Appl Microbiol Biotechnol (2006) 70: 612–617 DOI 10.1007/s00253-005-0102-y
APPLIED MICRO BIAL AND CELL PHYSIOLOGY
Marisa S. Garro . Laura Aguirre . Graciela Savoy de Giori
Biological activity of Bifidobacterium longum in response to environmental pH Received: 5 May 2005 / Revised: 11 July 2005 / Accepted: 13 July 2005 / Published online: 9 August 2005 # Springer-Verlag 2005
Abstract The influence of environmental pH on biological activity of Bifidobacterium longum CRL 849 grown in MRS-raffinose was evaluated. At pH 6.0, 5.5 and 5.0, raffinose was completely consumed by this microorganism, showing different consumption rates at each pH value (between 3.03 and 0.76 mmol l−1 h−1). At pH 4.5, the growth was lowest. The removal of raffinose was due to the α-galactosidase (α-gal) activity of this bifidobacteria, which was highest at pH 6.0–5.5 (1,280–1,223 mU ml−1). The production of β-glucosidase (β-glu) showed a similar pattern to α-gal activity with major values. The yield of organic acids produced during raffinose consumption was also highest at pH 6.0–5.5. The results of this study will allow the selection of the optimum growth conditions of B. longum CRL 849, with elevated levels of α-gal to be used in the reduction of nondigestible α-oligosaccharide in soy products and β-glu activities involved in isoflavone conversion to bioactive forms when used as starter culture.
Introduction Bifidobacterium species are a major component of the intestinal flora of healthy humans. It is reported that these M. S. Garro (*) . L. Aguirre . G. Savoy de Giori Centro de Referencia para Lactobacilos (CERELA-CONICET), Chacabuco 145, 4000 Tucumán, Argentina e-mail:
[email protected] Tel.: +54-0381-4310465 Fax: +54-0381-4005600 L. Aguirre Cátedra de Bioquímica, Facultad de Medicina Universidad Nacional de Tucumán (UNT), Tucumán, Argentina G. Savoy de Giori Cátedra de Microbiología Superior, Facultad de Bioquímica, Química y Farmacia, Universidad Nacional de Tucumán (UNT), Tucumán, Argentina
organisms can exert beneficial effects including the reduction of serum cholesterol, activation of the immune system and inhibition of the growth of potential pathogens that may cause infectious disease in the host. In addition, metabolites produced by bifidobacteria are considered to play an important role in the prevention of hormonedependent disorders (Anthony 2000; Kurzer 2000). These microorganisms are involved in the bioactivation of isoflavones (Hendrich 2002; Turner et al. 2003), possibly due to hydrolysis of isoflavone glycosides into bioactive aglycones. Aglycones are structurally similar to estrogen and therefore can mimic the functions of estradiol in the human body (Setchell and Cassidy 1999). The cleavage of isoflavone glucosides by bifidobacteria depends on their capability to express the enzyme β-glucosidase (β-glu). Today, industrial strategies focus on the selective stimulation of growth and activity of bifidobacteria in the intestine by supplementation of the food with specific, nondigestible carbohydrates (bifidogenic factors, prebiotics) (Biavati et al. 2000). Some oligosaccharides can be considered bifidogenic factors such as fructose-oligosaccharides, transβ-D-galactosyl-oligosaccharides and α-D-galactose-oligosaccharides (raffinose, stachyose), the latter being found in foods (legumes, beans) and as endogenous carbohydrates (mucus, glycoproteins). The adaptation of bifidobacteria to these α-D-galactose-oligosaccharides may be achieved by their capability to express specific hydrolytic enzymes. The degradation of α-galactosidic linkages is catalyzed by the enzyme α-galactosidase (EC. 3.2.1.22, α-gal). Since the recognition of the beneficial effects of bifidobacteria, there has been considerable interest in their use in foods. Food manufacturers claim that following ingestion, these bacteria colonize the gastrointestinal tract and exert their beneficial effects on the host. Bifidobacteria are often incorporated in fermented dairy products to increase their therapeutic value (Baron et al. 2000). Bifidobacteria might similarly be incorporated in soymilk (SM). In previous studies, Garro et al. (1994, 1998, 1999a,b, 2002) reported the growth characteristics and the end-products formation by different lactic acid bacteria in SM. The authors found that Lactobacillus fermentum, Streptococcus
613
salivarius subsp. thermophilus and Bifidobacterium longum, which utilize sucrose, raffinose or stachyose, exhibited significant growth and produced substantial amounts of acids in SM. The production of α-gal from L. fermentum and B. longum in SM at different temperatures was also evaluated (Garro et al. 2004a). B. longum showed maximum enzyme activity in SM at 37°C. However, few studies have described the effect of environmental pH on the biological activity of bifidobacteria when cultivated for industrial applications. The purpose of the present work was to determine the optimal growth kinetic parameters and enzymes production (α-gal and β-glu) of B. longum under pH-controlled conditions. These results will contribute to a better understanding of the technological aspects of this probiotic microorganism and could be important in the design and operation of manufacturing processes for probiotic soy food production.
Materials and methods Microorganism and growth conditions The strain B. longum CRL 849 used in this study was obtained from the Culture Collection (CRL) of the Centro de Referencia para Lactobacilos (CERELA, Tucumán, Argentina). Before experimental use, cultures were propagated (2%, v/v) twice in sterile modified MRS medium and incubated at 37°C for 18 h without agitation. Modified MRS medium consisted of MRS broth with 1% raffinose instead of glucose, supplemented with 0.05% L-cysteine hydrochloride, 0.0005% hemin and 0.00005% vitamin K. All aggregates were sterilized separately (0.22 μm filtration), then added to the MRS base.
Independent fermentation experiments were carried out in triplicate, and analytical determinations were performed in triplicate on samples coming from three different replicate fermentations. Analytical procedure Bacterial growth was measured turbidimetrically at 560 nm and by determination of cell dry weight (dcw, biomass). Biomass was determined by filtering an aliquot of culture broth through a pre-dried 0.45-μm polysulphone filter (Gelman Sciences, USA), washing it with a fivefold excess of distilled water and drying it to a constant weight. Specific growth rate (μ, h−1) was estimated during exponential phase by linear regression of ln X/X0 dcw vs time. For evaluation of residual sugars and organic acids production, samples were centrifuged at 10,000×g for 10 min at 4°C and supernatants were stored at −20°C until analysed. Lactic and acetic acids were determined by HPLC (Isco model 2360, NE, USA) using an Aminex HPX87H column (Bio Rad Laboratories, USA) at room temperature, with a flow rate of 0.6 ml min−1 of 5 mM sulfuric acid following absorbance at 210 nm. Raffinose, sucrose, melibiose, glucose, galactose and fructose were quantified by HPLC coupled to a differential refractometer (LKB, model 2142, Bromma, Sweden) using REZEX RSO oligosaccharides column (200×10 mm, Phenomenex, Torrance, CA, USA) at a column temperature of 70°C, using HPLC grade water as the eluant at a flow rate of 0.3 ml min−1. Samples were deproteinized before sugar determination as described previously (Garro et al. 2004c). Preparation of crude extract and enzymes activities determination
Fermentation conditions Batch culture experiments were performed in a 1.0-l fermentor (BioFlo New Brunswick Scientific Co. Inc., Edison, NJ, USA) filled with modified MRS medium and maintained under anaerobic conditions using filter-sterilized nitrogen gas with gentle agitation (50 rpm). The temperature was controlled at 37°C and automatically buffered to desired pH values of 6.0, 5.5, 5.0 or 4.5·by addition of sterile 1 M NH4OH or HCl. Inoculation was carried out using an 18-h culture grown on the same medium (the cells were harvested by centrifugation and washed two times with a saline solution) at 2% [v/v, i.e., 108 colony-forming units (cfu) ml−1]. Fermentation was allowed to proceed for 24 h, and samples were aseptically withdrawn at 0, 2, 4, 6, 8, 10, 12, 16 and 24 h from the fermentation vessel and immediately cooled on ice to determine α-gal and β-glu activities, bacterial growth, endproducts, residual carbohydrate (raffinose) and their hydrolysis products (sucrose, melibiose, glucose, galactose and fructose).
Samples (5 ml) were centrifuged at 10,000×g for 10 min at 4°C, and the cell pellet was washed twice with 5 mM McIlvaine buffer (Na2HPO4-citric acid, pH 5.8) (McIlvaine 1921) and resuspended in 0.5 ml of the same buffer (final OD560 nm=20). Cells were disrupted with 500 mg glass beads (0.10–0.11 mm, Sigma, USA) and with shaking at maximum speed on a vortex mixer ten times during 1 min, with 1-min intervals on ice in between. Cellular debris were removed by centrifugation at 10,000×g for 10 min at 4°C, and the supernatant was kept on ice until analysis of intracellular enzymes (α-gal and β-glu) activities. Protein concentration was determined using the Bio-Rad Protein Assay based on the method of Bradford (1976) using bovine serum albumin as a standard. Glycosidase activity was measured by the release of p-nitrophenol (pNP) from p-nitrophenyl-α-D-galactopyranoside (pNPGal) for α-gal or p-nitrophenyl-β-D-glucopyranoside (pNPGlu) for β-glu (Garro et al. 2004b). Briefly, to a 45-μl sample, 15 μl of 10 mM of specific substrate solution (pNPGal or pNPGlu) was added and incubated at 37°C for 15 min. Na2CO3 (900 μl 0.25 M) was added to stop the reaction. Absor-
614
bance at 400 nm was measured using a VersaMax Tuneable Microplate Reader (Molecular Devices, USA). One enzyme unit (U) was defined as the amount of enzyme that released 1.0 μmol of pNP from its substrates (pNPGal or pNPGlu) per minute under the given assay conditions. The results are expressed as U ml−1. Reproducibility All results presented in this paper are the average of three independent assays. The variations among results were less
Fig. 1 Effect of environmental pH on the biological activity during growth of Bifidobacterium longum CRL 849 on modified MRSraffinose medium. a pH 6.0, b pH 5.5, c pH 5.0, d pH 4.5 and e without pH control. Dashed lines: biomass (filled square), αgalactosidase (filled circle), β-glucosidase (filled triangle). Solid
than 10%. Results were expressed as mean±standard deviation, and their significance was analysed using Student’s t test.
Results Bifidobacterum longum CRL 849 cultured in modified MRS broth at fixed pH values of 4.5, 5.0, 5.5 or 6.0 in separate fermenter runs showed differences in growth rate, enzymatic activities and fermentation products (Fig. 1a–d). When the pH was maintained at 6.0 (Fig. 1a), the strain
lines: residual raffinose (filled diamond), lactic acid (empty circle) and acetic acid (empty triangle) produced. Results are expressed as (X-X0) where X is the concentration at a given time (T) and X0 at the start time (T0)
615
displayed a slow specific growth rate (μ=0.18 h−1) during the first 8 h, which increased between 8 and 12 h (μ= 0.46 h−1). By contrast, at pH 5.5 (Fig. 1b), the highest specific growth rate (μ=0.37 h−1) was attained before 8 h. Under the former growth conditions, the rates of raffinose consumption (3.0 mmol l−1 h−1) and lactic acid and acetic production (7.5 mmol l−1 h−1 and 20.2 mmol l−1 h−1, respectively) were 1.16, 1.15 and 1.70 times higher, respectively, with respect to uncontrolled pH (Fig. 1e; Table 1). The behaviour of CRL 849 at pH 5.0 (Fig. 1c) was quite different from that at pH 4.5 (Fig. 1d): cultures showed a reduced growth rate (μmax=0.24 h−1 and 0.07 h−1, respectively) and a minor biomass production (2,000 mg l−1 and 1,200 mg l−1, respectively) (Fig. 1c,d, Table 1). The scarce growth rate obtained could be due to the low raffinose consumption rate (0.8–0.1 mmol l−1 h−1) during the exponential growth phase compared to that observed at pH 6.0 (3.0 mmol l−1 h−1, Fig. 1a) or even at pH 5.5 (2.5 mmol l−1 h−1, Fig. 1b), although after 12 h of incubation time, the growth rate was similar to that observed at pH 5.0. At pH 4.5, the consumption of raffinose was nearly 20 times lower than at the other assayed pH conditions. At this pH, the maximum biomass was reached after 24 h, which was similar to that achieved at pH 5.0 at 24 h of incubation (data not shown). B. longum CRL 849 did not grow at pH 4.0 (data not shown). Residual raffinose concentration was below detectable levels after 11 h of fermentation at pH 5.5 (Fig. 1b), whereas about 0.92 mmol l−1 (6%) residual sugar was observed at pH 6.0. However, raffinose was completely consumed within 24 h at this later pH value (Fig. 1a) and at pH 5.0 (Fig. 1c). In contrast, about 12.20 mmol l−1 residual raffinose was still found at pH 4·5 after 24 h of fermentation. Sucrose, melibiose, galactose, glucose and fructose (hydrolysis products) were
Table 1 pH effect on growth kinetic parameters and enzyme activity of Bifidobacterium longum CRL 849 inoculated on MRS-raffinose
pH 6.0, 5.5, 5.0, 4.5 and without pH control (n/c) a The variations among results were less than 10% b Yx/s indicates the growth yield, i.e. gram biomass produced per mmol raffinose consumed c Yp/s indicates the product yield, i.e. mmol lactic and acetic acid produced per mmol raffinose consumed
found in trace amounts in the medium at the end of fermentations. The enzymatic activities (α-gal and β-glu) produced by B. longum are shown in Fig. 1a–d. The maximum specific activity for both enzymes was reached during the late exponential growth phase (10–12 h incubation) at pH 6.0, 5.5 and 5.0 (Table 1), while at pH 4.5, this same behaviour was observed after 24 h. Total α-gal and β-glu production by B. longum CRL 849 was 15.0- and 13.0-fold higher, respectively, in cultures at pH 5.5 compared to pH 4.5, either after 12 h (exponential growth phase) or 24 h incubation (stationary growth phase). The enzyme production kinetic was similar for the majority of pH-controlled experiments, although a slight difference of this pattern was observed at pH 6.0 (Fig. 1a). Here, when the cultures reached the stationary growth phase, enzyme production continued during the incubation period. All cultures produced more acetate than lactate at all pH tested (Table 1).The maximum production of organic acids was found at the end of exponential growth phase (Fig. 1). Acetate production was 20.2 mmol l−1 h−1 at pH 6.0, while at pH 4.5, the production was significantly lower (7.1 mmol l−1 h−1). An uncoupling between acetate production and growth of the microorganism (biomass formed) was observed after 12 h at pH 5.5 (Fig. 1b). Ethanol and formate were not detected during the different fermentations. The molar ratio of fermentation products by the culture did not change noticeably at controlled pH 6 (Fig. 1a) and 5.5 (Fig. 1b) (acetate to lactate ratio was 2.6– 2.4, respectively). This was constant during the fermentation period. By contrast, at controlled pH 5 (Fig. 1c) and 4.5 (Fig. 1d), the values were variable during the incubation time (between 2.6 to 4.2 and 1.2 to 5.4, respectively).
pH 6.0 Growth μmax (h−1) Biomass max (mg l−1) α-Galactosidase Rate production (U ml−1 h−1) Specific activity (U mg−1) β-Glucosidase Rate production (U ml−1 h−1) Specific activity (U mg−1) Rate consumption (mmol l−1h−1) of Raffinose Rate production (mmol l−1 h−1) of Lactic acid Acetic acid Yx/s (8 h)b Yp/s (8 h)c
5.5
5.0
4.5
n/c
0.46a 2,200
0.37 2,500
0.24 2,000
0.07 1,200
0.41 2,000
0.22 0.96
0.31 0.51
0.07 0.67
0.02 0.36
0.20 0.93
0.27 1.1 3.0
0.25 1.1 2.4
0.13 0.6 0.8
0.02 0.3 0.1
0.28 1.1 2.5
7.5 20.2 0.53 21.5
5.7 12.5 0.36 10.3
1.1 9.9 0.47 15.4
0.6 7.1 0.94 3.8
6.5 11.8 0.44 14.5
616
Among the different batch cultures, the main differences in molar product yield corresponded to the first 8 h of fermentation. The Yp/s was low at pH 4.5 (only 3.8 mmol lactate and acetate per mmol of raffinose consumed), while it was 15.4 mmol and 21.5 mmol at pH 5.0 and 6.0, respectively (Table 1). The results, summarized in Table 1, illustrate that at the higher pH values, studied enzyme activities, biomass growth and product yield were superior; the highest α-gal and β-glu production by B. longum CRL 849 was observed at pH 5.5.
Discussion The physiology of probiotic bacteria such as B. longum is of interest for two reasons. First, during growth in the lower intestines, these organisms compete with other bacteria for available substrates, and the metabolic products (acetate and lactate), which inhibit pathogens and/or are absorbed by the host, act to lower the intestinal pH. Secondly, in the food industry, these bacteria are cultivated, either in a fermenter from which they may be harvested and added to food as supplement or in situ in food. In this case, it is useful to know the kinetic parameters during growth on a particular substrate. In the present work, the experiments were carried out in a nutritionally complex medium using raffinose as substrate, since this is the situation in most of soyfoods. The use of raffinose for biomass production was affected by the environmental pH, although similar biomass yields for several pH-controlled and free-pH conditions were obtained (Table 1). In general, the substrate was the limiting factor for growth, since little or no residual carbohydrates could be detected after fermentation. When the pH was maintained at 5.5, growth of B. longum stopped after about 11 h of fermentation, at which time raffinose disappeared. At pH 5.0 and 4.5, this microorganism was able to grow. By contrast, Reilly and Gilliland (1999) reported that some strains of B. longum do not grow at pH 5.0 and that greater growth occurred at pH 6.0. The molar ratio (acetate to lactate) was constant (2.6) in function of incubation time at pH 5.5 and 6.0, but varied at pH 5.0 and 4.5 (between 1.2 and 5.4). The metabolic capacity of B. longum seems to be affected at low pHcontrolled conditions (5.0–4.5); in these conditions, the growth is scarce and the formation of end-products is abnormal, indicating a change in carbohydrate metabolism. It is well known that bifidobacteria produce mainly acetate and lactate from carbohydrate catabolism by the bifidus pathway, theoretically yielding 3 mol of acetic acid and 2 mol of lactic acid per 2 mol of glucose in synthetic medium (Biavati et al. 1992). The molar product yields increased significantly, from 14.54 to 21.51 mmol lactic and acetic acid produced per mmol raffinose consumed under uncontrolled pH with respect to pH 6.0, respectively. At pH 5.5, the acetic acid production rate decreased 1.6fold when compared with pH 6.0, where a similar decrease
(1.2-fold) in the production rate of lactic acid was observed. The first condition is desirable for the production of soy foods with agreeable taste since low acetic acid amounts were obtained. It has been shown that bifidobacteria strains with β-glu activity are potentially important in the production of compounds with higher estrogenicity and better absorption, facilitating the bioavailability of isoflavones (Hur et al. 2000). B. longum CRL 849 showed higher levels of β-glu activity than those found by Tsangalis et al. (2002) among the different strains of B. longum. These authors reported that B. longum-b did not produce β-glu in the presence of any other sugar except glucose (as a fermentable substrate). Our results indicate that B. longum CRL 849 is able to produce this enzyme in the presence of raffinose used as the principal energy source. Thus, this strain is considered a promising strain for soymilk fermentation and a heterogeneous source of carbohydrates, including predominantly raffinose and stachyose (Garro et al. 1999a,b, 2004a). At pH 6.0, 5.5 and 5.0, the specific activity for β-glu was highest at 12 h of incubation, which corresponded to the exponential growth phase. These results are in agreement with those of Scalabrini et al. (1998). Recently, Tsangalis et al. (2002) screened three strains of bifidobacteria (Bifidobacterium psedolongum, Bifidobacterium animalis and B. longum) for β-glu activity and reported that B. animalis has the highest β-glu activity. However, B. longum CRL 849 used in the present study showed 1.4and 26-fold higher β-glu activities than B. animalis and B. longum used by Tsangalis et al. (2002). Moreover, Pyo et al. (2005) reported that Lactobacillus plantarum KFRI 00144 showed higher β-glu activity than any other strains tested. However, the levels of activity of this L. plantarum were 85-fold lower than our strain. This discrepancy of β-glu-producing bifidobacteria is mostly attributed to the differences in the bacterial strains used and not in the composition of growth medium. α-Gal activity, detected in cell-free extracts, increased during the exponential growth phase under environmental pH conditions. In most cases, the activity decreased during the stationary growth phase, in agreement with the results published by Roy et al. (1991). According to Scalabrini et al. (1998), B. longum CRL 849 maintained high activities during 24 h of fermentation at pH 6.0, suggesting that this behaviour is independent of the growth phase. In conclusion, the behaviour of B. longum CRL 849 under controlled pH 5.5 resulted not only in a maximum biomass production (attained up to 8 h of fermentation) but also in enhanced lactate production and higher α-gal and β-glu activities. These conditions should be used for starter cultures production for fermented soy foods. Furthermore, B. longum CRL 849 has the potential to be used as a functional starter culture for fermented soymilk with higher estrogenicity and better absorption, facilitating the bioavailability of isoflavones.
617 Acknowledgements This study was partly supported by grants from Consejo Nacional de Investigaciones Científicas y Tecnológicas (CONICET), Agencia Nacional de Promoción Científica y Tecnológica (ANPyCT-FONCYT) and Consejo de Ciencia y Técnica de la Universidad Nacional de Tucumán (CIUNT), Argentina. The authors are grateful to Mr. Oscar Peinado Reviglione for excellent technical assistance in the HPLC analyses and to Master Jean Guy LeBlanc for constant discussion and suggestions.
References Anthony MS (2000) Soy and cardiovascular disease: cholesterol and beyond. J Nutr 130:662S–663S Baron M, Roy D, Vuillemard J-C (2000) Biochemical characteristics of fermented milk produced by mixed-cultures of lactic acid starters and bifidobacteria. Lait 80:465–478 Biavati B, Sgorbati B, Scardovi V (1992) The genus Bifidobacterium. In: Balows A, Truper HG, Dourkin M, Harder W, Schleifer KH (eds) The prokaryotes, 2nd edn. Springer, Berlin Heidelberg New York, pp 816–833 Biavati B, Vescovo M, Torriani S, Bottazzi V (2000) Bifidobacteria: history, ecology, physiology and applications. Ann Microbiol 50:117–131 Bradford MM (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 72:248–254 Garro MS, de Giori GS, de Valdez GF, Oliver G (1994) Utilization of soymilk galactosaccharides by lactobacilli. Microbiol Aliment Nutr 12:61–66 Garro MS, de Valdez GF, Oliver G, de Giori GS (1998) Growth characteristics and fermentation products of Streptococcus salivarius subsp. thermophilus, Lactobacillus casei and L. fermentum in soymilk. Z Lebensm Unters Forsch A 206:72–75 Garro MS, de Valdez GF, Oliver G, de Giori GS (1999a) Hydrolysis of soya milk oligosaccharides by Bifidobacterium longum CRL 849. Z Lebensm Unters Forsch A 208:57–59 Garro MS, de Valdez GF, Oliver G, de Giori GS (1999b) Starter culture activity in refrigerated fermented soymilk. J Food Prot 62:808–810 Garro MS, de Valdez GF, de Giori GS (2002) Application of conductimetry for evaluation of lactic starter cultures in soymilk. J Food Sci 67:1175–1178 Garro MS, de Valdez GF, de Giori GS (2004a) Temperature effect on the biological activity of Bifidobacterium longum CRL 849 and Lactobacillus fermentum CRL 251 in pure and mixed cultures grown in soymilk. Food Microbiol 21:511–518
Garro MS, de Valdez GF, de Giori GS (2004b) α-Galactosidase assay in fermented soymilk products. In: Spencer JFT, Ragout de Spencer AL (eds) Environmental microbiology: methods and protocols. Humana, New Jersey, pp 121–124 Garro MS, de Valdez GF, de Giori GS (2004c) Determination of oligosaccharides in fermented soymilk products by high-performance liquid chromatography. In: Spencer JFT, Ragout de Spencer AL (eds) Environmental microbiology: methods and protocols. Humana, New Jersey, pp 135–138 Hendrich S (2002) Bioavailability of isoflavones. J Chromatogr B Analyt Technol Biomed Life Sci 777:203–210 Hur JW, Lay JO, Beger RD, Freeman JP, Ra F (2000) Isolation of human intestinal bacteria metabolizing the natural isoflavone glycosides, daidzin and genistin. Arch Microbiol 174:422–428 Kurzer MS (2000) Hormonal effects of soy isoflavones: studies in premenopausal and postmenopausal women. J Nutr 130:660S– 661S McIlvaine TC (1921) A buffer solution for colorimetric comparison. J Biol Chem 49:183–186 Pyo YH, Lee TC, Lee YC (2005) Enrichment of bioactive isoflavones in soymilk fermented with β-glucosidase-producing lactic acid bacteria. Food Res Int 38:551–559 Reilly SS, Gilliland SE (1999) Bifidobacterium longum survival during frozen and refrigerated storage as related to pH during growth. J Food Sci 64:714–718 Roy D, Chevalier P, Savoie L (1991) Sugars fermented by Bifidobacterium infantis ATCC27920 in relation to growth and α-galactosidase activity. Appl Microbiol Biotechnol 34: 653–655 Scalabrini P, Rossi M, Spettoli P, Matteuzzi D (1998) Characterization of Bifidobacterium strains for use in soymilk fermentation. Int J Food Microbiol 39: 213–219 Setchell KDR, Cassidy A (1999) Dietary isoflavones: biological effects and relevance to human health. J Nutr 129(Suppl):758S– 767S Tsangalis D, Ashton JF, McGill AEJ, Shah NP (2002) Enzymic transformation of isoflavone phytoestrogens in soymilk by βglucosidase-producing bifidobacteria. J Food Sci 67:3104– 3113 Turner NJ, Thomson BM, Shaw IC (2003) Bioactive isoflavones in functional foods: the importance of gut microflora on bioavailability. Nutr Rev 61:204–213
