné et al. 1998) and some strains are used as ingredients of probiotic food (Molin 2001; de Vries et al. 2006). The probiotic and immunomodulating properties of lactobacilli seem closely linked to their cell envelope composition (Kleerebezem and Vaughan 2009; Lebeer et al. 2010). Indeed by studying a L. plantarum dlt cell wall mutant that synthesizes modified teichoic acids, Grangette et al. (2005) demonstrated that a specific cell surface biochemical feature may enhance its probiotic effect. Similar results were also observed with lipoteichoic-deficient strains of Lactobacillus acidophilus (Mohamadzadeh et al. 2011) and Lactobacillus rhamnosus GG (Claes et al. 2010). Even some bacterial stressrelated proteins can be relevant in interaction with the host and modulate its immune reaction as indicated by recent studies on recombinant lactobacilli that produce anti-oxidant enzymes (Watterlot et al. 2010; LeBlanc et al. 2011). We have recently generated strains of L. plantarum carrying null mutations in three genes encoding stress-related proteins (Fiocco et al. 2009; Capozzi et al. 2011), namely: CtsR—a transcriptional repressor of stress responsive genes (Derré et al. 1999), FtsH—a ubiquitous membrane-anchored protease (Ito and Akiyama 2005) and the molecular chaperone Hsp18.55 (Spano et al. 2005). Preliminary studies suggested that these defective strains might have distinctive cell surface features (Fiocco et al. 2010; Bove et al. 2012; Capozzi et al. 2011), therefore implying possible effects on their potential probiosis. In this study, we improved an in vitro system that reproduces the physiological events of ingestion and digestion of the human GI tract, including those occurring in the oral cavity, the stomach and the small intestine. In such a model, the survival potential of L. plantarum wild-type and related mutants was analysed and compared with the commercial probiotic strain of L. acidophilus LA5. Moreover, adhesion and immunomodulatory properties of L. plantarum strains were evaluated on Caco-2 cells.
Materials and methods Bacterial strains and culture conditions The bacterial strains used in this work were L. plantarum WCFS1 (Kleerebezem et al. 2003), related mutants ΔctsR, ΔftsH and Δhsp18.55 (Fiocco et al. 2009; Capozzi et al. 2011) and L. acidophilus LA5 (Chr. Hansen, Hörsholm Denmark). Lactobacilli were propagated on De Man Rogosa Sharpe (MRS, Oxoid, UK) broth (pH 6.2). The medium was supplemented with 0.1 % Tween and 0.05 % L-cysteine (Merck, Darmstad, Germany) for L. acidophilus
Appl Microbiol Biotechnol (2012) 96:431–441
433
culture. All incubations were performed at 28°C, except for L. acidophilus which was grown at 37°C. For preparation of milk–cell suspension, all strains were grown until they reached the mid-exponential phase (corresponding to a bacterial concentration between 109 and 1010 colony-forming units (CFU)/mL for L. plantarum and between 108 and 109 CFU/mL for L. acidophilus). Bacterial cells were sedimented by centrifugation (1,500g, 10 min) and resuspended in 10 % reconstituted skim milk powder (autoclaved 110°C, 15 min). For adhesion and stimulation experiments, mid-exponential phase bacteria were sedimented as above and resuspended in the appropriate volume of Dulbecco’s modified Eagle medium (DMEM, Sigma-Aldrich, St. Louis, MO, USA). Oro-gastro-intestinal transit assay The procedure used to mimic the GI transit is a modified version of a previous system (Fernández de Palencia et al. 2008) and is schematically represented in Fig. 1. The orogastrointestinal solutions were prepared fresh daily following previous protocols and data from human physiology literature (Marteau et al. 1997; Huang and Adams 2004). All incubation steps were performed at 37 °C and under shaking. To simulate the in vivo dilution of saliva, a gastric Fig. 1 Scheme of the in vitro system simulating the human oro-gastro-intestinal tract. Bacteria were resuspended in milk solution and subject to the sequential conditions indicated in the picture and described in the experimental section. Oral stress was mimicked by addition of a lysozyme-containing electrolyte solution (step 1, sample G1). Gastric stress was simulated by addition of pepsin and progressive pH reduction (steps 2–7, samples G2–G7). Samples of gastric-stressed bacteria (from steps 2–7) were adjusted to pH 6.5 and supplemented with bile salts and pancreatin to simulate intestinal stress (samples I1G2– I1G7 and I2G2–I2G7). Incubations were performed for the time indicated, at 37 °C and under shaking. Unstressed milk–bacterial suspension (sample G0) served as internal control
electrolytic solution (Marteau et al. 1997), containing 150 mg/L lysozyme (Sigma-Aldrich), was tenfold diluted in the milk– bacterial cell suspension. Gastric environment was reproduced by progressive acidification (addition of 1 M HCl) from the initial pH value of 6.0 to 5.0, 4.0, 3.0, 2.0 and 1.5 (samples G1–G7) and the suspension was sequentially incubated for 10, 10, 30, 30 and 10 min at each pH value, respectively. Intestinal stress was mimicked by treating for 1 h 5-mL aliquots, from each gastric pH value, as previously described (Fernández de Palencia et al. 2008) but reducing the bile salt concentration to 3 g/L. Then, 1-mL aliquots were withdrawn and analysed (samples I1G2–I1G7). Subsequently, residue samples were diluted (1:1) with intestinal electrolyte solution (Marteau et al. 1997) so to mimic the dilution and adsorption phenomena of the last tract of the small intestine. After 1 h of further incubation, samples (I2G2–I2G7) were recovered for analysis. All samples were immediately analysed. Appropriate dilutions from control and treated suspensions were plated on MRS agar plates and incubated. CFU were counted and percent survival was determined with respect to unstressed control. Caco-2 cell culture and adhesion test The Caco-2 epithelial cell line was employed for the adhesion experiments; these cells were used in their differentiated state
Milk-bacterial cell suspension
untreated control pH 6.0
G0
Oralstress
lysozyme electrolyte solution 5 min, pH 6.0
Intestinalstress 1 G1
pepsin pH 6.0 10 min
bile salts pancreatin pH 6.5
1:1 diluition
2 G2
pH 5.0 10 min
I1G2
I2G2
I1G3
I2G3
I1G4
I2G4
I1G5
I2G5
I1G6
I2G6
I1G7
I2G7
3 G3
Gastric stress
pH 4.0 10 min pH 3.0 30 min
pH 2.0 30 min
4 G4 5 G5 6 G6
pH 1.5 10 min
7 G7
434
to mimic small intestine mature enterocytes (Pinto et al. 1983). Caco-2 cells were grown in Men-Alpha Medium (GIBCO, Carlsbad, CA, USA) supplemented with 10 % (v/v) heatinactivated fetal bovine serum (Sigma-Aldrich), 2 mM L-glutamine (Sigma-Aldrich), 50 U/mL penicillin and 50 μg/mL streptomycin (GIBCO) at 37 °C in a humidified atmosphere containing 5 % CO2. Cells were seeded in 96-well tissue culture plates (Falcon Microtest, Becton Dickinson, NJ, USA) at 1.25×104 cells per well and grown as monolayers for 10 to 15 days to obtain differentiation (Fernández de Palencia et al. 2008). The medium (0.1 mL/well) was changed every 2 days; 24 h before an adhesion assay, an antibiotic-free medium was used. In post-confluent cultures, the viable cell number, as counted in a Burker chamber, was about 4.5×104 cells per well. To study the adhesion of each strain, Caco-2 cells were overlaid with bacteria resuspended in DMEM (0.1 mL/well) to a final concentration of approximately 5.0× 108 CFU/mL (ratio ≥1,000:1 bacteria to Caco-2 cells). Preliminary experiments indicated that such bacterial concentration was saturating in terms of adhesion. After 1 h of incubation at 37 °C under 5 % CO2 atmosphere, wells were washed three times with phosphate-buffered saline (PBS; pH 7.4) to remove unbound bacteria. Caco-2 cells and adherent bacteria were then detached by trypsin-EDTA 0.05 % treatment and resuspended in PBS (all from GIBCO). Serial dilutions of the samples were plated onto MRS agar plates to determine the number of cell-associated bacteria (viable counts) expressed as CFUs. CFU counts from control unwashed wells provided total bacterial load. Experiments were performed in triplicate. Caco-2 cell stimulation assay For immune stimulation experiments, Caco-2 cells were seeded at a density of 1.4×104 cells per well in 24-well tissue-treated culture plates (Iwaki, Tokio, Japan). The culture medium was changed every 2 days. Post-confluent cells were incubated with serum- and antibiotic-free medium for at least 12 h before bacterial stimulation test in order to avoid any interference with immune gene expression and with bacterial viability. The viable cell number, as counted in a Burker chamber, was about 2×105 cells per well. Caco-2 cells were incubated with either live or heatinactivated (1 h at 65 °C) bacteria at a concentration of 5×108 CFU/mL (1 mL/well). RNA isolation, cDNA synthesis and transcript profiling Total RNA was isolated from untreated Caco-2 cells (control) and after 1, 3 and 5 h of bacterial stimulation. Cells were washed with PBS and harvested with TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the supplier’s protocol. RNA integrity and concentration were determined by
Appl Microbiol Biotechnol (2012) 96:431–441
gel electrophoresis and spectrophotometric analysis. One microgram of total RNA was reverse-transcribed using QuantiTect Reverse Transcription kit (Qiagen, Valencia, CA, USA), which includes a DNase I treatment to remove genomic DNA. Absence of DNA contamination was confirmed by real-time PCR on corresponding DNase I-treated nonretrotranscribed RNAs. The transcriptional level of genes encoding interleukin-6 (IL-6), interleukin-8 (IL-8), macrophage inflammatory protein 3α (MIP-3α), human β-defensin-2 (HBD-2), lysozyme (LYZ) and mucin-2 (MUC-2) was analysed by quantitative real-time PCR (ABI 7300; Applied Biosystems, Foster City, CA, USA) using SYBR green I detection. Each reaction mixture, containing 5 μL of 20-fold diluted cDNA, 10 μL of QuantiFast SYBR Green PCR Master Mix (Qiagen) and 100 nM of each sense and antisense primer (Table S1 of “Electronic supplementary material”), was subject to amplification as previously described (Fiocco et al. 2009). Fluorescence data were analysed by applying the ΔΔCt method (Livak and Schmittgen 2001). Untreated Caco-2 cells corresponded to the calibrator condition. The expression of two potential housekeeping genes, encoding glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and β-actin, was monitored. Both of them exhibited steady expression; therefore, the GAPDH level was arbitrarily chosen to normalize the expression of target genes. Statistics Statistical analysis was performed using two-tailed, nonpaired Student’s t-test. Any P-value
