Evaluation of Prebiotic Effects of High-Purity Galactooligosaccharides in vitro and in vivo

There has been an increasing interest in the regulation of colonic microfl ora in order to improve the host’s health. This has been achieved traditionally by dietary inclusion of live microbes as food supplements known as probiotics. An alternative approach involves the consumption of food ingredients known as prebiotics. Prebiotics may provide advantages to probiotic bacteria in the gastrointestinal tract and additionally exert direct eff ects on the microfl ora in the large intestine (1). Galactooligosaccharides (GOS) that consist of 3–10 molecules of galactose and glucose are known to facilitate the growth of desirable intestinal microfl ora and are considered as potent non-digestible prebiotics (2). Commercial GOS that contain complex mixtures of oligosaccharides with diff erent glycosidic linkages and degrees of polymerization are usually synthesized by enzymatic transgalactosylation of lactose by β-galactosidases from various sources such as yeast, fungi or bacteria (3,4). In addition, these kinds of products can also contain transgalactosylated oligosaccharides, unreacted lactose, gluISSN 1330-9862 original scientifi c paper


Introduction
There has been an increasing interest in the regulation of colonic microfl ora in order to improve the host's health.This has been achieved traditionally by dietary inclusion of live microbes as food supplements known as probiotics.An alternative approach involves the consumption of food ingredients known as prebiotics.Prebiotics may provide advantages to probiotic bacteria in the gastrointestinal tract and additionally exert direct eff ects on the microfl ora in the large intestine (1).
Galactooligosaccharides (GOS) that consist of 3-10 molecules of galactose and glucose are known to facilitate the growth of desirable intestinal microfl ora and are considered as potent non-digestible prebiotics (2).Commercial GOS that contain complex mixtures of oligosaccharides with diff erent glycosidic linkages and degrees of polymerization are usually synthesized by enzymatic transgalactosylation of lactose by β-galactosidases from various sources such as yeast, fungi or bacteria (3,4).In addition, these kinds of products can also contain transgalactosylated oligosaccharides, unreacted lactose, glu-cose and galactose, which do not have prebiotic properties but only the caloric value (5).
The most effi cient process to produce high-purity GOS (HP-GOS) is yeast fermentation of sugars such as glucose, galactose and lactose (6).Cardelle-Cobas et al. (7) fi rst reported the production of non-monosaccharides and HP-GOS by repeated batch fermentation with immobilized yeast cells.A GOS syrup of an increased purity was produced by immobilized β-galactosidase from Penicillium expansum F3 and subjected to fermentation by Saccharomyces cerevisiae L1 and Kluyveromyces lactis L3 (8).This was a feasible industrial process to produce high-purity GOS.
Among numerous non-digestible carbohydrate-based prebiotics, convincing scientifi c evidence for suitability for use as prebiotics exists only for inulin/oligofructose and GOS (9).Studies of the prebiotic eff ects of GOS and FOS in humans have shown that a daily dose of 4-20 g signifi cantly increases the population of lactobacilli and bifi dobacteria in the gut (10).Other eff ects, such as hypocholesterolemic eff ects, prevention of colon cancer, and enhancement of calcium absorption have been described (11,12).Numerous studies have reported data on the effects of non-digestible oligosaccharides (NDOs) and dietary fi bre content on serum cholesterol and lipid levels; however, only a limited number of reports indicate positive eff ects of GOS (13) or inulin (14) on serum cholesterol metabolism in humans.Yamashita et al. (15) suggested fi rst that dietary inclusion of fructooligosaccharides demonstrated hypocholesterolemic eff ect in diabetic subjects.
In this paper, the ability of HP-GOS to support the in vitro growth of selected strains of probiotic bacteria, namely Lactobacillus acidophilus, Lactobacillus casei, Bifi dobacterium bifi dum and Bifi dobacterium longum, is investigated.In addition, it was determined whether oral administration of HP-GOS aff ected the growth of bifi dobacteria, as well as serum cholesterol and nitrogen levels and the expression of gene encoding glucagon-like peptide-1 (GLP--1) and tyrosine-tyrosine peptide (PYY), which act as signifi cant modulators of appetite via their peripheral eff ects (on the vagus nerve) and/or by infl uencing directly the arcuate nucleus in the rat.

Production of GOS
Batch reactions were performed by incubating β-galactosidase with a 40-45 °Brix lactose solution in a 100-litre incubator shaker at 150 rpm.Lactose (25 kg) was dis-solved in distilled water (60 L), and 0.08 % (by mass) β-galactosidase from Bacillus circulans was added to synthesize GOS at 55 °C and pH=6.0 for 24 h.All reactions were terminated by incubation at 100 °C for 10 min and the sugar profi le was analyzed by high-performance liquid chromatography.To increase the GOS content, 100 mL of 20 % (by mass) GOS syrup produced by enzymatic hydrolysis with β-galactosidase were fermented by 8 % (by mass) of fresh yeast (Saccharomyces cerevisiae L1).The fermentation process was carried out in a shaking incubator at 100 rpm and 30 °C for 24 h.Aft er each fermentation cycle, cells were transferred to 20 % (by mass) GOS syrup, which was then fi ltered and treated with active carbon for decolourization.Ion exchange chro matography using Amberlite ® CG-120-II column (Sigma-Aldrich, Buchs, Switzerland) was utilized for further purifi cation.The pooled fractions were evaporated to 45 °Brix and dried with a spray dryer.Lactose and HP-GOS were determined by an HPLC system with Waters ® 2414 refractive index detector (RID) (Waters Corporation, Milford, MA, USA) equipped with YMC-Pack Polyamine II column (4 mm×250 mm; YMC Co. Ltd, Kyoto, Japan), column heater (30 °C), and RID detector.Acetonitrile (64 %) was used as mobile phase.

Animals and diet
The experimental protocol was reviewed and approved by Institutional Animal Care and Use Committ ee of Korea University.Four-week-old male Sprague Dawley ® (SD) rats were purchased from Daehan Biolink Co. Ltd. (Cheongju, Republic of Korea).The animals were kept in a room at 24 °C and constant atmosphere with 60 % humidity and a 12-hour light/dark cycle.Rats were fed an AIN-93G diet based on the main ingredients of corn starch (40 %) and casein (20 %) (17).Aft er an adaptation period, the rats were randomly divided into four groups (N=8): the control group (oral administration of saline), the HP-GOS group (oral administration of HP-GOS), the HP-GOS+BB group (oral administration of HP-GOS and bifi dobacteria), and the BB group (oral administration of bifi dobacteria).The groups were orally administrated HP-GOS (1.5 mL of the solution of 1 g of HP-GOS and/or 10 9 CFU bifi dobacteria) daily for 5 weeks.
Fresh faecal samples were collected weekly (equal mass from four rats per pool) in sterile fl asks, and kept at -80 °C for microbiological analysis.To count the total number of bifi dobacteria, 3 g of faecal sample were diluted in 25 mL of dilution solution and an aliquot of 0.2 mL was spread on Petri dishes using BL agar.Colonies were incubated anaerobically during 2-3 days at 37 °C under anaerobic conditions and results were measured as the log CFU per gram of faecal sample.
At the end of the study, the rats were euthanized using CO 2 asphyxiation and the liver, kidney and spleen were removed and weighed immediately.The body mass of each rat was measured every week for 5 weeks and the mass of each organ was expressed as 100 g of body mass.

Blood analysis
Blood samples were collected into non-heparinized serum separator tubes.Serum triglycerides (TGs), total cholesterol (TC), and high-density lipoprotein cholesterol (HDL-C) levels were measured by using a FUJI Dri-Chem 3500 system (Fuji Photo Film Co., Osaka, Japan).Concentration of low-density lipoprotein cholesterol (LDL-C), in mg per 100 mL, was calculated according to the method of Friedewald et al. (18) as follows:

Statistical analysis
All statistical analyses were performed using the Statistical Package for Social Sciences, v. 12.0 (SPSS Inc., Chicago, IL, USA).The statistical signifi cance of diff erences was determined using one-way ANOVA at a signifi cance level of p<0.05.All data were signifi cant at 95 % level and reported as the mean value±standard deviation (S.D.).

GOS content before and after enzymatic hydrolysis and fermentation
Table 1 summarizes the content of saccharides during diff erent stages of GOS preparation, i.e. the enzymatic hydrolysis by β-galactosidase and fermentation by S. cerevisiae.Aft er treatment by β-galactosidase, GOS content increased from 0 to 51.0 g per 100 g, and lactose mass fractions decreased from 99.2 to 21.3 g per 100 g (Table 1).Monosaccharide (glucose and galactose) mass fractions also increased from 0.8 to 27.7 g per 100 g aft er the enzymatic reaction; however, these monosaccharides were completely removed by fermentation with S. cerevisiae, whereas lactose levels were slightly increased.In order to increase the purity of GOS, fermentation process was utilized.Aft er fermentation, glucose and galactose mass fractions decreased from 27.7 to 0.0 g per 100 g (Fig. 1 and Table 1).The purity of the GOS preparation increased  from 51.0 to 73.6 g per 100 g aft er fermentation with S. cerevisiae (Table 1), verifying that HP-GOS were successfully produced using the two-step enzymatic hydrolysis and fermentation by yeast.
For the production of HP-GOS, selective fermentation for a given microorganism was characterized.During the fermentation, ethanol may be produced depending on the used microorganism, which at toxic volume fractions may compromise the activity of the microorganism.S. cerevisiae was used to increase the purity of the mixture of GOS obtained by enzymatic hydrolysis with β-ga lac tosidase from Bacillus circulans, with complete removal of the monosaccharides (21).The same approach was adopted for the production of GOS mixture from B. bifi dum biomass (9).In a previous research, Kluyveromyces marxianus improved the purity of a GOS mixture produced by β-galactosidase from B. circulans from 38 to 97 % by selective fermentation of mono-and disaccharides (including lactose) (6).With a combined use of S. cerevisiae and K. lactis the purity of a GOS mixture produced by β-galac tosidase from Penicillium expansum increased from 29 to 98 % (8).In this work, HP-GOS were produced by fermentation with S. cerevisiae (Table 1 and Fig. 1), which increased the GOS content from 51.0 to 73.6 g per 100 g aft er fermentation with S. cerevisiae.

Utilization of GOS by intestinal bacteria in vitro
HP-GOS comprise substances with structural diff erences and have considerably greater prebiotic potential compared to a commercially available GOS mixture.Fig. 2 shows the growth of intestinal bacteria in a medium containing various commercial GOS.In order to confi rm that the growth was dependent on GOS utilization, strains were also inoculated into PYF basal medium containing 1 % HP-GOS, Y-GOS, C-GOS or Q-GOS as the carbohydrate source.During the early stages of growth (after 12 h), all strains exhibited a similar growth rate without signifi cant diff erences.All strains (L.acidophilus, L. casei, B. longum and B. bifi dum) in a medium containing HP-GOS had a higher cell growth rate than the strains grown in the media containing commercial GOS aft er 12 h of culture.These results suggest that HP-GOS serve as a good substrate and carbon source for supporting the growth of probiotic bacteria.
Growth curves for the given strains grown anaerobically in the medium without GOS or with 0.5, 1, 2 or 4 % (by mass) GOS are shown in Fig. 3. L. acidophilus and L. casei were able to grow at all the tested GOS mass fractions, both reaching a maximum A 600 nm of 1.15 aft er 36 h of growth.Cell growth was increased with an increase in HP-GOS mass fraction.Growth of B. bifi dum and B. longum increased rapidly until 12 h and then continued increasing slowly.Therefore, HP-GOS mass fractions above 1 % were found to be acceptable for the growth of probiotic bacteria.
Structural diff erences of GOS vary notably depending on the conditions and source of enzymes used for their synthesis (HP-GOS and Y-GOS were produced from 4'-or 6'-galactosyl-lactose, and C-GOS and Q-GOS from 4'-galactosyl-lactose) (4), which aff ects the fermentation process as well as their prebiotic properties.Studies of GOS utilization by bacteria have shown that diff erent strains vary in their ability to ferment GOS, with individual strains exhibiting specifi c substrate preferences (22,23).
We additionally analyzed the utilization of GOS by various probiotic bacteria (Fig. 2).L. acidophilus, L. casei, B. bifi dum and B. longum had a higher cell growth rate when utilizing HP-GOS in comparison with commercial GOS (Fig. 2), suggesting that 4'-or 6'-galactosyl-lactose may be a more suitable substrate for these strains.The utilization of GOS by a number of probiotic bacteria has been extensively analyzed (24,25).It has been shown that 4'-galactosyl-lactose is selectively utilized by all the Bifi dobacterium strains tested, compared with lactulose and raffi nose, whose specifi city is less noticeable.Other studies have also shown that some strains of Lactobacillus, Bacteroides and Clostridium ferment GOS, and that transgalactosylated disaccharides may even be bett er substrates for these bacteria (26).Some bacterial species can ferment both 4'and 6'-galactosyl-lactose, although there are some diff erences and the growth of bacteria is dose-dependent.
As shown in Fig. 3, B. bifi dum and B. longum utilized HP-GOS more rapidly than L. acidophilus and L. casei.It is considered that the utilization of non-digestible oligosaccharides (NDOs) by bifi dobacteria is mediated by the hydrolyzing enzymes produced by these strains.Many Bifidobacterium strains produce glycolytic enzymes that hydrolyze a wide range of monosaccharides and various glycosidic bonds, while the activities of the enzymes from other enteric bacteria such as Lactobacillus, Escherichia coli and Streptococcus are less varied and are weaker than those from Bifi dobacterium (27).

Changes in body mass, organ mass and serum parameters
Changes in body mass and organ mass of the rats in control, HP-GOS, HP-GOS+BB and BB groups are shown in Table 2.No signifi cant diff erences in body mass were found among the three groups.The changes in the masses of the liver and other internal organs of the rats aft er a 5-week administration of each diet are shown in Table 3.No diff erences were observed in the masses of the liver, spleen and kidney in relation to the body mass in all groups.
Serum glucose, protein, aspartate aminotransferase (AST), alanine aminotransferase (ALT) and lipid profi les are shown in Table 4. Groups administered HP-GOS+BB had signifi cantly diff erent AST and ALT levels compared with the control group or the group administered HP--GOS, respectively (p<0.05).However, the ALT and AST levels of the group administered HP-GOS+BB were in the normal range for SD rats (17.30-19.77and 74.4-80.4U/L, respectively) (28).There were no signifi cant diff erences in the lipid profi les (total cholesterol, HDL cholesterol and triacyglycerol) among the groups.

Enumeration of bifi dobacteria in rat faeces
Fig. 4 shows the total bifi dobacteria counts in rat faeces, expressed in log CFU per g of faeces, during four periods of the study.The groups that were orally administered HP-GOS and bifi dobacteria had signifi cantly (p<0.05)higher counts during all four periods, while the group receiving HP-GOS had higher counts in the 2nd and 4th period only in comparison with the control and BB group.In the 2nd and 3rd periods, the same trend was observed in groups administered HP-GOS and HP--GOS+BB.During the entire period, oral administration of HP-GOS+BB resulted in higher bifi dobacteria counts than the oral administration of control or single administration of bifi dobacteria.There was no signifi cant diff erence in bifi dobacteria counts between the HP-GOS+BB group and the HP-GOS group in the 2nd and 4th periods.
Bifi dobacteria in rat faecal samples were counted in the HP-GOS groups with or without bifi dobacteria.Our results indicated a positive eff ect in the synbiotic group (HP-GOS+BB) during the test periods.It is known that most bifi dobacteria strains of human intestinal origin can readily use galactooligosaccharides; however, only a few strains from other genera, such as lactobacilli, possess this ability (24,26).As far as the eff ects of probiotic consumption on the bifi dobacterial population are concerned, similar results have been observed in children and adults.Benno and Mitsuoka (29) reported an increase in the counts of bifi dobacteria as well as a remarkable decrease in the counts of clostridia in adult human subjects consuming a daily dose of B. longum.

Expression of genes encoding GLP-1 and PYY in ileum
There was an approx.1.6-fold increase in PYY mRNA levels in the ileum of rats administered GOS+BB (Fig. 5), which was signifi cantly higher than in the other groups (p<0.05).Similarly, GLP-1 mRNA levels in the ileum of rats administered GOS+BB and only bifi dobacteria were 1.5-and 1.6-fold higher, respectively, than in rats fed normal diet.There was a signifi cant upregulation of GLP-1 and PYY mRNA with GOS+BB intake.
Endocrine cells in the intestinal mucosa secrete peptides involved in the regulation of food intake and/or pancreatic function; the latt er are known as incretins (30,31).Endocrine L cells are distributed throughout the intestinal tract, and are predominantly present in the caeco-colon, where fermentation of inulin-type fructans occurs (32).GLP-1 and PYY, which are released from intestinal L cells, modulate appetite and thus reduce food intake (33).The ability of prebiotic fi bre to increase the proglucagon mRNA levels and GLP-1 secretion is well supported (34,35).Increased proglucagon expression typically takes place when caecal mass is increased due to the markedly increased bacterial fermentation that occurs with the consumption of prebiotics (34,35).It is interesting to note that the fi bre-containing prebiotic diet appears to have acute and lasting eff ects on the ability of L cells to produce and secrete GLP-1 and PYY.In the current study, measurements of GLP-1 and PYY levels were performed  at the end of a 5-week period, when GOS as prebiotics were administered as part of the diet; however, we additionally found higher expression of genes encoding GLP--1 and PYY in the HP-GOS group with or without B. bifidum at the end of the period in which the high fat-based diet was administered (Fig. 5).
It is acknowledged that an optimum balance in microbial populations in the digestive tract is associated with good nutrition and health, and that this may be achieved by the consumption of probiotics.In particular, HP-GOS exhibit a greater prebiotic activity than other commercial GOS.

Conclusion
High-purity galactooligosaccharides (HP-GOS) were successfully produced using a two-step process, enzymatic hydrolysis and fermentation by yeast.They were found to serve as a good substrate and carbon source for supporting the growth of enteric bacteria, compared with other commercial GOS.The benefi cial eff ect of regular intake of HP-GOS is att ributed to the intestinal survival of probiotic Lactobacillus and Bifi dobacterium strains.The health benefi ts associated with the consumption of HP--GOS in humans include improvement of intestinal tract health, increase in the expression of genes encoding GLP--1 and PYY, decrease in the risk of certain cancers, blood pressure control, and reduction of serum cholesterol levels.On the basis of our fi ndings, we propose that the prebiotic properties of HP-GOS are valuable in the production of potential health-enhancing foods and supplements, and that HP-GOS may be used as a functional food ingredient for human consumption.

Fig. 1 .
Fig. 1.Carbohydrate profi le: a) before and b) aft er the fermentation with Saccharomyces cerevisiae L1

Fig. 4 .
Fig. 4. Number of total bifi dobacteria in rat faeces, expressed as mean values of log colony-forming units (CFU) per g of wet faeces, N=6.Bars represent the standard error of triplicate measurements.Diff erent lett ers indicate signifi cant diff erences at p<0.05.HP-GOS=high-purity galactooligosaccharides, BB=bifido bacteria

Table 2 .
Changes in body mass of rat aft er treatment with galactooligosaccharides and/or bifi dobacteria Values are expressed as mean±standard deviation.Control=oral administration of saline, HP-GOS=oral administration of high-purity galactooligosaccharides, BB=oral administration of bifi dobacteria, HP-GOS+BB=oral administration of HP-GOS and bifi dobacteria

Table 4 .
Serum levels of glucose, total protein, AST, ALT, and lipid profi les of rats