Food Research International 42 (2009) 612–621

Contents lists available at ScienceDirect

Food Research International

journal homepage: www.elsevier.com/locate/foodres

Selection of probiotic bacteria for the fermentation of a soy beverage

in combination with Streptococcus thermophilus
C.P. Champagne a,*, J. Green-Johnson b, Y. Raymond a, J. Barrette a, N. Buckley c
a
Food R&D Centre, Agriculture and Agri-food Canada, 3600 Casavant, St. Hyacinthe QC, Canada J2S 8E3
b
Faculty of Science, University of Ontario Institute of Technology, 2000 Simcoe Street North, Oshawa ON, Canada L1H 7K4
c
Canadian Space Agency, 6767 Route de l’Aéroport, Longueuil QC, Canada J3Y 8Y9

a r t i c l e i n f o a b s t r a c t

Article history: Lactobacillus delbrueckii subsp. lactis R0187, Lactobacillus helveticus R0052, Lactobacillus rhamnosus R0011
Received 1 May 2008 and Bifidobacterium longum R0175 were examined for their ability to grow in combination with Strepto-
Accepted 25 December 2008 coccus thermophilus cultures in milk and a laboratory soy beverage (LSB; both standardized to 4.5% pro-
tein and 2.3% fat). Strains R0011 and R0187 did not rapidly acidify the soy beverage despite good growth
rates on soy carbohydrates. The S. thermophilus populations in the LSB were similar to that of milk even
Keywords: though milk had 30% more buffering capacity. In milk but not in soy, symbiosis with respect to acidifi-
Milk
cation rate was observed between S. thermophilus and L. helveticus or B. longum. The populations of L. helv-
Soymilk
Lactobacillus helveticus
eticus in the fermented products were similar in pure cultures or in the presence of the streptococci.
Bifidobacteria However B. longum did not compete well in the mixed culture. Fermentation conditions varied as a func-
Redox level tion of the ability of S. thermophilus strains to acidify media to a pH of 4.65 (between 8 and 24 h). The
Buffering capacity probiotic populations in the mixed culture were influenced by the S. thermophilus strain and by the time
Carbohydrates of fermentation. Variations in growth rates of the bacteria did not appear to be linked to differences in
initial redox or a-amino nitrogen levels. Strain selection enabled the preparation of a mixed starter, pro-
biotic-fermented soy beverage containing 1.1  108 CFU/mL of L. helveticus R0052, which represented
approximately 13% of the total final population.
Crown Copyright Ó 2009 Published by Elsevier Ltd. All rights reserved.

1. Introduction ing products have been launched, particularly in fruit-based drinks

and cereals. Soy is an excellent candidate for such products (De
Up to 93% of consumers believe certain foods have health ben- Valdez & Giori, 1993). A first benefit of soy beverage fermentation
efits that may reduce the risk of disease (Clydesdale, 2004). This is the reduction of its ‘‘beany” flavour (Blagden & Gilliland, 2005;
has created a strong demand for functional foods which are foods Desai, Small, McGill, & Shah, 2002; Stern, Hesseltine, Wang, &
containing specific ingredients that can reduce the risk of having Konishi, 1977). Soy is also considered a good substrate for func-
certain diseases (for example, diarrhoea, cancer, cardiovascular tional foods since fermentation by probiotics has the potential to
conditions). Worldwide, the dairy sector, which is strongly linked (1) reduce the levels of some carbohydrates which can be respon-
to probiotics, is the largest functional food market, accounting for sible for gas production in the intestinal system, (2) increase free
nearly 33% of the broad market, while cereal products have just isoflavone levels (Chien, Huang, & Chou, 2006; Wei, Chen, & Chen,
over 22% (Leatherhead Food International, 2006). Probiotic bacteria 2007) and (3) favour desirable changes in bacterial populations in
are the main ‘‘bioactive ingredients” added to the dairy matrix in the gastrointestinal tract (Benno, Endo, Shiragami, Samaya, & Mits-
order to generate this health benefit. Probiotics are defined as ‘‘live uoka, 1987; Bouhnik et al., 2004). Soy also benefits bone health
microorganisms which, when administered in adequate amounts, (Messina, Gugger, & Alekel, 2001, chap. 5), which is a concern for
provide a health benefit the host” (Araya et al., 2002). Probiotic- ageing people. It is noteworthy that bone health is also of concern
containing drinks are the fastest-growing dairy product in Europe to astronauts. Indeed, it was observed that astronauts exposed to a
and data show that the global market for probiotic functional foods microgravity environment on extended space flights demonstrated
has grown by 19% in recent years and is expected to grow by 5% significant bone mass loss (Lane, LeBlanc, Putcha, & Whitson,
annually between 2006 and 2011 (Nutraingredients-USA, 2007). 1993). Therefore, a food which would benefit bone metabolism
Picking up on the trend in dairy products, new probiotic-contain- would be helpful to the elderly as well as the astronauts in the
space or outpost (Moon, Mars) environments. Furthermore,
* Corresponding author. Tel.: +1 450 768 3238.
reduced activity of the immune system is also a common result
E-mail address: ChampagneC@agr.gc.ca (C.P. Champagne). of ageing and extended space flight (Stein, 2001). Since some

0963-9969/$ - see front matter Crown Copyright Ó 2009 Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.foodres.2008.12.018
 C.P. Champagne et al. / Food Research International 42 (2009) 612–621 613

probiotic bacteria have been shown to enhance immune functions strates were made under standardized conditions. Redox level
(Perdigon, Vintini, Alvarez, Medina, & Medici, 1999), fermentation has also been shown to influence the growth of probiotic bacteria
with probiotics also has the potential of benefiting the elderly or in milk, particularly oxygen-sensitive species (Dave & Shah,
the astronauts on extended microgravity duty. One of the aims of 1997b), however, it is not known if milk or soy beverages have dif-
this study is to develop a fermented soy product containing probi- ferent redox levels. Another objective of this study was to compare
otics with potential as a functional food on Earth and in space. growth of cultures in soy and milk substrates on a standardized
Numerous studies have been done on the growth of lactic cul- protein and fat basis, and to determine the impact of differences
tures in soy beverages (Mital & Steinkraus, 1979). A feature of soy in redox level, buffering capacity, total solids level and carbohy-
fermentation by probiotics is the strain-linked variability of the drate level.
acidification rate. The study by Stern et al. (1977), which included
eight Lactobacillus acidophilus cultures, and similar studies using bif-
idobacteria (Scalabrini, Rossi, Spettoli, & Matteuzzi, 1998; Tsangalis, 2. Materials and methods
Ashton, McGill, & Shah, 2002), revealed sharp differences between
strains in the rate of acid production on carbohydrates found in 2.1. Cultures
soy. This literature shows that for pure cultures, strain selection is
essential to obtain adequate acidification rates. Even then, with The following probiotic strains were graciously supplied by
pure probiotic cultures, fermentation times required to attain a Institute Rosell-Lallemand (Montréal, Canada): S. thermophilus
pH below 4.5 are typically 10 h or more at 37 °C (Angeles & Marth, R0083, Lactobacillus delbrueckii subsp. lactis R0187 (L. lactis
1971; Blagden & Gilliland, 2005; Chien et al., 2006; Garro, De Valdez, throughout the text), L. helveticus R0052, L. rhamnosus R0011 and
& De Giori, 2004; Murti, Lamberet, Bouillanne, Desmazeaud, & Lan- Bifidobacterium longum R0175. S. thermophilus R0083 can also be
don, 1993;Kamaly, 1997; LeBlanc et al., 2004). Industrially, short obtained from Abiasa (St-Hyacinthe, QC Canada). Strains ST5,
fermentation times are preferable in order to increase plant output Y12S and Y24S of S. thermophilus, were taken from the Food Re-
as well as reduce unwanted contaminating microorganisms. Fur- search and Development Center (FRDC) of Agriculture and Agri-
thermore, probiotic cultures alone can generate products with food Canada (St. Hyacinthe, QC, Canada) culture collection. Stock
unpleasant flavours (Macedo, Soccol, & Freitas, 1998). A potential cultures were prepared by mixing MRS-grown (lactobacilli and B.
solution to these two problems is the use of mixed cultures with a longum) or M17-grown (streptococci) cultures with sterile rehy-
yoghurt strain. However, little is known of how typical yoghurt drated skim milk 20% (w/w) and glycerol 20% (w/v) in a 2:5:5 ratio,
starters mixed with probiotics will behave in soy beverages. Pairing placing 1 mL in Cryovials (Wheaton, Milleville NJ, USA) and freez-
of probiotic cultures can be very disadvantageous to some strains ing at 80 °C. The glycerol and milk solutions were previously ster-
(Macedo et al., 1998), and data show that combinations with Strep- ilized, respectively at 121 °C for 10 min and 110 °C for 5 min (real
tococcus thermophilus can be detrimental to L. acidophilus (De Valdez temperature and holding time). Strains R0011, R0052, R0175 and
& Giori, 1993) and bifidobacteria (Murti, Bouillanne, Landon, & Des- R0187 were selected because of their potential to improve health
mazeaud, 1993). Thus, the development of a fermented soy product by modulating the immune system (Easo, Measham, Munroe, &
containing probiotics will require strain selection for ability to grow Green-Johnson, 2002; Wallace, Bradley, Buckley, & Green-Johnson,
in the substrate as well as ability to compete or even establish a syn- 2003; Wood, Keeling, Bradley, Johnson-Green, & Green-Johnson,
ergy between strains. The main probiotic bacteria studied in the 2007), preventing infection (Johnson-Henry et al., 2004; Johnson-
past for growth in soy beverages are L. acidophilus, Lactobacillus fer- Henry et al., 2005) reducing the symptoms of stress (Gareau, Jury,
mentum and the bifidobacteria. Little is known of probiotic Lactoba- MacQueen, Sherman, & Perdue, 2007; Zareie et al., 2006) and pro-
cillus rhamnosus, Lactobacillus helveticus, Lactobacillus delbrueckii ducing bioactive compounds (Fiander, Bradley, Johnson-Green, &
ssp. lactis. This study looked at the growth in soy of various bacteria Green-Johnson, 2005). In clinical trials, these microbes have been
included these lesser-used but promising probiotics and describes used to treat acute gastrointestinal infections (Kocian, 1994), re-
the development of a starter culture that can ferment soy rapidly duce pain and bloating associated with irritable bowel syndrome
while maintaining significant probiotic yields. (Benes, Kretk, & Tompkins, 2006) and maintain remission in pa-
Pure cultures are rarely used in milk fermentations. To produce tients with ulcerative colitis (Haskey & Dahl, 2006). The data sup-
cheese, blends of Lactococcus lactis strains are used, while in yo- port the definition of these strains as probiotics.
ghurt, combinations of streptococci and lactobacilli serve the pur- For the preparation of the inocula, 100 mL of sterile MRS (Difco,
pose. In cheese making, blends of L. lactis strains must be carefully Sparks, USA) and M17 (EMD, Darmstadt, Germany) broths were
selected because of bacteriophage issues as well as potential antag- inoculated with 1 mL of a thawed stock culture and incubated at
onism (Hugenholtz, 1986; Lawrence, Thomas, & Terzaghi, 1976). 37 °C until a pH of 4.5 was reached. The incubation time varied be-
Similarly, pairing of yoghurt cultures must be carried out keeping tween 11 and 16 h as a function of the strain. The MRS medium
in mind that not all S. thermophilus and Lactobacillus bulgaricus was prepared as proposed by the manufacturer, but lactose content
blends are successful (Moon & Reinbold, 1976). While data are of M17 was adjusted to 1% (w/v) instead of the customary 0.5%. For
available on mixed cultures of S. thermophilus and L. bulgaricus in the bifidobacteria, 0.1% of ascorbic acid (BioShop, Burlington ON,
soy beverages (De Valdez & Giori, 1993), this study detailed growth Canada) was added to the MRS broth by adding 1 mL of a sterile
and acidification by probiotic bacteria when combined with S. ther- solution of 10% ascorbic acid (as an antioxidant) to 100 mL of
mophilus strains. pre-sterilized MRS broth.
With the exception of Murti, Lamberet, et al. (1993), there has
been no standardization of protein or total solids when milk and 2.2. Milk and soy beverages
soy beverages are compared as growth substrates (Angeles &
Marth, 1971; Macedo et al., 1998; Mital, Steinkraus, & Naylor, All commercial soy beverages (CSB) contain supplements. For
1974). Protein (Bury, Jelen, & Kimura, 1998) or carbohydrate con- our studies, the commercial product with the fewest supplements
centration (Chang & Stone, 1990) both affect growth of lactic cul- was used: natur-aTM (Nutrisoya, St-Hyacinthe QC, Canada). The la-
tures. Therefore comparisons in un-standardized media raise the bel of natur-aTM reported the following ingredients: filtered water,
concern that observed differences could be due to the chemical soybeans (certified biological OCPP/PRO-CERT Canada), calcium
nature of the ingredients in the matrices and their concentration. carbonate, natural flavours, palmitate of vitamin A, vitamins B2,
In this study, evaluations of mixed cultures in milk and soy sub- D2 and B12. It contained 3.2% protein and 1.6% fat.
614 C.P. Champagne et al. / Food Research International 42 (2009) 612–621

For a second series of assays, a soy beverage was prepared in counts between 1.0 and 3.0  107 CFU/mL (Fig. 1). The freshly inoc-
our laboratory (LSB). Soybeans (250 g) were initially soaked in 1 ulated cell suspension was fractionated in 10 mL portions placed
L of water for 16 h at 25 °C. Subsequently, 190 g of soaked beans into 15 mL conical tubes (Sarstedt, Newton NC, USA) and incubated
in 500 mL of water were ground with a commercial blender unit at 37 °C for up to 24 h. Every 2 h, a tube was analysed for pH and
(Waring, New-Hartford, USA) for 3 min, filtered on Whatman #4 viable counts.
membranes and boiled for 5 min. This LSB contained 4.5% protein With the LSB, the media were prepared as described for CSB
and 2.3% fat. (ascorbic acid, pH). Hundred milliliter of medium (milk or soy) in
To obtain the same protein and fat levels as the LSB, a milk solu- 120 mL bottles were inoculated with 2  107 CFU/mL for the pure
tion was prepared by combining 3.25% milk fat commercial milk cultures and 1  107 CFU/mL of each of the strains for the mixed
(Lactantia Pur Filtre, Parmalat, Victoraville QC, Canada) with skim cultures and incubated at 37 °C. Cultures were sampled periodi-
milk powder (Crino, Agropur, Granby QC, Canada). This powder cally for pH measurement and viable counts. The fermentations
contained 36.4% protein. Finally, the 4.5% protein/2.3% fat-milk were carried out until a pH of 4.65 ± 0.05 was reached. Once this
was prepared by adding 61.9 g of powder to 715 mL of commercial pH was attained, the fermented product was analysed for viable
milk containing 3.27% protein and adding water to make 1 L. This counts, HPLC and a-amino nitrogen.
milk blend was then boiled for 5 min, cooled in an ice bath and
keep at 4 °C until used. 2.5. Analyses
Ascorbic acid was added to the soy or milk media, and MRS
broths used to prepare inocula of B. longum R0175, because it is Viable counts were obtained by spreading a 0.1 mL sample of an
very sensitive to oxygen. appropriate dilution onto the surface of plates of M17 agar, for S.
thermophilus, or MRS agar at pH 5.2 for lactobacilli and B. longum.
2.3. Carbohydrate assimilation Plates were incubated at 37 °C for 48 h in aerobic conditions for
S. thermophilus, L. rhamnosus and L. helveticus. The two other strains
API 50CH and API-zym (Biomérieux, Canada, St-Laurent QC, were incubated in an anaerobic hood containing 85% N2, 5% CO2
Canada) were used respectively to test for the assimilation of sug- and 10% H2.
ars and the presence of specific enzymes related to the use of sug- The protein content of the various soy and milk products was
ars as instructed by the manufacturer. A fresh Petri dish culture of estimated by multiplying the total nitrogen values by a factor of
each strain was prepared 48 h before each of the three indepen- 6.25 for the LSB and 6.38 for the milk solution. Total nitrogen
dent repetitions. determination was done using a FP-428 LECO apparatus (LECO Cor-
The effect of carbohydrate on growth rates was ascertained by poration, Saint-Joseph, MI, USA) operated under the following con-
automated spectrophotometry (AS) as described by Champagne, ditions: sample size, 300 lL; oxidation furnace temperature,
Gaudreau, Chartier, Conway, and Fonchy (1999). The base medium
contained, per L: 10 g yeast extract, 6 g K2HPO4, 6 g KH2PO4, 6 g tri-
sodium citrate, 0.4 g MgSO4
7H2O, 0.1 g MnSO4
H2O, 2 g tween 80
and 20 g of the appropriate carbohydrate (pH 6.5). A solution con-
9.5
taining ascorbic acid (10% p/v) and cysteine (5% p/v) was sterilized
Population (Log CFU/ ml)

by filtration and was added to each minimal medium at 1% (v/v). 9.0

The probiotic cultures (40 mL each) were centrifuged at 5300 g
for 20 min (KA-18.50, Dupont Instruments, Sorvall, Newton, CT). 8.5
The supernatant was removed; the cell pellet was resuspended in
40 mL of sugarless base medium, and then diluted 1/10 in sugar- 8.0
less minimal medium. The growth of the strains was followed on
a 100-well plate (10  10 rows; ThermoElectron Oy, Vantaa, Fin- 7.5
L. helveticus R0052
land) with each well containing 200 lL of base medium w/o carbo-
7.0 S. thermophilus ST5
hydrate (two wells per carbohydrate per strain) inoculated with
B. longum R0175
20 lL of the diluted cultures. Incubation was carried out at 37 °C
6.5 L. lactis R0187
in the spectrophotometer (Bioscreen CTM, Thermo Lab system, Fin-
land). The optical density (OD) at 600 nm was recorded every
15 min for 24 h. Just prior to OD readings the plates were shaken
at medium speed for 10 s. The average growth rate (l average) 7.0
was obtained by determining the slope of the growth curve (Ln val-
ues) between the end of the lag phase and the beginning of the sta- 6.5
tionary growth phase. The AS experimental plan was designed to
have one negative control and two positive controls. The negative 6.0
pH

control was the base medium without any carbohydrate (identified

as ‘‘None” in Table 2) while the first positive control was base med- 5.5
ium with glucose and the second positive control was medium
MRS. In addition to having glucose in the medium, MRS is richer 5.0
in growth factors than the base medium and serves as a good com-
mercial medium. 4.5

2.4. Fermentations
0 5 10 15 20 25 30
The CSB was used for initial strain selection as follows. CSB Time (h)
(150 mL) was placed into 200 mL bottles and the pH was adjusted Fig. 1. Growth and acidification of the cultures in nature-aTM commercial soy
to 6.5 with 1 N HCl or NaOH using a calibrated and sterilized elec- beverage. Error bars are presented on only one plot, for reasons of clarity, and
trode. The inocula were then added at 1% (V/V) which gave initial represent the SEM. Results are the average of two independent assays.
 C.P. Champagne et al. / Food Research International 42 (2009) 612–621 615

900 °C; oxidation standby temperature, 650 °C; purge cycles, 3; the literature (Lin, Chiu, & Pan, 2004; Wang, Yu, & Chou, 2004), al-
minimum timeout, 30 s; comparator level, 1.00; loop select low beit in the high range which may be explained by the high level of
range, flow constants at high: gases, oxygen 99.99% and helium protein and other solids. This study is the first report of compara-
99.99%. The calibration standard was composed of 150 mg EDTA tive buffering ability of soy and milk at identical protein levels. On
(No. 502-092), LECO Corporation, Saint-Joseph, MI, USA). an identical protein basis, the milk formulation had about 30%
The a-amino nitrogen of LSB and milk was determined by titra- higher buffering capacity which could have a significant effect on
tion following reaction with formaldehyde (USP, 1985). The media viable counts. Indeed, the drop in pH usually generates an uncou-
were adjusted to pH 7.0 with 0.1 M NaOH or 0.1 M HCl. pling between growth and acid production (Chou & Hou, 2000;
Lipids were measured by the method of Mojonnier (Marshall, Chumchuere & Robinson, 1999) and media that are more highly
1992). buffered (with higher solids) tend to have higher populations of
For sugar and acid analyses, an HPLC (Waters 510 pump and lactic acid bacteria (Chang & Stone, 1990).
Waters 717-plus auto sampler, Milford, MA, USA) was used under The redox value recorded for fresh LSB was +105 mV (SD = 45),
the following conditions: a 20 lL injection on a separation column while that of the milk preparation was +46 mV (SD = 32): statisti-
ICE-Ion 300 (Transgenomic, Omaha, NE, USA), kept at temperature cal analysis failed to reveal significant differences (P = 0.12). Media
of 40 °C; mobile phase of H2SO4 0.02 N injected at 0.4 mL/min; RI supplemented with ascorbic acid were not analysed, but the redox
detector (Waters 410, Milford, MA, USA) at 40 °C coupled with a would presumably be lower (Dave & Shah, 1997c). The data with
PDA detector (Waters 996, Milford, MA, USA) at a wavelength of respect to milk fall into the range observed in the literature (Bol-
208 nm. The injection solutions were prepared by placing 5.0 mL duc, Raymond, Fustier, Champagne, & Vuillemard, 2006). Although
of sample (milk or LSB) in a 15 mL test tube, adding 2.50 mL of there is one report on the antioxidant properties of soy extracts
H2SO4 0.1 N, mixing for 1 min, incubating at 85 °C in a water bath (Hu et al., 2004) this is the first report demonstrating comparative
for 30 min, cooling to room temperature, centrifuging for 20 min at redox values of milk and soy at similar protein levels. Oxygen con-
3500 rpm (IEC, Centra CP 8R, USA) and filtering on a Millipore filter tent and low redox potential value have been shown to be impor-
(Millex-HV, Bedford, MA, USA) of 0.45 lm in a vial for HPLC. tant factors for growth of Bifidobacterium species (Bolduc et al.,
The buffering capacity of the media was also analyzed in tripli- 2006; Klaver, Kingma, & Weerkamp, 1993), and these analyses
cate. Using a titration unit (TitraLab, TIM 865, Radiometer Analyt- served to examine if strong variations in redox levels could explain
ical, Villeurbanne Cedex, France) the amount of 0.1 N HCl required differences in growth patterns of probiotic bacteria. Our data do
to lower the pH of 10 mL of the media from 6.5 to 4.0 was not link differences in growth rates in the two products to differ-
determined. ences in redox levels of the substrates.
The oxidoreduction potential (Redox) was determined with a The percentage of solids calculated for the LSB was 9.2%, much
platinum electrode Metrohm 6.0402.100 MC (Brinkman lower than the 13.9% found for the standardized milk solution. This
#2092170-6, Westbury NY, USA) connected to a pH meter (Oakton is mostly due to the lower carbohydrate concentration in the LSB at
pH-100, Vernon Hills IL, USA) and calibrated using a 250 mV oxido- only 1.7% versus 6% lactose in milk. Soy beverage carbohydrate
reduction buffer (Brinkmann #2009612-8). After the electrode content affects the acidification level which can be attained (Chang
reading on the fresh unfermented products had been stable for at & Stone, 1990), and it was initially our concern that the low sugar
least 20 min, readings were taken. The oxygen content of the med- level in LSB was insufficient for fermentation. However, data from
ium can strongly influence redox values and vary as a function of the buffering ability of LSB suggest that as little as 0.6% lactic acid
the sample treatment (heating, agitation, nitrogen gassing of the would be sufficient for fermentation to pH 4.0. Thus, a sucrose-po-
medium, vacuum). Products were left for 1 h at room temperature sitive culture would have sufficient carbohydrate assuming a typ-
in an aerobic environment without agitation to simulate industrial ical 90% conversion level by homolactic cultures of the 0.82%
conditions for yoghurt manufacture, where no attempts are made sucrose in the LSB. However, should a strain only be able to assim-
to remove oxygen (except for heat treatment and absence of agita- ilate carbohydrates other than sucrose in the LSB, acidification
tion during the fermentation). would not reach pH 4.0. This explains why carbohydrate supple-
The total solids were determined by dry weight, following an mentation of soy beverages has been shown to improve the growth
incubation of 18 h at 100 °C.1 and acidification of lactic cultures (Chang & Stone, 1990). There-
fore, although there are theoretically ample carbohydrates in LSB
3. Results and Discussion to carry out a fermentation that would attain the pH of yoghurt-
type products (pH 4.2–4.6), the acidification pattern will depend
3.1. Media characteristics on the sugar assimilation properties of the probiotics. Thus, this
property was measured in order to compare growth and acidificat-
The weight of the soybeans increased by a factor of 2.3 during ion of soy beverages by probiotics.
soaking in water. Thus, 690 g of slurry originally contained 83 g
of beans. On a dry bean basis, Chang and Stone (1990) obtained 3.2. Use of carbohydrates by the probiotic cultures
yields similar to ours. Hou, Yu, and Chou (2000) had slightly lower
yields than Chang and Stone (1990) since they obtained 2.9% pro- Milk mostly contains lactose, while soy beverages contain, by
tein from an initial 1:10 bean:water ratio. Therefore, there are vari- order of importance, sucrose, stachyose, raffinose, glucose and
ations in the literature with respect to protein yields, but data fructose (Table 1). Therefore, to grow in milk, the ability to use lac-
suggest that our methodology gave comparable yields and tose in critical, but cultures have a variety of options in soy sub-
composition. strates. It was therefore deemed important to assess the ability
In order to lower the pH of the products from 6.5 to 4.0, 6.26 mL of the strains to use the carbohydrates in order to explain growth
of HCl 0.1 N were necessary for the LSB, while 9.04 mL were patterns in the soy and milk substrates.
needed for the milk formulation. Data for the LSB are in line with Results between the two API kits were generally in agreement
with each other and with reports in the literature (Chien et al.,
2006; Hou et al., 2000; Scalabrini et al., 1998; Stern et al., 1977;
1
The cultures and products used in this study were selected for research purposes
Wei et al., 2007). In the API-zym the only strain having a-galacto-
only. Their selection does not constitute an endorsement of any nature by Agriculture sidase activity was B. longum R0175, while the only strain not
and Agri-Food Canada. showing b-galactosidase was L. rhamnosus R0011. Thus, with the
616 C.P. Champagne et al. / Food Research International 42 (2009) 612–621

Table 1
Residual sugars (g/L) and organic acids (g/L) prior to and after the fermentations of the laboratory soy beverage (LSB) and milk blend.
LSB fermentations Stachyose Raffinose Sucrose Glucose Fructose Lactic acid Acetic acid
LSB non-fermented 8.8 a 2.1 a 9.3 a 1.7 a 1.1 a 0.1 a 0.1 a
LSB S. thermophilus ST5 9.7 a 1.7 a 5.9 bc 0.7 b 0.4 b 4.5 c 0.2 a
LSB L. helveticus R0052 9.3 a 1.5 a 4.6 c 1.0 b 0.4 b 5.2 d 0.4 a
LSB B. longum R0175 0.4 b 0.1 b 8.0 ab 1.1 b 0.7 ab 2.9 b 2.4 b
LSB mix R0052 + ST5 8.9 a 1.5 a 4.5 c 0.7 b 0.3 b 4.7 cd 0.3 a
LSB mix R0175 + ST5 8.7 a 1.5 a 5.9 bc 0.7 b 0.4 b 4.4 c 0.4 a
Milk fermentations Lactose Glucose Galactose Lactic acid Acetic acid
Milk non-fermented 65.4 a 0.2 a 0.4 a 0.6 a 0.9 a
Milk S. thermophilus ST5 59.9 b 0.2 a 3.4 b 3.4 c 1.1 a
Milk L. helveticus R0052 58.2 b 0.1 a 0.5 a 6.5 d 1.1 a
Milk B. longum R0175 63.2 a 0.3 a 0.4 a 1.1 b 0.5 a
Milk mix R0052 + ST5 57.6 b 0.2 a 0.3 a 6.3 d 1.1 a
Milk mix R0175 + ST5 57.8 b 0.2 a 0.3 a 4.2 c 2.2 b

Data are the means of three independent assays.

abc: For a given carbohydrate in a given medium (Milk or LSB), values followed by the same letter are not significantly different (P > 0.05).

exception of L. rhamnosus R0011, cultures which were b-galactosi- are in agreement with the literature. However, the AS method used
dase positive in API-zym were lactose-positive in API50CH (Table in this study provides more information than the simple ‘‘positive
2). The same was noted with a-galactosidase activity and stachy- or negative” use of sugars and was useful in examining the effects
ose/raffinose assimilation. It is noteworthy that all lactobacilli of carbohydrate on growth rates and biomass.
shared the same API carbohydrate assimilation patterns. Except In the base medium, the highest optical densities were always
for the utilisation of sucrose by L. lactis R0187, data from this study obtained with glucose. Biomass levels were lower than those ob-
tained in glucose for L. lactis R0187 on fructose, L. helveticus
Table 2 R0052 on fructose and L. rhamnosus R0011 on fructose, sucrose
Use of carbohydrates by the probiotic bacteria as ascertained by automated or lactose (data not shown). Lower biomass yields as a function
spectrophotometry or APITM. of carbohydrate, particularly fructose, have been reported for Bifi-
Strain Carbohydrate API l (h 1) dobacterium (Kim, Song, Kang, & Oh, 2003; Mlobeli, Gutierrez, &
Maddox, 1998; Perrin, Warchol, Grill, & Scheider, 2001; Shene,
Lactobacillus delbrueckii None NEG
subsp. lactis R0187 Glucose + 0.15 ab
Mardones, Zamora, & Bravo, 2005). However, this is the first report
Stachyose ND NEG of a similar finding with probiotic lactobacilli and it is of interest to
Lactose + 0.13 a see reduced biomass in all the lactobacilli tested on fructose.
Fructose + 0.12 a While the lmax values were unaffected by carbohydrate source
Sucrose + 0.14 ab
(data not shown), differences were noted with respect to average
Raffinose – NEG
Galactose + ND l. The growth rates obtained under our experimental conditions
MRS 0.19 b were lower than those reported for some bifidobacteria (Marti-
Lactobacillus helveticus None NEG nez-Villaluenga & Gomez, 2007; Mlobeli et al., 1998), but similar
R0052 Glucose + 0.13 ab to others (Perrin et al., 2001; Shene et al., 2005). Because l values
Stachyose ND NEG vary so widely in the literature as a function of strains or media
Lactose + 0.15 bc
conditions, their usefulness is best limited to relative comparisons
Fructose + 0.08 a
Sucrose + 0.08 a
within a specific study. Slower growth rates were observed on su-
Raffinose – NEG crose than on lactose or glucose for L. helveticus R0052 and L.
Galactose + ND rhamnosus R0011, respectively (Table 2). Different growth and fer-
MRS 0.19 c mentation rates of probiotics as a function of carbohydrates have
Lactobacillus rhamnosus None NEG been noted in the literature for L. brevis (Zhang & Lovitt, 2006), B.
R0011 Glucose + 0.18 b longum (Mlobeli et al., 1998; Stern et al., 1977), Bifidobacterium
Stachyose ND NEG
infantis (Perrin et al., 2001) and L. delbrueckii subsp. bulgaricus
Lactose + 0.11 ab
Fructose + 0.17 b (Chervaux, Ehrlich, & Maguin, 2000). Thus, data from this study
Sucrose + 0.07 a add L. helveticus and L. rhamnosus to the growth rate literature.
Raffinose – NEG The only discrepancy between the AS and the API results was in
Galactose + ND the use of sucrose by B. longum R0175, which was negative in the
MRS 0.17 b
AS assays and positive in API. Data from the literature suggests that
Bifidobacterium longum None NEG
most B. longum strains do indeed assimilate sucrose (Scalabrini
R0175 Glucose + 0.06 a
Stachyose ND 0.06 a
et al., 1998).
Lactose + 0.07 a
Fructose – NEG 3.3. Commercial soy beverages
Sucrose + NEG
Raffinose + 0.06 a
Two series of assays were carried out on soy beverages. The ini-
Galactose + ND
MRS 0.08 a tial assays in the CSB were designed to rapidly select cultures
which would have the ability to grow well in a soy substrate as
NEG = not enough growth. The l values were not calculated if DO increases were
well as to ascertain the effect of the supplements typically found
lower than 0.3 over 24 h of incubation. ND = not determined. l = Growth rate.
Data are the means of three independent assays. in the commercial soy beverages.
abc: For a given strain, values followed by the same letter are not significantly Because L. rhamnosus R0011 did not acidify the CSB, and the pH
different (P > 0.05). remained above 6.0 after 24 h of incubation, its data are not pre-
 C.P. Champagne et al. / Food Research International 42 (2009) 612–621 617

sented in Fig. 1 with the other strains. Farnworth et al. (2007) re- Contrary to literature that suggest that L. helveticus grows much
ported growth of L. rhamnosus in soy to a population of less in soy extracts compared to milk ( Murti, Lamberet et al.,
2  108 CFU/mL, and that growth was better in a commercial soy 1993), strain R0052 grew well in both (Table 3) making it a good
beverage than in milk. We had also reported previously that L. candidate for the development of probiotic-containing fermented
rhamnosus grew slowly in milk (Gaudreau, Champagne, & Jelen, soy products.
2005), but a discrepancy occurs in the data on soy. This may be In CSB, the most rapid pH drops and rapid growth were ob-
linked to strains used or to slight differences in the composition served with S. thermophilus R0083 and B. longum R0175 (Fig. 1).
of the supplemented CSB. Not all strains were able to acidify the medium to pH 4.5 within
The growth of L. lactis R0187 over 24 h was slow and constant 24 h due to either slow acidification (L. helveticus R0052 and L. lac-
(Fig. 2), which is reflected in the acidification curve which shows tis R0187) or stabilization of pH above 4.5 (B. longum R0175). The
an equally slow acidification rate in CSB, with a pH of 4.8 eventu- inability of B. longum R0175 to further drop the pH was likely
ally reached after 24 h (Fig. 1). This is in agreement with literature due to the exhaustion of all fermentable carbohydrates, as sup-
reports of pH values under 4.5 found in fermented soy extracts ported by our carbohydrate analyses showing B. longum R0175 as
(Blagden & Gilliland, 2005; Pyo, Lee, & Lee, 2005). Data from the being unable to assimilate sucrose, the major carbohydrate of LSB.
API carbohydrate assays (Table 2) showed that this strain could Considerable differences in the ability of B. longum strains to
not use stachyose or raffinose, but could use sucrose. This was pre- grow on soy are found in the literature (Garro, De Valdez, Oliver,
sumably the main substrate used in the fermentation. In nutrient- & De Giori, 1999a, 1999b; Garro, De Valdez, & De Giori, 2001 and
rich media, the growth rate of L. lactis R0187 on sucrose was iden- Garro et al., 2004; Kamaly, 1997; Tsangalis, Ashton, McGill, & Shah,
tical to that found with glucose and l values were in the same 2003; Abd El-Gawad, El-Sayed, Hafez, El-Zeini, & Saleh, 2005;
range as other lactobacilli (Table 2). Therefore the growth and acid- Wang, Yu, & Chou, 2002; Chou & Hou, 2000). Data from this study
ification patterns of L. lactis R0187 in CSB do not appear to be car- suggests that strain R0175 is very well adapted to soy, largely due
bohydrate-linked. A population close to109 cfu/mL was reached to its ability to grow as well on stachyose and raffinose as on glu-
after 24 h, comparable to the values reported by of Pyo et al. cose (Tables 1 and 2). One reason for our observed superior growth
(2005); however, this study also includes acidification and growth may be the inclusion of ascorbic acid in the medium. Actions to im-
profiles throughout the 24 h incubation period. Our results indicate prove the growth and stability of bifidobacteria in milk with anti-
that L. lactis R0187 is a strain with relatively good growth rates on oxidants (Dave & Shah, 1997b), as have been reported for soy
soy carbohydrates, but that its growth in CSB is slower than other beverages (Kamaly, 1997) may explain our findings. The reports
lactobacilli. Other studies will have to ascertain if this is strain or of poor growth of B. longum in soy (Abd El-Gawad et al., 2005; Scal-
species-related. Based on this result, this strain was eliminated abrini et al., 1998; Tsangalis et al., 2002; Wang et al., 2002), may be
from subsequent assays in the mixed cultures. related to unfavourable redox levels of the soy media or the lack of
available nutrients for these strains.
We frequently observed uncoupling of growth and acid produc-
tion when a certain pH value was obtained: pH 5.4 for S. thermophi-
7.0 Milk lus, and pH 5.5 for L. helveticus. To a lesser extent, uncoupling
between growth and acid production also occurred with B. longum
6.5 at pH 4.9. These data suggest that fermentation to the pH of yo-
ghurt, typically around pH 4.3, is unnecessary and does not result
6.0 in the highest probiotic population in the fermented soy products.
In fact, our studies provide evidence that acidification below pH 4.9
pH

5.5 would be unnecessary to ensure maximal probiotic yields. How-

ever, with respect to sensory properties, milk coagulation would
not be advanced enough at pH 4.9 and it was decided to select a
5.0
pH of 4.70 for LSB assays.
These data from the assays on commercial SB indicated that B.
4.5
longum R0175 and L. helveticus R0052 were the best candidates
for growth with S. thermophilus for the final series of
fermentations.
7.0 LSB ST5
R0175 3.4. Growth and acidification in milk
6.5 R052
ST5 and R0175 The S. thermophilus ST5 strain used in these assays did not acid-
6.0 ST5 and R0052 ify milk very fast. Perhaps this is due to the absence of the L. del-
pH

brueckii ssp. bulgaricus strains that are typically used in yogurt

5.5 manufacture. In fact, none of the pure cultures showed a rapid
acidification of milk, although L. helveticus R0052 was superior to
5.0 the other strains (Fig. 2).
In all mixed cultures, streptococci dominated with counts that
4.5 reached 8.7  108 CFU/mL even though S. thermophilus ST5 grew
to slightly greater numbers in milk when in pure cultures as com-
pared to mixed cultures. Mixed cultures resulted in faster acidifi-
0 5 10 15 20 25 cation rates (Fig. 2), suggesting some type of acidifying
Time (h) symbiosis. This has frequently been observed with traditional yo-
gurt starter cultures, but until now, has not been reported between
Fig. 2. Acidification of milk or of the laboratory soy beverage (LSB) by pure or mixed
cultures of Streptococcus thermophilus ST5, Lactobacillus helveticus R0052 and
S. thermophilus and L. helveticus or B. longum. However, this did not
Bifidobacterium longum R0175. Data are means of three independent assays. Error translate in higher probiotic populations. The growth of L. helveti-
bars are SEM. cus R0052, as a pure culture, was three times lower than for the
618 C.P. Champagne et al. / Food Research International 42 (2009) 612–621

Table 3
Effect of fermentation time, probiotic or streptococci cultures as well as strain pairing on the viable counts of fermented milks.

Strains Initial (CFU/mL) Milk Soy

Fermented (CFU/mL) Fermentation time (h) Fermented (CFU/mL) Fermentation time (h)
Pure R0083 S. thermophilus 2.0  107 1.2  109 8 1.1  109 8
Pure ST5 S. thermophilus 2.3  107 8.7  108 24 8.5  108 24
Pure R0175 B. longum 2.3  107 3.6  108 24 4.7  108 24
Pure R0052 L. helveticus 1.9  107 2.8  108 24 2.8  108 24
Mixed S. thermophilus 1.0  107 1.1  109 8 4.9  108 8
R0083 + R0175 B. longum 9.1  106 5.4  107 8 1.0  107 8
Ratio St/Bl 1 20 49
Mixed S. thermophilus 9.6  106 7.4  108 20 1.0  109 20
ST5 + R0175 B. longum 9.1  106 6.8  107 20 4.7  107 20
Ratio St/Bl 1 11 21
Mixed S. thermophilus 1.0  107 1.4  109 8 7.1  108 8
R0083 + R0052 L. helveticus 1.2  107 2.6  108 8 7.4  107 8
Ratio St/Lh 1 5 10

streptococci (approximately 0.5 log CFU), and was largely un- R0175 did not compete well with S. thermophilus ST5. For the most
changed by the presence of the streptococci. The lower lactobacilli part, our finding with mixed cultures in soy media support previ-
population in mixed culture appear simply related to the lower ous reports describing reduced populations of probiotic strains in
growth rates of L. helveticus as compared to S. thermophilus (Ange- mixed cultures with S. thermophilus (Chumchuere & Robinson,
les & Marth, 1971). This was not the case for B. longum R0175, 1999; De Valdez & Giori, 1993; Murti, Bouillanne, et al., 1993).
where the presence of the streptococci was very detrimental to However, there is one report of a mixed culture of S. thermophilus
growth (Table 3). Thus, it appears that the mixed cultures with and B. longum where there is no inhibition of B. longum when inoc-
the increased acidification rate only benefited the streptococci. ulation levels are very low (104 CFU/mL) and the total population
Ascorbic acid was added to milk and soy media used for cultiva- of S. thermophilus in the fermented product remains under
tions of bifidobacteria since its omission led to negligible growth 108 CFU/mL (Wang et al., 2002). This suggests that one way of
but was not used in media for lactobacilli and streptococci, which improving probiotic counts in fermented soy beverages would be
grew well without addition of ascorbic acid. The potential growth to select streptococci capable of producing only a set, limited bio-
promoting effect of this antioxidant on bifidobacteria has to be mass in soy. Commercial yoghurts typically contain 107 CFU/mL of
considered when comparing the results obtained for different probiotics, which enables a 109 CFU/100 mL portion. In our assays,
microorganisms in this study. However, the results suggest that the fermented products with mixed cultures containing L. helveti-
ascorbic acid be considered as a potential additive for the improve- cus R0052 had more than 108 CFU/mL probiotics, more than 10
ment of probiotic preparations. times those of many commercial yoghurts. This suggests that our
selection process was very effective in obtaining fermented prod-
3.5. Growth and acidification in the laboratory soy beverage (LSB) ucts with high densities of probiotics. We recognize that viability
losses commonly occur during storage in yoghurt (Dave & Shah,
All three pure cultures were able to lower the pH below 5.0 1997a) and that it remains to be determined if viability of blends
within 16 h of incubation (Fig. 2). In all cases, acidification was fas- containing L. helveticus R0052 would be maintained on the grocery
ter in the soy beverage than in milk. Although this would suggest shelves. Furthermore the effects of higher populations of probiotic
that acid production was greater in the soy beverage, it could re- bacteria on sensory properties remain to be ascertained.
flect the lower buffering capacity of the soy beverage as compared HPLC analyses (Table 1) basically confirmed what the results of
to that of milk. As opposed to what was observed in milk, the API and AS (Table 2) carbohydrate assimilation tests. The lactoba-
mixed cultures in soy showed the same acidification rate as the cilli and streptococci mainly used sucrose, glucose and fructose as
pure culture of S. thermophilus. Thus, no symbiosis seems to occur carbohydrates, while the bifidobacteria mainly used stachyose and
in the soy product. raffinose. These data support the literature with respect to carbo-
All three cultures grew faster in the soy beverage than in milk hydrate assimilation in LSB by B. longum (Scalabrini et al., 1998)
(Fig. 2). The growth curves (CFU/mL) in the soy beverage corre- and S. thermophilus (Chumchuere & Robinson, 1999; Mital et al.,
spond with data observed for acidification curves (data not 1974) and their mixed cultures (Wang, Yu, Yang, & Chou, 2003),
shown). From a commercial standpoint, an important aspect of and provide new data with respect to L. helveticus. In mixed cul-
the acidification patterns in soy is that a pH of 4.5 could be reached tures in milk, the probiotic bacteria also seem to have used the gal-
well within 10 h of incubation at 37 °C when S. thermophilus ST5 actose excreted by S. thermophilus (Table 1). With respect to
was present. organic acids, significant increases in acetic (Table 1) acid corre-
The S. thermophilus ST5 population in the fermented LSB was lated with high populations of B. longum R0175 (Table 3).
similar to that of milk (Table 3). This was unexpected in view of To determine if the growth of probiotic bacteria is enhanced by
the fact the soy beverage has a lower buffering capacity than milk, the presence of free peptides and amino acids as reported previ-
and that bacterial populations are generally lower in media with ously (Gaudreau et al., 2005; Lê Nguyen, Champagne, Lee, & Goulet,
low buffering capacity (Champagne, St. Gelais, & Audet, 1996). 2003) the levels of a-amino nitrogen in the LSB and milk products
As was the case in milk, L. helveticus R0052 populations were were measured. Milk and LSB shared similar initial a-amino nitro-
largely unaffected by the presence of the streptococci (Table 3). gen content (Table 4) suggesting that differences in growth of lac-
The biggest differences in bacterial populations between milk tic cultures in milk and soy beverages were independent of free
and the soy beverage were observed with B. longum R0175. In a amino compounds concentration. It is possible that the amino acid
pure culture, this strain grew much faster and better in the soy composition and not its concentration, may have been responsible
beverage than in milk (Fig. 2). This was not due to a faster assim- for growth stimulation (Gaudreau et al., 2005; Lê Nguyen et al.,
ilation rate of the carbohydrates (Table 2). Moreover, B. longum 2003) however, this was not assessed. Results show slight in-
 C.P. Champagne et al. / Food Research International 42 (2009) 612–621 619

creases in a-amino nitrogen levels following fermentation, the It was decided to select S. thermophilus R0083 for further assays
highest levels being when L. helveticus was present (Table 4), and on mixed cultures, because this strain had an acidification pattern
particularly in milk, suggesting that proteolysis occurred. This of LSB very similar to that of strain ST5 and because its acidification
may in part explain the apparent symbiosis between S. thermophi- pattern in milk was similar to that in LSB (Fig. 3). Therefore, this
lus and L. helveticus in milk (Fig. 2). These levels might have in- would enable us to obtain data on probiotic populations following
creased during sample storage prior to analysis (Donkor, similar fermentation time.
Henriksson, Vasiljevic, & Shah, 2005 and Donkor, Henriksson, Results show that cultures with S. thermophilus R0083 resulted
Vasiljevic, & Shah, 2007). Data from this study are in line with re- in a significantly higher proportion of B. longum R00175 in fer-
ports (Donkor, Henriksson, Vasiljevic, & Shah, 2005; Murti, Lamb- mented milks and soy as compared to those obtained with strain
eret et al., 1993 and Donkor, Henriksson, Vasiljevic, & Shah, ST5 (Table 3). This also occurred with L. helveticus R0052, but to
2007) of small increases in free nitrogen compounds following fer- a lower extent (data not shown). Therefore, with the strains used,
mentation with probiotic cultures, and lower apparent proteolysis a longer fermentation time slightly improved the content in bifido-
with streptococci versus lactobacilli. With equal total N levels, soy bacteria but did not account alone for the low counts of probiotic
beverages did not have significantly higher amounts of free nitro- bacteria in LSB. It is unclear if growth inhibition of S. thermophilus
gen than milk and it would appear that this might limit growth, as by strain R0083 played a role. In yoghurt, the literature shows that
is the case in milk. Indeed, bifidobacteria biomass was slightly im- stain composition of the starter culture significantly affects the
proved by supplementation with carbohydrates, but the highest evolution of probiotic bacteria during manufacture as well as their
gains were reported when peptones or yeast extracts were added survival during storage (Dave & Shah, 1997a). However, little is
(Chou & Hou, 2000). This suggests that supplementing the soy bev- known of the interactions between S. thermophilus and probiotic
erage with amino acids or peptides in some fashion might be a suc- bacteria in soy. This is the first report of the effect of different S.
cessful strategy to improve probiotic counts in mixed cultures. thermophilus strains on the populations of B. longum cultures in fer-
Our hypothesis that the selection of probiotic strains for their mented soy beverages. With respect to growth rates, the relation-
ability to grow rapidly in soy would enable higher populations of ship seems to be more complex than a simple correlation with
the probiotic strains in the fermented soy products as compared acidification rates. It remains to be seen if the observed reduction
to milk, alas, was not observed (Table 3). The study design was of probiotics in the presence of streptococci is linked to competi-
to try to simulate yoghurt manufacturing practices and, therefore, tion for substrates, production of inhibitor compounds, proteolysis
fermentation was stopped at a set pH rather than after a set incu- or the evolution of redox levels during the fermentation.
bation time. This meant that fermentation in LSB was stopped after With both probiotic cultures, S. thermophilus R0083 counts were
only 8 h as compared to 20 h in milk (Fig. 2). These data suggested reduced by up to 50% in mixed cultures as compared to pure cul-
that the lack of improvement in probiotic population in fermented tures (Table 3). However, with S. thermophilus ST5, counts were
LSB might simply be due to the shorter fermentation time in soy.
To explore this possibility, three other S. thermophilus strains were
examined for their ability to acidify milk and LSB, in order to find
one which would acidify milk as rapidly as soy.
R0083 - Milk
3.6. Fermentations with different S. thermophilus strains 6.5 Y12S - Milk
Y24S - Milk
The four S. thermophilus strains in this study had similar acidifi- 6.0 ST5 - Milk
cation patterns in LSB, but not milk (Fig. 3). The acidification rate in
milk never exceeded that of LSB which suggests that S. thermophi-
pH

5.5
lus is well suited to grow in soy and may be less dependant on
symbiosis with the lactobacilli. Indeed, assays with S. thermophilus
5.0
ST5 showed faster acidification in milk in both instances where
mixed cultures were used, but not in LSB where acidification rates
were identical in the presence of the two probiotic cultures (Fig. 2). 4.5
Although this suggests that soy supports very well the growth of
streptococci, there are many reports of S. thermophilus population
below 108 CFU/mL in soy (Angeles & Marth, 1971; Blagden & Gilli- 0 5 10 15 20 25
R0083 - LSB
land, 2005; Donkor et al., 2005;Garro, De Valdez, Oliver, & De Giori, 6.5 Time (h)
Y12S - LSB
1999a; Wang et al., 2002). This supports the importance of strain
Y24S - LSB
selection for efficient fermentation.
6.0 ST5 - LSB
pH

5.5
Table 4
a-Amino nitrogen (AAN) content (g/L) of the fermented milk or laboratory soy
beverage (LSB). 5.0
Culture Milk LSB
Control (initial) 0.56 a 0.52 a 4.5
S. thermophilus ST5 0.71 bc 0.56 ab
L. helveticus R0052 0.77 c 0.62 b
B. longum R0175 0.62 ab 0.57 ab
0 5 10 15 20 25
S. thermophilus + L. helveticus 0.78 c 0.62 b
S. thermophilus + B. longum 0.72 bc 0.60 ab Time (h)
Data are the means for at least three independent assays. Fig. 3. Acidification of milk or of the laboratory soy beverage (LSB) by four
abc: For a given column, values followed by the same letter are not significantly Streptococcus thermophilus strains. Data are means of three independent assays.
different (P > 0.05). Error bars represent SEM.
620 C.P. Champagne et al. / Food Research International 42 (2009) 612–621

actually higher in the presence of B. longum (Table 3). This shows Acknowledgments
the variability in relationships between streptococci and probiotic
cultures (Dave & Shah, 1997a; Mital et al., 1974; Wang et al., 2002). The authors wish to thank Thomas Tompkins, Vanessa De Carv-
alho and Institut Rosell – Lallemand for providing the cultures and
carrying out strain identification analyses. This work was sup-
4. Conclusions ported by the Ontario Ministry of Agriculture and Food (OMAF).

Scientific studies contribute to knowledge in the academic and References

industrial communities by providing data of a confirmatory nature
and by providing new findings. This experimental plan confirmed Abd El-Gawad, I. A., El-Sayed, E. M., Hafez, S. A., El-Zeini, H. M., & Saleh, F. A. (2005).
The hypocholesterolaemic effect of milk yoghurt and soy–yoghurt containing
numerous data from the literature: (1) soy beverage yields and bifidobacteria in rats fed on a cholesterol-enriched diet. International Dairy
buffer properties, (2) sugar assimilation profiles and enzymatic Journal, 15, 37–44.
activities of probiotic bacteria, (3) the effect of carbohydrate source Angeles, A. G., & Marth, E. H. (1971). Growth and activity of lactic-acid bacteria in
soymilk. 1. Growth and acid production. Journal of Milk and Food Technology, 34,
on biomass yields of streptococci, (4) comparative growth of S.
30–36.
thermophilus, probiotic lactobacilli and B. longum on soy and milk, Araya, M., Morelli, L., Reid, G., Sanders, M. E., Stanton, C., & Pineiro, M. (2002).
(5) the benefit of ascorbic acid supplementation of media on Guidelines for the evaluation of probiotics in food. Joint FAO/WHO Working
Group report on drafting guidelines for the evaluation of probiotics in food,
growth of bifidobacteria, (6) the detrimental effect of streptococci
London (ON, Canada) April 30 and May 1. <ftp://ftp.fao.org/es/esn/food/
of growth of bifidobacteria in fermented soy or milks, (7) sugar wgreport2.pdf>.
assimilation and acid production profiles in milk and soy by S. ther- Benes, K., Kretk, A., & Tompkins, T. A. (2006). A probiotic combination for IBS. A pilot
mophilus and B. longum, (8) a-amino nitrogen levels in fermented clinical study. NutraFoods, 5(1), 20–27.
Benno, Y., Endo, K., Shiragami, N., Samaya, K., & Mitsuoka, T. (1987). Effects of
soy beverages. Furthermore this study provides the following nov- raffinose intake on human fecal microflora. Bifidobacteria Microflora, 6, 59–63.
elties: (1) the comparisons in redox level and buffering power be- Blagden, T. D., & Gilliland, S. E. (2005). Reduction of levels of volatile components
tween soy beverages and milk at equal protein levels, (2) the effect associated with the ‘beany’ flavor in soymilk by lactobacilli and streptococci.
Journal of Food Science, 70, M186–M189.
of carbohydrate source on growth rates and biomass yields of pro- Bolduc, M. P., Raymond, Y., Fustier, P., Champagne, C. P., & Vuillemard, J. C. (2006).
biotic lactobacilli, (3) the symbiosis between S. thermophilus and L. Sensitivity of bifidobacteria to oxygen and redox potential of non-fermented
helveticus or B. longum in milk, (4) the detrimental effect of strep- pasteurized milk. International Dairy Journal, 16, 1038–1048.
Bouhnik, Y., Raskine, L., Simoneau, G., Vicaut, E., Neut, C., Flourie, B., et al. (2004).
tococci of growth of bifidobacteria in fermented soy beverages, The capacity of nondigestible carbohydrates to stimulate fecal bifidobacteria in
(5) sugar assimilation and acid production profiles in milk or soy healthy humans: A double-blind, randomized, placebo-controlled, parallel-
by L. helveticus, (6), the effects of S. thermophilus strain of growth group, dose-response relation study. American Journal of Clinical Nutrition, 80,
1658–1664.
of L. helveticus or B. longum, in a soy beverage.
Bury, D., Jelen, P., & Kimura, K. (1998). Whey protein concentrate as a nutrient
Preparing the fermented soy product with a yoghurt starter in or- supplement for lactic acid bacteria. International Dairy Journal, 8, 149–151.
der to control flavour attributes and shorten the fermentation time Champagne, C. P., Gaudreau, H., Chartier, N., Conway, J., & Fonchy, E. (1999).
Evaluation of yeast extracts as growth media supplements for lactococci and
results in lower populations of probiotics. When mixed cultures con-
lactobacilli using automated spectrophotometry. Journal of General and Applied
taining S. thermophilus are prepared the probiotic fraction of the Microbiology, 45, 17–21.
mixed culture population is often in the 1–30% range, and our results Champagne, C. P., St. Gelais, D., & Audet, P. (1996). Starters produced on whey
fall in this range. This suggests that our strategy to select fast-grow- protein concentrates. Milchwissenschaft, 51, 561–564.
Chang, C. Y., & Stone, M. B. (1990). Effect of total soymilk solids on acid production
ing probiotics in LSB was a success with respect to L. helveticus. by selected Lactobacilli. Journal of Food Science, 55, 1643–1647.
Many manufacturers of probiotic-containing yoghurt guarantee Chervaux, C., Ehrlich, S. D., & Maguin, E. (2000). Physiological study of Lactobacillus
1 billion probiotic cells per portion, which generally amounts to delbrueckii subsp. bulgaricus strains in a novel chemically defined medium.
Applied and Environmental Microbiology, 66, 5306–5311.
107 CFU/mL. Data from this study show that much higher levels Chien, H. L., Huang, H. Y., & Chou, C. C. (2006). Transformation of isoflavone
could be reached if the probiotic strains are inoculated at the same phytoestrogens during the fermentation of soymilk with lactic acid bacteria and
level as the starter, and at the beginning of an 8 h fermentation. bifidobacteria. Food Microbiology, 23, 772–778.
Chou, C. C., & Hou, J. W. (2000). Growth of bifidobacteria in soymilk and their
In the past, many probiotic cultures were selected for their abil- survival in the fermented soymilk during storage. International Journal of Food
ity to grow and survive during storage in yoghurt. Now, the tech- Microbiology, 56, 113–121.
nological properties of the probiotic strains are becoming less Chumchuere, S., & Robinson, R. K. (1999). Selection of starter cultures for the
fermentation of soya milk. Food Microbiology, 16, 129–137.
important than their demonstrated health effects. In this situation,
Clydesdale, F. (2004). IFT expert report. Food Technology, 58(12), 35–40.
specific probiotic strains are targeted and the company must de- www.ift.org.
velop a process that in turn, enable their successful use. In the past, Dave, R. I., & Shah, N. P. (1997a). Viability of yoghurt and probiotic bacteria in
yoghurts made from commercial starter cultures. International Dairy Journal, 7,
the probiotic culture had to be adapted to the technological starter,
31–41.
but this may now change. If a company wished to use a given pro- Dave, R. I., & Shah, N. P. (1997b). Effect of cysteine on the viability of yoghurt and
biotic strain, in the aim of obtaining demonstrated health effects, probiotic bacteria in yoghurts made with commercial starter cultures.
then data from this study suggest that the selection strategy would International Dairy Journal, 7, 537–545.
Dave, R. I., & Shah, N. P. (1997c). Effectiveness of ascorbic acid as an oxygen
have to shift towards the selection of a compatible starter whether scavenger in improving viability of probiotic bacteria in yoghurts made with
the substrate is milk or soy. commercial starter cultures. International Dairy Journal, 7, 435–443.
One of the goals of this project was to develop functional foods De Valdez, G. F., & Giori, G. S. (1993). Effectiveness of soy milk as food carrier for
Lactobacillus acidophilus. Journal of Food Protection, 56, 320–322.
for astronauts, in order to answer problems related to bone mass Desai, A., Small, D., McGill, A. E. J., & Shah, N. P. (2002). Metabolism of raffinose and
and immune dysfunction (Sonnenfeld, Butel, & Shearer, 2003). Pro- stachyose in reconstituted skim milk and of n-hexanal and pentanal in soymilk
biotics cells as well as soy fermented by probiotics can potentially by bifidobacteria. Bioscience Microflora, 21, 245–250.
Donkor, O. N., Henriksson, A., Vasiljevic, T., & Shah, N. P. (2005). Probiotic strains as
address both these problems (Chien et al., 2006; Ouwehand, Kirj- starter cultures improve angiotensin-converting enzyme inhibitory activity in
avainen, Shortt, & Salminen, 1999; Tang, Shah, Wilcox, Walker, & soy yogurt. Journal of Food Science, 70, M375–M381.
Stojanowska, 2007; Tsangalis & Shah, 2004; Wei et al., 2007), Donkor, O. N., Henriksson, A., Vasiljevic, T., & Shah, N. P. (2007). a-Galactosidase and
proteolytic activities of selected probiotic and dairy cultures in fermented
which could provide functional foods both on Earth and in space.
soymilk. Food Chemistry, 104, 10–20.
In order to further explore the potential of fermented soy bever- Easo, J., Measham, J. D., Munroe, J., & Green-Johnson, J. M. (2002).
ages for astronaut use, assays are planned to examine the function- Immunostimulatory actions of lactobacilli: Mitogenic induction of antibody
production and proliferation by Lactobacillus bulgaricus and Lactobacillus
ality of such fermented milks on immune functions as well as to
acidophilus. Food and Agricultural Immunology, 14, 73–83.
ascertain if fermentations are affected by microgravity.
 C.P. Champagne et al. / Food Research International 42 (2009) 612–621 621

Farnworth, E. R., Mainville, I., Desjardins, M. P., Gardner, N., Fliss, I., & Champagne, C. Mital, B. K., & Steinkraus, K. H. (1979). Fermentation of soy milk by lactic acid
P. (2007). Growth of probiotic bacteria and bifidobacteria in a soy yogurt bacteria. A review. Journal of Food Protection, 42, 895–899.
formulation. International Journal of Food Microbiology, 116, 174–181. Mital, B. K., Steinkraus, K. H., & Naylor, H. B. (1974). Growth of lactic acid bacteria in
Fiander, A., Bradley, S., Johnson-Green, P. C., & Green-Johnson, J. M. (2005). Effects of soy milks. Journal of Food Science, 39, 1018–1022.
lactic acid bacteria and fermented milks on eicosanoid production by intestinal Mlobeli, N. T., Gutierrez, N. A., & Maddox, I. S. (1998). Physiology and kinetics of
epithelial cells. Journal of Food Science, 70, M81–M86. Bifidobacterium bifidum during growth on different sugars. Applied Microbiology
Gareau, M. G., Jury, J., MacQueen, G., Sherman, P. M., & Perdue, M. H. (2007). and Biotechnology, 50, 125–128.
Probiotic treatment of rat pups normalises corticosterone release and Moon, N. J., & Reinbold, G. W. (1976). Commensalism and competition in mixed
ameliorates colonic dysfunction induced by maternal separation. Gut, 56, cultures of Lactobacillus bulgaricus and Streptococcus thermophilus. Journal of
1522–1528. Milk and Food Technology, 39, 337–341.
Garro, M. S., De Valdez, G. F., & De Giori, G. S. (2001). Application of conductimetry Murti, T. W., Bouillanne, C., Landon, M., & Desmazeaud, M. J. (1993). Bacterial
for evaluation of lactic starter culture in soymilk. Journal of Food Science, 67, growth and volatile compounds in yoghurt-type products from soymilk
1175–1178. containing Bifidobacterium ssp. Journal of Food Science, 58, 153–157.
Garro, M. S., De Valdez, G. F., & De Giori, G. S. (2004). Temperature effect on the Murti, T. W., Lamberet, G., Bouillanne, C., Desmazeaud, M. J., & Landon, M. (1993).
biological activity of Bifidobacterium longum CRL 849 and Lactobacillus Croissance des lactobacilles dans l’extrait de soja. Effets sur la viscosité, les
fermentum CRL 251 in pure and mixed cultures grown in soymilk. Food composés volatils et la protéolyse. Science des Aliments, 13, 491–500.
Microbiology, 21, 511–518. Nutraingredients-USA (2007). <http://www.nutraingredients-usa.com> Accessed
Garro, M. S., De Valdez, G. F., Oliver, G., & De Giori, G. S. (1999a). Starter culture 16.07.07 and Datamonitor report <http://www.nutraingredients-usa.com>
activity in refrigerated fermented soymilk. Journal of Food Protection, 62, Accessed 09.02.07.
808–810. Ouwehand, A. C., Kirjavainen, P. V., Shortt, C., & Salminen, S. (1999). Probiotics:
Garro, M. S., De Valdez, G. F., Oliver, G., & De Giori, G. S. (1999b). Hydrolysis of soya Mechanisms and established effects. International Dairy Journal, 9, 43–52.
milk oligosaccharides by Bifidobacterium longum CRL 849. Z Lebensmittel Unters Perdigon, G., Vintini, E., Alvarez, S., Medina, M., & Medici, M. (1999). Study of the
Forschung A, 208, 57–59. possible mechanisms involved in the mucosal immune system activation by
Gaudreau, H., Champagne, C. P., & Jelen, P. (2005). The use of crude cellular extracts lactic acid bacteria. Journal of Dairy Science, 82, 1108–1114.
of Lactobacillus delbrueckii ssp. bulgaricus 11842 to stimulate growth of a Perrin, S., Warchol, M., Grill, J. P., & Scheider, F. (2001). Fermentations of fructo-
probiotic Lactobacillus rhamnosus culture in milk. Enzyme and Microbial oligosaccharides and their components by Bifidobacterium infantis ATCC 15697
Technology, 36, 83–90. on batch culture in semi-synthetic medium. Journal of Applied Microbiology, 90,
Haskey, N., & Dahl, W. J. (2006). Synbiotic therapy: A promising new adjunctive 859–865.
therapy for ulcerative colitis. Nutrition Reviews, 64(3), 132–138. Pyo, Y. H., Lee, T. C., & Lee, Y. C. (2005). Enrichment of bioactive isoflavones in
Hou, J. W., Yu, R. C., & Chou, C. C. (2000). Changes in some components of soymilk soymilk fermented with b-glucosidase-producing lactic acid bacteria. Food
during fermentation with bifidobacteria. Food Research International, 33, Research International, 38, 551–559.
393–397. Scalabrini, P., Rossi, M., Spettoli, P., & Matteuzzi, D. (1998). Characterization of
Hu, C. C., Hsiao, C. H., Huang, S. Y., Fu, S. H., Lai, C. C., Hong, T. M., et al. (2004). Bifidobacterium strains for use in soymilk fermentation. International Journal of
Antioxidant activity of fermented soybean extract. Journal of Agricultural and Food Microbiology, 39, 213–219.
Food Chemistry, 52, 5735–5739. Shene, C., Mardones, M., Zamora, P., & Bravo, S. (2005). Kinetics of Bifidobacterium
Hugenholtz, J. (1986). Population dynamics of mixed starter cultures. Netherlands longum ATCC 15707 fermentations: Effect of the dilution rate and carbon
Milk and Dairy Journal, 40, 129–140. source. Applied Microbiology and Biotechnology, 67, 623–630.
Johnson-Henry, K. C., Mitchell, D. J., Avidzur, Y., Galindo-Mata, E., Jones, N. L., & Sonnenfeld, G., Butel, J. S., & Shearer, W. T. (2003). Effects of space flight on the
Sherman, P. M. (2004). Probiotics reduce bacterial colonization and gastric immune system. Review Environmental Health, 18, 1–17.
inflammation in H. Pylori-infected mice. Digestive Diseases and Sciences, 49(7/8), Stein, T. P. (2001). Nutrition in the space station era. Nutrition Research Reviews, 14,
1095–1102. 87–117.
Johnson-Henry, K. C., Nadjafi, M., Avitzur, Y., Mitchell, D. J., Ngan, B. Y., Galindo- Stern, N. J., Hesseltine, C. W., Wang, H. L., & Konishi, F. (1977). Lactobacillus
Mata, E., et al. (2005). Amelioration of the effects of Citrobacter rodentium acidophilus utilization of sugars and production of a fermented soybean
infection in mice by pretreatment with probiotics. The Journal of Infectious product. Canadian Institute of Food Science and Technology Journal, 10, 197–200.
Diseases, 191, 2106–2117. Tang, A. L., Shah, N. P., Wilcox, G., Walker, K. Z., & Stojanowska, L. (2007).
Kamaly, K. M. (1997). Bifidobacteria fermentation of soybean milk. Food Research Fermentation of calcium-fortified soymilk with Lactobacillus: Effects on calcium
International, 30, 675–682. solubility, isoflavone conversion, and production of organic acids. Journal of
Kim, T. B., Song, S. H., Kang, S. C., & Oh, D. K. (2003). Quantitative comparison of Food Science, 72(9), M431–M436.
lactose and glucose utilization in Bifidobacterium longum cultures. Biotechnology Tsangalis, D., Ashton, J. F., McGill, A. E. J., & Shah, N. P. (2002). Enzymatic
Progress, 19, 672–675. transformation of isoflavone phytoestrogens in soymilk by b-glucosidase-
Klaver, F. A. M., Kingma, F., & Weerkamp, A. H. (1993). Growth and survival of producing bifidobacteria. Journal of Food Science, 67, 3104–3113.
bifidobacteria in milk. Netherlands Milk Dairy Journal, 47, 151–164. Tsangalis, D., Ashton, J. F., McGill, A. E. J., & Shah, N. P. (2003). Biotransformation of
Kocian, J. (1994). Lactobacili in the treatment of dyspepsia in dysmicrobia of various isoflavones by bifidobacteria in fermented soymilk supplemented with D-
aetiology. Internal Medicine, 40, 79–83. glucose and L-cysteine. Journal of Food Science, 68, 623–631.
Lane, H. W., LeBlanc, A. D., Putcha, L., & Whitson, P. A. (1993). Nutrition and human Tsangalis, D., & Shah, N. P. (2004). Metabolism of oligosaccharides and aldehydes
physiological adaptations to space flight. American Journal of Clinical Nutrition, and production of organic acids in soymilk by probiotic bifidobacteria.
58, 583–588. International Journal of Food Science and Technology, 39, 541–554.
Lawrence, R. C., Thomas, T. D., & Terzaghi, B. E. (1976). Reviews of the progress of USP XXI (1985). Protein hydrolysate injection. The US Pharmacopiae, United States
dairy science: Cheese starters. Journal of Dairy Research, 43, 141–193. Pharmacopeia Convention.
Lê Nguyen, T., Champagne, C. P., Lee, B. H., & Goulet, J. (2003). Growth of Wallace, T., Bradley, A., Buckley, N., & Green-Johnson, J. (2003). Interactions of lactic
Lactobacillus paracasei ssp. paracasei on tofu whey. International Journal of Food acid bacteria with human intestinal epithelial cells: Effects on cytokine
Microbiology, 89, 67–75. production. Journal of Food Protection, 66, 466–472.
Leatherhead Food International (2006). The international market for functional Wang, Y. C., Yu, R. C., & Chou, C. C. (2002). Growth and survival of bifidobacteria and
foods (3rd ed.). Functional Food Market Report. ISBN: 1 904007-82-1. lactic acid bacteria during the fermentation and storage of cultured soymilk
LeBlanc, J. G., Garro, M. S., Silvestroni, A., Connes, C., Piard, J.-C., Sesma, F., et al. drinks. Food Microbiology, 19, 501–508.
(2004). Reduction of a-galactooligosaccharides in soymilk by Lactobacillus Wang, Y. C., Yu, R. C., & Chou, C. C. (2004). Viability of lactic acid bacteria and
fermentum CRL 722: In vitro and in vivo evaluation of fermented soymilk. Journal bifidobacteria in fermented soymilk after drying, subsequent rehydration and
of Applied Microbiology, 97, 876–881. storage. International Journal of Food Microbiology, 93, 209–217.
Lin, F. M., Chiu, C. H., & Pan, T. M. (2004). Fermentation of a milk–soymilk and Wang, Y. C., Yu, R. C., Yang, H. Y., & Chou, C. C. (2003). Sugar and acid content in
Lycium chinense Miller mixture using a new isolate of Lactobacillus paracasei soymilk fermented with lactic acid bacteria alone or simultaneously with
subsp. paracasei NTU101 and Bifidobacterium longum. Journal of Industrial bifidobacteria. Food Microbiology, 20, 333–338.
Microbiology and Biotechnology, 31, 559–564. Wei, Q. K., Chen, T. R., & Chen, J. T. (2007). Using of Lactobacillus and Bifidobacterium
Macedo, R. F., Soccol, C. R., & Freitas, R. J. S. (1998). Production of low cost beverage to product the isoflavone aglycones in fermented soymilk. International Journal
with soya milk, Buffalo cheese whey and cow milk fermented by Lactobacillus of Food Microbiology, 117, 121–124.
casei Shirota and Bifidobacterium adolescentis. Journal of Scientific & Industrial Wood, C., Keeling, S., Bradley, S., Johnson-Green, P., & Green-Johnson, J. M. (2007).
Research, 57, 679–685. Vasoactive intestinal peptide modulates the down-regulatory action of
Marshall, R. T. (1992). Standard methods for the examination of dairy products. Lactobacillus rhamnosus on LPS-induced interleukin-8 production by intestinal
Washington: American Public Health Association. pp 474–479. epithelial cells. Microbial Ecology in Health and Disease, 19, 191–200.
Martinez-Villaluenga, M., & Gomez, R. (2007). Characterization of bifidobacteria as Zareie, M., Johnson-Henry, K., Jury, J., Yang, P. C., Ngan, B. Y., McKay, D. M., et al.
starters in fermented milk containing raffinose family of oligosaccharides from (2006). Probiotics prevent bacterial translocation and improve intestinal barrier
lupin as prebiotic. International Dairy Journal, 17, 116–122. function in rats following chronic psychological stress. Gut, 55, 1–8.
Messina, M., Gugger, E. T., & Alekel, D. L. (2001). Soy protein, soy isoflavones and Zhang, D., & Lovitt, R. W. (2006). Performance assessment of malolactic fermenting
bone health: A review of the animal and human data. In R. E. C. Wildman (Ed.), bacteria Oenococcus oeni and Lactobacillus brevis in continuous culture. Applied
Handbook of nutraceuticals and functional foods (pp. 77–98). Boca Raton: CRC Microbiology and Biotechnology, 69, 658–664.
Press.