Journal Pre-proof

Effect of Kappa carrageenan on acid-induced gelation of whey protein aggregates.

Part I: Potentiometric titration, rheology and turbidity

Dasong Liu, Peng Zhou, Taco Nicolai

PII: S0268-005X(19)32358-6
DOI: https://doi.org/10.1016/j.foodhyd.2019.105589
Reference: FOOHYD 105589

To appear in: Food Hydrocolloids

Received Date: 9 October 2019

Revised Date: 25 November 2019
Accepted Date: 10 December 2019

Please cite this article as: Liu, D., Zhou, P., Nicolai, T., Effect of Kappa carrageenan on acid-induced
gelation of whey protein aggregates. Part I: Potentiometric titration, rheology and turbidity, Food
Hydrocolloids (2020), doi: https://doi.org/10.1016/j.foodhyd.2019.105589.

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.

© 2019 Published by Elsevier Ltd.

 Graphical abstract
Effect of Kappa Carrageenan on Acid-induced Gelation of Whey Protein Aggregates. Part I:
Potentiometric Titration, Rheology and Turbidity

Dasong et al.

103

Elastic modulus (Pa)

Optical absorbance
2

pH 102
1

OA
Gel
whey protein κ -carrageenan mixed gel 101 0
aggregates -4 -2 0 2 4 6
net charge density
 1 Effect of Kappa Carrageenan on Acid-induced Gelation of Whey Protein
2 Aggregates. Part I: Potentiometric Titration, Rheology and Turbidity

4 Dasong Liua,b, Peng Zhou a, Taco Nicolaib,*

a
5 State Key Laboratory of Food Science and Technology, Jiangnan University, Wuxi, Jiangsu Province
6 214122, China

b
7 Le Mans Université, IMMM UMR-CNRS 6283, Polymères, Colloïdes et Interfaces, 72085 Le Mans
8 cedex 9, France

9 Email: [email protected]

10

11 Abstract

12 Gelation of suspensions of whey protein aggregates (WPA) was induced by lowering the pH in-situ
13 using GDL. The effect of adding κ-carrageenan (KC) on the relationship between the net charge
14 density of the proteins (α) and the pH was determined by potentiometric titration. It showed that
15 complexes were formed for α > -2. The effect of adding KC on the storage shear modulus and the
16 turbidity was studied by oscillatory shear rheology and UV-Vis spectrometry. The KC concentration
17 dependence was investigated up to 6 g/L at a fixed protein concentration of 20 g/L. Gels were
18 formed for α > -3 in the absence of KC and α > -3.7 with 2 g/L KC, which was accompanied by an
19 increase of the turbidity starting at the same α. The elastic modulus and the turbidity of the gels at
20 steady state increased with increasing α. For a given α, a strong decrease of the elastic modulus and
21 a strong increase of the turbidity were observed with increasing KC concentration. The effect of the
22 gelation temperature was studied between 20°C and 80°C. The gelation rate increased strongly with
23 increasing temperature and was controlled by the degradation rate of GDL, but it was not influenced
24 by the KC concentration. Increasing the gelation temperature led to an increase of the elastic
25 modulus and the turbidity at steady state.

26 Keywords whey proteins; carrageenan; gel; complex; rheology; titration; turbidity

27

28 1. Introduction

1
29 Food products often contain proteins and polysaccharides that can be induced to aggregate
30 and gel in aqueous solution, e.g. by heating, adding salt or changing the pH, which is exploited to give
31 them desired textures. It is clear that when both types of biopolymers are present, gelation of
32 proteins and/or polysaccharides will be influenced by the interaction between the two biopolymers
33 (Kasapis, 2008; Nicolai, 2019; Norton and Frith, 2001; van de Velde, de Hoog, Oosterveld and Tromp,
34 2015). When the proteins and polysaccharides are both strongly charged with the same sign,
35 repulsive interactions can drive segregative phase separation into a protein rich and polysaccharide
36 rich phase (Corredig, Sharafbafi and Kristo, 2011; Frith, 2010; Turgeon, Beaulieu, Schmitt and
37 Sanchez, 2003). However, when the sign of the charge is opposite, proteins and polysaccharides bind
38 together leading to formation of complexes or coacervates (De Kruif, Weinbreck and de Vries, 2004;
39 Kayitmazer, Seeman, Minsky, Dubin and Xu, 2013; Schmitt and Turgeon, 2011; Turgeon and
40 Laneuville, 2009) .

41 Here we report on an investigation of the effect of adding the polysaccharide κ-carrageenan

42 (KC) on gelation of whey protein aggregates (WPA) induced by acidification. Whey proteins are a
43 mixture of globular proteins consisting mainly of β-lactoglobulin and α-lactalbumin (Guo, 2019). If
44 whey proteins are heated in aqueous solution at neutral pH they form irreversibly bound aggregates
45 with a size that increases with increasing concentration and a gel above a critical concentration
46 (Brodkorb, Croguennec, Bouhallab and Kehoe, 2016; Nicolai, Britten and Schmitt, 2011). Stable
47 suspensions of WPA at low temperatures can be induced to form gels by reducing the electrostatic
48 repulsion either by adding salt or by lowering the pH towards the isoionic point (pI) (Bryant and
49 McClements, 1998; Kharlamova, Chassenieux and Nicolai, 2018a; Kharlamova, Nicolai and
50 Chassenieux, 2018b). Potentiometric titrations showed that pI = 4.8 for native whey proteins and
51 slightly higher for WPA (pI = 5.0) (Kharlamova, Inthavong, Nicolai and Chassenieux, 2016).
52 Carrageenan is a family of anionic sulfated polysaccharides isolated from algae. The most common
53 types are κ-, ι- and λ-carrageenan, respectively, with 1, 2 and 3 sulfate groups per sugar unit. KC is
54 widely used in the food industry for its property to form a gel in aqueous solution below a critical
55 temperature that depends on the type and concentration of added salt (Piculell, 2006). Gel
56 formation is induced by a transition of the chain conformation at the critical temperature from a coil
57 to a helix. However, here we consider only conditions where KC is in the coil conformation.

58 Mixtures of whey proteins and KC have been studied in some detail at pH 7. It was reported
59 that, whereas mixtures with native proteins are miscible over a wide range of concentrations,
60 segregative phase separation occurs above a critical KC concentration in mixtures with WPA
61 (Baussay, Durand and Nicolai, 2005; Cakir and Foegeding, 2011; Capron, Nicolai and Smith, 1999;
62 Croguennoc, Durand and Nicolai, 2001; Nguyen, Phan-Xuan, Benyahia and Nicolai, 2014). The critical

2
63 KC concentration was found to decrease strongly with increasing size of the protein aggregates and
64 more weakly with increasing WPA concentration. The proteins were strongly partitioned between
65 the two phases and KC was depleted to some extent form the protein rich domains (Nguyen, Nicolai,
66 Chassenieux and Benyahia, 2014). As far as we are aware, the effect of reducing the pH on the phase
67 diagram has not yet been investigated. Reducing the pH leads to a reduction of the net negative
68 charge of the proteins (α) and it becomes positive for pH < pI. The negative charge of KC is caused by
69 sulphate groups that remain fully deprotonated for pH > 4, which means that charge density of KC
70 does not depend significantly on the pH in this range. The implication is that repulsive interactions
71 between KC and the proteins decreases with decreasing pH. Below a critical pH (pHc) the effective
72 interaction between the proteins and the polysaccharides is expected to become attractive. As a
73 consequence, it is expected that the tendency for KC and WPA to phase separate segregatively
74 decreases with decreasing pH and that below pHc complexes containing both biopolymers are
75 formed.

76 Formation of complexes in dilute mixtures of native whey protein isolate (WPI) and
77 carrageenan has been studied as a function of the pH and the biopolymer composition (Stone and
78 Nickerson, 2012; Lam and Nickerson, 2014); Weinbreck, Nieuwenhuijse, Robijn, de Kruif, &
79 Chemistry, 2004). Formation of complexes between the proteins and the different types of
80 carrageenan was observed when the pH was decreased below approximately pH 5.4 causing a sharp
81 increase of the turbidity. Complexation starts at pH > pI, because the negatively charged carrageenan
82 can interact with positive patches of the net positively charged proteins. pHc did not depend much on
83 the protein/polysaccharide ratio, but the extent of complexation as measured by the turbidity
84 increased sharply with increasing protein/polysaccharide ratio up to maximum beyond which it
85 decreased gradually.

86 De Jong et al. (De Jong, Klok and Van de Velde, 2009; de Jong and van de Velde, 2007)
87 studied acid-induced gelation of WPA with a hydrodynamic radius Rh = 20 nm by reducing the pH to
88 4.8 with GDL at CWP= 30 g/L in the presence of KC. The mixtures were prepared at neutral pH and at
89 low KC concentrations so that they were homogeneous. Light scattering and rheology measurements
90 showed a steep increase of the scattering intensity, the apparent hydrodynamic radius and the
91 oscillatory storage modulus below pH 5.7 caused by aggregation and gelation, both in the absence
92 and presence of 0.5 g/L KC. During acidification, the microstructure of individual WPA gels observed
93 by confocal laser scanning microscopy (CLSM) remained homogeneous, but in the presence of 0.5 g/L
94 KC it became heterogeneous at pH ≤ 5.6. The gel containing 0.5 g/L KC was slightly stiffer than
95 without KC. The heterogeneity increased with increasing KC concentration between CKC= 0.25 and 3
96 g/L and for CKC > 1.5 g/L a network of dense protein microdomains was formed. The liquid mixtures at

3
 97 neutral pH were homogeneous and the heterogeneous structure appeared during acidification at the
98 same pH as where the gel was formed. The effect of KC on the microstructure was attributed to
99 segregative phase separation between KC and WPA leading to formation of protein dense
100 microdomains. Subsequently, the protein domains connect to each other to form a system spanning
101 network, which arrests the process of phase separation. Since the mixtures were initially
102 homogeneous it was assumed that segregative phase separation was induced by the growth of the
103 WPA when the interaction with KC was still net repulsive. Such a phenomenon has been reported for
104 gelation of WPA in the presence of KC at neutral pH induced by adding NaCl (Ako, Durand and
105 Nicolai, 2011). At neutral pH there is definitely no complex formation between KC and WPA and
106 segregative phase separation could clearly be identified during the gelation process.

107 Acid-induced gelation of WPA has also been studied in the presence of other types of
108 polysaccharides (De Jong et al., 2009; Van den Berg, Rosenberg, Van Boekel, Rosenberg and Van de
109 Velde, 2009; Van den Berg, Van Vliet, Van der Linden, Van Boekel and Van de Velde, 2007; Zhang,
110 Hsieh and Vardhanabhuti, 2014a; Zhang and Vardhanabhuti, 2014b). For polysaccharides with a low
111 charge density at the relevant pH such as locust bean gum, phase separation was observed on larger
112 length scales and was more important when the acidification rate was slower. For mixtures with a
113 higher charge density than KC, such as ι-carrageenan the observed microstructure of the gels was
114 homogeneous. The general picture that has emerged from these studies is that the microstructure is
115 determined by the interplay between segregative phase separation between WPA and the
116 polysaccharides on one hand and gelation, which stops the progression of phase separation, on the
117 other. The effect of complexation between WPA and anionic polyelectrolytes, which unavoidably
118 occurs at low pH, was not explicitly addressed.

119 It is clear, however, that anionic polysaccharides are bound to the protein network at pH < pI.
120 No anionic polysaccharide was found in serum released from compressed gels (Van den Berg et al.,
121 2007) and SEM images of mixed gels at pH < pI show that the anionic polysaccharide was bound to
122 the protein network (Van den Berg et al., 2009; Zhang et al., 2014b). An important issue is therefore
123 whether the pH at which WPA start to bind to each other is higher than the pH at which the
124 polysaccharides start to bind to the proteins. Only in that case can segregative phase separation
125 occur in the pH range where WPA bind to each other, but not yet to KC. Comparison of results on
126 formation of complexes between κ-car and native WPI as a function of decreasing pH (Stone et al.,
127 2012) and acid induced gelation of WPA (Ahmed, Aschi and Nicolai, 2018) shows that these
128 phenomena both start at almost the same pH ≈ 5.4. An investigation of the association between
129 WPA and KC during acidification also showed that flocculation of the aggregates in the absence of KC
130 and complexation of KC and the aggregates started at approximately the same pH (Khemissi et al.,

4
131 2018). A precise quantitative comparison requires a detailed investigation of complex formation and
132 gelation for the same type and size of WPA. If complexation occurs before gelation the
133 heterogeneous microstructure of the gels may be caused by associative phase separation leading to
134 microdomains rich both in protein and KC.

135 The interaction between proteins and anionic polysaccharides depends on their net charge
136 density, which for proteins can be varied by varying the pH. The effect of adding polysaccharide on
137 acid-induced gelation has in general been studied as a function of the pH. However, for a given pH
138 the net charge density of proteins (α) varies when the protein concentration is varied (Kharlamova et
139 al., 2016) or when polysaccharides are added (Ahmed et al., 2018; Wen and Dubin, 1997). This means
140 that it is better to study the effect of adding polysaccharides at constant α rather than at constant
141 pH. For a given α, changes in the hydronium concentration corresponding to changes in the pH are
142 less than 0.1 mM for 4 < pH < 10, which means that the effect on the ionic strength and thus on
143 electrostatic interactions between the biopolymers is negligible.

144 Another aspect that has received only limited attention so far is the effect of temperature. It
145 was reported recently that the rate of acid-induced gelation increases strongly with increasing
146 temperature (Kharlamova et al., 2018a). Furthermore, it is well known that the rate of GDL
147 degradation increases strongly with increasing temperature (Pocker and Green, 1973). De Jong et al.
148 (2009) looked at the effect of temperature on the microstructure of mixtures of WPA at 30 g/L and
149 KC at 0.5 g/L at 6, 25 and 37°C. Not much difference was seen between gels formed at 25°C and 37°C,
150 but the microstructure remained homogeneous at 6°C. The latter was explained by the coil-helix
151 transition of KC in the mixture during cooling below 17°C. The effect of the gel temperature on the
152 gel stiffness has not yet been investigated.

153 The aim of the present investigation was to study in more detail the interaction between KC
154 and whey proteins during acidification. We have done potentiometric titrations in the presence of
155 different KC concentrations to investigate the effect of the latter on the relationship between α and
156 the pH. Such measurements can also show the onset and degree of complexation (Ahmed et al.,
157 2018; Wen et al., 1997). The effect of KC on the shear moduli and the turbidity was studied at
158 different temperatures as a function of time during acidification and as a function of the steady state
159 protein charge density. The effect of KC on the microstructure was investigated by CLSM using
160 different specific labeling for the proteins and the polysaccharides, which allowed us to determine
161 the distribution of each biopolymer in the mixtures. These latter results will be reported and
162 discussed in part II.

5
163

164 2. Materials and methods

165 2.1. Sample preparation

166 Whey protein isolate (BiPRO®, batch JE 254-3-420) was purchased from Davisco Foods
167 International, Inc. (Le Sueur, MN, USA). The composition of the powders was reported by (Nguyen,
168 Nicolai, Chassenieux, Schmitt and Bovetto (2016). Size exclusion chromatography (SEC) showed
169 that the BiPRO® sample contained 65% β-lg and 20% α-lac. The remaining 15% consisted of other
170 whey proteins and a small amount of caseins. The powders were dissolved in Milli-Q water
171 containing 200 ppm NaN3 to protect against bacterial growth by stirring overnight. The protein
172 concentration was determined by UV absorption at 278nm using ε=1.01 g/L/cm.

173 Whey protein aggregates were prepared by heating a solution of 62 g/L WPI at pH 7 at 80°C
174 during 24 hours. This heating time is largely sufficient to reach a steady state in which all native WPI
175 is converted into aggregates. The average molar mass (Mw = 1.4×106 kg/mol) and the hydrodynamic
176 radius (Rh = 26 nm) was determined in highly diluted solutions using static and dynamic light
177 scattering as described in detail elsewhere (Inthavong, Kharlamova, Chassenieux and Nicolai, 2016).
178 Mixtures were prepared by adding a concentrated solution of KC to the WPA suspensions under
179 stirring. Subsequently, gelation was induced by addition of aliquots of a freshly prepared
180 concentrated GDL solution to the mixtures under stirring.

181 The κ-carrageenan sample used for this study was purchased from Sigma. It was purified by
182 extensive dialysis first against 0.1 M NaCl in order to exchange other cations for sodium and then
183 against Milli-Q water to remove excess salt. The purified KC was subsequently freeze-dried. The
184 water content of the purified KC powder was 8% and the concentrations of Na+ and K+ were
185 40 mg/g and 0.1 mg/g, respectively, as determined by flame spectroscopy, whereas no Ca2+
186 was detected.

187 Stock solutions of KC were prepared by dissolving the freeze-dried KC at a

188 concentration of 30 g/L in Milli-Q water with 200 ppm sodium azide added as a
189 bacteriostatic agent by stirring for few hours at room temperature. The pH of the stock
190 solution was 7. For observation with CLSM a small amount of labeled KC was added. KC
191 concentrations are expressed in gram KC powder per liter. The molar mass distribution of
192 the KC was determined by SEC with on-line refractive index (Shodex RI 71, Showa Denko

6
193 K.K., Tokyo, Japan) and light scattering (Dawn EOT™,Wyatt technology, Santa Barbara, CA,
194 USA) detection taking for the refractive index increment dn/dC = 0.145 ml/g. A TSK gel
195 G6000PW column was used with 0.1 M NaNO3 at pH 7 as the mobile phase. In this manner
196 we found for the weight and number average molar mass: Mw = 4.5× 105 g/mol and Mn= 2.7 ×
197 105 g/mol.

198

199 2.2. Gel formation

200 Mixtures of WPA and KC were prepared from stock solutions at the desired concentration. A
201 GDL solution at 1 M was prepared by adding water to the GDL powder. After stirring for 1 min to
202 dissolve all the GDL powder, aliquots of the GDL solution were added to the biopolymer solutions
203 while stirring. To slow down the degradation of GDL during the sample preparation icy cold water
204 was added to the GDL powder and the solution was stirred on ice.

205

206 2.3 Methods

207 2.3.1 Potentiometric titration. Titration experiments were conducted at room temperature
208 using an automatic titrator (TIM 856, Radiometer Analytical) equipped with a combined pH electrode
209 and a temperature probe. The pH electrode was calibrated by a three-point calibration using
210 standard buffers. Dilute biopolymer solutions (CWP = 1 g/L) were titrated to pH 4 with a standard
211 solution of HCl ([HCL] = 0.1M or 1M) at a rate of 0.1 mL/min. Concentrated biopolymer solutions (CWP
212 = 20 g/L) could not be titrated in this manner as they formed gels and were therefore prepared
213 separately at different [HCl] and their pH was measured after breaking up the gel. Changes in the
214 number of charges per protein were calculated as the molar concentration of added HCl divided by
215 the molar concentration of whey proteins:

216 ∆α=[HCl].MWP/CWP 1

217 using for the average molar mass of the whey proteins MWP = 1.75×104 g/mol. The amount of added
218 H+ that remained in solution was negligible down to pH 4 and is therefore not considered in eq. 1.
219 Titration curves of pH vs ∆α obtained at different NaCl concentrations [NaCl] cross at the isoionic
220 point, because the pH is independent of [NaCl] when α = 0, which has been discussed in detail
221 elsewhere (Kharlamova et al., 2016). In this manner it was found that pI = 5.1 ± 0.1 for the WPA used
222 for this study, see figure S1 of the supplementary information. Knowing the pI, it is straightforward to

7
223 calculate α as a function of the pH as α(pH) = ∆α(pH) - ∆α(pI). Here we have ignored the fraction of
224 HCl that binds to KC, which is negligible for pH > 4.

225 If α is varied by hydrolysis of added GDL one needs to consider the dissociation constant of δ-
226 gluconic acid: HGA⇄ GA- + H+. We have determined the pKa of this reaction by measuring the pH at
227 different GDL concentrations, see figure S2 of the supporting information. The results showed that
228 pKa= 3.75 both in Milli-Q water and at 0.1 M NaCl, which is within the experimental error the same
229 as the value reported by Pocker et al. (pKa = 3.77). Knowing the pKa, the GA- concentration was
230 calculated at each pH and ∆α was calculated using eq. 1 by replacing [HCl] with the amount of H+
231 produced by the gluconic acid at each pH, i.e. [GA-].

232 2.3.2.Rheology. The loss (G") and storage (G') oscillatory shear moduli were measured using
233 a stress-controlled rheometer (AR2000, TA Instruments, Crawley, UK) equipped with a cone-plate
234 geometry (diameter: 40 mm; angle: 2º; truncation: 54 μm). Evaporation was avoided by covering the
235 sample with a silicon oil. Measurements were done as a function of time at an oscillation frequency
236 of 0.1 Hz. After the time sweep the temperature was decreased to 20°C at a rate of 5°C per min while
237 monitoring the shear moduli at 0.1 Hz. After reaching 20°C the frequency dependence of the shear
238 moduli was measured.

239 2.3.3. Optical density. The light transmittance (T) was measured as a function of time at a
240 wavelength of 600 nm at different temperatures in cells with a path length of 0.1 cm using a UV-
241 visible spectrometer Varian KC y - 50 Bio. The optical density was calculated as OD = log(1/T).

242

243 3. Results and Discussion

244 3.1 Relationship between the pH and the protein charge density

245 Figure 1 shows the pH as a function of α for solutions at a fixed protein concentration CWP = 1
246 g/L and different KC concentrations (CKC). A clear difference between titration curves of WPA with
247 and without KC on the pH was observed for pH < 5.7, which was caused by complexation between
248 the two biopolymers. Complexation started when the net charge of the proteins was still negative
249 (α≈ -2), due to the presence of positively charged patches on the proteins. Complexation with KC
250 caused an increase of α for a given pH, as the negative charge of KC does not change significantly for
251 pH > 4. The effect of complexation on the titration curves increased with increasing KC concentration
252 until CKC = 0.4 g/L. It appears that binding of KC to WPA saturates at weight ratio WPA:KC ≈ 2:1.

8
 8

0 g/L
0.1 g/L
7 0.2 g/L.
0.4 g/L
0.6 g/L
1 g/L

pH
6

4
-5 0 5 10 15

253 α

254 Fig. 1 pH as a function of the net charge density for solutions containing 1 g/L WPA and different
255 amounts of KC as indicated in the figure.

256

257 The objective of the present study was to investigate the effect of KC on acid-induced
258 gelation of WPA by adding GDL. Preliminary measurements showed that self-supporting gels were
259 formed after addition of GDL for WPA suspensions at CWP =10 g/L and higher, whereas at CWP = 5 g/L
260 precipitation was observed. For this study we have varied the KC concentration and the amount of
261 GDL and fixed the WPA concentration at 20 g/L.

262 As was mentioned above, the properties of the gels depend on the net charge of the
263 proteins. The value of α that is reached at steady state is not only related to the GDL concentration,
264 but also to the pH as δ-gluconic acid (HGA) is a weak acid. We verified that the pKa of HGA was not
265 influenced by the presence of KC and WPA, by comparing the pH at steady state with that obtained
266 in the presence of HCl. This is shown in figure 2 where the pH at steady state is plotted as a function
267 of α for a solution containing 20 g/L WPA and different amounts of KC. The pH was set either by
268 adding HCl or GDL and α was calculated in the latter case by replacing [HCl] in eq.1 with [GA-] using
269 pKa=3.75 to calculate [GA-] from [GDL]. Figure 2 shows that for a given pH, the calculated values of α
270 were the same within the experimental uncertainty. The relationship between the GDL concentration
271 and α is shown in figure S3 of the supplementary information.

9
 7 0 g/L
2 g/L
4 g/L
12 g/L

6
pH

4
-8 -6 -4 -2 0 2 4 6 8

272 α

273 Fig. 2 pH as a function of the net charge density of the proteins for solutions containing 20 g/L WPA
274 and different amounts of KC as indicated in the figure. The pH was set either by adding HCl (red
275 symbols) or by adding GDL (green symbols). Solid lines represent guides to the eye.

276

277 At high pH, addition of KC caused a small decrease of the pH, which is caused by screening of
278 electrostatic interactions by the counterions from KC. A similar decrease was also observed when
279 NaCl was added (Kharlamova et al., 2016). This effect of KC was not observed in Fig. 1, because in
280 that case only very small amounts of KC were added. At α ≈ -2 the titration curves of WPA in the
281 presence of KC cross those in the absence of KC and the effect of complexation becomes important
282 for larger α. The effect of KC on the titration curve increased with increasing KC concentration, but
283 became small for CKC > 4 g/L at least in the range of α that was investigated in this study. Comparison
284 of Figs 1 and 2 shows that at both WPA concentrations the effect of complexation started at
285 approximately α ≈ -2. However, the corresponding pH depended weakly on the WPA concentration:
286 pH 5.4 at CWP = 20 g/L versus pH 5.7 at CWP = 1 g/L.

287

288 3.2. Rheology

289 3.2.1 Effect of the KC concentration

10
290 Gel formation was studied by measuring the shear moduli as a function of time after adding
291 GDL. Figure 3 shows the evolution of the storage shear modulus (G') for suspensions at CWP =20 g/L
292 and at different concentrations of KC at 20°C and at 80°C after adding 9 mM GDL. At this
293 concentration of GDL the steady state value of α was close to zero and very little dependent on the
294 KC concentration, see figure S3 of the supplementary information. The pH at steady state varied
295 between 4.9 and 5.1, see figure 2. G' increased steeply when a gel was formed and became much
296 larger than the loss modulus. The gel time (tg) defined as the time when G' starts to increase steeply
297 clearly did not depend significantly on the KC concentration, but was much larger at 20°C than at
298 80°C. Notice that the time to reach 80°C was about one minute. Though gelation started at
299 approximately the same time, the evolution of G' was more gradual when more KC was present.

103
20°C 0 g/L
0.5
1.0
102 1.5
1.625
G' (Pa)

1.75
1.875
101 2.0
4.0
6.0

100

102 103

300 Time (min)

103 0 g/L
80°C
0.25
0.5
102 0.75
1.0
G' (Pa)

1.5
101 2.0
2.5
3.0
100 4.0
5.0
6.0
10-1
101 102
Time (min)

301

11
302 Fig. 3 Storage modulus at 0.1 Hz as a function of time after adding 9 mM GDL (α≈ 0) at 20°C or 80°C
303 for WPA solutions at 20 g/L and different KC concentrations indicated in the figure. The final value of
304 G’ decreased with increasing KC concentration.

305

306 For all gels, the frequency dependence of G' was small and G' > G" over the whole explored
307 frequency range (0.01 - 10 Hz) (results not shown). At the end of the heating ramp, the gels formed
308 at 80°C were cooled to 20°C, which led to a significant increase of G'. The relative increase by about a
309 factor of 4.5 was approximately the same for all KC concentrations, see figure S4a for the
310 supplementary information. For a fair comparison between the dependence of the elastic modulus
311 on the KC concentration for gels formed at 20°C and at 80°C, we need to take into account the
312 difference in the kinetics. Therefore we have taken G' at the same value of t/tg≈ 4 corresponding to
313 the final time at 20°C. We also need to consider the effect of the temperature on G'. Therefore we
314 multiplied the values of G' obtained at 80°C by 4.5 so that they correspond to those after cooling to
315 20°C. The results are plotted in fig. 4 and show that G' depended only weakly on the KC
316 concentration for CKC < 1 g/L. Between CKC = 1 g/L and 2 g/L G' decreased steeply in particular for gels
317 formed at 20°C followed by a more gradual decrease at higher KC concentrations. Interestingly, the
318 decrease of G' occurred when the KC chains started to overlap (Croguennoc, Meunier, Durand and
319 Nicolai, 2000). The elastic moduli formed at 80°C were systematically larger than those formed at
320 20°C.

103 80°C
20°C

102
G' (Pa)

101

0 1 2 3 4 5 6

CKC (g/L)
321

12
322 Fig. 4 Storage modulus at 20°C of gels at t/tg≈ 4 formed at 20°C or 80°C after adding 9 mM GDL (α≈ 0)
323 to WPA solutions at 20 g/L as a function of the KC concentration.

324

325 3.2.3 Effect of the temperature

326 It is clear that there is an important effect of temperature on the gelation process, which we
327 studied in more detail for a few KC concentrations. The effect of temperature on the evolution of G'
328 after adding 9 mM GDL is shown in fig. 5a for WPA solutions containing 2 g/L KC. The gelation rate
329 decreased strongly with decreasing temperature and there was also an effect on the elastic modulus.
330 Master curves could be obtained after time-temperature superposition using both horizontal and
331 vertical shifts. Fig. 5b shows the evolution of G'/Gel as a function of t/tg, where Gel is defined as G' at
332 t/tg = 20.

a
102
G' (Pa)

101

100
101 102 103

333 time (min)

100 b

10-1
G'/Gel

80°C
65°C
10-2 50°C
35°C
20°C
10-3
100 101
t/tg
334

13
335
336 Fig. 5a Storage modulus at 0.1 Hz as a function of time for solutions of WPA at 20 g/L and 2 g/L KC
337 after adding 9 mM GDL (α≈ 0) measured at different temperatures indicated in fig. 5b. The
338 corresponding master curve is shown in Fig.5b.

339

340 Similar measurements were done at other KC concentrations and in all cases master curves
341 could be obtained, see fig. 6. G' increased more gradually with time, i.e. with increasing α, when
342 more KC was present. Notice that values of G' that were shown in fig. 4 were obtained at t/tg = 4,
343 whereas Gel was the value of G’ obtained at t/tg = 20. Since G’ increases weakly between t/tg = 4 and
344 t/tg = 20 the values of G’ shown in fig. 4 are 2 to 4 times smaller than Gel.

100

10-1
G'/Gel

0 g/L
10-2 1 g/L
1.5 g/L
2 g/L
4 g/L
10-3
100 101
t/tg
345

346 Fig. 6 Master curves of G' vs t after normalization by Gel and tg, respectively, for WPI gels at 20 g/L
347 with 9 mM GDL (α≈ 0) and different concentrations of KC as indicated in the figure.

348

349 An Arrhenius plot of the temperature dependence of tg is shown in fig. 7a. tg was not
350 significantly influenced by the presence of KC as was already shown in figure 3 for gels formed at
351 80°C or at 20°C. Log (tg) had a linear dependence on 1/T and from the slope an activation energy Ee =
352 61 kJ/mol was obtained. This value is close to that reported by Pocker et al. (Ea = 59 kJ/mol) for the
353 degradation rate of GDL. We determined the degradation rate of GDL at 20°C and 50°C by measuring
354 the pH as a function of time after adding water to the GDL powder, see figure S5a of the

14
355 supplementary information. The fraction of degraded GDL increased linearly with time up to about
356 50% and then increased more slowly to reach 95% after 10 h at 20°C. Degradation was faster at 50°C
357 and superimposed with that at 20°C using the same shift factor as was found in the rheology
358 measurements, see Figure S5b of the supplementary information. We may therefore conclude that
359 the gelation rate was controlled by the rate at which GDL degrades and therefore the rate at which α
360 increases.

103
a

102
tg (min)

101 0 g/L
1 g/L
1.5 g/L
2 g/L
4 g/L
100
2.8 2.9 3.0 3.1 3.2 3.3 3.4 3.5
-1
361 1000/T (K )

103
Gel (Pa)

102

101
10 20 30 40 50 60 70 80 90

362 T (°C)

363 Fig. 7a Arrhenius representation of the dependence of the gel time for WPA solutions at 20 g/L with
364 9 mM GDL (α≈ 0) and different KC concentrations indicated in the figure. The solid line represents
365 linear least squares fit.
366 Fig. 7b Dependence of Gel on the temperature of gel formation measured after cooling the gels to
367 20°C.

368

15
369 Figure 7b compares the dependence of Gel measured at 20°C on the temperature at which
370 the gels were formed. At all temperatures Gel decreased with increasing KC concentration, but the
371 decrease was stronger at lower gelation temperatures in agreement with the comparison between
372 20°C and 80°C shown in figure 4. In all cases, Gel was larger if the gels were formed at higher
373 temperatures in particular for gels containing 2 or 4 g/L KC. The effect of the gelation temperature
374 was small in the absence of KC as had already been reported elsewhere (Kharlamova et al., 2018a)
375 and it remained small up to CKC = 1.5 g/L. However, at larger CKC when the effect of KC on Gel became
376 important, the effect of the gelation temperature also became important. One possible explanation
377 is that phase separation was less important at higher temperatures leading to a weaker decrease of
378 Gel with increasing CKC.

379

380 3.2.3 Effect of the net charge density

381 The effect of α was investigated at 80°C for WPA solutions without KC and with 2 g/L KC, see
382 figure 8. With increasing α stiffer gels were formed and tg decreased. Syneresis occurred for gels
383 formed in the absence of KC at α ≥ 1.8 causing a decrease of G', because part of the gel lost contact
384 with the geometry. Strong effects of syneresis were also observed in the presence of KC at α > 3.1
385 (not shown).

103
a -3
-2.7
-2.4
-2.0
102 -1.3
-0.1
G' (Pa)

0.7
1.8
101

100

101 102

386 Time (min)

16
 103
b -3.6
-3.2
-2.7
-2.4
102 -2.0
-1.2

G' (Pa)
-0.6
0.1
1.2
101 3.1

100
101 102

387 Time (min)

388

389 Fig. 8 Storage modulus at 0.1 Hz as a function of time at 80°C for WPA solutions at 20 g/L without KC
390 (a) and with 2 g/L KC (b) after adding different GDL concentrations. The corresponding steady state
391 values of α are indicated in the figures.

392

393 After heating at 80°C during 100 min, the systems were cooled to 20°C, which led to an
394 increase of G', see figure S4b of the supplementary information. The relative increase of G' with
395 decreasing temperature did not depend on α. Figure 9a compares the dependence of G' at 20°C on α
396 at steady state for gels with and without KC. In both cases, G' increased initially rapidly with
397 increasing α and then stagnated, but In the presence of KC the increase of G' with increasing α was
398 more gradual and weaker. In the presence of KC, the value of α where gelation was first observed
399 was slightly smaller (αg = -3.7) than in the absence of KC (αg = -3.0). It is clear that gelation started at
400 a significantly lower α than at which the effect of complexation on the titration curves became
401 apparent (α ≈ -2), see figure 2. This suggests that when WPA starts to form a gel KC is not yet bound
402 to the proteins in significant quantities. Nevertheless, the presence of KC influences the gelation
403 process. One possible cause is that KC induces formation of dense protein domains during the
404 growth of the aggregates that drives gelation at lower α and modifies the gel structure. In this
405 scenario, the proteins form a network first and then KC binds to the already formed protein network
406 as α is increased.

407

17
 a

103

G' (Pa)
102

no KC
2 g/L KC
101

-4 -2 0 2 4

408 α

102
b
tg (min)

101

100
-4 -2 0 2 4

409 α

410 Fig. 9 Dependence of G' (a) and of tg (b) on α for WPA gels at 20 g/L formed at 80°C without KC and
411 with 2 g/L KC.
412

413 The dependence of the gel time on α is shown in fig. 9b. tg decreased rapidly with increasing
414 α up to α = -2 and more slowly at larger α. tg corresponds to the time needed to reach αg during GDL
415 degradation, which is shorter if αg is smaller. This explains why tg was smaller with (αg = -3.7) than
416 without (αg = -3.0) KC for α < -2. When more GDL is added, α reaches αg rapidly and the difference in
417 time becomes very small. Kharlamova et al. (2018a) reported on the rate of acid-induced gelation of
418 WPA at a given α between -5 and -3.4 set by adding HCl. For α < -5.0 gelation was too slow even at
419 80°C (tg> 24h) and for α > -3.4 it was too fast even at 5°C. tg was found to decrease sharply with

18
420 increasing α, WPA concentration or temperature. The temperature dependence was controlled by an
421 activation energy Ea = 155 kJ/mol independent of α and the WPA concentration. In the investigation
422 by Kharlamova et al. (2018a), α was set very rapidly and tg represented the time needed to gel at a
423 given α. Here, α was decreased slowly and gelation was fast when αg was reached. Therefore, in the
424 present investigation the temperature dependence was controlled by Ea of GDL degradation,
425 whereas in that of Kharlamova et al. it was controlled by Ea of WPA gelation at fixed α.

426

427 3.3. Optical density

428 3.3.1 Effect of the KC concentration

429 Fig. 10 shows the optical density as a function of time at 20°C after adding 9 mM GDL for
430 WPA solutions at CWP = 20 g/L and different amounts of KC. The mixtures were transparent at these
431 KC concentrations before addition of GDL. A sharp increase was observed at a critical time (tc)
432 followed by a weak increase at longer times.

0 g/L
0.25
1 0.5
0.75
1.0
1.5
OD

2.0
0.1 3.0
4.0
5.0
6.0

0.01
100 1000

433 time (min)

434 Fig. 10 Evolution with time at 20°C of the optical density at λ = 600 nm for WPA solutions at 20 g/L
435 and different KC concentrations as indicated in the figure after addition of 9 mM GDL (α≈ 0).

436

437 Figure 11a shows the dependence of tc, taken as the time where OD = 0.01, on CKC at 20°C. tc
438 decreased with increasing KC concentration and became significantly shorter than tg at 20°C, which

19
439 was between 350 and 400 min almost independent of CKC, see figure 3. The increase of the turbidity
440 is caused by the formation of large aggregates that percolate into a network. It appears that the time
441 between the onset of association of the WPA and their gelation increased with increasing KC
442 concentration.

2.0
300 a b

1.5
250
tc (min)

ODend
1.0
200

150 0.5

100 0.0
0 1 2 3 4 5 6 0 1 2 3 4 5 6
443 CKC (g/L) CKC (g/L)

444 Fig. 11 Dependence on the KC concentration of the critical time (a) and the optical density after 50h
445 (b) for mixtures containing 20 g/L WPA in the presence of 9 mM GDL (α≈ 0) at 20°C.

446

447 Fig. 11b shows the optical density of the gels at the end (ODend) when steady state was
448 reached in most situations. ODend increased sharply with increasing KC concentration up to CKC = 0.5
449 g/L followed by a very weak increase at higher KC concentrations. The change in the turbidity reflects
450 a change in the microstructure of the gels. It is remarkable, however, that a rapid increase of ODend
451 occurred at low KC concentrations where the elastic modulus of the gel was not yet influenced much,
452 see figure 4. On the other hand, the strong decrease of the elastic modulus for CKC> 1 g/L
453 corresponds to a weak progressive increase of ODend. It is obvious that changes of the elastic modulus
454 are not only caused by changes in the microstructure, but also by changes in the connectivity and
455 bond strength. Therefore one cannot directly correlate rheology and turbidity measurements.

456

457 3.3.2 Effect of the net charge density and the temperature

458 The effect of α on ODend and tc was determined for WPA solutions without KC and with 2 g/L
459 KC at 20°C, 50°C and 80°C. Examples of the evolution of OD at different α and temperature are
460 shown in figure S6 of the supplementary information. The dependence of ODend on α is shown in

20
461 figure 12a. ODend increased initially sharply with increasing α followed by a weaker increase. The
462 increase was much stronger in the presence of KC as was already shown in fig.11b for α≈ 0 and it
463 started at smaller α when KC was present. We may compare the values of α at which ODend started
464 to increase with that at which G' started to increase shown in figure 9a for 80°C. The values of α at
465 which ODend started to increase in the presence of CKC = 2 g/L and without KC were within the
466 experimental error the same as those at which G' started to increase: α = -3.8 and α = -3.0,
467 respectively. It follows that formation of dense protein domains and gelation started at almost the
468 same charge densities.

469 The increase of ODend was stronger at higher temperatures, which is not caused by the
470 difference in the kinetics as the values of ODend were taken after sufficiently long times where they no
471 longer depended significantly on time. When gels formed at higher temperatures were subsequently
472 cooled to 20°C, ODend did not change significantly. The implication is that at different temperatures
473 different microstructures were formed and that these microstructures did not depend significantly
474 on the temperature once they had reached steady state.

3.0
a
2.5

2.0
ODend

1.5

1.0

0.5

0.0
-4 -2 0 2 4 6

475 α

21
 b 20°C
50°C
103
80°C

tc (min)
102

101

100
-4 -2 0 2 4 6

476 α

477 Fig. 12 Dependence of ODend (a) and the critical time (b) as a function of α for WPA solutions at 20 g/L
478 without KC (filled symbols) and with 2 g/L KC (open symbols) at different temperatures indicated in
479 the figure.

480

481 Figure 12b shows tc as a function of α at steady state. tc decreased with increasing α both in
482 the presence and absence of KC, but the dependence was weaker in the latter case. As a
483 consequence the difference between the values of tc with and without KC decreased with increasing
484 α. The strong increase of tc with decreasing temperature was caused by the strong decrease of the
485 degradation rate of GDL with decreasing temperature. The difference between the values of tc with
486 and without KC appears to become smaller at higher temperatures. Comparison of tc as a function of
487 α with tg at 80°C, see figure 9b, shows that tc is systematically shorter, which can be explained by the
488 fact that tc characterizes the onset of formation of large aggregates, which some time later percolate
489 into a gel. Notice that it took about 30 s to reach 80°C, which is not negligible at large α.

490 The rheology and turbidity measurements reported here do not give information about
491 whether and to what extent microphase separation between WPA and KC occurs and how much KC
492 binds to the proteins at different α. To address these issues we have probed the microstructure of
493 the mixed gels with confocal laser scanning microscopy using different fluorescent labels for the
494 proteins and the polysaccharides. The results of this study will be discussed in detail in part II.

495

22
496 4. Conclusion

497 Potentiometric titration measurements indicate that complexation between KC and whey
498 protein aggregates is important when α > -2. The effect of KC on the titration curves increases until a
499 WPA:KC weight ratio of about 2:1 and adding more KC does not make a significant difference.
500 Gelation of 20 g/L WPA induced by acidification starts at α ≈ -3.0 without KC and at α ≈ -3.7 in the
501 presence of 2 g/L KC. The turbidity increases sharply at the same charge density at which the gels are
502 formed. The turbidity at steady state increases strongly with increasing CKC up to approximately 0.5
503 g/L and weakly at higher concentrations whereas the elastic modulus changes little up to CKC = 1 g/L
504 and decreases strongly at higher KC concentrations. The absence of a clear correlation between the
505 change in structure probed by turbidity and the elastic modulus demonstrates the importance of
506 changes in the connectivity and bond strength that is not probed by turbidity measurements. The
507 gelation rate increases strongly with increasing temperature and is controlled by the rate of GDL
508 degradation. The elastic modulus increases with increasing temperature of preparation. The relative
509 increase is weak for CKC ≤ 1.5 g/L, but strong for CKC ≥ 2 g/L, i.e. at concentrations where KC strongly
510 influences the elastic modulus. The turbidity also increases with increasing temperature of
511 preparation. The turbidity of gels that have reached steady state does not change significant during
512 heating or cooling, but the elastic modulus decreases with increasing temperature, which
513 demonstrates again the importance of changes in the connectivity and bond strength that is not
514 probed by turbidity measurements.

515 Acknowledgements

516 Christophe Chassenieux is thanked for critically reading and commenting this article. This research
517 was partly supported byt he National Natural Science Foundation of China (31901613),the China
518 Postdoctoral Science Foundation (2018M640457) and the Jiangsu Province Postdoctoral Science
519 Foundation (2018K089C).

520

521

522 References

523 Ahmed, K. F., Aschi, A. & Nicolai, T. (2018), Formation and characterization of chitosan-protein
524 particles with fractal whey protein aggregates, Colloids and Surfaces B: Biointerfaces, 169, 257-264.
525 Ako, K., Durand, D. & Nicolai, T. (2011), Phase separation driven by aggregation can be reversed by
526 elas\city in gelling mixtures of polysaccharides and proteins† Soft Matter, 7, 2507-2516.

23
527 Baussay, K., Durand, D. & Nicolai, T. (2005), Effect of the Cluster Size on the Micro Phase Separation
528 in Mixtures of β-Lactoglobulin Clusters and κ-Carrageenan, Biomacromolecules, 7, 304-309..
529 Brodkorb, A., Croguennec, T., Bouhallab, S. & Kehoe, J. J. (2016). Heat-Induced Denaturation,
530 Aggregation and Gelation of Whey Proteins, Advanced Dairy Chemistry(pp. 155-178), Springer
531 Bryant, C. M. & McClements, D. J. (1998), Molecular basis of protein functionality with special
532 consideration of cold-set gels derived from heat-denatured whey, Trends in Food Science &
533 Technology, 9, 143-151.
534 Cakir, E. & Foegeding, E. A. (2011), Combining protein micro-phase separation and protein-
535 polysaccharide segregative phase separation to produce gel structures, Food Hydrocolloids, 25, 1538-
536 1546.
537 Capron, I., Nicolai, T. & Smith, C. (1999), Effect of addition of κ-carrageenan on the mechanical and
538 structural properties of β-lactoglobulin gels, Carbohydrate Polymers, 40, 233-238.
539 Corredig, M., Sharafbafi, N. & Kristo, E. (2011), Polysaccharide-protein interactions in dairy matrices,
540 control and design of structures, Food Hydrocolloids, 25, 1833-1841.
541 Croguennoc, P., Durand, D. & Nicolai, T. (2001), Phase Separation and Association of Globular Protein
542 Aggregates in the presence of Polysaccharides: 1. Mixtures of Preheated β-Lactoglobulin and κ-
543 Carrageenan at room temperature, Langmuir, 17, 4372-4379.
544 Croguennoc, P., Meunier, V., Durand, D. & Nicolai, T. (2000), Characterization of semidilute κ-
545 carrageenan solutions, Macromolecules, 33, 7471-7474.
546 De Jong, S., Klok, H. J. & Van de Velde, F. (2009), The mechanism behind microstructure formation in
547 mixed whey protein-polysaccharide cold-set gels, Food Hydrocolloids, 23, 755-764.
548 de Jong, S. & van de Velde, F. (2007), Charge density of polysaccharide controls microstructure and
549 large deformation properties of mixed gels, Food Hydrocolloids, 21, 1172-1187.
550 De Kruif, C. G., Weinbreck, F. & de Vries, R. (2004), Complex coacervation of proteins and anionic
551 polysaccharides, Current opinion in colloid & interface science, 9, 340-349.
552 Frith, W. J. (2010), Mixed biopolymer aqueous solutions–phase behaviour and rheology, Advances
553 in Colloid and Interface Science, 161, 48-60.
554 Guo, M. (2019), Whey Protein Production, Chemistry, Functionality, and Applications. Wiley.
555 Heilig, A., Göggerle, A. & Hinrichs, J. (2009), Multiphase visualisation of fat containing β-lactoglobulin-
556 к-carrageenan gels by confocal scanning laser microscopy, using a novel dye, V03-01136, for fat
557 staining., LWT - Food Science and Technology, 42, 646–653.
558 Inthavong, W., Kharlamova, A., Chassenieux, C. & Nicolai, T. (2016), Structure and flow of dense
559 suspensions of protein fractal aggregates in comparison with microgels, Soft matter, 12, 2785-2793.

24
560 Kasapis, S. (2008), Phase separation in biopolymer gels: a low-to high-solid exploration of structural
561 morphology and functionality, Critical Reviews in Food Science and Nutrition, 48, 341-359.
562 Kayitmazer, A. B., Seeman, D., Minsky, B. B., Dubin, P. L. & Xu, Y. (2013), Protein-polyelectrolyte
563 interactions, Soft Matter, 9, 2553-2583.
564 Kharlamova, A., Chassenieux, C. & Nicolai, T. (2018a), Acid-induced gelation of whey protein
565 aggregates: Kinetics, gel structure and rheological properties, Food Hydrocolloids, 81, 263-272.
566 Kharlamova, A., Inthavong, W., Nicolai, T. & Chassenieux, C. (2016), The effect of aggregation into
567 fractals or microgels on the charge density and the isoionic point of globular proteins, Food
568 Hydrocolloids, 60, 470-475.
569 Kharlamova, A., Nicolai, T. & Chassenieux, C. (2018b), Calcium-induced gelation of whey protein
570 aggregates: Kinetics, structure and rheological properties, Food Hydrocolloids, 79, 145-157.
571 Khemissi, H., Bassani, H., Aschi, A., Capron, I., Benyahia, L. & Nicolai, T. (2018), Exploiting complex
572 formation between polysaccharides and protein microgels to influence particle stabilization of W/W
573 emulsions, Langmuir, 34, 11806.
574 Lam, R. S. H. & Nickerson, M. T. (2014), Effect of the Biopolymer Mixing Ratio on the Formation of
575 Electrostatically Coupled Whey Protein−κ- and ι‑Carrageenan Networks in the Presence and Absence
576 of Oil Droplets, J. Agric. Food Chem., 62, 8730−8739
577 Nguyen, B. T., Nicolai, T., Chassenieux, C. & Benyahia, L. (2014), The effect of protein aggregate
578 morphology on phase separation in mixtures with polysaccharides, Journal of Physics: Condensed
579 Matter, 26, 464102.
580 Nguyen, B. T., Nicolai, T., Chassenieux, C., Schmitt, C. & Bovetto, L. (2016), Heat-induced gelation of
581 mixtures of whey protein isolate and sodium caseinate between pH 5.8 and pH 6.6, Food
582 Hydrocolloids, 61, 433-441.
583 Nguyen, T. B., Phan-Xuan, T., Benyahia, L. & Nicolai, T. (2014), Combined effects of temperature and
584 elasticity on phase separation in mixtures of κ-carragheenan and β-lactoglobulin aggregates, Food
585 Hydrocolloids, 34, 138-144.
586 Nicolai, T. (2019), Gelation of food protein-protein mixtures, Advances in Colloid and Interface
587 Science, 270, 147-164
588 Nicolai, T., Britten, M. & Schmitt, C. (2011), b-Lactoglobulin and WPI aggregates: formation, structure
589 and applications, Food Hydrocolloids, 25, 1945-1962.
590 Norton, I. T. & Frith, W. J. (2001), Microstructure design in mixed biopolymer composites, Food
591 Hydrocolloids, 15, 543-553.
592 Piculell, L. (2006). Gelling Carrageenans, In A. M. Stephen, G. O. Philips &P. A. Williams Food
593 Polysaccharides and their Applications(pp. 239), Boca Raton: CRC Press

25
594 Pocker, Y. & Green, E. (1973), Hydrolysis of D-glucono-. delta.-lactone. I. General acid-base catalysis,
595 solvent deuterium isotope effects, and transition state characterization, Journal of the American
596 Chemical Society, 95, 113-119.
597 Schmitt, C. & Turgeon, S. L. (2011), Protein/polysaccharide complexes and coacervates in food
598 systems, Advances in Colloid and Interface Science, 167, 63-70.
599 Stone, A. K. & Nickerson, M. T. (2012), Formation and functionality of whey protein isolate–(kappa-
600 , iota-, and lambda-type) carrageenan electrostatic complexes, Food Hydrocolloids, 27, 271-277.
601 Turgeon, S. L., Beaulieu, M., Schmitt, C. & Sanchez, C. (2003), Protein-polysaccharide interactions:
602 phase-ordering kinetics, thermodynamic and structural aspects, Current opinion in colloid & interface
603 science, 8, 401-414.
604 Turgeon, S. L. & Laneuville, S. I. (2009), Protein+ polysaccharide coacervates and complexes: from
605 scientific background to their application as functional ingredients in food products, Modern
606 biopolymer science, 327-363.
607 van de Velde, F., de Hoog, E. H. A., Oosterveld, A. & Tromp, R. H. (2015), Protein-polysaccharide
608 interactions to alter texture, Annual review of food science and technology, 6, 371-388.
609 Van den Berg, L., Rosenberg, Y., Van Boekel, M. A. J. S., Rosenberg, M. & Van de Velde, F. (2009),
610 Microstructural features of composite whey protein/polysaccharide gels characterized at different
611 length scales, Food Hydrocolloids, 23, 1288-1298.
612 Van den Berg, L., Van Vliet, T., Van der Linden, E., Van Boekel, M. & Van de Velde, F. (2007),
613 Breakdown properties and sensory perception of whey proteins/polysaccharide mixed gels as a
614 function of microstructure, Food Hydrocolloids, 21, 961-976.
615 Weinbreck, F., Nieuwenhuijse, H., Robijn, G. W., de Kruif, C. G. (2004), Complexation of whey
616 proteins with carrageenan. Journal of AgriculturaJ and Food Chemistry, 52, 3550-3555.
617 Wen, Y. P. & Dubin, P. L. (1997), Potentiometric studies of the interaction of bovine serum albumin
618 and poly (dimethyldiallylammonium chloride), Macromolecules, 30, 7856-7861.
619 Zhang, S., Hsieh, F.-H. & Vardhanabhuti, B. (2014a), Acid-induced gelation properties of heated whey
620 protein-pectin soluble complex (Part I): Effect of initial pH, Food Hydrocolloids, 36, 76-84.
621 Zhang, S. & Vardhanabhuti, B. (2014b), Acid-induced gelation properties of heated whey protein-
622 pectin soluble complex (Part II): Effect of charge density of pectin, Food Hydrocolloids, 39, 95-103.
623
624

26
- Kappa carrageenan forms complexes with whey protein aggregates during acidification

-Potentiometric titration shows at what protein charge density complexation becomes important

- Complexation starts at the same protein charge density as aggregation and gelation of WPA

-The gelation rate increases with temperature and is determined by the degradation rate of GDL

-The gel stiffness decreases with increasing carrageenan concentration and gelation temperature
Dasong Liu : experimental invesitgation, reviewing, editing. Peng Zhou: supervision. Taco Nicolai:
conceptualization, writing, supervision.
Declarations of interest: none