Complexation with whey protein fibrils and chitosan: A potential vehicle for curcumin with improved aqueous dispersion stability and enhanced antioxidant activity

Journal Pre-proof Complexation with whey protein fibrils and chitosan: A potential vehicle for curcumin with improved aqueous dispersion stability and enhanced antioxidant activity Yu Hu, Chengxin He, Chengjia Jiang, Yang Liao, Hua Xiong, Qiang Zhao PII:

S0268-005X(19)31543-7

DOI:

https://doi.org/10.1016/j.foodhyd.2020.105729

Reference:

FOOHYD 105729

To appear in:

Food Hydrocolloids

Received Date: 10 July 2019 Revised Date:

7 January 2020

Accepted Date: 30 January 2020

Please cite this article as: Hu, Y., He, C., Jiang, C., Liao, Y., Xiong, H., Zhao, Q., Complexation with whey protein fibrils and chitosan: A potential vehicle for curcumin with improved aqueous dispersion stability and enhanced antioxidant activity, Food Hydrocolloids (2020), doi: https://doi.org/10.1016/ j.foodhyd.2020.105729. 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. © 2020 Published by Elsevier Ltd.

Graphic abstract

Complexation with whey protein fibrils and chitosan improved the aqueous dispersion stability of curcumin significantly.

1

Complexation with whey protein fibrils and chitosan: A potential vehicle

2

for curcumin with improved aqueous dispersion stability and enhanced

3

antioxidant activity

4 5

Yu Hu, Chengxin He, Chengjia Jiang, Yang Liao, Hua Xiong, Qiang Zhao*

6 7

State Key Laboratory of Food Science and Technology, Nanchang University, Jiangxi

8

330047, China

9 10

*Corresponding author. Tel/Fax: +86-791-86634810

11

E-mail address: [email protected], [email protected]

1

12

ABSTRACT: Whey protein fibrils (WPF) formed at 6, 12, 18, and 24 h (80 °C, pH 2) and

13

chitosan were utilized as vehicles to enhance curcumin dispersion stability at pH 3.5. Their

14

antioxidant activity and release behavior in vitro were investigated. Compared with whey

15

protein isolates (WPI), WPF possessed higher surface hydrophobicity and ζ-potential without

16

reducing solubility. The solubility of curcumin was improved to 297.8 ± 3.3 µg/mL (400

17

µg/mL added) via complexation with chitosan and WPF-18 (WPF formed at 18 h). The

18

hydrophobic groups of fibrils, rather than those exposed due to the hydrolyzed peptides,

19

accounted for major binding sites of curcumin. Chitosan and fibrils combined with curcumin

20

formed a bicontinuous polymer through electrostatic interaction and increased the repulsive

21

force between fibrils, resulting in a delivery system with increased stability. Compared with

22

curcumin alone, the complexes showed significantly improved antioxidant activity (DPPH

23

radical scavenging activity and reducing power). Moreover, the delivery systems further

24

provided opportunities for curcumin to release in the intestine. This potential vehicle may

25

contribute further to introduce curcumin into fat-free acidic functional beverages.

26 27

Keywords: Curcumin; Whey protein fibril; Chitosan; Dispersion stability; Antioxidant

28

activity

2

29

1. Introduction

30

Curcumin, a phenolic constituent extracted from the rhizome of turmeric, has attracted

31

considerable interest in many fields due to its health-promoting properties, such as

32

anti-inflammatory, antitumor, antimicrobial, and antioxidant activities (Peng et al., 2018b).

33

However, the applications of curcumin as a functional ingredient in foods have been

34

restricted because of its low aqueous solubility and poor chemical stability, resulting in

35

inferior oral bioavailability and reduced consumer acceptance (Zheng et al., 2018). Thus,

36

numerous delivery systems (e.g., emulsions (Zheng et al., 2018), nanoparticles (Patel et al.,

37

2010; Peng et al., 2018b), micelles (Khanji et al., 2015; Liu et al., 2014), and liposomes

38

(Peng et al., 2017)) have been proposed to introduce curcumin into foods with improving

39

product quality. However, incorporating curcumin into fat-free beverages and foods remains

40

challenging. Recently, complexation with proteins has been developed as an efficient and

41

promising approach to improve the dispersion, chemical stability, and bioavailability of

42

curcumin in an aqueous environment. Protein materials including zein (Patel et al., 2010),

43

casein (Pan et al., 2014), soy protein (Chen et al., 2015a), and egg white protein (Chang et al.,

44

2019), have been combined with curcumin, mainly relying on the hydrophobic interaction or

45

the incorporation of curcumin into protein clusters.

46

Whey protein isolate (WPI), a usual by-product of cheese processing, has been widely

47

applied to foods because of its excellent nutritional value and functional properties (e.g.,

48

emulsifying, thickening, and gelling). WPI is also used as a vehicle for bioactive substances

49

(Livney, 2010). In addition, WPI can self-assemble into many supramolecular structures

50

under certain conditions (Nicolai et al., 2011). WPI fibrils with reduced diameters 3

51

(approximately 10 nm) and lengths in the order of microns can be formed via an

52

uncomplicated treatment referring to heating above denaturation temperatures for several

53

hours under strongly acidic condition (usually pH 2.0) and low ionic strength (Akkermans et

54

al., 2008a; Akkermans et al., 2008b; Jones & Mezzenga, 2012;). From a food engineering

55

perspective, these quintessential high aspect ratio (length/diameter) fibril structures can

56

provide unique opportunity for rheology modification (Peng et al., 2018a), enhancement to

57

the emulsification and foaming property (Oboroceanu et al., 2014; Peng et al., 2016), and

58

improved antioxidant activity (Mohammadian & Madadlou, 2016) compared with native

59

proteins. There has been reported the utilization of protein fibrils for microcapsules

60

(Ansarifar et al., 2017; Humblet-Hua et al., 2011) and emulsion (Chang et al., 2018) to

61

encapsulate or stabilize active ingredients. In addition, researchers have also investigated the

62

interaction between fibrils and active ingredients, for instance, Hu et al. (2018) designed

63

lysozyme fibrils binding with polyphenol into reversible hydrogels for biomedical

64

applications and Shen et al. (2017) utilized β-lactoglobulin fibril systems to deliver nanosized

65

iron. Hence, protein fibrils may serve as potential candidates utilized in delivery systems.

66

Based on hydrophobic interaction between protein fibrils and active ingredients, the

67

aqueous solubility of curcumin at pH 3.2 could be promoted with whey protein nanofibrils

68

formed at 85 °C and pH 2 for 5 h (Mohammadian et al., 2019). However, in our previous

69

research (Hu et al., 2019), we found whey protein isolate fibril (WPF) formation in a mixture

70

of whey protein system would be delayed within 24 h, resulting in significant variation in

71

fibril conversion rate, structure, and morphology. For these reasons, we plan to investigate the

72

possibility of using WPFs from different reaction times to improve the aqueous dispersion 4

73

stability of curcumin under acidic conditions (pH 3.5).

74

Polysaccharides are often introduced to enhance the stability of protein-based aqueous

75

delivery systems, which can interact with proteins electrostatically, and the electrostatic

76

interaction depends on their charge properties, pH, and ionic strength (Jones et al., 2011;

77

Hosseini et al., 2015). Wang et al. (2019) also reported the complex of hen egg white

78

lysozyme fibrils and polysaccharides enhanced the stability of high internal phase emulsions.

79

Chitosan, which is produced by the removal of N-acetyl from chitin, is a class of linear

80

molecule consisting of N-acetylglucosamine and β-linked glucosamine units and usually

81

presents a cation under acidic conditions (Chang et al., 2018). Gilbert, Campanella, and Jones

82

(2014) demonstrated the addition of chitosan could improve β-lactoglobulin fibril stability in

83

acidic conditions to increased pH values (pH 3-7). Accordingly, to the best of the authors'

84

knowledge, this is the first study on the formation and stability of aqueous dispersion system

85

stabilized by chitosan containing a mixture of WPFs and curcumin.

86

Therefore, in the present work, WPI solution (4 %) was heated at 80 °C for 6, 12, 18, and

87

24 h in pH 2 acidic environment to prepare the protein nano-fibrils. These WPFs were

88

complexed with curcumin at pH 3.5 with further adding chitosan or not. The stability of the

89

aqueous dispersed systems was determined using various techniques and the interaction

90

between WPF and curcumin was confirmed with fluorescence quenching and Van't Hoff

91

equation. In addition, the functional properties of the complex were evaluated on the basis of

92

their antioxidant activity (DPPH scavenging activity and reducing power) and the release

93

behavior of curcumin in vitro.

94 5

95

2. Materials and methods

96

2.1. Chemicals and proteins

97

Whey protein isolate powder (Hilmar 9410, protein > 93%) was purchased from Hilmar

98

Corp. (California, USA); curcumin (purity > 95%) was purchased from Yuanye

99

Biotechnology Co., Ltd. (Shanghai, China) and chitosan (deacetylation degree > 96%,

100

viscosity 100-200 mPa.s) was purchased from Maclean Biotechnology Co., Ltd. (Shanghai,

101

China). Pepsin (P7000, ≥250 units/mg) and pancreatin (P7545, 8× USP) were purchased from

102

Sigma-Aldrich Corp. (St. Louis, MO, USA); 1-anilino-8-naphthalene sulphonate (ANS) and

103

2,2-diphenyl-1-picrylhydrazyl (DPPH) were purchased from Aladdin Corp. (Shanghai,

104

China). Other chemicals were of analytical grade.

105 106

2.2. Preparation of fibrils

107

WPI powder (4%, w/v) was dissolved in deionized water and adjusted to pH 2.0 with 6 M

108

HCl, and the solutions were stored at 4 °C overnight for a full dissolution. Then the protein

109

solution was filtered through a filter (Hydrophilic PES 0.45 µm, Millipore Millex-HP) to

110

remove the precipitating materials. Samples were heated at 80 °C for 6, 12, 18, and 24 h in a

111

silicone oil bath respectively, and then freeze-dried. And fibrils formed at 6, 12, 18, and 24 h

112

were named WPF-6, WPF-12, WPF-18, and WPF-24, respectively.

113 114

2.3. Th T fluorescence

115

Th T fluorescence of WPI and WPF was characterized by a spectrofluorometry (F-7000,

116

HITACHI, Tokyo, Japan) according to Nilsson (2004). Protein samples (40 µL) were 6

117

thoroughly blended with the Th T working solution (4 mL). The emission spectrum from 460

118

to 560 nm was attained at the excitation wavelength of 440 nm. Meanwhile, Th T working

119

solution was deducted as a background.

120 121

2.4. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE)

122

The test was operated in a discontinuous buffering system (12% separating gel, 5%

123

stacking gel) according to Laemmli (1970). The protein samples were blended with the

124

reducing loading buffer containing DTT. After heating in boiling water for 5 min, each

125

sample (10 µL) was loaded to the gel. Electrophoresis was performed under a constant

126

current of 15 mA in the concentrated gel and at 25 mA in the separation gel. Then the gel was

127

stained in 0.25% Coomassie Blue (R-250) solution and destained in ethanol/acetic acid/water

128

solution (50:75:875, v/v/v).

129 130

2.5. Surface hydrophobicity (H0)

131

Surface hydrophobicity of the protein samples was determined referring to the method

132

described by Zhao et al. (2012) using ANS as the hydrophobic fluorescence probes at the

133

excitation wavelength of 390 nm. Protein samples were diluted to a series of concentrations

134

from 0.005% to 0.025%（w/v）with deionized water pre-adjusted to pH 3.5 and were

135

thoroughly blended with 20 µL of ANS solution (8 mmol/L; solvent PBS, 10 mM, pH 7.0).

136

The fluorescence intensity of each sample at 470 nm versus protein concentration (mg/mL)

137

by linear regression analysis and initial slope presented as an index of H0. The excitation and

138

emission slits were both 5 nm. 7

139 140

2.6. Preparation of the Complex

141

Chitosan (dissolved in 2% acetic acid) was added or not to WPI and WPF solutions and

142

adjusted to pH 3.5. Then the curcumin (dissolved in ethanol) in continuous stirring conditions

143

was added. The final concentration of protein, chitosan, and curcumin in solution were 40

144

mg/mL, 1 mg/mL, and 0.4 mg/mL, respectively. It’s worth noting that the amount of ethanol

145

added didn’t exceed 1% (v/v), at which the protein does not degenerate (Kanakis et al., 2013).

146

The complexes of WPI or WPF and curcumin were named WPI-Cur or WPF-Cur, and the

147

complexes of WPI or WPF, chotisan and curcumin were named WPI-CS-Cur or

148

WPF-CS-Cur.

149 150

2.7. Curcumin solubility

151

Curcumin solubility in various delivery solutions was determined using the method

152

described by Tapal and Tiku (2012). Solutions were centrifuged (5000 ×g, 10 min) to remove

153

the free curcumin and the supernatant was diluted with ethanol to separate curcumin.

154

Subsequently, the absorbance at 420 nm of each diluted sample was recorded and the content

155

of curcumin in the supernatant was determined by the standard curve of curcumin dissolved

156

in ethanol at 420 nm (r2=99.87%).

157 158

2.8. Fluorescence quenching

159

The emission spectra from 315 to 450 nm of the complex were recorded at the excitation

160

wavelength of 295 nm to elucidate the effects of curcumin on WPF at 304 and 310 K, 8

161

respectively. The excitation and emission slits were both 10 nm.

162 163

2.9. Curcumin fluorescence

164

Curcumin fluorescence intensities of the complexation samples were characterized to

165

reveal the binding of curcumin to protein and chitosan further. Before measurement, each

166

sample was diluted 10 times with pre-adjusted deionized water (pH 3.5). Then the emission

167

spectrum from 450 to 650 nm was attained at the excitation wavelength of 440 nm.

168

Meanwhile, protein or chitosan solution was deducted as a background.

169 170

2.10. Particle size and zeta-potential

171

The particle size and zeta-potential of each sample were determined using a particle size

172

potentiometer (Nano-ZSE, Malvern Corp., U.K.). Each sample was diluted with deionized

173

water pre-adjusted to pH 3.5.

174 175

2.11. Transmission electron microscope (TEM)

176

Sample solutions were diluted to 0.01% (w/v) with deionized water pre-adjusted to pH 3.5.

177

Formvar-coated copper grids were coated with 10 µL of diluted samples and negatively

178

stained with 1% uranyl acetate solution. The morphology of the sample was observed with

179

TEM (JEM-2100, JEOL, Japan) at 200 kV.

180

9

181

2.12. Antioxidant activity

182

2.12.1 DPPH scavenging activity

183

DPPH radical scavenging activity of samples containing 10 mg/mL protein with or without

184

0.1 mg/mL curcumin was determined according to the previous method (Shimada et al., 1992;

185

Zhao et al., 2012) with slight modifications. 1 mL of each sample solution was blended with

186

4 ml of DPPH solution (0.1 mM, dissolved in ethanol) and incubated for 30 min in the dark.

187

Subsequently, the absorbance of each sample was recorded. The radical scavenging activity

188

was calculated according to eqn. (1):

189

DPPH scavenging activity (%) = [Ac-(As-Ab)]/Ac×100

190

where Ac, As, and Ab is the absorbance at 517 nm of the control, the sample, and the sample

191

without DPPH solution, respectively.

（1）

192 193

2.12.2 Reducing power

194

Reducing power of samples containing 40 mg/mL protein with or without 0.4 mg/mL

195

curcumin was determined according to the method described by Zhao et al. (2012). 1 mL of

196

each sample solution was blended with 2.5 mL PBS (0.2 M, pH 6.6) and 2.5 mL of

197

K3Fe(CN)6 solution (1%, w/v), and incubated for 20 min in a 50 °C water bath. Then 2.5 mL

198

of TCA (10%, w/v) was added and centrifuged (3000 ×g, 10 min). 2.5 ml of supernatant was

199

blended with deionised water (2.5 mL) and 0.1% (w/v) FeCl3 solution (0.5 ml), and stored for

200

10 min at room temperature. Subsequently, the absorbance at 700 nm of each sample was

201

recorded. A higher absorbance indicates a stronger antioxidant activity.

202 10

203

2.13. In vitro release behavior

204

To probe the effect of the addition of Chitosan on the slow release of curcumin, in vitro

205

release behaviors of curcumin from the complex of WPF and Chitosan were determined by

206

the previous method (Mohammadian et al., 2019) with slight modifications. In brief, 3 mL of

207

each sample solution was mixed with 3 mL simulated gastric fluid (SGF; containing 2 g/L

208

NaCl, 7 mL/L hydrochloric acid, and 3.2 g/L pepsin, pH 1.2) and adjusted to pH 1.2, and then

209

loaded into a dialysis bag (MWCO, 3500 Da). Subsequently, the dialysis bag was placed in

210

150 mL of release medium containing ethanol and SGF without pepsin (1:1, v/v), and

211

incubated at 37 °C for 2 h with a shaking of 100 rpm. Then 6 mL of simulated intestinal fluid

212

(SIF; containing 6.8 g/L KH2PO4 and 1% pancreatin, pH 7.5) was added to the mixture and

213

adjusted to 7.5. The dialysis bag was placed in 150 mL of release medium containing ethanol

214

and SIF without pepsin (1:1, v/v), and incubated for another 4 h under the same conditions.

215

The release medium at a specific time (1, 2, 3, 4, 5, 6 h) was collected with a supplement of

216

equal fresh medium, and the concentration of released curcumin was analyzed at 420 nm

217

according to the curcumin standard curve performed in the same release medium.

218 219

2.14. Statistical analyses.

220

All analyses were performed on triplicate samples at least. Data were analyzed by an

221

analysis of variance (ANOVA) using Origin 2018 statistics program (Origin Lab Corporation,

222

Northampton, MA, USA) and presented as the mean ± standard deviations. Statistical

223

differences were defined as P < 0.05 with Turkey’s test.

224 11

225

3. Results and discussion

226

3.1. Characterization of WPF

227

Th T fluorescence is generally accepted as an indicator for fibrils because Th T can bind to

228

the β-sheet structures existing in the fibrils (Nilsson, 2004). As shown in Fig. 1A, compared

229

with WPI in which almost no fluorescence was observed, other WPF samples (WPF-6,

230

WPF-12, WPF-18, and WPF-24) showed gradually increased Th T fluorescence intensity

231

with reaction time, indicating that additional fibrils were formed. However, the growth rate in

232

Th T fluorescence of WPF-18 to WPF-24 decreased. In general, fibril formation derived from

233

globular proteins involves the following sequential steps of lag phase, elongation phase, and

234

mature phase referring to hydrolysis, nucleation, and growth of the hydrolyzed peptides by

235

β-sheet alignment (Dave et al., 2014; Mohammadian & Madadlou, 2018). Therefore, WPF-18

236

to WPF-24 may be gradually entering the mature phase during which the rate of

237

self-assembly declines.

238

Combining the result of SDS-PAGE (WPI and WPF samples) under reducing conditions

239

(Fig. 1B), WPI showed a main broad band of β-lactoglobulin approximately 18.4 kDa and the

240

band intensity decreased continuously from WPF-6 to WPF-24 due to hydrolysis during fibril

241

formation. Similar result was supported by size exclusion chromatography in Fig. S1, the

242

molecular weight gradually decreased with the reaction time. These hydrolyzed peptides

243

instead of intact proteins have been suggested as the building blocks for the creation of fibrils

244

(Akkermans et al., 2008b), and the ability to form the fibril of pre-hydrolyzed whey proteins

245

with different kinds of proteases was unequal (Gao et al., 2013). The resulting new band was

246

either fibril or peptide, and the bottom band intensity from WPF-6 to WPF-24 gradually 12

247

enhanced. However, distinguishing them from the bands alone was difficult.

248

The size distribution (volume frequency, %) and morphology of WPI and WPF samples

249

imaged with TEM are depictured in Fig. 1C. Given its acid stability, WPI composed of

250

globulin still exhibited a nano-diameter spherical morphology with two main volume peaks at

251

pH 3.5, whereas spherical morphology was almost not observed in WPF-6, WPF-12, WPF-18,

252

and WPF-24 and turned into fibril structure with micron length. The volume peaks gradually

253

moved to a larger scale from WPF-6 to WPF-24. Compared with fibrils formed at 12, 18, and

254

24 h, WPF-6 presented hydrolyzed proteins and newly formed fibrils with a slightly larger

255

diameter (~ 20 nm), and a similar result was previously reported (Gao et al., 2013). The

256

diameter of WPF-12 reduced to approximately 12 nm, whereas those of WPF-18 and WPF-24

257

were reduced to roughly 10 nm. Their lengths were also not uniform but increased with

258

fibrillation.

259 260

3.2. Curcumin solubility

261

The curcumin solubility in various delivery solutions referring to water, WPI, WPF-6,

262

WPF-12, WPF-18, and WPF-24, with or without 0.1% chitosan is illustrated in Fig. 2.

263

Curcumin was almost insoluble in deionized water (pH 3.5) or 0.1% chitosan solution (pH

264

3.5), whereas the solubility of curcumin was significantly enhanced (p < 0.05) via

265

complexing with WPI or WPF and was improved further after the addition of chitosan. In

266

particular, the solubility of curcumin in complexes of WPF-18 without or with chitosan was

267

increased to 190.3 ± 7.8 and 297.8 ± 3.3 µg/mL (400 µg/mL curcumin added), respectively.

268

The appearance of various delivery systems stored at room temperature for one month is 13

269

presented in Fig. 3. After one day of production, curcumin dissolved in deionized water and

270

WPI solutions with or without 0.1% chitosan began to flocculate, whereas curcumin

271

solubility in deionized water and chitosan solution had no significant difference (p > 0.05).

272

The flocculation rates of curcumin in these two solutions were different. The relative slow

273

flocculation rate in 0.1% chitosan solution under natural conditions rather than high-speed

274

centrifugation may be attributed to the viscosity or thickening effect of chitosan. After one

275

week of production, curcumin in the first four solutions flocculated further and curcumin in

276

WPF-6 and WPF-12 solutions flocculated slightly. Their flocculation states after one month

277

of preparation were consistent with the result of solubility measurement. Particularly,

278

curcumin in complexes of WPF-18 with chitosan expressed improved stability, indicating the

279

possibility of being applied to functional acid beverages.

280 281

3.3. Surface hydrophobicity

282

ANS was used to detect the exposed nonpolar surfaces in WPI and WPF, and the result of

283

surface hydrophobicity (H0) is shown in Fig. 4. Compared with the native WPI, fibrillation

284

significantly enhanced the surface hydrophobicity of protein (p < 0.05). The fibrillation

285

involves hydrolysis and the hydrolysis of protein into lower molecular weight peptides might

286

lead to an increase of hydrophobicity as a result of the exposure of hydrophobic groups that

287

are folded inside the intact native protein molecule (Zhao et al., 2012). In addition,

288

hydrophobic interaction in the interior gap is an important property of amino acid to promote

289

β-sheet formation (Jones & Mezzenga, 2012). Combining the result of curcumin solubility

290

(Fig. 3), these exposed hydrophobic groups could provide additional binding sites for 14

291

curcumin and the hydrophobic interaction was recognized as the main driving force for

292

protein-curcumin complexes (Chen et al., 2015b). However, although WPF-6 and WPF-12

293

possessed higher surface hydrophobicity than WPF-24, their combinations with curcumin

294

were not as good as that with WPF-24. Hence, the hydrophobic groups of fibrils, rather than

295

those exposed due to the hydrolyzed peptides, could be hypothesized to account for the major

296

binding sites of curcumin.

297 298

3.4. Fluorescence quenching and Curcumin fluorescence

299

The interaction between WPF and curcumin can be indicated by fluorescence spectroscopy

300

because of the high sensitivity of fluorophore to its polarity surrounding (Ye, Woo, &

301

Selonmulya, 2019). As shown in Fig. 5A, the fluorescence intensity of WPF-18 decreased

302

with rising curcumin concentrations. And the fluorescence quenching can be described by the

303

Stern-Volmer equation as eqn. (2):

304

F0 /F=1+kq τ0 [Cur]=1+ksv [Cur]

305

Where F0 and F are the fluorescence intensity of WPF-18 without and with different

306

concentrations of curcumin (from 0 to 90 µM), respectively; Kq represents the quenching rate

307

constant and τ0 represents the fluorophore lifetime (10−8 s for most biomolecules); Ksv

308

indicates the Stern-Volmer quenching constant and [Cur] is the curcumin concentration.

(2)

309

From Fig. 5B, Kq was calculated as 5.55 × 1012 M-1·s-1at 304 K while 3.81 × 1012 M-1·s-1 at

310

310 K, which indicated the static quenching was the main quenching mechanism. Therefore,

311

the relationship between the fluorescence quenching intensity and curcumin concentration

312

can be described according to eqn. (3): 15

313

lg  F0 -1 =lgKa +nlgC

314

where Ka and n indicate the binding constant and the number of binding sites, respectively.

F

(3)

315

From Fig. 5C, Ka was calculated as 2.61 × 103 M-1 at 304 K while 2.03 × 103 M-1 at 310 K.

316

The interaction between WPF and curcumin can be confirmed by Ka and thermodynamic

317

parameters (Ross & Subramanian, 1981) calculated by the Van't Hoff equation as eqn. (4),

318

and (5) to (6):

319

∆G = ∆H -T∆S

(4)

320

K2 ⁄K1 = 1⁄T1 - 1⁄T2  ∆H⁄R

(5)

321

∆G = -RTln

(6)

322

where ∆G, ∆H, and ∆S indicate the free energy change, the enthalpy change, and the entropy

323

change, respectively.

324

∆H was calculated as 2.5 kJ/mol and ∆S calculated as 73.81 J·mol-1·K-1. The values of ∆H

325

and ∆S were both positive, indicating the interaction between WPF and curcumin may be

326

attributed to the hydrophobic forces dominate according to Ross & Subramanian (1981).

327

Additionally, Hu et al. (2018) found the polyphenol molecule which is more hydrophobic

328

shows higher affinity to the fibrils; Mohammadian et al. (2019) presented WPI fibrils with

329

higher surface hydrophobicity compared to WPI could significantly bind more curcumin.

330

In addition, as presented in Fig. 6, compared with WPI-Cur and WPI-CS-Cur, the

331

maximum emission wavelength of other complexes blue shifted, and their maximum

332

fluorescence intensity increased. This result indicates that the microenvironment around

333

curcumin was excessive from the hydrophilic to a relatively hydrophobic one (Li et al., 2013).

334

The outcome could also react to the hydrophobicity of the delivery system from the side. 16

335 336

3.5. Zeta-potential

337

Zeta-potential was measured to reflect the stability of the dispersion in the aqueous

338

delivery system. As demonstrated in Fig. 7, fibrillation significantly expanded the positive

339

zeta-potential of protein to approximately 38.3 mV (WPF-6), 39.9 mV (WPF-12), 41.3 mV

340

(WPF-18), and 40.9 mV (WPF-24), respectively (p < 0.05) compared with that of WPI (~

341

24.7 mV). This result may be attributed to the unfolding and hydrolysis of proteins as

342

confirmed by the result of surface hydrophobicity and SDS-PAGE. Moreover, electrostatic

343

interaction is one of the non-covalent effects during fibril formation (Jones & Mezzenga,

344

2012). After complexation with curcumin, the positive zeta-potential was further enhanced.

345

Similar results have been reported in the complexes of SPI-Cur (Chen et al., 2015a) and

346

NaCas-Cur (Pan et al., 2014). Neutral curcumin molecules bind primarily to hydrophobic

347

groups on the surface of proteins, and proteins further binding curcumin molecules tend to

348

attract each other to form a large size through curcumin as a bridge, increasing net positive

349

charge (Chen et al., 2015b). In the current study, the addition of chitosan improved the

350

positive zeta-potential, particularly when WPF-18 was compounded with curcumin and

351

chitosan possessed the maximum potential (~ 54.6 mV), indicating a relatively stable

352

dispersion at acidic conditions. Of note, although the net charge of the protein was positive at

353

pH 3.5, it still possessed positive and negative regions on its surface. Thus, chitosan can

354

participate in attracting electrostatic interactions.

355

17

356 357

3.6. TEM Only TEM images of the complexes referring to WPF-18-Cur (A) and WPF-18-CS-Cur (B)

358

were observed (Fig. 8) because of their excellent delivery effect. Complexation with

359

curcumin and chitosan did not affect the morphology of fibrils. Numerous curcumin

360

molecules (2-8 nm) were bound to fibrils, whereas partially free curcumin was observed both

361

in WPF-18-Cur and WPF-18-CS-Cur. Fig. 8A shows that WPF-18 presented mutual

362

aggregation. By contrast, WPF-18 followed a certain order with mutual exclusivity, as

363

presented in Fig. 8B. Among heterogeneous biopolymer systems (protein and polysaccharide),

364

core-shell type, bicontinuous type, and dispersion type are three kinds of possible internal

365

structures used to incorporate lipophilic compounds (Matalanis et al., 2011). Therefore,

366

combing with the results in TEM images and other findings, linear chitosan and fibrils

367

combined with curcumin can be hypothesized to form a bicontinuous polymer by electrostatic

368

interaction and increase the repulsive force (steric hindrance or electrostatic interaction)

369

between fibrils, resulting in a stable delivery system.

370 371

3.7. Antioxidant activity

372

The antioxidant activity of each sample was evaluated by measurements of DPPH radical

373

scavenging activity and reducing power (reduction of ferricyanide complex/Fe3+ to Fe2+). As

374

presented in Fig. 9A, compared with WPIs, WPFs with different reaction times significantly

375

improved the reducing power to supply electrons to Fe3+ (p < 0.05), and similar results were

376

reported by Mohammadian and Madadlou (2016). The antioxidant activity of WPI may be

377

mainly attributed to the sulfhydryl groups on its surface and other amino acid residues (Tyr, 18

378

Trp, Met, and Lys) (Mohammadian & Madadlou, 2018). Fibril assembly from globular

379

protein involves steps of acid hydrolysis, and these hydrolyzed peptides and fibril structure

380

(high aspect ratio, length/diameter) may promote its oxidative enhancement. The complexes

381

of protein and curcumin with or without chitosan possessed higher reducing power (p < 0.05)

382

compared with curcumin dispersed in deionized water at identical concentrations. The

383

addition of chitosan did not significantly increase the reducing power of the complexes (p >

384

0.05). The antioxidant mechanism of curcumin is sequential proton loss electron transfer

385

(Litwinienko & Ingold, 2004), and its binding to protein through hydrophobic interaction

386

promoted the electron transfer to Fe3+. Chitosan combined with them through electrostatic

387

interaction, which may not have a visible effect on electron transfer when faced with this

388

oxidant.

389

As depicted in Fig. 9B, the result of DPPH radical scavenging activity was consistent with

390

the reducing power. Compared with the complexes of WPI and curcumin, the complexes of

391

WPF and curcumin possessed higher DPPH radical scavenging activity (p < 0.05), and the

392

addition of chitosan further improved the DPPH radical scavenging activity of the complexes

393

(p < 0.05), possibly depending on the degree of complexation confirmed by the result of

394

fluorescence, solubility, and potential.

395 396

3.8. In vitro release behavior

397

As illustrated in Fig. 10, after releasing for 2 h in the SGF, the percentages of curcumin

398

released from WPF-6-CS-Cur, WPF-12-CS-Cur, WPF-18-CS-Cur, and WPF-24-CS-Cur were

399

approximately 24.8%, 24.7%, 20.3%, and 22.1% respectively, providing further opportunities 19

400

for curcumin to be released in the intestine. The cumulative releases of curcumin from the

401

above complexes were approximately 47.1%, 51.9%, 38.8%, and 43.1%, respectively, after

402

another 4 h in the SIF. This result indicates that curcumin was continuously rather than

403

explosively released from the four complexes whether in the SIF or in the SGF, which might

404

be attributed to the strong hydrophobic and electrostatic interactions among WPF, curcumin,

405

and chitosan. The sustained release results of the four samples were consistent with their

406

effects on enhancing the aqueous solubility of curcumin (see Fig. 2).

407 408

4. Conclusion

409

WPF formation of WPF involves steps of hydrolysis, nucleation, and growth of the

410

hydrolyzed peptides by β-sheet alignment. Compared with WPI, WPF possessed higher

411

surface hydrophobicity and zeta-potential. WPF formed at 6, 12, 18, and 24 h (80 °C, pH 2)

412

and chitosan were utilized as a vehicle for curcumin. The solubility of curcumin was

413

improved to a maximum of 297.8 ± 3.3 µg/mL (400 µg/mL added) via complexing with

414

chitosan and WPF-18. Hydrophobic groups of fibrils instead of those exposed due to the

415

hydrolyzed peptides accounted for major binding sites of curcumin. Chitosan plays an

416

essential role in increasing the repulsive force by forming a bicontinuous polymer with fibrils.

417

Compared with curcumin, the complexes showed significantly improved antioxidant activity

418

(DPPH radical scavenging activity and reducing power). In addition, the delivery systems

419

further provided opportunities for curcumin to be released in the intestine. This potential

420

vehicle can contribute to introduce curcumin into fat-free acidic functional beverages and

421

provide a reference for delivery systems based on fibrils derived from other proteins. With 20

422

this study as a basis, further research can investigate the influence of environmental factors

423

(temperature, illumination, ionic strength, and pH) on its stability and the release behavior in

424

vivo using cell and animal models.

425 426

Acknowledgments

427

This research was supported by the National Natural Science Foundation of China

428

(31860451), Natural Science Fund for Distinguished Young Scholars (20192BCB23006),

429

Major Science and Technology Program (S2018ZDYFE0040) of Jiangxi Province, Major

430

Science and Technology Project of Jiangxi Academy of Sciences (2018-YZD1-05), and

431

Graduate Innovation Fund Project of Nanchang University (CX2018109).

432 433

References

434

Akkermans, C., van der Goot, A. J., Venema, P., van der Linden, E., & Boom, R. M. (2008a).

435

Formation of fibrillar whey protein aggregates: Influence of heat and shear treatment,

436

and resulting rheology. Food Hydrocolloids, 22 (7), 1315-1325.

437

Akkermans, C., Venema, P., van der Goot, A. J., Gruppen, H., Bakx, E. J., Boom, R. M., &

438

van der Linden, E. (2008b). Peptides are building blocks of heat-induced fibrillar

439

protein aggregates of β-lactoglobulin formed at pH 2. Biomacromolecules, 9 (5),

440

1474-1479.

441

Ansarifar, E., Mohebbi, M., Shahidi, F., Koocheki, A., & Ramezanian, N. (2017). Novel

442

multilayer microcapsules based on soy protein isolate fibrils and high methoxyl pectin:

443

Production, characterization and release modeling. International Journal of Biological

444

Macromolecules, 97, 761-769.

445

Chang, C. H., Meikle, T. G., Su, Y. J., Wang, X. T., Dekiwadia, C., Drummond, C. J., Conn, C.

446

E., & Yang, Y. J. (2019). Encapsulation in egg white protein nanoparticles protects 21

447

anti-oxidant activity of curcumin. Food Chemistry, 280, 65-72.

448

Chang, H. W., Tan, T. B., Tan, P. Y., Abas, F., Lai, O. M., Wang, Y., Wang, Y. H., Nehdi, I. A.,

449

& Tan, C. P. (2018). Physical properties and stability evaluation of fish oil-in-water

450

emulsions stabilized using thiol-modified beta-lactoglobulin fibrils-chitosan complex.

451

Food Research International, 105, 482-491.

452

Chen, F.-P., Li, B.-S., & Tang, C.-H. (2015a). Nanocomplexation between curcumin and soy

453

protein isolate: Influence on curcumin stability/bioaccessibility and in vitro protein

454

digestibility. Journal of Agricultural and Food Chemistry, 63 (13), 3559-3569.

455

Chen, F.-P., Li, B.-S., & Tang, C.-H. (2015b). Nanocomplexation of soy protein isolate with

456

curcumin: Influence of ultrasonic treatment. Food Research International, 75,

457

157-165.

458

Dave, A. C., Loveday, S. M., Anema, S. G., Jameson, G. B., & Singh, H. (2014). Modulating

459

β-lactoglobulin nanofibril self-Assembly at pH 2 using glycerol and sorbitol.

460

Biomacromolecules, 15 (1), 95-103.

461

Gao, Y. Z., Xu, H. H., Ju, T. T., & Zhao, X. H. (2013). The effect of limited proteolysis by

462

different proteases on the formation of whey protein fibrils. Journal of Dairy Science,

463

96 (12), 7383-7392.

464

Gilbert, J., Campanella, O., & Jones O. G. (2014). Electrostatic stabilization of

465

β-lactoglobulin fibrils at increased pH with cationic polymers. Biomacromolecules,

466

15(8), 3119-3127.

467

Hosseini, S. M. H., Emam-Djomeh, Z., Sabatino, P., & Van der Meeren, P. (2015).

468

Nanocomplexes arising from protein-polysaccharide electrostatic interaction as a

469

promising carrier for nutraceutical compounds. Food Hydrocolloids, 50, 16-26.

470

Hu, B., Shen, Y., Adamcik, J., Fischer, P., Schneider, M., Loessner, M. J., & Mezzenga, R.

471

(2018). Polyphenol-binding amyloid fibrils self-assemble into reversible hydrogels

472

with antibacterial activity. ACS Nano, 12 (4), 3385-3396.

473

Hu, Y., He, C. X., Woo, M. W., Xiong, H., Hu, J. W. & Zhao, Q. (2019). Formation of fibrils

474

derived from whey protein isolate: structural characteristics and protease resistance.

475

Food & Function, 10 (12), 8106-8115. 22

476

Humblet-Hua, K. N. P., Scheltens, G., van der Linden, E., & Sagis, L. M. C. (2011).

477

Encapsulation systems based on ovalbumin fibrils and high methoxyl pectin. Food

478

Hydrocolloids, 25 (4), 569-576.

479

Jones, O. G., Handschin, S., Adamcik, J., Harnau, L., Bolisetty, S., & Mezzenga R. (2011). Complexation

481

Biomacromolecules, 12 (8), 3056-3065.

482 483

of

β-lactoglobulin

480

fibrils

and

sulfated

polysaccharides.

Jones, O. G., & Mezzenga, R. (2012). Inhibiting, promoting, and preserving stability of functional protein fibrils. Soft Matter, 8 (4), 876-895.

484

Kanakis, C. D., Tarantilis, P. A., Polissiou, M. G., & Tajmir-Riahi, H. A. (2013). Probing the

485

binding sites of resveratrol, genistein, and curcumin with milk β-lactoglobulin.

486

Journal of Biomolecular Structure and Dynamics, 31 (12), 1455-1466.

487

Khanji, A. N., Michaux, F., Jasniewski, J., Petit, J., Lahimer, E., Cherif, M., Salameh, D.,

488

Rizk, T., & Banon, S. (2015). Structure and gelation properties of casein micelles

489

doped with curcumin under acidic conditions. Food & Function, 6 (12), 3624-3633.

490 491

Laemmli, U. K. (1970). Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 227, 680-685.

492

Li, M., Ma, Y., & Ngadi, M. O. (2013). Binding of curcumin to β-lactoglobulin and its effect

493

on antioxidant characteristics of curcumin. Food Chemistry, 141 (2), 1504-1511.

494

Litwinienko, G., & Ingold, K. U. (2004). Abnormal solvent effects on hydrogen atom

495

abstraction. 2. Resolution of the curcumin antioxidant controversy. The role of

496

sequential proton loss electron transfer. The Journal of Organic Chemistry, 69 (18),

497

5888-5896.

498

Liu, J., Chen, F., Tian, W., Ma, Y., Li, J., & Zhao, G. (2014). Optimization and

499

characterization of curcumin loaded in octenylsuccinate oat β-Glucan micelles with an

500

emphasis on degree of substitution and molecular weight. Journal of Agricultural and

501

Food Chemistry, 62 (30), 7532-7540.

502 503 504

Livney, Y. D. (2010). Milk proteins as vehicles for bioactives. Current Opinion in Colloid & Interface Science, 15 (1), 73-83. Matalanis, A., Jones, O. G., & McClements, D. J. (2011). Structured biopolymer-based 23

505

delivery systems for encapsulation, protection, and release of lipophilic compounds.

506

Food Hydrocolloids, 25 (8), 1865-1880.

507

Mohammadian, M., & Madadlou, A. (2016). Characterization of fibrillated antioxidant whey

508

protein hydrolysate and comparison with fibrillated protein solution. Food

509

Hydrocolloids, 52, 221-230.

510

Mohammadian, M., & Madadlou, A. (2018). Technological functionality and biological

511

properties of food protein nanofibrils formed by heating at acidic condition. Trends in

512

Food Science & Technology, 75, 115-128.

513

Mohammadian, M., Salami, M., Momen, S., Alavi, F., Emam-Djomeh, Z., &

514

Moosavi-Movahedi, A. A. (2019). Enhancing the aqueous solubility of curcumin at

515

acidic condition through the complexation with whey protein nanofibrils. Food

516

Hydrocolloids, 87, 902-914.

517

Nicolai, T., Britten, M., & Schmitt, C. (2011). β-Lactoglobulin and WPI aggregates:

518

Formation, structure and applications. Food Hydrocolloids, 25 (8), 1945-1962.

519

Nilsson, M. R. (2004). Techniques to study amyloid fibril formation in vitro. Methods, 34 (1),

520

151-160.

521

Oboroceanu, D., Wang, L. Z., Magner, E., & Auty, M. A. E. (2014). Fibrillization of whey

522

proteins improves foaming capacity and foam stability at low protein concentrations.

523

Journal of Food Engineering, 121, 102-111.

524

Pan, K., Luo, Y. C., Gan, Y. D., Baek, S. J., & Zhong, Q. X. (2014). pH-driven encapsulation

525

of curcumin in self-assembled casein nanoparticles for enhanced dispersibility and

526

bioactivity. Soft Matter, 10 (35), 6820-6830.

527 528

Patel, A., Hu, Y. C., Tiwari, J. K., & Velikov, K. P. (2010). Synthesis and characterisation of zein-curcumin colloidal particles. Soft Matter, 6 (24), 6192-6199.

529

Peng, J. F., Calabrese, V., Veen, S. J., Versluis, P., Velikov, K. P., Venema, P., & van der

530

Linden, E. (2018a). Rheology and microstructure of dispersions of protein fibrils and

531

cellulose microfibrils. Food Hydrocolloids, 82, 196-208.

532 533

Peng, J. F., Simon, J. R., Venema, P., & van der Linden, E. (2016). Protein fibrils induce emulsion stabilization. Langmuir, 32 (9), 2164-2174. 24

534

Peng, S., Li, Z., Zou, L., Liu, W., Liu, C., & McClements, D. J. (2018b). Enhancement of

535

curcumin bioavailability by encapsulation in sophorolipid-coated nanoparticles: An in

536

vitro and in vivo study. Journal of Agricultural and Food Chemistry, 66 (6),

537

1488-1497.

538

Peng, S., Zou, L., Liu, W., Li, Z., Liu, W., Hu, X., Chen, X., & Liu, C. (2017). Hybrid

539

liposomes composed of amphiphilic chitosan and phospholipid: Preparation, stability

540

and bioavailability as a carrier for curcumin. Carbohydrate Polymers, 156, 322-332.

541 542

Ross, P. D., & Subramanian, S. (1981). Thermodynamics of protein association reactions: forces contributing to stability. Biochemistry, 20 (11), 3096-3102.

543

Shen, Y., Posavec, L., Bolisetty, S., Hilty, F. M., Nyström, G., Kohlbrecher, J., Hilbe, M.,

544

Rossi, A., Baumgartner, J., Zimmermann M. B., & Mezzenga, R. (2017). Amyloid

545

fibril systems reduce, stabilize and deliver bioavailable nanosized iron. Nature

546

Nanotechnology, 12, 642–647.

547

Shimada, K., Fujikawa, K., Yahara, K., & Nakamura, T. (1992). Antioxidative properties of

548

xanthan on the autoxidation of soybean oil in cyclodextrin emulsion. Journal of

549

Agricultural and Food Chemistry, 40 (6), 945-948.

550

Tapal, A., & Tiku, P. K. (2012). Complexation of curcumin with soy protein isolate and its

551

implications on solubility and stability of curcumin. Food Chemistry, 130 (4),

552

960-965.

553

Wang, X. Y., Nian, Y. Q., Zhang, Z. J., Chen, Q., Zeng X. X., & Hu, B. (2019). High internal

554

phase emulsions stabilized with amyloid fibrils and their polysaccharide complexes

555

for encapsulation and protection of β-carotene. Colloids and Surfaces B: Biointerfaces,

556

110459.

557

Ye, Q. Y., Woo, M. W., & Selonmulya, C.. (2019). Modification of molecular conformation of

558

spray-dried whey protein microparticles improving digestibility and release

559

characteristics. Food Chemistry, 280, 255-261.

560

Zhao, Q., Xiong, H., Selomulya, C., Chen, X. D., Zhong, H., Wang, S., Sun, W., & Zhou, Q.

561

(2012). Enzymatic hydrolysis of rice dreg protein: Effects of enzyme type on the

562

functional properties and antioxidant activities of recovered proteins. Food Chemistry, 25

563

134 (3), 1360-1367.

564

Zheng, B., Peng, S., Zhang, X., & McClements, D. J. (2018). Impact of delivery system type

565

on curcumin bioaccessibility: Comparison of curcumin-loaded nanoemulsions with

566

commercial curcumin supplements. Journal of Agricultural and Food Chemistry, 66

567

(41), 10816-10826.

26

568

Figures:

569

Fig. 1 Th T fluorescence spectroscopy (A), SDS-PAGE analysis (B), and volume distribution

570

and TEM images (C) of samples including WPI, WPF-6, WPF-12, WPF-18, and WPF-24.

571

Fig. 2 Curcumin solubility in different delivery solutions including water, WPI, WPF-6,

572

WPF-12, WPF-18, and WPF-24, with or without 0.1% chitosan.

573

Fig. 3 The appearance of various curcumin delivery solutions including water, WPI, WPF-6,

574

WPF-12, WPF-18, and WPF-24, with or without 0.1% chitosan, which was stored at room

575

temperature during one month.

576

Fig. 4 Surface hydrophobicity (H0 values) of samples including WPI, WPF-6, WPF-12,

577

WPF-18, and WPF-24. Different superscript letters indicate significant differences at the p <

578

0.05 level.

579

Fig. 5 Fluorescence spectra of WPF-18 (0.1 mg/mL, pH 3.5) in the presence of different

580

concentrations of curcumin (0–90 µM) at an excitation wavelength of 295 nm (A); plot of

581

F0/F versus [Cur] as per the Stern-Volmer equation, i.e. F0/F = 1 + Kqτ0[Cur] = 1 + Ksv[Cur]

582

(B); lg[(F0−F)/F] vs lg[Cur] as per lg(F0-F)/F = lgKa + nlg[Cur] (C).

583

Fig. 6 Curcumin fluorescence spectroscopy of various delivery solutions including water,

584

WPI, WPF-6, WPF-12, WPF-18, and WPF-24, with or without 0.1% chitosan.

585

Fig. 7 ζ-potential of various curcumin delivery solutions including water, WPI, WPF-6,

586

WPF-12, WPF-18, and WPF-24, with or without 0.1% chitosan. Different superscript letters

587

indicate significant differences at the p < 0.05 level.

588

Fig. 8 TEM images of the complex: WPF-18-Cur (A); WPF-18-CS-Cur (B). And the particles

589

shown in the images indicates curcumin (~ 10 nm). 27

590

Fig. 9 Reducing power (Abs at 700 nm) (A) and DPPH scavenging activity (B) of samples

591

including curcumin, WPI, WPF, WPI-Cur, WPF-Curs, WPI-CS-Curs, and WPF-CS-Curs

592

Fig. 10 Cumulative curcumin release of various delivery solutions including WPF-6-CS,

593

WPF-12-CS, WPF-18-CS, and WPF-24-CS.

28

Highlight 1. Whey protein fibrils and chitosan were utilized as vehicles for curcumin 2. The aqueous solubility of curcumin was improved to the maximum 297.8 µg/mL 3. Hydrophobic groups of fibrils account for major binding sites of curcumin 4. Antioxidant activity of the complex significantly enhanced compared with curcumin 5. The delivery systems provided more opportunities for curcumin to release

Conflict of interest The authors declare no conflicts of interest.