Food Chemistry 141 (2013) 1716–1721

Contents lists available at SciVerse ScienceDirect

Food Chemistry
journal homepage: www.elsevier.com/locate/foodchem

Acid induced gelation of soymilk, comparison between gels prepared

with lactic acid bacteria and glucono-d-lactone
A. Grygorczyk, M. Corredig ⇑
Department of Food Science, University of Guelph, Guelph, Ontario, Canada N1K 2W5

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

Article history: The objective of this work was to compare the gelation of soymilk particles induced by the acidification of
Received 16 November 2012 a commercial starter culture with that resulting by addition of glucono-d-lactone (GDL). Structure
Received in revised form 24 March 2013 formation was followed using rheology, and the microstructure was observed by confocal microscopy.
Accepted 25 March 2013
Acidification of lactic acid bacteria resulted in a higher gelation pH (pH 6.29 ± 0.05) compared to that
Available online 25 April 2013
of a gel induced by GDL (pH 5.9 ± 0.04). This difference was attributed to the longer time available
for rearrangements of the soymilk particles in soymilk with starter cultures compared to the fast
Keywords:
acidification by GDL. In spite of the earlier gelation pH, there were no observed differences in the final
Soy protein
Lactic acid bacteria
gel stiffness measured at pH 5.1, the value of tan d, the frequency dependence and the linear viscoelastic
Glucono-d-lactone range of the gels measured at the final pH. Microstructural observations also showed a similar protein
Soymilk network structure.
Acid induced gel Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction curds are prepared, acidification is usually carried out using

glucono-d-lactone (GDL), as the pH decreases gradually, creating a
In addition to the use of soymilk as a base for traditional soy homogeneous gel (Malaki Nik, Alexander, Poysa, Woodrow, &
based products such as soy beverages and tofu, soymilk is Corredig, 2011). Unlike in heat-induced gelation of soy protein
increasingly employed as a protein matrix to design novel food isolates (Renkema & van Vliet, 2002), except for the bonds that
products. For this reason, a better understanding of the details of develop during heating, covalent bonds do not play a major role in
its colloidal stability and gelation behaviour are needed. Soymilk the network formation of acid-induced gels. The driving forces
is a beverage produced by grinding soaked soybeans with water, behind acid gelation of soy proteins are non-covalent in nature
followed by cooking around boiling temperatures for around 15 min (Kohyama, Sano, & Doi, 1995), and include salt bridging (Zhang,
and removal of insoluble fibre (okara) by filtration or centrifugation Liang, Tian, Chen, & Subirade, 2012) and short range interactions
(Canabady-Rochelle, Sanchez, Mellema, & Banon, 2009; such as hydrogen bonding and Van der Waals forces (Ringgenberg,
Prabhakaran, Perera, & Valiyaveettil, 2006). Heating of soymilk is Alexander, & Corredig, 2013).
an essential step to denature anti-nutritional compounds, and to As the pH of soy protein suspensions decreases to values around
modify the structure of soymilk particles to improve their colloidal 6, it has been observed that the basic subunits of glycinin and the b
stability and decrease their size distribution to a few micrometre in subunit of b-conglycinin are the first to destabilise. As the system
diameter (Malaki Nik, Tosh, Woodrow, Poysa, & Corredig, 2009). It approaches a net neutral charge, the acidic subunits of glycinin
has also been shown that soy protein denaturation induced by and the a and a0 subunits of b-conglycinin then also begin to
heating is a necessary step for gel formation (Renkema & Van Vliet, participate in gel network formation (Ringgenberg et al., 2013).
2002). During heating, soluble protein aggregates, composed of The use of lactic acid bacteria to produce soymilk curds has
acidic and basic polypeptides of glycinin linked via disulfide bonds been evaluated in the past (Liu et al., 2009; Mital & Steinkraus,
form, as well as a small amount of a and a0 subunits of b-conglycinin 1975). Lactic acid bacteria are known to primarily ferment sucrose
(Ren, Tang, Zhang, & Guo, 2009). These aggregates, of submicron in soy products. However, some lactic acid cultures are also capa-
sizes, interact by hydrophobic interactions and hydrogen bonding ble of fermenting other low molecular weight carbohydrates found
to make protein particles with basic subunits of glycinin in the in soybeans, such as raffinose and stachyose (Mital & Steinkraus,
interior and the acidic glycinin subunits and a and a0 subunits of 1975). Although several studies have focused on lactic acid
b-conglycinin on the exterior (Ren et al., 2009). When acid bacteria metabolism of soy products (Mital & Steinkraus, 1975;
Mital, Steinkraus, & Naylor, 1974), and on large deformation
⇑ Corresponding author. Tel.: +1 (519) 824 4120x56101; fax: +1 (519) 824 6631. rheological properties and microstructure of final gel products
E-mail address: [email protected] (M. Corredig).
(Donkor, Henriksson, Vasiljevic, & Shah, 2007; Ghosh, Chattoraj,

0308-8146/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.foodchem.2013.03.096
 A. Grygorczyk, M. Corredig / Food Chemistry 141 (2013) 1716–1721 1717

& Chattopadhyay, 2011; Yang, Fu, & Li, 2012), no data exist in the light scattering (Mastersizer S; Malvern Instruments Inc.,
literature regarding the formation of the structure of soymilk acid- Southborough, MA), using 1.46 as the refractive index of the soy-
gels produced using lactic acid bacteria. milk particles and 1.333 as the refractive index of the dispersant
Several studies exist on the acid-induced aggregation of soymilk (water) as previously reported (Malaki Nik et al., 2009). The ratio
particles acidified with GDL (Kohyama et al., 1995; Malaki Nik of 11S:7S protein in the soy flour and soymilk was determined
et al., 2011; Tay & Perera, 2006). The gel pH ranges between 5.6 using SDS–PAGE, using conditions published in the literature
and 5.8 depending on the variety of soybean used, and different (Keerati-u-rai & Corredig, 2009). Gel analysis was carried out using
concentrations of GDL do not result in differences in the gelation a Gel Doc™ EZ Imager (Bio-Rad, Mississauga, ON, Canada). Crude
pH (Malaki Nik et al., 2011). However, acidification using lactic oil content of the soybeans was determined using the soxhlet oil
acid bacteria is known to be far slower than GDL acidification extraction method with petroleum ether (AOAC method 945.16)
(Lucey, Tamehana, Singh, & Munro, 1998). with a Labline multi-unit extraction heater (Barnstead Labline;
In consideration of the complex composition of soymilk parti- Thermo Scientific, Asheville, NC). The fat content of soymilk was
cles, it may be hypothesised that allowing more time for protein determined using the Babcock method (AOAC method 989.04,
rearrangements to occur may influence the gelation behaviour of 2000) with a Babcock System for Fat Analysis (Cole-Parmer, Mon-
soymilk. Although studies exist comparing the acidification of treal, QC, Canada) as previously reported in the literature (Buono,
cow’s milk using GDL or lactic acid cultures (Lucey et al., 1998), Erickson, Fung, & Jeon, 1990). Mineral analysis was determined
no such studies are reported for soymilk. Thus the purpose of this using inductively coupled plasma optical emission spectroscopy
study was to evaluate the difference in the gelation behaviour of by the advanced analytical laboratories at the University of Guelph
soymilk particles during acidification induced with GDL or bacte- (Guelph, Ontario, Canada).
rial culture.
2.3. Gelation experiments

2. Materials and methods

Soymilk was acidified at 40 °C with either 0.6% glucono-d-lac-
tone (GDL) (Sigma–Aldrich Co., St. Louis, MO) or YO-MIX™ 511
2.1. Soymilk preparation
LYO 375 DCU (Danisco Canada Inc., Scarborough, ON, Canada).
The starter culture was pre-diluted by adding 0.2 g of the freeze-
Soymilk was prepared using food grade materials with a proce-
dried starter culture to 30 mL of warm soymilk (40 °C) and stirring
dure that would resemble that of a household process. Soybeans
for 30 s before adding 157 lL of the diluted culture to 30 mL of
were obtained from a local grocery store and characterised as
soymilk sample for a final concentration of 0.00349% starter cul-
described in the following section. A portion of 175 g of beans
ture. Yo-mix is a starter culture commonly employed in the
was washed with filtered water (BritaÒ Faucet Filtration System
fermentation of milk for yoghourt production, containing Strepto-
(Model FF-100), Brita Canada Corp., Brampton, ON, Canada) and
coccus thermophilus and Lactobacillus delbrueckii subsp. Bulgaricus.
soaked overnight in water. The hydrated soybeans were rinsed
Soymilk was pre-warmed at 40 °C for 5 min. After addition of
once again with water and placed into a household soymilk maker
GDL or bacterial culture, the soymilk was mixed for 30 s and then
(Soyquick™ Premier Milk Maker Model SQ930P, Kitchen’s Best
immediately placed in the rheometer.
Manufacturing Group Ltd., Nanaimo, BC, Canada) with 933 mL of
Aliquots of the same samples were kept at 40 °C in a circulating
filtered water to produce 4% protein soymilk. The soymilk maker
water bath to measure pH in parallel to the rheology experiments.
cycles involved the following steps: soybeans and water were
The pH was recorded on line using an Accumet AR15 pH metre
heated to 80 °C (in approximately 5 min) and once the temperature
(Fisher Sci., Mississauga, ON, Canada) connected to a computer,
was reached, the soybeans were ground for 5 s. After this short
using AR15 pH recorder software (Mediavention Engineering Inc.,
grinding cycle, the soymilk temperature was brought up to just be-
Guelph, ON, Canada). Gel formation was followed using an
low boiling temperature (1 min) and then four grinding cycles
Advanced Rheometer AR 1000 with Rheology Advantage Instru-
were performed, each cycle lasting 40 s with a 5-s pause between
ment Control AR software v5.4.0 (TA Instruments Ltd., New Castle,
cycles. After grinding, the soymilk maker continued to hold the
DE) at a constant strain of 0.01, and a frequency of 1 Hz. The
soymilk just below boiling temperature for 10 min. The hot
temperature was controlled with an external water bath and kept
soymilk was then immediately poured through a strainer (Kitch-
at 40 °C. Once the samples reached a pH of 5.1, a frequency sweep
en’s Best Manufacturing Group Ltd., Vancouver, BC, Canada) to
was performed, using a constant strain of 0.01 and frequencies of
remove the okara, and then passed twice through a cheese cloth
0.01–10 Hz. Finally, a strain sweep was carried out at a frequency
(Kitchen’s Best Manufacturing Group Ltd, Vancouver, BC, Canada).
of 1 Hz and oscillation stress of 0.5–150 Pa, to determine a yield
After filtration the soymilk was cooled to refrigeration
strain, defined as the value of strain at which the elastic modulus
temperatures.
deviated by 10% from its value in the linear viscoelastic range.

2.2. Soybean and soymilk characterisation 2.4. SDS–PAGE of fermented soymilk

To determine the amount of protein as well as the polypeptide Samples of fermented soymilk were collected at pH 5.8 and 5.5
distribution of the soybeans, the beans were milled with a grain and analysed by SDS–PAGE to determine if there was any protein
mill (IKAÒ Works Inc., Wilmington, NC) and the ground flour was hydrolysis occurring during fermentation. Fermented soymilk
analysed using the Dumas combustion method in a LECO FP-528 was analysed for proteolysis at pH values soon after the gel point,
(Leco Corp., St. Joseph, MI) with 6.25 as a conversion factor for % since if proteolysis was occurring at an extent sufficient to alter the
protein from % N (AACC method 46–30.01, 1999). The protein con- gel point, it is expected that proteolysis would have been detect-
centration in soymilk was also measured with Dumas, while solids able by such pH values. The method employed has been previously
were determined by weighing 1 mL soymilk in dry aluminium described others (Keerati-u-rai & Corredig, 2009). The gels were
pans containing dry sand (Fisher Sci., Mississauga, ON, Canada) scanned using a Bio-Rad Gel Doc™ EZ Imager (Bio-Rad, Mississau-
as a dispersing agent. The pans were placed into an IsoTemp forced ga, ON, Canada). The protein composition of fermented soymilk
air oven (Fisher Sci., Mississauga, ON, Canada) at 105 °C for 24 h. was compared to that of unfermented soymilk and of unfermented
Particle size distribution of the soymilk was determined using laser soymilk incubated at 40 °C for 2.5 h.
1718 A. Grygorczyk, M. Corredig / Food Chemistry 141 (2013) 1716–1721

2.5. Gel microstructure Woodrow, & Corredig, 2008) that after centrifugation at 8000g,
soymilk exhibited a monomodal distribution with an average par-
Images of the gel microstructure were taken using an inverted ticle size of 0.2 lm. The distribution was mostly due to soymilk
confocal scanning laser microscope (Leica TCS SP2, model Leica protein particles. In the present work centrifugation was not car-
DM IRE2, Leica Microsystems CMS GmbH, Mannheim, Germany) ried out, hence fat globules were present, together with larger pro-
with an Ar/Kr visible light laser and 63 (oil) objective. Resolution tein particles and some cell wall debris. It was indeed
of the acquired digital images was 1024  1024 pixels. A fluores- demonstrated that the centrifugation step decreases the amount
cence dye (rhodamine B) (Fisher Sci., Fairlawn, NJ) (excitation of total solids and causes creaming of the fat globules in soymilk
and emission wavelengths 543 and 625 nm, respectively) was used without decreasing significantly the amount of protein (Malaki
for staining the protein. Twenty microlitre of rhodamine B (0.2% w/ Nik et al., 2008).
v in milli-Q water) were added to 5 mL of sample for staining. Two Acidification was conducted with addition of 0.6% GDL or the
drops of sample were placed into grooves of a concave microscope addition of lactic acid bacteria. The differences in the pH profiles
slide and a cover slip was placed over the top and sealed. Slides between the treatments are shown in Fig. 2. Similar data have been
were incubated at 40 °C until pH 5.1 was reached, and then cooled shown previously, for example, in milk matrices (Amice-Queme-
in a refrigerator for at least 30 min at 4 °C before viewing under the neur, Haluk, Hardy, & Kravtchenko, 1995; Lucey et al., 1998). The
microscope. lactic acid bacteria culture was able to acidify soymilk to pH 5.1
in about 240 min. The decrease in pH caused by bacterial acidifi-
2.6. Statistical analysis cation began slowly, most probably due to an adapting lag phase
of the bacteria, which lasted about 100 min. After this initial phase,
Measurements were carried out in triplicate and treatments where the pH was constant at about 6.6, the pH started to decrease.
were tested for equal variances then analysed using an unpaired After approximately 4 h, a pH of 5.1 was reached. In contrast with
student’s t-test assuming equal variances, using the using the data the slow acidification by lactic cultures, when 0.6% GDL was added
analysis function in Microsoft Office Excel 2007. the pH drop was very rapid due to the fast hydrolysis of GDL. After
about 50 min, acidification slowed down as GDL began to approach
equilibrium and the soymilk reached pH 5.1 after approximately
3. Results and discussion 2 h. This value of pH was therefore used to compare the final gels.
A comparison between the two modes of acidification allows the
In the present study the soymilk was prepared using common determination of possible differences in the acid-induced gelation
household practises, using only food-grade ingredients. The soy- behaviour of the soymilk particles with varying rates of
beans, purchased from a local grocery store contained 36.8 ± 0.2% acidification.
(n = 3) protein and 15.32 ± 0.01% fat (n = 3) (dry basis), and had a The development of the gel structure during acidification was
ratio between glycinin and b-conglycinin polypeptides of approxi- followed using rheology and a summary of the measured rheolog-
mately 1.7, as determined by SDS–PAGE and densitometry.
The soymilk obtained had a total solids content of 8.55 ± 0.09%,
4.00 ± 0.04% protein and 1.0 ± 0.3% fat. Mineral analysis revealed 7.0
that the soymilk contained the following amounts of minerals
6.8
per kilogramme of soymilk: 250 mg calcium, 330 mg magnesium,
6.6
630 mg phosphorus, 2100 mg potassium, 87 mg sodium and
360 mg sulfur. 6.4
The particle size distribution of the soymilk (Fig. 1) was mea- 6.2
sured using integrated light scattering. The soymilk particles were
pH

6.0
characterised by a bimodal distribution, with two populations at
5.8
approximately 0.2 and 10 lm. These two populations represented
mostly soymilk protein particles and fat globules. Incubation of the 5.6
samples with 1% SDS did not affect the size distribution (data not 5.4
shown). It has been previously reported (Malaki Nik, Tosh, Poysa, 5.2
5.0
0 50 100 150 200 250
4 Time (min)

Fig. 2. Acidification profile of soymilk with GDL (triangles) or bacterial culture

3 (circles). The results are the average of 3 replicates and bars represent the standard
deviation.
% Volume

2
Table 1
Rheological parameters measured during acidification and at a final pH of 5.1 for
soymilk acidified with GDL or with lactic acid bacteria. Gelation pH (measured as tan
1 d = 1), values at tan d at plateau; value of G0 at pH 5.1, slope of frequency sweep
dependence of G0 ; strain at yield. Values are the average of three independent
experiments, with standard deviation. Within a row, different letters indicate
0 statistical difference at p < 0.05.

Soymilk + GDL Soymilk + lactic acid bacteria

0.001 0.01 0.1 1 10 100 1000 10000
Gelation pH 5.90 ± 0.04a 6.29 ± 0.05b
Particle size (µm) Tan d at plateau 0.138 ± 0.003a 0.142 ± 0.002a
G0 at pH 5.1 (Pa) 412 ± 72a 467 ± 62a
Fig. 1. Particle size distribution, measured by integrated light scattering, of soymilk Log/log slope 0.073 ± 0.004a 0.072 ± 0.002a
prepared using a household soymilk maker. Distribution is representative of at least Strain at yield 0.128 ± 0.046a 0.113 ± 0.063a
three independent experiments.
 A. Grygorczyk, M. Corredig / Food Chemistry 141 (2013) 1716–1721 1719

500 neutral, as well as pH 5.8 and 5.5 was analysed by SDS–PAGE. No

apparent proteolysis was observed in the samples: all the subunits
1.0 were still present below the gelation pH (data not shown).
400
It was then concluded that the difference in gel point with bac-
0.8 terial acidification compared to GDL acidification may be the
300
difference in rate of pH reduction. When bacterial acidification is
G' (Pa)

Tan δ
0.6 used, the soymilk approaches the isoelectric point of soymilk very
200 slowly, allowing more time for proteins to rearrange and interact
0.4 with each other, resulting in a higher gelation point. When GDL
100 is used, the isoelectric point of soy proteins is surpassed very rap-
0.2 idly so it is possible that the pH must drop further before proteins
have time to interact sufficiently to generate a gel structure.
0
0.0 Indeed, the complex composition of the soymilk protein may cause
6.6 6.4 6.2 6.0 5.8 5.6 5.4 5.2 5.0 heterogeneities in the structure of the soymilk particles, and with
pH more time, particles may be able to orient themselves more opti-
mally for short-range attractions. It has been previously shown
Fig. 3. Development of the elastic modulus (G0 ) (filled symbols) and tan d (empty
(Ringgenberg et al., 2013) that the basic subunits of glycinin and
symbols) during acidification of soymilk with bacterial cultures (circles) or GDL
(triangles). The results are representative of three replicates. For statistical analysis b subunits of b-conglycinin begin to aggregate above the gel point,
see Table 1. probably because of their elevated isoelectric points (pH 6.8–8.5
and 5.7–6.0, respectively) (Staswick, Hermodson, & Nielsen,
ical parameters is shown in Table 1. Fig. 3 illustrates the changes in 1981; Thanh & Shibasaki, 1977). As these protein subunits ap-
the elastic modulus (G0 ) and tan d (where d is the phase angle) as a proach a net neutral charge and aggregate, the other subunits of
function of pH for soymilk acidified with GDL or with bacterial cul- soy proteins, which are covalently linked to basic glycinin subunits
tures. The environmental pH is critical to soy protein stability: dur- and b subunits of b-conglycinin, may also approach each other.
ing acidification, protons gradually neutralise surface charges. When these subunits are in closer proximity to each other, short-
When the net charge approaches zero, soymilk particles destabi- range interactions may begin to take place and contribute to the
lise, aggregate and form a particulate gel network (Kohyama development of a gel network. However, until the pH is reduced,
et al., 1995). The gelation pH was defined as the point where tan some repulsion still exists as these subunits have lower isoelectric
d = 1. points (Staswick et al., 1981; Thanh & Shibasaki, 1977). With the
In both cases, the initial value of the elastic modulus (G0 ) ample time provided by bacterial acidification for proteins to rear-
remained very low and constant until the point of gelation, which range into optimal conformations for interaction, the initial aggre-
was earlier in the case of lactic acid bacteria fermentation. The gation of basic subunits of glycinin and b-subunits of b-conglycinin
sudden increase in G0 , which occurs around the isoelectric point may be sufficient to make a network before the remaining protein
of soy proteins, is in full agreement with earlier reports, indicating subunits are incorporated as well. Fig. 3 also shows that during
that the gelation of the soymilk particles is a result of short range structure development, there was a statistically significant differ-
interactions, as gelation occurs when protein particles have ence in the slope of the increase of G0 as a function of pH. Although
reduced repulsive charges and can be in close proximity to each bacterial acidification results in a higher gelation pH, the presence
other (Malaki Nik et al., 2008; Ringgenberg et al., 2013). However, still of a number of proteins with repulsive charges would prevent
there was a significant difference in the pH of gelation of the a rapid development of the gel network.
soymilk particles depending on the mode of acidification (Table As previously mentioned, although the gelation occurred earlier
1). The gelation pH for soymilk acidified with lactic acid bacteria for soymilk acidified with bacteria, and the structure development
was 6.29 ± 0.05. After this pH, the G0 began to increase and contin- also appeared earlier, the small deformation properties of the gel at
ued to increase in a smooth upward curve until the end of the pH 5.1 did not differ. Fig. 4 illustrates the frequency dependence of
gelation experiment (which was arbitrarily set at pH 5.1). The aver- soymilk acid gels produced using a bacterial culture and using GDL.
age final G0 at pH 5.1 was 467 ± 62 Pa. On the other hand, when The curves using both modes of acidification are parallel to each
soymilk was acidified using GDL the gel point was pH 5.9 ± 0.04, other and rise with increasing frequency. The log/log slopes of fre-
in agreement with previous researchers (Malaki Nik et al., 2008,
2011; Ringgenberg et al., 2013). The average final G0 of soymilk
acidified with GDL was 412 ± 72 Pa. This was not significantly dif- 1000
ferent from the final G0 when bacterial acidification was used. Both
gels also showed a similar final tan d value. Although the values of
G0 and tan d for the two gels at pH 5.1 were similar, there seemed
to be a quicker gel development with pH for the gels prepared with
GDL. The values of tan d also reached a similar final value, regard-
G' (Pa)

less of the mode of acidification. The slower fermentation time of

soymilk by bacteria allowed for more protein rearrangements dur-
ing acidification and may have been the cause of the higher pH of
destabilisation in those gels. However, the molecular details of this
early gelation are not known, as this effect has never been reported
before.
Although soymilk contains fermentable carbohydrates (Mital & 100
0.01 0.1 1 10
Steinkraus, 1975), previous literature has shown that some strains
of lactic acid bacteria may produce proteolytic enzymes (Aguirre, Frequency (Hz)
Garro, & Savoy de Giori, 2008). To test the occurrence of proteolysis Fig. 4. Frequency dependence of soymilk gels prepared with bacterial culture and
during fermentation, the composition of the polypeptides of GDL. Results are the average of three replicates and bars represent standard
soymilk incubated with or without lactic acid bacteria, at pH near deviations.
1720 A. Grygorczyk, M. Corredig / Food Chemistry 141 (2013) 1716–1721

Fig. 5. Confocal microscopy of the soymilk gels prepared with bacterial culture (A) or GDL (B). Bar size 50 lm.

quency dependence on G0 for the GDL-induced soymilk gel and the Acknowledgements
bacteria-induced soymilk gel (0.073 and 0.072, respectively) were
not statistically different (Table 1). The positive slope suggested This research was supported by the National Science and
that the gels are particulate in nature, as has been previously Engineering Council of Canada and the Hannam Soybean Utiliza-
reported for soymilk acid gels (Malaki Nik et al., 2011). The average tion Fund. The authors would like to thank Danisco Canada
yield strain of soymilk gels formed using a bacterial culture was (Scarborough, ON, Canada) for providing the bacterial culture.
0.113 ± 0.063 and the same value for the gel made by GDL acidifi-
cation was 0.128 ± 0.046. These values were also not found to be
significantly different, again suggesting that the final gel formed
References
by either GDL or bacterial acidification did not differ.
The confocal microscopy images are also in full agreement with Aguirre, L., Garro, M. S., & Savoy de Giori, G. (2008). Enzymatic hydrolysis of
the conclusion that at the final pH of 5.1, the gels were very similar. soybean protein using lactic acid bacteria. Food Chemistry, 111, 976–982.
Amice-Quemeneur, N., Haluk, J. P., Hardy, J., & Kravtchenko, T. P. (1995). Influence of
Fig. 5 shows confocal microscopy images of soymilk samples gelled
the acidification process on the colloidal stability of acidic milk drinks prepared
using either lactic acid bacteria or GDL. The microstructure of the from reconstituted nonfat dry milk. Journal of Dairy Science, 78, 2683–2690.
protein gel appears densely packed and composed of many small Buono, M. A., Erickson, L. E., Fung, D., & Jeon, I. J. (1990). Carbohydrate utilization
structures assembled into a network of protein aggregates. and growth kinetics in the production of yogurt from soymilk part I:
Experimental methods. Journal of Food Processing and Preservation, 14, 135–153.
Canabady-Rochelle, L., Sanchez, C., Mellema, M., & Banon, S. (2009). Study of
calcium-soy protein interactions by isothermal titration calorimetry and pH
cycle. Journal of Agricultural and Food Chemistry, 57, 5939–5947.
Donkor, O., Henriksson, A., Vasiljevic, T., & Shah, N. (2007). Rheological properties
and sensory characteristics of set-type soy yogurt. Journal of Agricultural and
Food Chemistry, 55, 9868–9876.
Ghosh, D., Chattoraj, D., & Chattopadhyay, P. (2011). Studies on changes in
4. Conclusions microstructure and proteolysis in cow and soy milk during fermentation
using lactic cultures for improving protein bioavailability. Journal of Food
Science and Technology. http://dx.doi.org/10.1007/s13197-011-0421-1.
Comparison of gelation of soymilk induced by acidification Keerati-u-rai, M., & Corredig, M. (2009). Effect of dynamic high pressure
using commercial lactic cultures or GDL hydrolysis revealed that homogenization on the aggregation state of soy protein. Journal of Agricultural
bacterial fermentation resulted in a significantly earlier gel point. and Food Chemistry, 57, 3556–3562.
Kohyama, K., Sano, Y., & Doi, E. (1995). Rheological characteristics and gelation
The slower rate of acidification compared to that of samples chem- mechanism of tofu (soybean curd). Journal of Agricultural and Food Chemistry, 43,
ically acidified may be allowing enough time for the protein sub- 1808–1812.
units with higher isoelectric points to rearrange for more optimal Liu, C., Hu, C., Chiang, S., Tseng, K., Yu, R., & Pan, T. (2009). Beneficial preventive
effects of gastric mucosal lesion for soy-skim milk fermented by lactic acid
interaction with each other, resulting in a change in the elastic bacteria. Journal of Agricultural and Food Chemistry, 57, 4433–4438.
modulus at an earlier pH. When the rate of acidification is much Lucey, J. A., Tamehana, M., Singh, H., & Munro, P. A. (1998). A comparison of the
higher, all the subunits will be incorporated much faster, as their formation, rheological properties and microstructure of acid skim milk gels
made with a bacterial culture or glucono-d-lactone. Food Research International,
lower isoelectric points will be reached faster. It was hypothesised 31, 147–155.
that because of the range of isoelectric points of soy protein Malaki Nik, A., Tosh, S., Poysa, V., Woodrow, L., & Corredig, M. (2008).
subunits, at the pH where some subunits become destabilised Physicochemical characterization of soymilk after step-wise centrifugation.
Food Research International, 41, 286–294.
and start to aggregate, other subunits still exhibit repulsive Malaki Nik, A., Tosh, S. M., Woodrow, L., Poysa, V., & Corredig, M. (2009). Effect of
charges, hindering the ability of destabilising proteins to form a soy protein subunit composition and processing conditions on stability and
network and when there is rapid acidification a gel network will particle size distribution of soymilk. Food Science and Technology, 42,
1245–1252.
appear only after more protein species have reached their
Malaki Nik, A., Alexander, M., Poysa, V., Woodrow, L., & Corredig, M. (2011). Effect of
isoelectric point. soy protein subunit composition on the rheological properties of soymilk during
Although the pH of gelation is significantly different with the acidification. Food Biophysics, 6, 26–36.
two modes of acidification, the final gel structure of the gels did Mital, B. K., & Steinkraus, K. H. (1975). Utilization of oligosaccharides by lactic acid
bacteria during fermentation of soymilk. Journal of Food Science, 40, 114–118.
not appear to be different based on small oscillatory Mital, B. K., Steinkraus, K. H., & Naylor, H. B. (1974). Growth of lactic acid bacteria in
measurements, yield strain and visual microstructure. soymilks. Journal of Food Science, 39, 1018–1022.
 A. Grygorczyk, M. Corredig / Food Chemistry 141 (2013) 1716–1721 1721

Prabhakaran, M. P., Perera, C. O., & Valiyaveettil, S. (2006). Effect of different Tay, S. L., & Perera, C. O. (2006). Physicochemical properties of 7S and 11S protein
coagulants on the isoflavone levels and physical properties of prepared firm mixtures coagulated by glucono-d-lactone. Journal of Food Science, 69,
tofu. Food Chemistry, 99, 492–499. FEP139–FEP143.
Ren, C., Tang, L., Zhang, M., & Guo, S. (2009). Structural characterization of heat- Thanh, V. H., & Shibasaki, K. (1977). Beta-conglycinin from soybean proteins.
induced protein particles in soy milk. Journal of Agricultural and Food Chemistry, Isolation and immunological and physicochemical properties of the monomeric
57, 1921–1926. forms. Biochimica Et Biophysica Acta, 490, 370–384.
Renkema, J., & van Vliet, T. (2002). Heat-induced gel formation by soy proteins at Yang, M., Fu, J., & Li, L. (2012). Rheological characteristics and microstructure of
neutral pH. Journal of Agricultural and Food Chemistry, 50, 1569–1573. probiotic soy yogurt prepared from germinated soybeans. Food Technology and
Ringgenberg, E., Alexander, M., & Corredig, M. (2013). Effect of concentration and Biotechnology, 50, 73–80.
incubation temperature on the acid induced aggregation of soymilk. Food Zhang, J., Liang, L., Tian, Z., Chen, L., & Subirade, M. (2012). Preparation and in vitro
Hydrocolloids, 30, 463–469. evaluation of calcium-induced soy protein isolate nanoparticles and their
Staswick, P. E., Hermodson, M. A., & Nielsen, N. C. (1981). Identification of the acidic formation mechanism study. Food Chemistry, 133, 390–399.
and basic subunit complexes of glycinin. Journal of Biological Chemistry, 256,
8752–8755.