Food Research International 55 (2014) 207–214

Contents lists available at ScienceDirect

Food Research International

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

Associative interactions between chitosan and soy protein fractions:

Effects of pH, mixing ratio, heat treatment and ionic strength
Yang Yuan, Zhi-Li Wan, Xiao-Quan Yang ⁎, Shou-Wei Yin
Research and Development Center of Food Proteins, College of Light Industry and Food, South China University of Technology, Guangzhou 510640, PR China

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 paper is to explore the complexation between the soy protein fractions (glycinin and β-
Received 30 May 2013 conglycinin) and chitosan (CS) and to investigate the influence of pH, mixing ratio, heat treatment and ionic
Accepted 10 November 2013 strength. Phase behavior and microstructure showed that soluble complex and coacervate were obtained in
glycinin/CS and β-conglycinin/CS mixtures at specific pHs, following a nucleation and growth mechanism. More-
Keywords:
over, the coacervates showed higher thermal stability than protein alone. Specially, the glycinin/CS mixture
Soy protein fractions
Chitosan
displayed a gel-like network structure at pH 5.5 and 6.0, and this structure kept the mixture soluble at a long
Associative interaction pH region. The turbidity versus ζ-potential pattern showed that, independent of protein, the self aggregation
Complex of soy protein fractions and the coacervation of glycinin/CS and β-conglycinin/CS mixtures were all obtained at
Coacervate charge neutralization pH, indicating that the ζ-potential is the most critical parameter to understand the stability
of soy protein/chitosan mixture. This predictive parameter was less affected by mixing ratio and heating but was
significantly affected by ionic strength because mixing ratio and heating only changed the equilibrium between
repulsive and attractive forces in colloid system while sodium chloride destroyed the predictability of colloidal
stability via shielding charged reactive sites on both biopolymers to disrupt electrostatic interactions.
© 2013 Elsevier Ltd. All rights reserved.

1. Introduction
Proteins and polysaccharides are widely used in food industry, and
understanding what structure can be designed by protein and polysac-
charide requires a comprehensive consideration of evaluating the inter-
action between protein and polysaccharide in their mixture.
Consequently, the investigation of protein/polysaccharide interaction
seems to be a fundamental thing to conduct. Mixing proteins and poly-
saccharides in an aqueous environment results in two types of interac-
tions, segregative and associative. In general, segregative interaction
occurs at high concentrations of biopolymers and has already been
used to predict the microstructure and physical properties of protein/
polysaccharide mixed gel in many references (Çakır & Foegeding,
2011). In recent years, research has been focused on the protein/poly-
saccharide associative interaction in which protein can interact with
polysaccharide to form electrostatic complexes by following a two-
step nucleation and growth-type kinetic process (Elmer, Karaca, Low,
& Nickerson, 2011). Depending on the biopolymer characteristics and
solvent conditions, different types of structures can be tailored upon as-
sociative interaction to create different functionalities. Complex (solu-
ble and interpolymeric complexes), coacervate (liquid in nature) and
electrostatic gel (network structure) were mostly investigated in recent
years (Klein, Aserin, Ben Ishai, & Garti, 2010; Liu et al., 2010; Schmitt &
⁎ Corresponding author. Tel.: +86 20 87114262; fax: +86 20 87114263.
E-mail addresses: fexqyang@163.com, fexqyang@scut.edu.cn (X.-Q. Yang).
Turgeon, 2011; Turgeon, Schmitt, & Sanchez, 2007; Weinbreck, Tromp,
& de Kruif, 2004).
However, creating the structure of protein/polysaccharide com-
plexes is not only based on the solvent conditions, but also takes into ac-
count the characteristic of the biopolymers. According to this
perspective, soy protein and cationic polysaccharide complexation
should be a valuable research topic to conduct. Indeed, chitosan (CS)
[(1–4)-2-amino-2-deoxy-β-d-glucan], as a linear cationic polysaccha-
ride, began to be noticed in recent years. Soluble complex, nanoparticle
and coacervate between chitosan and protein were sufficiently investi-
gated (Anal, Tobiassen, Flanagan, & Singh, 2008; Lee & Hong, 2009;
Mounsey, O'Kennedy, Fenelon, & Brodkorb, 2008).
Soy proteins are the most important representative of legume pro-
teins due to their high protein level and well-balanced amino-acid com-
position, showing great potential in the substitution of meat and dairy
proteins (van Vliet, Martin, & Bos, 2002). Takeuchi, Morita, Saito,
Kugimiya, and Fukamizo (2006) indicated that the thermal stability of
β-conglycinin subunit was improved by interacted with chitosan
through electrostatic interactions. Murakami and Takashima (2003) re-
ported that the mechanical strength of the chitosan/soy globulin cap-
sule was found to depend on the pH values of the chitosan and
globulin solutions. Most recently, Huang, Sun, Xiao, and Yang (2012) in-
vestigated the formation of soybean protein isolate/chitosan coacervate
as a function of pH, temperature, time, ionic strength, total biopolymer
concentration and mixing ratio.
However, β-conglycinin (7S) and glycinin (11S), which are two
main fractions of soy protein isolate, have different structures and

0963-9969/$ – see front matter © 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.foodres.2013.11.016
208 Y. Yuan et al. / Food Research International 55 (2014) 207–214
properties. β-Conglycinin is a trimeric glycoprotein consisting of at least
six combinations of three subunits: α, α′ and β (Thanh & Shibasaki,
1977). Glycinin consists of two polypeptide components linked via di-
sulfide bonds (AB), one (A, Mw = 34–44 kDa) with acidic and the
other (B, Mw = 20 kDa) with basic isoelectric points (Staswick,
Hermodson, & Nielsen, 1981). From this standpoint, the interaction be-
tween soy protein and chitosan should be investigated individually, in-
volving the interaction of 7S/CS and 11S/CS. As far as we know, only Liu
et al. (2011) discussed the light scattering properties of 7S globulin/chi-
tosan and 11S globulin/chitosan complexes under hydrothermal condi-
tions as a function of pH. Our previous research has also investigated the
formation of 11S/CS soluble complex at acidic pH (Yuan et al., 2013).
Since the associative interactions of 7S/CS and 11S/CS mixtures have
not been reported yet, the objective of this paper is to explore the com-
plexation between the soy protein fractions and chitosan and to inves-
tigate the influence of pH, mixing ratio, heat treatment and ionic
strength. The study can be an available reference for the utilization of
chitosan in soybean-based food.
2. Materials and methods
2.1. Materials
The defatted soybean flakes were purchased from Shandong
Xinjiahua Industrial and Commercial Co. Ltd., China. The chitosan (mo-
lecular weight about 300 kDa; degree of deacetylation of 95%; moisture
8.0%; ash content 0.7%) used for the experiment was purchased from
Shandong Aokang Industrial and Commercial Co. Ltd., China. The fluo-
rescent dye Rhodamine B was purchased from Sigma Chemical Co. (St.
Louis, MO, USA). All chemicals used in this work were of analytical or
better grade.
2.2. Preparation of glycinin and β-conglycinin
Glycinin and β-conglycinin were prepared according to Nagano's
procedure (Nagano, Hirotsuka, Mori, Kohyama, & Nishinari, 1992). The
obtained precipitate (glycinin or β-conglycinin) was adjusted to
pH 7.5 with 2 M NaOH, dialyzed (18% concentration) against distilled
water for 48 h and then lyophilized to yield glycinin or β-conglycinin
powders. Protein contents of glycinin and β-conglycinin lyophilized
powder were 93.6 ± 0.9% and 89.7 ± 0.4% determined by the Dumas
method (N × 6.25), respectively.
2.3. Preparation of mixtures
Soy protein fractions (11S and 7S, 2%, w/v) and chitosan (2%, w/v)
stock solutions were prepared by dissolving the powders in 100 mM
acetic acid/sodium acetate buffer (pH 3.0) with stirring (600 rpm) at
room temperature (25 °C) for 3 h. The 11S/CS and 7S/CS mixtures
were obtained by mixing corresponding stock solutions. Various con-
centrations of mixtures were obtained by diluting the mixtures with
acetic acid/sodium acetate buffer at the desired pH and ionic strength
(NaCl). As for the heat treatment, the mixtures were heated at different
pH (3.0–8.0) in a series of tightly capped glass tubes in a water bath at
95 °C for 30 min. After heating, the mixtures were vibrated vigorously
and cooled to room temperature immediately.
2.4. Confocal Laser Scanning Microscope (CLSM)
The CLSM-images of mixtures were recorded at room temperature
with a Leica TCS SP5 Confocal Laser Scanning Microscope (Leica
Microsystems Inc., Heidelberg, Germany), equipped with an inverted
microscope (Model Leica DMI6000). Before imaging, the mixtures (pro-
tein content, 5 mg/mL) were mixed with Rhodamine B solution (20 μL
of a 0.2 wt.% solution per mL sample). After that, a drop of the stained
mixture was placed on a concave slide, which was then covered with
a glycerol-coated cover slip. The incident light was emitted by a laser
beam at 488 nm (Rhodamine B).
2.5. Differential scanning calorimetry (DSC)
DSC experiments were performed on a TA Q100 differential scan-
ning calorimeter (TA Instruments, New Castle, DE). Indium standards
were used to calibrate the temperature and energy. Lyophilized samples
(2.0 mg) were accurately weighed into aluminum liquid pans, and
10 μL of water was added. Before analysis, the aluminum pans were
sealed and equilibrated at room temperature overnight. The samples
were analyzed at the scanning rate of 5 °C/min for a temperature
range from 30 to 120 °C. A sealed empty pan was used as a reference.
The determinations of the transition temperature at peak maximum
as well as the calorimetric enthalpy change were computed by the uni-
versal analyzer 2000, version 4.1D (TA Instrument-Waters LLC, New
Castle, DE). Thereinto, Tm was defined as the peak maximum. ΔH was
obtained by integration of the area of the excess heat capacity endo-
therm. All experiments were conducted in duplicate.
2.6. ζ-Potentials
The ζ-potentials of the samples were determined using a Nanosizer
ZS instrument (Malvern Instruments Ltd., Worcestershire, UK). Aque-
ous solutions (protein content, 5 mg/mL) of samples were filled into a
disposable ζ-potential folded capillary cell (DTS1060). The instrument
determined the electrophoretic mobility. The Henry equation was
then applied for calculating the ζ-potential. A dielectric constant of
78.5 and a viscosity of 0.8872 mPa·s were also used for continuous
phase. All measurements were conducted in triplicate.
2.7. Turbidimetric analysis
Turbidity of soy protein/chitosan mixtures (protein content, 5 mg/
mL) was measured using a UV-2000 spectrophotometer at 600 nm
with 10 mm path length quartz cuvettes. The mixtures were measured
at the protein concentration of 5 mg/mL (w/v). All measurements were
conducted in triplicate.
2.8. Statistical analysis
All the data were subjected to one way analysis of variance and cor-
relations between the results were carried out using Statistical Analysis
System Software (SAS version 9.0, SAS Institute, Cary, NC); significant
differences were determined by Duncan's multiple range test and ac-
cepted at P b 0.05 (Duncan, 1955).
3. Results and discussion
3.1. Phase behavior, ζ-potential and microstructure of the 11S/CS and 7S/CS
mixtures as a function of pH
Fig. 1 shows the phase diagram and ζ-potential of the 11S/CS and 7S/
CS mixtures as a function of pH. The mixtures were settled overnight
and the phase behavior of the mixtures was assessed by visual observa-
tion of a transparent, a translucent, a turbid, a partly or a totally phase
separated solution (Fig. 1.A). As a control, chitosan showed no phase
separation during all tested pHs while 7S and 11S showed phase sepa-
ration at pHs around 4.5 and 5.0 respectively. According to Fig. 1.B, it
is indicated that chitosan showed high ζ-potential at most tested pHs,
keeping it soluble at a lone pH region. Soy protein fractions lost their
electrostatic repulsive force at pH 4.3 and 5.1 respectively due to the re-
duced ζ-potential attained to zero, which was usually referred as the
isoelectric point (pI). It is noteworthy that 7S and 11S showed
distinguishing ζ-potential–pH curves, and this phenomenon can be
 Y. Yuan et al. / Food Research International 55 (2014) 207–214
Fig. 1. Phase diagram and ζ-potentials of the 11S/CS and 7S/CS mixtures as a function of pH. (A) Phase diagram and (B) ζ-potentials (protein content = 5 mg/mL; chitosan content = 0.5
mg/mL; NaCl content = 0 mM).
ascribed to the different amino acid composition and colloidal proper-
ties in the two soy protein fractions (Wolf, 1970).
Mixtures showed significantly different phase diagram. According to
Fig. 1.A, the phase separated pH region shifts to higher pH after mixing
chitosan, and coacervation phase separation of the 7S/CS and 11S/CS
mixtures was found at pHs around 5.5 and 7.0 respectively. Concomi-
tantly, the ζ-potential–pH curves of two mixtures settled between that
of two individual polymers and the zero value of ζ-potential of the
11S/CS and 7S/CS mixtures shifted to 5.7 and 7.0, respectively.
The microstructures of the 11S/CS and 7S/CS mixtures as a function
of pH are shown in Fig. 2. As a control, soy protein fractions showed an
aggregate at the pHs close to their pIs (data not shown). Both the 11S/CS
and 7S/CS mixtures showed more homogenized structure and less in
size and number of aggregates than protein alone at that pHs. At coacer-
vation pHs of the 11S/CS and 7S/CS mixtures, however, dense and large
coacervates were observed in Fig. 2, resulting in the phase separation in
macro phase. Furthermore, the 11S/CS mixture showed a gel-like net-
work structure at pH 5.5 and 6.0 but the 7S/CS mixture was not detect-
ed. This network structure probably increased the stability of the 11S/CS
mixture and kept it stable at a long pH region (Fig. 1.A, pH 4.5–6.0).
According to the above results, soluble complexes were formed in
the 11S/CS and 7S/CS mixtures at low pH because the interaction mainly
occurred between the positively charged chitosan and negatively
charged part of protein. Similar phenomenon was recorded previously
in whey protein/chitosan and pea protein/chitosan system (Elmer
et al., 2011; Mounsey et al., 2008). It should be realized that the acidic
soluble complex could be used in acidic beverage, making the clear
juice product which contains soy protein become possible. At neutral
209
-
pH, coacervates were obtained because the strong interaction between
the two oppositely charged biopolymers induced the charge neutraliza-
tion of mixtures, resulting in the decrease of electrostatic repulsive force
and increase of polymer size. It also can be seen clearly from the micro-
structure (Fig. 2) that the primary complexes formed initially at pH 4.0
and 4.5, then newly formed interpolymeric complexes coarsen as the
pH increased, following a nucleation and growth mechanism. Most in-
terestingly, the formation of network structure in the 11C/CS mixture
suggested that the coarse stranded also could be found at such a low
protein concentration and this microstructure was dependent on the
protein. Similarly, Laneuville, Turgeon, Sanchez, and Paquin (2006) in-
vestigated kinetics of the electrostatic gelation by using native β-
lactoglobulin and xanthan. They formed a new hydrogel at β-
lactoglobulin concentration of 0.1% and they explained their results by
using an improved nucleation and growth mechanism appropriately.
Most recently, Le and Turgeon (2013) proceed their research with re-
gard to β-lactoglobulin/xanthan electrostatic gel by focusing on the mi-
crostructure of this gel. However, it is still lacking report on how to
utilize this finding in the food industry. We also found that the coacer-
vates between soy protein fractions and chitosan were not liquid in na-
ture. According to Turgeon's definition (Turgeon & Laneuville, 2009),
they should be called as a type of charge induced interpolymeric com-
plex between protein and polysaccharide. However, we still used the
terms coacervate and coacervation throughout this paper. In addition,
Weinbreck et al. (2004) reported that coacervates began to breakdown
since the polysaccharides will lose their charge at pH close to
polysaccharide's pKa and this breakdown was dependent on the charge
of polysaccharide. During all tested pH in this paper, however, the
210 Y. Yuan et al. / Food Research International 55 (2014) 207–214

Fig. 2. Microstructure of the 11S/CS and 7S/CS mixtures as a function of pH (protein content = 5 mg/mL; chitosan content = 0.5 mg/mL; NaCl content = 0 mM). Red color represented
the protein phase in the CLSM image.

breakdown of complex coacervate was only occurred in the 7S/CS mix-
ture (partly phase separation in Fig. 1.A), suggesting that this break-
down should be dependent not only on the charge of polysaccharide,
but also on the charge of protein.
3.2. Thermal properties of the 11S/CS and 7S/CS coacervates
Fig. 3 shows the heat flow-temperature pattern of soy protein frac-
tions and their coacervates with chitosan. Also, the thermal transition
characteristics of these samples are listed in Table 1. All tested samples
exhibited a prominent endothermic peak in Fig. 3. The major endother-
mic peaks of 7S and 11S were clearly attributed to the thermal denatur-
ation of soy protein fraction which were normally reported as 74 °C and
90 °C for 7S and 11S, respectively (Sorgentini & Wagner, 1999). In this
paper, the Tm and ΔH of 7S and 11S were 76.625 °C, 5.494 J/g and
89.905 °C, 10.575 J/g, respectively (Table 1). The effects of polysaccha-
rides on the protein's thermal stability were investigated in recent pub-
lication. However, these effects seem to be different in different protein
and polysaccharide systems. Vardhanabhuti, Yucel, Coupland, and
Foegeding (2009) applied nanoDSC to complexation of β-lactoglobulin
 Y. Yuan et al. / Food Research International 55 (2014) 207–214
Fig. 3. Typical DSC thermograms of soy protein fractions and their coacervates with chito-
san (chitosan/protein ratio = 0.1 g g−1).
and dextran sulfate and observed a reduction in Tm at neutral pH. Lee
and Hong (2009) reported that the α-lactalbumin/CS complex showed
higher thermal stability than α-lactalbumin alone while the β-
lactoglobulin/CS complex showed reversed result. Plashchina et al.
(2001) suggested the chitosan increased the thermal stability of faba
bean legumin. Most recently, Huang et al. (2012) investigated the for-
mation of SPI/CS coacervate and they found that the thermal denatur-
ation temperature of SPI was enhanced by coacervation with chitosan.
However, the information of the effect of chitosan on the thermal stabil-
ity of soy protein fractions is still lacking and needs further research. Ac-
cording to Fig. 3 and Table 1, both the 7S/CS and 11S/CS coacervates
showed significant (P b 0.05) enhancement in Tm and ΔH than protein
alone, indicating that the coacervation between soy protein fractions
and chitosan significantly enhanced the thermal stability of soy protein
fractions. The reason could be attributed to the dense internal structure
of coacervate associated with protein/polysaccharide interaction. More-
over, the electrostatic interaction between polymers in aqueous phase
was reported as an endothermic reaction (Abe, Ohno, Nii, & Tsuchida,
1978). The increased ΔH suggested that the interaction between soy
protein fractions and chitosan was mainly electrostatic induced.
3.3. The effects of pH and mixing ratio on stability of the 11S/CS and 7S/CS
mixtures
Mounsey et al. (2008) remarked that an increased turbidity develop-
ment can therefore be attributed to an increase in both the number and
the size of the complex aggregates. Fig. 4 shows the effect of mixing
ratio on turbidity and ζ-potential of the 7S/CS and 11S/CS mixtures as
a function of pH. Generally, the polymer mixing ratio influences the
charge balance of their complexes (Elmer et al., 2011; Weinbreck
et al., 2004). In this paper, mixing ratio showed similar effect on the tur-
bidity and ζ-potential of the 11S/CS and 7S/CS mixtures. In detail, 7S
alone showed high turbidity in the pH region of 3.5–5.0 and the 11S
alone evidenced self-aggregation in the pH region of 4.0–6.0. The tur-
bidity at the aforementioned pH was considerably decreased depending
Table 1
Thermal transition characteristics of soy protein fraction and their coacervate with
chitosan (chitosan/protein ratio = 0.1 g g−1).
Samples Tm (°C) ΔH (J/g)
11S 89.905 ± 0.045b 10.575 ± 0.405b
11S/CS 98.465 ± 0.015a 13.425 ± 0.245a
7S 76.625 ± 0.115d 5.494 ± 0.111d
7S/CS 80.895 ± 0.035c 8.664 ± 0.268c
Superscript letters (a–d) indicate significant difference (P b 0.05) in Tm and ΔH among all
samples.
211
on the CS/soy protein ratio, while the turbidity at the pH above 6.0 was
on the contrary markedly increased (Fig. 4.A and C). Concomitantly, the
maximum turbid point shifted gradually toward a more alkaline pH
upon the mixing ratio elevation, which was in agreement with the ζ-
potential data. The magnitude of ζ-potential increased gradually at pH
lower than the pI of 7S and 11S, while the ζ-potential changed from neg-
ative to positive number at pH 4.0–6.0 and decreased from −35 and −
40 mV to about −5 and −20 mV at high pH, as the CS/soy protein ra-
tios were increased from 0 to 0.2 g g−1 (Fig. 4.B and D). Concomitantly,
the charge neutralization of the 7S/CS and 11S/CS mixtures was shifted
to 5.2, 5.7, 6.4 and 6.2, 7.0, 7.8, respectively. Most of the researchers
showed the effect of mixing ratio on mixture properties in their publica-
tions. Elmer et al. (2011) said the effect which chitosan caused a shift in
turbidity curve was independent of the mixing ratio in pea protein iso-
late/chitosan mixtures. However, the magnitude of this shift was ratio
dependent. Our previous study (Yuan et al., 2013) also found that the
phase boundaries (pHc, pHϕ) of glycinin/CS mixture were related to
the mixing ratio.
The strength of the electrostatic interaction (SEI) between soy pro-
tein fractions and chitosan is shown in Fig. 5. The SEI between opposite-
ly charged polyelectrolytes can be estimated by calculating the absolute
value of the product of the measured ζ-potential of both polymers at
each pH (Espinosa-Andrews et al., 2013; Weinbreck et al., 2004). The
SEI–pH curves displayed a reversed U-shape and the trend of these
curves indicated that the interactions between 11S and chitosan as
well as 7S and chitosan were strongest at pH 6.0 and 6.5 respectively.
These pHs were closed to the pH range of charge neutralization of the
mixtures (Fig. 4.B and D). Furthermore, 7S showed higher interaction
affinity with chitosan than 11S did.
The turbidity of mixtures against their ζ-potentials is also shown in
Fig. 5. Previous research indicated that particles with ζ-potentials more
positive than + 30 mV or more negative than −30 mV are normally
considered stable, and the transition state was obtained at the ζ-potential
region between −30 mV and +30 mV (van Nieuwenhuyzen & Szuhaj,
1998). Along with following this hypothesis, the turbidity curve displayed
a reversed U-shape ζ-potential dependence and this phenomenon was
seemingly independent on the type of protein used and concentration
of chitosan added. Generally, mixing ratio mainly changes the net charge
of the mixture system. In other words, the stability of the 11S/CS and 7S/
CS mixtures was changed as a function of charge. Under certain condition,
changing the type of protein and concentration of polysaccharide mainly
changed the equilibrium between repulsive (steric and electrostatic) and
attractive (electrostatic, van der Waals and hydrophobic) forces in colloid
system. Thus, the stronger the electrostatic interaction was, the lower the
ζ-potential got and the higher the turbidity obtained (Weinbreck et al.,
2004). As a result, the stability of the soy protein and chitosan mixture
should be called charge dependent instead of pH or mixing ratio depen-
dent. That means the ζ-potential is the most important parameter to un-
derstand the stability of the soy protein/chitosan mixture in an untreated
condition. It is noteworthy in some data point that they showed different
turbidities even if they had similar ζ-potentials. This phenomenon that
can be ascribed to the ζ-potential measurement of the mixture was the
total charge in mixture. It should contain the charge of mixture and also
may include the charge of surplus soy protein and/or chitosan. The
surplus soluble soy protein and/or chitosan will contribute to the
ζ-potential measurement but showed less contribution to the
turbidity measurement in an opaque sample.
3.4. The effects of ionic strength and heat treatment on stability of the 11S/
CS and 7S/CS mixtures
According to previous research, heat treatment and ionic strength
were reported as other two effective effects in controlling the charge
balance within an electrostatic complex system (Mounsey et al., 2008;
Vardhanabhuti & Foegeding, 2008). The effects of heat treatment and
sodium chloride (NaCl) on Turbidity–ζ-potential pattern are shown in
212 Y. Yuan et al. / Food Research International 55 (2014) 207–214

Fig. 4. Turbidity and ζ-potentials of the 11S/CS and 7S/CS mixtures as a function of pH and mixing ratio. Panels A and B: 11S/CS mixtures (protein content = 5 mg/mL; chitosan/11S
mixing ratio from 0 to 0.2 g g−1; NaCl content = 0 mM); Panels C and D: 7S/CS mixtures (protein content = 5 mg/mL; chitosan/7S mixing ratio from 0 to 0.2 g g−1; NaCl
content = 0 mM).

Fig. 6. As shown in Fig. 6.A and B, the Turbidity–ζ-potential plots of both
11S/CS and 7S/CS were affected by heat treatment but still displayed a
reversed U-shape. Exposing hydrophobic group caused by heating fa-
vored hydrophobic interaction in soy protein and chitosan system.
That probably be the reason why mixtures showed high turbidity with
high ζ-potential. However, it is hard to change the reversed U-shape
by heating since an electrostatic repulsive force still plays an important
role in this condition. In contrast, NaCl showed significant effect on
Turbidity–ζ-potential plot as can be seen in Fig. 6. It is hard to tell a U-
shape in Turbidity–ζ-potential plot after 100 mM NaCl was added,

Fig. 5. The effect of associative interaction on the stability of the 11S/CS and 7S/CS mixtures. Left Y versus bottom X: Turbidity of the 11S/CS and 7S/CS mixtures with different mixing ratios
as a function of ζ-potentials; Right Y versus upper X: The strength of the electrostatic interaction as a function of pH.
 Y. Yuan et al. / Food Research International 55 (2014) 207–214
Fig. 6. The effect of heat treatment and ionic strength on the Turbidity–ζ-potential pattern. Panel A: 11S/CS mixtures (protein content = 5 mg/mL; chitosan content = 0.5 mg/mL); Panel
B: 7S/CS mixtures (protein content = 5 mg/mL; chitosan content = 0.5 mg/mL).
indicating that ionic strength showed stronger influence on electrostatic
complex because it destroyed the predictability of colloidal stability via
shielding charged reactive sites on both biopolymers to disrupt electro-
static interactions. The result suggested that the main weakness of soy
protein/chitosan complexes and coacervates for food applications is
the narrow ionic strength range that ensures complex stability.
4. Conclusion
Overall, the formation of soy protein fractions/chitosan complex and
coacervate are electrostatic dependent, following a nucleation and
growth mechanism. Consequently, ζ-potential should be firstly detect-
ed to understand the stability of soy protein fractions/chitosan mixture.
The most interesting finding in this paper was that the gel-like network
structures were observed using confocal technique in 11S/CS soluble
complex. The network structure kept the complex stable at a long pH re-
gion. The acidic soluble soy protein fractions/chitosan complex could be
used in acidic beverage, making the clear juice product which contains
soy protein become possible. However, the main weakness of this com-
plex for food applications is the narrow pH and ionic strength range that
ensures complex stability. Several approaches could be evaluated to sta-
bilize the complex in future such as covalent crosslink.
213
Acknowledgments
This work is part of the research projects of the Chinese National
Natural Science Fund (Serial Numbers: 31371744 and 31130042, spon-
sored by the NSFC), the research project of the National Key Technology
Research and Development Program for the 12th Five-year Plan
(2012BAD33B10, 2012BAD34B04) and the Fundamental Research
Funds for the Central Universities (SCUT, 2012ZZ0082).
References
Abe, K., Ohno, H., Nii, A., & Tsuchida, E. (1978). Calorimetric study on the formation of
polymer complexes through electrostatic interaction and hydrogen bonding. Die
Makromolekulare Chemie, 179, 2043–2050.
Anal, A. K., Tobiassen, A., Flanagan, J., & Singh, H. (2008). Preparation and characterization
of nanoparticles formed by chitosan–caseinate interactions. Colloids and Surfaces B:
Biointerfaces, 64, 104–110.
Çakır, E., & Foegeding, E. A. (2011). Combining protein micro-phase separation and
protein-polysaccharide segregative phase separation to produce gel structures. Food
Hydrocolloids, 25(6), 1538–1546.
Duncan, D. B. (1955). Multiple range and multiple F tests. Biometrics, 11(1), 1–42.
Elmer, C., Karaca, A.C., Low, N. H., & Nickerson, M. T. (2011). Complex coacervation in pea
protein isolate–chitosan mixtures. Food Research International, 44(5), 1441–1446.
Espinosa-Andrews, H., Enriquez-Ramirez, K. E., Garcia-Marquez, E., Ramirez-Santiago, C.,
Lobato-Calleros, C., & Vernon-Carter, J. (2013). Interrelationship between the zeta po-
tential and viscoelastic properties in coacervates complexes. Carbohydrate Polymers,
95(1), 161–166.
214 Y. Yuan et al. / Food Research International 55 (2014) 207–214
Huang, G. Q., Sun, Y. T., Xiao, J. X., & Yang, J. (2012). Complex coacervation of soybean pro-
tein isolate and chitosan. Food Chemistry, 135(2), 534–539.
Klein, M., Aserin, A., Ben Ishai, P., & Garti, N. (2010). Interactions between whey protein
isolate and gum arabic. Colloids and Surfaces B: Biointerfaces, 79(2), 377–383.
Laneuville, S. I., Turgeon, S. L., Sanchez, C., & Paquin, P. (2006). Gelation of native
β-lactoglobulin induced by electrostatic attractive interaction with xanthan gum.
Langmuir, 22(17), 7351–7357.
Le, X. T., & Turgeon, S. L. (2013). Rheological and structural study of electrostatic
cross-linked xanthan gum hydrogels induced by β-lactoglobulin. Soft Matter, 9(11),
3063–3073.
Lee, A.C., & Hong, Y. H. (2009). Coacervate formation of α-lactalbumin–chitosan and
β-lactoglobulin–chitosan complexes. Food Research International, 42(5–6),
733–738.
Liu, S. H., Cao, Y. L., Ghosh, S., Rousseau, D., Low, N. H., & Nickerson, M. T. (2010). Intermo-
lecular interactions during complex coacervation of pea protein isolate and gum ara-
bic. Journal of Agricultural and Food Chemistry, 58(1), 552–556.
Liu, C., Yang, X. Q., Lin, M. G., Zhao, R. Y., Tang, C. H., Luo, L., et al. (2011). Complex coac-
ervation of chitosan and soy globulins in aqueous solution: A electrophoretic mobility
and light scattering study. International Journal of Food Science and Technology, 46,
1363–1369.
Mounsey, J. S., O'Kennedy, B. T., Fenelon, M.A., & Brodkorb, A. (2008). The effect of heating
on β-lactoglobulin–chitosan mixtures as influenced by pH and ionic strength. Food
Hydrocolloids, 22(1), 65–73.
Murakami, R., & Takashima, R. (2003). Mechanical properties of the capsules of chitosan–
soy globulin polyelectrolyte complex. Food Hydrocolloids, 17(6), 885–888.
Nagano, T., Hirotsuka, M., Mori, H., Kohyama, K., & Nishinari, K. (1992). Dynamic visco-
elastic study on the gelation of 7S globulin from soybeans. Journal of Agricultural
and Food Chemistry, 40(6), 941–944.
Plashchina, I. G., Mrachkovskaya, T. A., Danilenko, A. N., Kozhevnikov, G. O.,
Starodubrovskaya, N. Y., Braudo, E. E., et al. (2001). Complex formation of faba bean
legumin with chitosan: Surface activity and emulsion properties of complexes. Food Col-
loids (pp. 293–303). Cambridge: Royal Society of Chemistry.
Schmitt, C., & Turgeon, S. L. (2011). Protein/polysaccharide complexes and coacervates in
food systems. Advances in Colloid and Interface Science, 167(1–2), 63–70.
Sorgentini, D. A., & Wagner, J. R. (1999). Comparative study of structural characteristics
and thermal behavior of whey and isolate soybean proteins. Journal of Food
Biochemistry, 23(5), 489–507.
Staswick, P. E., Hermodson, M.A., & Nielsen, N. C. (1981). Identification of the acidic and basic
subunit complexes of glycinin. Journal of Biological Chemistry, 256(16), 8752–8755.
Takeuchi, T., Morita, K., Saito, T., Kugimiya, W., & Fukamizo, T. (2006). Chitosan–
soyprotein interaction as determined by thermal unfolding experiments. Bioscience,
Biotechnology, and Biochemistry, 70(7), 1786–1789.
Thanh, V. H., & Shibasaki, K. (1977). Beta-conglycinin from soybean proteins. Isolation
and immunological and physicochemical properties of the monomeric forms.
Biochimica and Biophysica Acta, 490, 370–384.
Turgeon, S. L., & Laneuville, S. I. (2009). Protein + polysaccharide coacervates and com-
plexes: From scientific background to their application as functional ingredients in
food products. In K. Stefan, T. N. Ian, & B. U. Johan (Eds.), Modern biopolymer science
(pp. 327–363). San Diego: Academic Press.
Turgeon, S. L., Schmitt, C., & Sanchez, C. (2007). Protein–polysaccharide complexes and
coacervates. Current Opinion in Colloid & Interface Science, 12(4–5), 166–178.
van Nieuwenhuyzen, W., & Szuhaj, B. F. (1998). Effects of lecithins and proteins on the
stability of emulsions. Lipid/Fett, 100(7), 282–291.
van Vliet, T., Martin, A. H., & Bos, M.A. (2002). Gelation and interfacial behaviour of veg-
etable proteins. Current Opinion in Colloid & Interface Science, 7(5–6), 462–468.
Vardhanabhuti, B., & Foegeding, E. A. (2008). Effects of dextran sulfate, NaCl, and initial
protein concentration on thermal stability of β-lactoglobulin and α-lactalbumin at
neutral pH. Food Hydrocolloids, 22(5), 752–762.
Vardhanabhuti, B., Yucel, U., Coupland, J. N., & Foegeding, E. A. (2009). Interactions be-
tween β-lactoglobulin and dextran sulfate at near neutral pH and their effect on ther-
mal stability. Food Hydrocolloids, 23(6), 1511–1520.
Weinbreck, F., Tromp, R. H., & de Kruif, C. G. (2004). Composition and structure of whey
protein/gum arabic coacervates. Biomacromolecules, 5(4), 1437–1445.
Wolf, W. J. (1970). Soybean proteins: Their functional, chemical, and physical properties.
Journal of Agricultural and Food Chemistry, 18(6), 969–976.
Yuan, Y., Wan, Z. L., Yin, S. W., Teng, Z., Yang, X. Q., Qi, J. R., et al. (2013). Formation and
dynamic interfacial adsorption of glycinin/chitosan soluble complex at acidic pH: Re-
lationship to mixed emulsion stability. Food Hydrocolloids, 31(1), 85–93.