Since EPS are macromolecules that cannot cross the cell membrane through the common transport systems, they first must be hydrolyzed to be metabolized in the cells. Glycosidases (EC 3.2.1) are enzymes that hydrolyze the glycosidic bonds of polysaccharides, hence facilitating the release of the smaller sugars involved in nutrient acquisition. Therefore, we have determined whether an increased glycosidase activity could augment any prebiotic character of the 2-substituted (1,3)-β-d-glucan synthesized by P. parvulus 2.6. Therefore, the β-glycosidase (bgl, lp_3629) gene from L. plantarum strain WCFS1, which was previously characterized in response to abiotic stress [20], was cloned into the pGIZ906 vector and overexpressed in L. plantarum. Overproduction of Bgl in the recombinant strain (named L. plantarum WCFS1β-gal) was confirmed by SDS-PAGE. Figure 1 shows that this strain produced large amounts of a 53 kDa protein, corresponding to the theoretical molecular mass of Bgl. Acebrón et al. [21] demonstrated biochemically that the lp_3629 gene encodes a functional β-galactosidase (EC 3.2.1.23). However, the amino acid sequence of Bgl shows that it belongs to the glycosyl hydrolase family 1. Enzymes in this family have a broad substrate specificity [22]. Therefore, the glucosidic bonds of the P. parvulus 2.6 β-glucan could be hydrolyzed by Bgl.

Furthermore, we analyzed the potential prebiotic effect of EPS from bacterial origin on the survival of three strains belonging to the Lactobacillus genus: L. plantarum WCFS1, the isogenic recombinant L. plantarum WCFS1β-gal and Lactobacillus acidophilus strain NCFM. These microorganisms are known for their beneficial contribution to human health and are widely distributed in fermented food [2,23,24].

The hypothetical role of the 2-substituted (1,3)-β-glucan produced by P. parvulus 2.6 as a sugar source for bacterial metabolism was evaluated by testing for microbial growth in a chemically defined medium which has been recently described [25]. To confirm that a source of carbohydrate was essential for bacterial development, the medium was initially prepared omitting d-glucose. Under this condition, no growth was detected for any of the strains examined (data not shown), in agreement with Terrade and Mira de Orduña [26], who did not observe biomass yield in the same medium without d-ribose as a unique sugar substrate. When only d-glucose was added to the medium as a sugar source, both L. plantarum WCFS1 and L. plantarum WCFS1β-gal strains showed a very similar growth kinetics, reaching the stationary phase after approximately 24 h (OD₆₀₀ = 2.00) (Figure 2). In contrast, L. acidophilus exhibited a much lower growth at all the monitored times. The maximum cell concentration was reached after 24 h (OD₆₀₀ = 0.75), followed by a death-phase after 30 h (OD₆₀₀ = 0.65) (Figure 2), as even confirmed by plate count analysis (data not shown). Since the medium used in this work was optimized to analyze LAB isolated from wine [25], the reduced growth observed for L. acidophilus may be related to the composition of the growth media.

The exopolysaccharide used in this study was obtained from P. parvulus 2.6 and the yield of purified EPS was about 200 mg per liter of microbial culture, as previously reported by Garai-Ibabe et al. [14]. Therefore, due to their low availability, the experimental assays in presence of EPS were performed in microtitre plates. The concentration of viable cells, expressed as colony forming units (cfu), was determined at incubation times 0, 3, 6, 9, 24, 27, and 30 h (Figure 3). In the L. plantarum strains analyzed (Figure 3A,B), a significant difference of more than 1-log was observed when bacteria were grown in a medium supplemented with only d-glucose or with EPS as carbon source. In addition, a beneficial effect was also observed in the viability of the two strains when inoculated in medium containing d-glucose, supplemented with EPS. In particular, L. plantarum WCFS1β-gal showed an increased biomass in all phases when both substrates (d-glucose and EPS) were in the medium, reaching a maximum cell viability of 9.8 × 10⁹ cfu mL⁻¹ after 24 h, followed by a subsequent slight decrease (Figure 3B). A similar pattern of growth was observed in presence of only glucose, but a reduced viability was detected in the logarithmic and stationary phase, with a maximum concentration corresponding to 7.7 × 10⁹ cfu mL⁻¹ after 24 h (Figure 3B). In contrast, L. plantarum WCFS1 showed an identical overlapping of the viability during the first 24 h (approximately 6.0 × 10⁹ cfu mL⁻¹), independently from the presence or absence of EPS in addition to d-glucose. Interestingly, after the critical time of 24 h of incubation, bacteria where still able to grow in EPS-containing media (around 9.0 × 10⁹ cfu mL⁻¹, at 30 h), while their number gradually declined in media supplemented with only d-glucose (5.4 × 10⁹ cfu mL⁻¹, at 30 h) (Figure 3A).

The growth rate observed for L. acidophilus in the presence of only EPS was lower compared to the assay carried out when only d-glucose was provided, with plate counts of 1.3 and 2.3 × 10⁸ cfu mL⁻¹, as maximum viable cell concentration, respectively (Figure 3C). Surprisingly, the availability of both substrates stimulated the growth of L. acidophilus by delaying its entry into stationary phase. Indeed, after 24 h the cfu were 2.5-fold more copious than in the other conditions, and at the last experimental time a value of 8.7 × 10⁸ cfu mL⁻¹ was reached (Figure 3C).

The results reported in this study revealed a positive effect of the β-d-glucan on the growth of the investigated strains, suggesting that its use as a prebiotic may positively modulate the growth of probiotic organisms. In particular, the availability of both d-glucose and EPS extended the logarithmic phase for both L. plantarum WCFS1 and L. acidophilus strains. To explain this result, we assume that after the immediate consumption of d-glucose, bacteria could use EPS as a further carbon source. The release of hydrolytic enzymes under conditions of starvation, or the lysis of bacteria occurring at late exponential or stationary-phase, would enhance the activity of glycosidases in the extracellular environment. It is well known that the microbial species investigated in this study have a genetic makeup that allows them to use a variety of carbohydrate sources. Indeed, in silico analysis of the L. plantarum WCFS1 genome predicted 25 complete phosphoenolpyruvate sugar-transferase systems (PTS) enzyme II complexes, several incomplete complexes and 30 transporter systems that are involved in the transport of carbon sources [27]; likewise, the genome of L. acidophilus encodes a large variety of genes related to carbohydrate utilization, including 20 PTS and five transporters of the ABC family [24]. In addition, the substrate specificity cannot be predicted for some PTS and other carbon-uptake systems, and various sugar transport systems are known to import more than one substrate, thereby expanding the carbon transport capacity of these species even further [27].

In the medium supplemented with only EPS, we always observed an increase in the number of viable cells after 30 h of incubation, corresponding to approximately 1 log for L. acidophilus and L. plantarum WCFS1, while for the engineered L. plantarum WCFS1β-gal strain, viable bacteria numbers increased from 1.3 × 10⁶ cfu mL⁻¹ to 3.8 × 10⁸ cfu mL⁻¹. In addition an efficient co-metabolism of glucose and EPS was observed for this strain, which resulted in an increase of cell viability. This confirms that the overexpression of the lp_3629 gene encoding a functional β-glycosidase [21] could lead to a more marked hydrolysis of EPS, thus enhancing bacterial survival. Expression of similar enzymes might be a widespread strategy of the gut synbiotic bacteria. Indeed, in order to survive in the lower intestinal tract, bifidobacteria produce various kinds of exo- and endoglycosidases in surface-bound and/or extracellular forms, by which they can utilize diverse carbohydrates [28]. Recent genome sequence analysis of the probiotic Bifidobacterium longum NCC2705 revealed that more than 8.5% of the total predicted proteins were involved in the degradation of oligo- and polysaccharides, suggesting a superior ability of this organism to adapt to the poor nutritional conditions of the low intestine [29].

We also found that the growth of L. plantarum WCFS1 and L. acidophilus was favored by the simultaneous availability of both substrates, suggesting a synergistic effect of EPS and d-glucose in promoting bacterial viability. However, further studies are necessary to explain the biochemical and molecular bases of these observations.

Because EPS are known to be involved in biofilm formation [30], several authors have investigated in vitro the link between EPS production and the capability to colonize the intestinal environment [13,14]. However, such studies did not contemplate the influence that exogenous EPS could have on the adhesion level of other microbial species. Therefore, in this work we investigated the effect of the 2-substituted (1-3)-β-d-glucan on the adhesive ability to human intestinal epithelial cells of L. plantarum WCFS1, a strain that does not show any mucous production in culture medium supernatants. In the absence of the purified EPS, approximately 4% of the bacteria adhered to Caco-2 cells. This is consistent with previous work reporting a percentage of adhesion ranging from 2% to 10% for commercial probiotic strains [31]. However, when L. plantarum was incubated with 0.5% w/v of EPS, the percentage of bound bacteria increased approximately five-fold (Figure 4). It has already been demonstrated that prebiotics such as oligosaccharides can influence adhesion of specific strains [32]. Intriguingly, to the best of our knowledge, no global positive effects of these molecules on microbial adhesion potential has been detected [32–35].

As components of the bacterial cell surface, the EPS produced and secreted by LAB are likely to underlie inter-cellular recognition mechanisms and/or the process of biofilm formation. In this regard, β-d-glucans might promote the initial steps of adhesion to host intestinal cells. This finding raises considerable interest for its potential applications in the design of symbiotic food products. Indeed, we can speculate that addition of specific EPS in the food matrix that vehicles probiotics might enhance microbial adhesion to the intestinal epithelium, thus improving gut colonization by beneficial microorganisms. A persistent or at least transient colonization of the gut is a desirable feature of probiotics, as a close and prolonged interaction with the host allows them to accomplish beneficial effects such as immune function modulation and pathogen exclusion [36]. Moreover, as observed for other prebiotic NDO, β-d-glucans might also contrast pathogen adhesion to the intestinal mucosa and/or act as antioxidants to protect probiotics from the intestinal hostile environment [37]. Further research shall be undertaken to address these issues and test such hypothesis.

The bacterial strains used in this study were L. plantarum strain WCFS1 [27], its recombinant isogenic strain that overexpresses the bgl gene encoding a β-galactosidase (L. plantarum WCFS1β-gal) and the L. acidophilus strain NCFM [24,38].

A chemically defined medium was prepared as reported by Terrade et al. [25] but without d-ribose and modifying the final pH at 6.2. The growth kinetics was evaluated either in absence of any carbon source or supplementing the same medium with d-glucose (10 g·L⁻¹), EPS (0.05% w/v) or both. The recombinant strain was grown in media that included 10 μg·mL⁻¹ of erythromycin. L. plantarum strains and L. acidophilus were always incubated without shaking at the optimal temperature of 30 °C and 37 °C, respectively.

P. parvulus 2.6 [9] was isolated from ropy cider at the Department of Applied Chemistry, Faculty of Chemistry (University of the Basque Country UPV/EHU, San Sebastián, Spain).

Strains were kept in de Man Rogosa Sharpe (MRS) broth (Pronadisa, Madrid, Spain) supplemented with 20% (v/v) glycerol for long-term storage at −80 °C.

To obtain a pre-inocula for EPS production, P. parvulus 2.6 was grown on MRS broth, supplemented with 0.05% (w/v) L-cysteine hydrochloride (Merck, Darmstad, Germany) and 0.1% (w/v) Tween 80 (Pronadisa, Barcelona, Spain). For production of the EPS, P. parvulus 2.6 was grown in the semi-defined media MST [39], containing glucose (50 g·L⁻¹) and ethanol 4.9% (w/v), in a 3-L fermenter (Bioflo 110, New Brunswick Scientific, Enfield, CT), at 30 °C for 96 h. The pH was controlled at 5.2 with 5 N NaOH, the agitation was kept at 50–70 rpm to keep the fermentation broth homogeneous, and nitrogen gas (0.2 L·h⁻¹) was sparged through the headspace continuously to maintain anaerobic conditions.

Bacterial cells were removed from fermented media by centrifugation (16,000× g, 4 °C, 30 min) after 96 h of culture. The clear supernatant was collected and the EPS precipitated by adding three volumes of cold absolute ethanol, and maintained overnight at 4 °C. The precipitate was recovered by centrifugation at 14,000× g for 10 min at 4 °C, the resulting EPS pellet was dissolved in deionized water, and the EPS was recovered by three cycles of precipitation with ethanol. The final precipitate was dissolved in and dialyzed against deionized water, using a dialysis membrane (Medicell International, Ltd., London, U.K.) having a cut-off of 3.5 kDa, for 2–3 days (water changed twice each day). After dialysis, the precipitate was lyophilized.

For the experimental assay, media were prepared by adding appropriate volumes of a 100× concentrated stock solution of EPS to obtain the desired final concentration.

All the plasmid constructions were performed in Escherichia coli TG1 strain. The bgl gene was cloned in the pGIZ906 vector, a pLAB1301-derivative with a 0.352-kb insert containing the ldhL gene expression signals of L. plantarum WCFS1 [27,40]. The bgl gene was amplified from L. plantarum WCFS1 genomic DNA using forward (5′-AAAACTGCAGGAGTTCCGGAAGGCTTT-3′) and reverse (5′-GCTCTAGATCAAAACCCATTCCGTTCCCCA-3′) primers, harbouring PstiI and XbaI sites (underlined sequence), respectively. The PCR fragment was digested with PstI–XbaI and located between the NsiI–XbaI sites of pGIZ906 vector by ligation to generate a PldhL-bgl transcriptional fusion. The recombinant plasmid was electroporated into L. plantarum WCFS1 and transformants were identified on MRS plates containing erythromycin. Plasmid DNA preparation was performed following standard procedures. The absence of mutations in the plasmidic bgl gene was confirmed by DNA sequencing.

Cells of L. plantarum β-gal grown in 30 mL of MRS to middle exponential phase were sedimented by centrifugation (5000× g, 10 min, 4 °C) and washed twice with 0.1 M sodium phosphate buffer pH 7.0. Bacteria were resuspended in 600 μL of cold phosphate buffer containing 0.1 mM phenylmethylsulfonyl fluoride (PMSF) (Roche Applied Science, Indianapolis, IN), mixed with glass beads (212–300 μm Ø, Biospec Products, Bartlesville, OK) and mechanically disrupted in a Mini Beadbeater™ (Biospec Products) (four cycles of 1 min at max speed followed by 1 min on ice). After removal of cellular debris by centrifugation (13,000× g, 10 min, 4 °C), supernatants were aliquoted and stored at −20 °C. Quantification of total protein extracts was determined with a Bio-Rad Protein Assay (Biorad, Milan, Italy), according to the manufacturer’s instructions. Protein fractionation was carried out by sodium dodecylsulphate-polyacrylamide gel electrophoresis (SDS-PAGE) at 150 V for 1 h in a 12% (w/v) resolving and 4 % (w/v) stacking gel. A prestained SDS-PAGE standards, low range (Biorad) was used as protein marker. After PAGE, protein bands were revealed by staining with coomassie brilliant blue (0.2% CBB R-250).

Strains from glycerol stocks were inoculated into MRS medium and grown to middle exponential phase (OD₆₀₀ = 0.8 and for L. plantarum and OD₆₀₀ = 0.6 for L. acidophilus strains). These cultures were used as inocula (dilution 1:100 in fresh defined medium), after harvest by centrifugation (5000× g, 5 min), two washes and resuspension in the corresponding chemically defined medium. The kinetics of bacterial growth were spectrophotometrically monitored during 24 h in media supplemented or not with d-glucose. The analysis was performed in the wells of microtitre plates containing 200 μL of chemically defined medium in the presence of only glucose, only EPS or both. The growth of all strains was analyzed by determination of colony forming units (cfu) by plate count of samples incubated for 30 h at optimal temperature. Experiments were performed in triplicate.

Adhesion assays were performed using the Caco-2 epithelial cell line, originated from human colonic carcinoma, and kindly gifted by Dr. C. Lamacchia, University of Foggia. Caco-2 cells were grown as previously described [14,31,41] and supplemented with 2 mM l-glutamine (Sigma-Aldrich, St. Louis, MO). In post-confluent cultures, the viable cell number, as counted in a Burker chamber, was about 4.3 × 10⁴ cells per well and bacteria were added to achieve a final concentration of approximately 4.3 × 10⁸ cfu/mL (ratio 1000:1 bacteria to Caco-2 cells). Adhesion experiments were then performed according to Bove et al. [41]. Serial dilutions of the samples were plated onto MRS-agar plates to determine the number of cell-associated bacteria (viable counts) expressed as cfu. The adhesion percentage was calculated by comparing the number of cfu from washed wells (cell-bound bacteria) with those from control unwashed wells (unbound and bound bacteria). Experiments were performed in triplicate.

Significant differences between averages of duplicate measurements were evaluated by performing t-tests after analysis of variance at a confidence level of p = 0.05.