Fermentability of a Novel Galacto-Oligosaccharide Mixture by Lactobacillus spp. and Bifidobacterium spp.

This study aimed to investigate the specific growth stimulation of certain desired intestinal bacteria by a novel galacto-oligosaccharide mixture, which was produced with a β-galactosidase from a potential probiotic Lactobacillus isolate that contained mainly oligosaccharides of β-1,3 and β-1,6 glycosidic linkages (termed Lb-GOS) using single-strain fermentations. The composition of this Lb-GOS mixture was 33.5% disaccharides, 60.5% trisaccharides, 4.8% tetrasaccharides, and 1.0% pentasaccharides with a negligible amount of monosaccharides, lactose, and lactobionic acid (0.3%). Eight Lactobacillus spp. strains and three Bifidobacterium spp. strains were used in single-strain fermentations to determine the fermentation activity scores of this Lb-GOS preparation compared to two commercially available prebiotic mixtures, 4′GOS-P and Vivinal GOS (V-GOS). The highest scores were obtained when L. reuteri Lb46 and the two Bifidobacterium strains, B. animalis subsp. lactis Bif1 and Bif3, were grown on these galacto-oligosaccharide mixtures. In addition, the Lb-GOS mixture was found to have higher fermentation activity scores; hence, it stimulated the growth of these probiotic strains more than 4′GOS-P and V-GOS, which may be attributed to the different glycosidic linkage types that are found in the Lb-GOS mixture compared to the other two commercial preparations. These findings suggested that the Lb-GOS mixture that is described in this work should be of interest for the formulations of new carbohydrate-based functional food ingredients.


Introduction
The human colonic microbiota is composed of more than 1000 different species [1]. Most of these species are bacteria, some of which have been related to the health and well-being of the host [2]. Among the beneficial gut bacteria, bifidobacteria and lactobacilli are numerically predominant, and are most frequently considered for health-promoting effects [3]. The concept of prebiotics was first introduced by Gibson and Roberfroid [4], was later revised by them [5,6], and has been

Effects of Various Galacto-oligosaccharides as the Main Carbohydrate Substrates on the Single-Strain Fermentations
Three strains of bifidobacteria, eight strains of lactobacilli, E. faecium En61, E. coli DSM 613, S. epidermis DSM 20044, K. oxytoca DSM 6673, and C. freundii DSM 30039 were grown on glucose (0.5% w/v), various oligosaccharides (0.5% w/v), and on their basal complex media without an added sugar. The optical densities of the cultures (OD600) were measured during the course of these cultivations, with the maximum values presented in Table 2. The Lactobacillus strains that were tested can be divided into two groups based on their ability to grow on galacto-oligosaccharides, glucose, and the blank medium. A group of five strains, L. reuteri Lb46, L. reuteri Lb21, L. acidophilus Lb19, L. acidophilus Lb71, and L. acidophilus Lb105, showed similar growth characteristics with good growth on all of the test substrates, while growth on the blank was low. These strains showed slightly better growth on glucose and Lb-GOS than that on the commercial GOS products (Table 2); in some

Effects of Various Galacto-oligosaccharides as the Main Carbohydrate Substrates on the Single-Strain Fermentations
Three strains of bifidobacteria, eight strains of lactobacilli, E. faecium En61, E. coli DSM 613, S. epidermis DSM 20044, K. oxytoca DSM 6673, and C. freundii DSM 30039 were grown on glucose (0.5% w/v), various oligosaccharides (0.5% w/v), and on their basal complex media without an added sugar. The optical densities of the cultures (OD 600 ) were measured during the course of these cultivations, with the maximum values presented in Table 2. The Lactobacillus strains that were tested can be divided into two groups based on their ability to grow on galacto-oligosaccharides, glucose, and the blank medium. A group of five strains, L. reuteri Lb46, L. reuteri Lb21, L. acidophilus Lb19, L. acidophilus Lb71, and L. acidophilus Lb105, showed similar growth characteristics with good growth on all of the test substrates, while growth on the blank was low. These strains showed slightly better growth on glucose and Lb-GOS than that on the commercial GOS products ( Table 2); in some instances, growth was even better on the oligosaccharide mixture than on glucose. Three other Lactobacillus strains, L. rhamnosus Lb29, L. paracasei subsp. paracasei Lb16, and L. paracasei subsp. paracasei Lb20, showed very good growth on glucose; in fact, it was the best growth on glucose out of all of the strains that were tested. However, they showed only moderately good growth on Lb-GOS and V-GOS, and slow growth on 4 GOS-P compared to that of the Lactobacillus strains in the first group. These three strains also showed the highest cell densities on carbohydrate-free basal media among the Lactobacillus strains tested. The three strains of Bifidobacterium showed comparable growth characteristics, with moderate to good growth on glucose and Lb-GOS, slow growth on V-GOS, and relatively poor growth on 4 GOS-P. E. faecium En61 showed the highest growth on Lb-GOS compared to the other test substrates. The strains of E. coli DSM 613, S. epidermis DSM 20044, K. oxytoca DSM 6673, and C. freundii DSM 30039 were grouped as 'enteric bacteria' in this study, and showed moderate growth when there was no C-source present in the medium, and slow growth in the presence of GOS. The concentration of the C-source was 0.5% w/v. V-GOS, Vivinal galacto-oligosaccharides; 4 GOS-P, β1→4 linked galacto-oligosaccharides; GOS, galacto-oligosaccharides produced by β-galactosidase from Lactobacillus sp. c A blank shows growth on the respective complex basal medium without an added sugar.
The maximum OD 600 values displayed in Table 2 were used for the calculation of the fermentation activity scores ( Figure 2) using Equation (1). The highest scores were for L. reuteri Lb46 and B. animalis subsp. lactis Bif1 grown on Lb-GOS (10.04 and 9.02, respectively), and for L. reuteri Lb46 paired with V-GOS and 4 GOS-P (9.00 and 8.65, respectively), followed by B. animalis subsp. lactis Bif3 grown on Lb-GOS and V-GOS (8.39 and 7.11, respectively). Low scores were found for the group of L. rhamnosus Lb29, L. paracasei subsp. paracasei Lb20, and L. paracasei subsp. paracasei Lb16, as well as for E. faecium, especially when grown on 4 GOS-P. In addition, the three Bifidobacterium strains grown on 4 GOS-P showed consistently lower fermentation activity scores compared with V-GOS and the novel Lb-GOS mixture.

L . p a r a c a s e i s u b s p . p a r a c a s e i L b 1 6 L . p a r a c a s e i s u b s p . p a r a c a s e i L b 2 0 B . a n im a li s s u b s p . la c t is B if 1 B . a n im a li s s u b s p . la c t is B if 3 B . lo n g u m B if 1 4 E . f a e c iu m E n 6 1
Fermentation activity score

Discussion
Prebiotics, which commonly are oligosaccharides, are metabolized selectively in the gastrointestinal tract by beneficial bacteria associated with health and well-being. These carbohydrates can thus positively modulate the colonic microbiota, which exerts an important influence on host health [5,8,9,22]. Different methods such as pure culture fermentations of single, selected strains [23][24][25][26] and in vitro fermentations of mixed bacterial populations, particularly fecal bacteria [19,27,28], have been used as preliminary screening tools for prebiotic activities, whereas in vivo fermentations of non-digestible carbohydrates in animals and human subjects can be used for evaluating the prebiotic effects of different oligosaccharide mixtures [5,29]. Pure culture fermentations are performed in appropriate basal media supplemented with the respective prebiotics, and the increase in cell numbers is quantified by the turbidimetry of the cultures or by viable cell count. A better model for investigating the interactions between the gut populations is pH-controlled batch cultures. Here, fermentation is again based on basal media, with the test carbohydrate being the sole fermentable substrate present, but the use of fecal bacterial populations allows for an investigation of the interactions, competition, and cross-feeding during growth on the selected substrate. Changes in the concentration of intestinal bacteria in feces are monitored using

Discussion
Prebiotics, which commonly are oligosaccharides, are metabolized selectively in the gastrointestinal tract by beneficial bacteria associated with health and well-being. These carbohydrates can thus positively modulate the colonic microbiota, which exerts an important influence on host health [5,8,9,22]. Different methods such as pure culture fermentations of single, selected strains [23][24][25][26] and in vitro fermentations of mixed bacterial populations, particularly fecal bacteria [19,27,28], have been used as preliminary screening tools for prebiotic activities, whereas in vivo fermentations of non-digestible carbohydrates in animals and human subjects can be used for evaluating the prebiotic effects of different oligosaccharide mixtures [5,29]. Pure culture fermentations are performed in appropriate basal media supplemented with the respective prebiotics, and the increase in cell numbers is quantified by the turbidimetry of the cultures or by viable cell count. A better model for investigating the interactions between the gut populations is pH-controlled batch cultures. Here, fermentation is again based on basal media, with the test carbohydrate being the sole fermentable substrate present, but the use of fecal bacterial populations allows for an investigation of the interactions, competition, and cross-feeding during growth on the selected substrate. Changes in the concentration of intestinal bacteria in feces are monitored using molecular techniques such as fluorescence in situ hybridization (FISH) or real-time PCR [25,30]. Alternatively, in vitro colonic models and the 13 C labeling of substrates can be used to study the prebiotic activity [31].
Comparative studies on different oligosaccharides are still limited. Results between studies are sometimes difficult to compare [32], and no overall conclusion concerning the prebiotic efficiency of different oligosaccharides or structure/function relationships have yet been found [32,33]. In a comparative in vitro study by Watson et al. [34], FOS, GOS, and lactulose were oligosaccharides with growth-promoting effects against lactobacilli and bifidobacteria, in which GOS possessed a superior potential to the majority of the observed strains over FOS or inulin. In another study, GOS, isomalto-oligosaccharides, lactulose, and FOS were found to be metabolized well by all of the tested bifidobacterial strains with varying growth rates, while the efficient utilization of xylo-oligosaccharides was limited to some strains e.g., B. lactis [35]. Bouhnik et al. [36] tested the capacity of different oligosaccharides to stimulate fecal bifidobacteria in a placebo-controlled in vivo study; FOS, GOS, soybean oligosaccharides, and type III resistant starch were found to be bifidogenic. The ability of oligosaccharide uptake generally seems to vary within the genus of Lactobacillus and Bifidobacterium; hence, different growth rates on various oligosaccharides can be observed. In a recent study, the growth of single strains of Bifidobacterium, Lactobacillus, and Streptococcus on various trisaccharides (including 4 -galactosyl-lactose, 6 -galactosyl-lactose, 4 -galactosyl-lactulose, and 6 -galactosyl-lactulose) was evaluated, and in general, these strains grew faster on the trisaccharides with a β(1→6)-galactosyl moiety [37]. According to Thongaram, et al. [38], the GOS-utilizing capacity varying among bifidobacteria and lactobacilli relied on the degree of polymerization of GOS, and was strain-dependent as well. However, the preferential fermentation of short-chain oligosaccharides by bifidobacteria was reported [39].
The strains of Bifidobacterium and Lactobacillus as well as E. faecium in this study were selected, since some isolates are already established as probiotic strains and are used in food/feed products, or they have potentially probiotic properties such as a positive impact on the establishment and balance of the normal microflora, protection from gastrointestinal diseases, the production of important digestive enzymes, the alleviation of symptoms of lactose intolerance, cholesterol-lowering effects, stimulation of the immune system, and managing inflammatory bowel disease [40,41]. When comparing the GOS mixtures in single-strain cultivations, the highest fermentation activity scores were found for the three Bifidobacterium strains tested and L. reuteri Lb46 grown on purified Lb-GOS. Significantly lower fermentation activity scores were obtained for these Bifidobacterium strains when paired with V-GOS, which contains significant amounts of glucose and galactose that can support growth, and especially 4 GOS-P. Presumably, the different glycosidic linkages of the oligosaccharides in the GOS mixtures influence the selective ability of probiotic bacteria to metabolize these carbohydrates, as it is also evident from the maximal optical density values obtained for lactobacilli and bifidobacteria, which are consequently higher when using Lb-GOS compared to 4 GOS-P. An explanation for the above-mentioned observation could be that β-galactosidases from bifidobacteria possess a preference for hydrolyzing β(1→6) and β(1→3) linkages; hence, their growth was stimulated better with Lb-GOS, containing mainly β(1→6) and β(1→3) linked galacto-oligosaccharides, rather than with 4 GOS-P or V-GOS, which are mainly β(1→4) linked galacto-oligosaccharides.
A large variation in the fermentation activity scores was found in the single-strain fermentations for the different strains of Lactobacillus. Interestingly, even different isolates within a single species, such as L. reuteri Lb46 and L. reuteri Lb21, exhibited significantly different fermentation activity scores. L. rhamnosus and L. paracasei subsp. paracasei, which are phylogenetically closely related [42], showed low fermentation activity scores on all three GOS mixtures. Especially L. rhamnosus Lb29 seems unable to utilize GOS, as growth on these substrates was only comparable to the blank on the complex basal medium without an added carbohydrate. It was discussed previously that these differences may be due to diversity among lactobacilli and the presence of genes coding for the metabolic systems that are necessary for the transportation and utilization of a particular prebiotic as a carbon source [10,24]. Among the substrates tested with Lactobacillus spp., the novel Lb-GOS mixture, which was produced with an enzyme from Lactobacillus sp., was the best-performing substrate, giving the highest fermentation activity scores and optical density values for all of the lactobacilli  Table 2). This is obvious, especially when comparing its growth and fermentation activity scores with 4 GOS-P, which is also free of monosaccharides. These results substantiate the idea that oligosaccharides produced with enzymes from probiotic strains can have a more pronounced effect on the growth of these probiotic strains [16][17][18]43].

Chemicals
All of the chemicals were purchased from Sigma (St. Louis, MO, USA) or Oxoid (Basingstoke, Hampshire, UK), and were of the highest quality available, unless otherwise stated. Glucose oxidase (GOD) from Aspergillus niger (lyophilized, 205 U/mg enzyme preparation) was from Fluka (Buchs, Switzerland) and horseradish peroxidase (POD) (lyophilized, 210 U/mg) was from Boehringer (Mannheim, Germany). The test kit for the determination of D-galactose/lactose was from Megazyme (Bray, Ireland).

Prebiotic Oligosaccharide Mixtures
The commercial prebiotic galacto-oligosaccharide mixture Vivinal (V-GOS), containing 40% monosaccharides and lactose and 60% oligosaccharides, was obtained from Borculo Domo Ingredients (Zwolle, the Netherlands). The mixture of 4 GOS-P from Yakult Honsha (Tokyo, Japan) is a purified product with a purity of 99.9% of galacto-oligosaccharides, which are mainly β(1→4) linked oligosaccharides. Lb-GOS, containing mainly β(1→3) and β(1→6) linked oligosaccharides, was produced and purified using β-galactosidase from Lactobacillus sp. Discontinuous GOS production was carried out at 23 • C using purified β-galactosidase (five U/mL) in a five-liter scale stirred tank reactor (total volume of six liters) and 206.5 g/L lactose dissolved in 50 mM of sodium phosphate buffer (pH 6.0) containing 2 mM of MgCl 2 . When the desired degree of lactose conversion of 73% was reached, the reaction mixture was heated to 98 • C to inactivate β-galactosidase and clarified by centrifugation. To remove non-converted lactose from the sugar mixture, the enzymatic conversion of lactose to lactobionic acid was carried out as described previously using the enzyme cellobiose dehydrogenase (CDH) from Sclerotium rolfsii [44]. After the enzymatic oxidation step, the GOS solution was centrifuged and filtered to remove insoluble material, and then applied onto two ion-exchange chromatography columns in series using a strong cation exchange resin, Lewatit ® S2528 (Bayer AG, Leverkusen, Germany), and a medium basic anion exchange resin, Lewatit ® S4328 (Bayer AG, Leverkusen) for the removal of ions as described previously [45]. For the separation of the GOS from D-glucose and D-galactose, the strong acidic cation exchange material Unibead UBK-530 (Mitsubishi Chemical Industries, Tokyo, Japan) was used.

Enzyme Activity Assay and Protein Measurement
The determination of β-galactosidase activity was carried out at 30 • C using 22 mM of o-nitrophenylβ-D-galactopyranoside (oNPG) in 50 mM of sodium phosphate buffer (pH 6.5) as substrate, as previously described [46]. The reaction was initiated by adding 20 µL of enzyme solution to 480 µL of the substrate solution, and then incubated for 10 min using an Eppendorf thermomixer compact (Eppendorf, Hamburg, Germany) with an agitation of 600 rpm. The reaction was stopped by adding 750 µL of 0.4 M of Na 2 CO 3 . The release of o-nitrophenol (oNP) was measured by determining the absorbance at 420 nm. One unit of oNPG activity was defined as the amount of enzyme releasing one µmol of oNP per minute under the described conditions. Protein was determined by the method of Bradford [47] with the BioRad Coomassie Blue reagent (Marnes-la-Coquette, France) using bovine serum albumin as the standard.

Sugar Analysis
Monosaccharide analysis. D-Glucose was measured enzymatically by the coupled GOD/POD assay, as described previously [20]. For the determination of D-galactose, the lactose/D-galactose test kit from Megazyme was used.
Oligosaccharide analysis. Capillary electrophoresis (CE) and high-performance anion exchange chromatography with pulsed amperometric detection (HPAEC-PAD) (Dionex, Sunnyvale, CA, USA) were used for the qualitative and quantitative analysis of galacto-oligosaccharides. A capillary electrophoresis system with a UV-DAD detector (Agilent Technologies, Palo Alto, CA, USA) together with a fused silica capillary (internal diameter of 25 µm) equipped with a bubble cell detection window (bubble factor of five) was used for carbohydrate analysis. Carbohydrate samples were derivatized with 2-amino pyridine for CE analysis, as given in detail in [20]. HPAEC-PAD analysis was carried out on a Dionex DX-500 system consisting of a GP50 gradient pump (Dionex), an ED 40 electrochemical detector with a gold working electrode (Dionex), and an Ag/AgCl reference electrode (Dionex). Separations were performed at room temperature on a CarboPac PA-1 column (4 × 250 mm) connected to a CarboPac PA-1 guard column (Dionex).

Single Strain Cultivations and Automated Turbidimetry
Frozen cultures of bifidobacteria and lactobacilli were activated by streaking onto BHI agar containing one g/L of glucose, or MRS agar supplemented with 2% (w/v) lactose, respectively, and incubating anaerobically at 37 • C for 24-48 h as required. E. coli, K. oxytoca, and C. freundii were streaked onto nutrient agar, and S. epidermis and E. faecium were streaked onto Corynebacterium medium containing 0.5% (w/v) glucose. These five strains were incubated aerobically at 37 • C for 24 h. After the given incubation times, one single colony from each plate was picked and transferred to an appropriate medium to obtain pure cultures. Each strain was then transferred from the agar plates into appropriate liquid media (without a carbohydrate source), and a series of dilutions was prepared to obtain the inoculum for further experiments. The respective inocula (100 µL) and 200 µL of the appropriate fresh medium (containing the C-source) were placed in each micro-plate well (Honeycomb 2 plates; Labsystem, Les Ulis, France). The final carbohydrate concentration (glucose or prebiotic oligosaccharides) in these media was 0.5% (w/v), and the starting optical densities (OD 600 ) were approximately 0.001. For the cultivation of bifidobacteria and lactobacilli, anaerobic conditions were maintained by adding the enzyme system Oxyrase ® (Oxyrase Inc., Mansfield, OH, USA) to a final concentration of 2% (v/v), and the reading plates were kept airtight. The inoculated honeycomb plates were then placed in the reading chamber of a Bioscreen C MBR (Labsystems, Vantaa, Finland) and incubated at 37 • C. The cultures were mixed for 30 s before each reading was taken by the setting 'medium-intensity shaking mode'. The optical densities (OD 600 ) of the cultures were measured with readings being taken every 30 min for 24 h. Growth of the cultures in the media without an added carbohydrate source was also monitored as a blank. All of the measurements were carried out in triplicate.

Fermentation Activity Score
Fermentation activity scores (FAS), i.e., a measure for the ability of one selected strain to utilize a sugar mixture for growth in comparison with other enteric bacteria, were determined using the following Equation (1): Pro max prebiotic and Ent max prebiotic are the highest optical densities OD 600 obtained during 24 h of the growth of probiotic bacteria and enteric bacteria, respectively, on prebiotic oligosaccharides; Pro min prebiotic and Ent min prebiotic are the lowest optical densities OD 600 obtained at the inoculation of probiotic bacteria and enteric bacteria, respectively, when using prebiotic oligosaccharides; Pro max blank and Ent max blank are the highest optical densities OD 600 obtained during 24 h of growth of probiotic bacteria and enteric bacteria, respectively, in the medium without added sugar; and Pro min blank and Ent min blank are the lowest optical densities OD 600 obtained at the inoculation of probiotic bacteria and enteric bacteria, respectively, for the medium without added sugar.
This equation assumes that an increase in the number of probiotic bacteria, bifidobacteria, and lactobacilli gives a positive effect, whilst an increase in the number of enteric bacteria, which in this study are E. coli DSM 613, S. epidermis DSM 20044, K. oxytoca DSM 6673, and C. freundii DSM 30039, gives a negative effect. Based on this equation, substrates with a high fermentation activity score support the growth of bifidobacteria, lactobacilli, and E. faecium, with the culture optical densities significantly higher than the cultivations on the complex medium without the added carbohydrate.

Statistical Analysis
All of the experiments and measurements were performed at least in triplicate, and the data are given as the mean ± standard deviation when appropriate. The data were analysed using SPSS (SPSS Inc. Chicago, IL, USA; Version 11.0.0).

Conclusions
The fermentation activity scores, as determined from single-strain fermentations on several galacto-oligosaccharide substrates, reflect the potential of a given carbohydrate to promote the selective growth of a specific microorganism. This approach of pairing a specific strain with a prebiotic seems especially valid when searching for synbiotic mixtures, in which a certain probiotic strain is paired with a prebiotic supporting its growth. Hence, this highly potential prebiotic GOS mixture described in this work should be of considerable interest for the formulations of new carbohydrate-based functional food ingredients.

Conflicts of Interest:
The authors declare no conflict of interest.