Foam and emulsion properties of potato protein isolate and purified fractions

Accepted Manuscript Foam and emulsion properties of potato protein isolate and purified fractions Jesper Malling Schmidt, Henriette Damgaard, Mathias Greve-Poulsen, Lotte Bach Larsen, Marianne Hammershøj PII:

S0268-005X(17)30431-9

DOI:

10.1016/j.foodhyd.2017.07.032

Reference:

FOOHYD 4001

To appear in:

Food Hydrocolloids

Received Date: 14 March 2017 Revised Date:

14 July 2017

Accepted Date: 18 July 2017

Please cite this article as: Schmidt, J.M., Damgaard, H., Greve-Poulsen, M., Larsen, L.B., Hammershøj, M., Foam and emulsion properties of potato protein isolate and purified fractions, Food Hydrocolloids (2017), doi: 10.1016/j.foodhyd.2017.07.032. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. 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.

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Potato protein

Emulsion rheology

AC C

EP

TE D

Foam

Surface tension

ACCEPTED MANUSCRIPT

Foam and emulsion properties of potato protein isolate and purified fractions

1 2 3

RI PT

4 5

Jesper Malling Schmidta, Henriette Damgaardb, Mathias Greve-Poulsenc, Lotte Bach Larsena, Marianne

6

Hammershøj a,*

7 8

a

9

b

10

c

KMC, Herningvej 60, 7330 Brande, Denmark

11 12

M AN U

AKV Langholt, Gravsholtvej 92, 9310 Vodskov, Denmark

SC

Department of Food Science, Aarhus University, Blichers Allé 20, DK-8830 Tjele, Denmark

*corresponding author: [email protected]

13

AC C

EP

TE D

14

1

ACCEPTED MANUSCRIPT Abstract

16

Spray dried potato protein and specific isolated fractions were used for foaming and emulsification studies.

17

The spray dried protein was separated into a patatin and a protease inhibitor (PI) rich fraction by ion

18

exchange chromatography (IEX), respectively, and these two fractions were purified by hydrophobic

19

interaction chromatography into a low (HIC 1) and a high (HIC 2) hydrophobic fraction. Foam overrun for

20

the spray dried powder and all patatin fractions were highest at pH 3, with gradually lower values at pH 5

21

and 7, while the PI fractions had highest overrun at pH 5 and equally lower values at pH 3 and 7. Relative

22

foam stability varied from 18-78 % of the initial foam at pH 3 while lower variation of 67-80 % was seen at

23

pH 5 and 7. The HIC fractions did generally perform better than the spray dried powder and IEX fractions,

24

with patatin HIC 1 and HIC 2 having superior performance at pH 3 and PI HIC 2 at pH 5 and 7. Emulsions

25

were characterized by emulsion stability and activity, emulsion droplet size and small scale dynamic

26

rheological measurements. The PI, and especially PI HIC 1 showed poor emulsion properties with low

27

stability, large droplet size and less texture. Interestingly, PI HIC 2 had much better emulsion properties on

28

par with the spray dried powder and patatin, while having a more frequency dependent textural response.

29

Overall, the best emulsion properties were obtained with patatin HIC 2, thus showing the importance of

30

hydrophobicity for protein functionality.

33

SC

M AN U

TE D

EP

32

Keywords: potato; protein: patatin; foam; emulsion; surface tension

AC C

31

RI PT

15

2

ACCEPTED MANUSCRIPT 34

1. Introduction Potato protein has in recent years drawn more and more attention as a protein source for human consumption

36

and as a unique food ingredient. Industrially potato starch is the main value-added ingredient from potatoes

37

with the proteins being located in a side stream from starch production known as potato fruit juice (PFJ). PFJ

38

contains roughly 2-5 % solids of which 35 % is N-containing substances i.e. protein, peptides and amino

39

acids (Knorr, Kohler, & Betschart, 1977). The potato proteins have been concentrated from PFJ by acid

40

precipitation, but to preserve the functional properties more gentle techniques needs to be used e.g. expanded

41

bed absorption (Lokra, Helland, Claussen, Straetkvern, & Egelandsdal, 2008), ion exchange chromatography

42

(van Koningsveld, et al., 2001) or ultrafiltration (Straetkvern & Schwarz, 2012).

43

The potato proteins can roughly be divided into three main groups, with the first being the patatins,

44

representing up to 40 % of the total protein. The patatins are glycoproteins with a molecular weight of 39-43

45

kDa (existing as native 80 kDa dimers) with different isoelectric points (pI) from pH 4.45-5.17 and

46

glycosylation patterns (1-3 glycosylations) (Barta, Bartova, Zdrahal, & Sedo, 2012). The second group of

47

potato proteins is the protease inhibitors (PI), which represents up to 50 % of the total protein and this group

48

can be further divided into seven sub-groups (Pouvreau, et al., 2001). The molecular weight of the PI

49

proteins varies widely, from 4.3-20.6 kDa, with pI values of pH 5.1-9.0 (Pouvreau, et al., 2001). The last

50

group of proteins is mainly composed of oxidative enzymes, like lipoxygenase, polyphenol oxidase and

51

enzymes associated with starch synthesis (Jorgensen, Stensballe, & Welinder, 2011).

52

Two important functional properties of food proteins are their ability to form and stabilize foams and

53

emulsions by adsorbing to the interface between air and water or oil and water, hence lowering the interfacial

54

tension and providing electrostatic and steric stabilisation of the interface film (Wierenga & Gruppen, 2010).

55

Only limited results have been published on foam and emulsification behavior of potato proteins and even

56

fewer on the purified fractions (Ralet & Gueguen, 2000, 2001; van Koningsveld, et al., 2002; van

57

Koningsveld, et al., 2006).

58

The aim of this study was to provide further insight into foam and emulsion properties of potato proteins and

59

purified fractions hereof. It was theorized that purification of a total protein isolate by first ion exchange and

60

then hydrophobic interaction chromatography would lead to unique fractions with more distinct pI and

AC C

EP

TE D

M AN U

SC

RI PT

35

3

ACCEPTED MANUSCRIPT 61

hydrophobicity characteristics that could be tailored for specific food applications or pH regimes e.g. with

62

fractions high in foaming and fractions high in emulsification properties. 2. Materials and Methods

64

2.1. Protein samples

RI PT

63

Spray dried potato protein isolate powders from the harvest years of 2014 and 2015 were provided by KMC,

66

Brande, Denmark. The spray dried powder had a protein content of 83.4 % (Kjeldahl, N × 6.25) and an ash

67

content of 1 % as determined by heating 5 g of sample stepwise from 200 °C to 900 °C for a total time of 17

68

hours. Fractionation of resolubilised spray dried protein isolate by ion exchange chromatography results in a

69

patatin fraction (Pat) and a protease inhibitor fraction (PI), further purification by hydrophobic interaction

70

chromatography (HIC) resulted in two patatin fractions (Patatin HIC 1 and Patatin HIC 2) and two protease

71

inhibitor fractions (PI HIC 1 and PI HIC 2). Fractionation was done as described previously (Schmidt, et al.,

72

2017). The purified fractions were concentrated and diafiltrated against destilled water on an Ultralab

73

tangential flow system with a 10 kDa Minimate™ membrane (Pall, New York, USA) and freeze-dried before

74

use.

TE D

M AN U

SC

65

75

2.2. Preparation of protein dispersions

EP

76

The powders were suspended in buffer for one hour before use. The used buffers were 30 mM trisodium

78

citrate dihydrate (target pH 3), 22 mM sodium acetate (target pH 5) and 7.5 mM disoudium

79

hydrogenphosphate dihydrate (target pH 7). All chemicals were of analytical grade and supplied by MERCK

80

(Darmstadt, Germany) or SIGMA (St. Louise, USA). Milli-Q ultrapure water was used throughout the

81

experiments All buffers were adjusted to an ionic strength of 50 mM with NaCl. For emulsions at pH 3 a 95

82

mM trisodium citrate dihydrate was used. Prior to foam, zeta potential, or tensiometry measurements

83

suspensions were centrifuged (4700 rpm, 15 min, 4 °C) in order to avoid any insolubilized material in the

84

sample for analysis.

AC C

77

85 4

ACCEPTED MANUSCRIPT 86

2.3. Protein composition by SDS-PAGE SDS-PAGE using CriterionTM TGXTM 8-16 % precast gels (Bio-Rad, Richmond, CA, USA) was performed

88

essentially as described earlier (Laemmli, 1970). The samples were mixed 1:1 with sample buffer (20 mM

89

Tris, 2 % SDS, 20 % glycerol, pyronin Y), and reduced with 1/10 vol 0.2 M DTE and boiled for 3 min. A 30

90

µL sample of concentration (c) 1 g/L were loaded onto the gel. Gels were stained with Coomassie Brilliant

91

blue G-250. Molecular mass was estimated by a prestained broad range molecular weight marker (Thermo

92

Scientific™ SpectraTM Multicolor Broad Range Protein Ladders).

SC

2.4. Protein concentration determination

M AN U

93 94

RI PT

87

The bicinchoninic acid assay (BCA, Thermo Scientific™ Pierce™), with bovine serum albumin (2 g/L) as

96

reference protein was used for protein determination. The working solution containing Cu2+ and BCA was

97

added to a microtiter plate containing reference or sample protein and incubation was carried out for 30 min

98

at 37 °C following measurement of absorbance at 562 nm measured in a Synergy 2 Microplate reader

99

(BioTek Instruments Inc, USA). Measurements were conducted in duplicates

100 101

2.5. Zeta-potential

EP

TE D

95

Zeta-potential (mV) of 10 mL protein solutions at a concentration of 1 g/L and at pH 3, 5 and 7 were

103

determined by a Stabino® (Particle Metrix, Meerbusch, Germany) fitted with a 200 µm piston by measuring

104

for 180 seconds and the endpoint recorded. The Stabino® produces an electric streaming potential by

105

moving a piston in an oscillating motion. Zeta potential is calculated based on the streaming potential

106

following calibrated to the electrophoresis potential of a known zeta potential standard of 0.01 N KCl. A

107

minimum of four measurements per sample were made.

AC C

102

108

5

ACCEPTED MANUSCRIPT 109

2.6. Turbidity measurement

110

Spray dried potato protein isolate powder and the purified fractions hereof were dispersed at c = 1 g/L for

111

one hour in 30 mM trisodium citrate dihydrate, 22 mM sodium acetate or 7.5 mM disoudium

112

hydrogenphosphate

113

spectrophotometrically as absorbance units (AU) at λ = 500 nm in a quartz cuvette. Two to three dispersions

114

were made per sample with three measurements per dispersion.

buffer.

Turbidity

of

unsettled

measured

2.7. Foam analysis

M AN U

116

were

SC

115

dispersions

RI PT

dihydrate

Foam was produced by shaking 20 mL of 1 g/L solution in 100 mL closed cylinders at a frequency of 4 Hz

118

for 45 seconds by hand essentially as described previously (Hammershoj, Peters, & Andersen, 2004). Visual

119

evaluation of the resulting foam and liquid volumes were conducted. Foam overrun (FO) were calculated

120

(equation 1) as a relative overrun volume based on Vfoam = volume of foam (L) and Vliquid = volume of liquid

121

(L). The total volume of foam and liquid as well as the liquid volume were recorded for two hours. The foam

122

stability was calculated (equation 2) as the relative foam volume (FV) (Hammershoj, et al., 2004). Minimum

123

three repetitions were made per sample.

TE D

117

126

AC C

125

EP

124



 =   [/]   



 =    ∗ 100 [%]  

(1)

(2)

2.8. Surface tension

127

Surface tension was determined using a platinum Wilhelmy plate (19.62 × 10 × 0.1 mm) controlled by a

128

Sigma 700 force tensiometer (Biolin Scientific AB, Västra Frölunda, Sweden) with the following measuring

129

conditions; 6 mm wetting depth, 4 s stabilization time, 4 s integration time, 20 mm/min pull speed.

130

To a small glass vessel filled with 38 mL buffer, 2 mL of 10 g/L protein solution resulting in a final protein

131

concentration of 0.5 g/L was applied to the bottom of the vessel with a syringe and the measurements started.

6

ACCEPTED MANUSCRIPT The initial surface tensions of the buffers were aimed to be ~72 mN/m (as for pure water at 20 °C)) and was

133

measured initially, i.e. indicating no surface active compounds present before adding the protein samples.

134

The decrease in surface tension was monitored for six hours and the resulting surface tension γ (N m-1) value

135

recorded. Data for the first three hours was fitted to an exponential one-phase decay function and the half-life

136

calculated as the time for the surface tension to reach 50 % of the total decrease, i.e. the lower half-life time,

137

the faster initial decrease in surface tension, and the higher surface activity of the actual sample.

138

Measurements were performed as a minimum in duplicates. Measurements were conducted at room

139

temperature without temperature control (~22 °C).

SC M AN U

140 141

RI PT

132

2.9. Interfacial tension

Interfacial tension was measured at the oil-water interface between 1 g/L protein solution (pH 3) and

143

sunflower oil using a Wilhelmy plate (as described in section 2.6.) for nine hours with the following

144

measurement conditions; 14 mm wetting depth, 4 s stabilization time, 4 s integration time, 20 mm/min pull

145

speed. Measurements were done in triplicate. Measurements were conducted at room temperature without

146

temperature control (~22 °C).

149

2.10. Emulsion preparation

AC C

148

EP

147

TE D

142

150

Emulsions were prepared by mixing 6 mL of a 20 g/L protein suspension (pH 3) with 18 mL sunflower oil

151

by an ultra turrax T25 (IKA Werke GmbH & Co. KG, Staufen, Germany) for 5 min at 8000 rpm followed by

152

4 min at 13500 rpm. Oil content was aimed to represent a typical mayonnaise emulsion. Two emulsions

153

were prepared per protein sample.

154

The emulsifying activity index (EAI) (equation 3) and emulsifying stability index (ESI) (equation 4) were

155

determined by the method originally developed by Pearce and Kinsella (1978) with later modifications

156

(Cheung, Wanasundara, & Nickerson, 2014) using UV absorbance at λ = 500 nm. A0 is absorbance of the 7

ACCEPTED MANUSCRIPT diluted sample at time = 0 and A10 at time = 10 minutes, N is a dilution factor (500), c is protein

158

concentration (0.02 g/mL) and  is the oil volume fraction (0.75). The emulsion was diluted 500 times in 0.1

159

% SDS in two steps by diluting 1 g emulsion in 100 mL 0.1% SDS, and 1 mL hereof was further diluted 5-

160

fold in 0.1% SDS. After mixing for 30 s, the diluted emulsion was measured in a plastic cuvette at 500 nm in

161

a UV-visible spectrophotometer. # % $

=

&×&.()(×* ×+ ,×-×.))))

162

 ! "

163

/! 01234 = *

(4)

SC

2.11. Emulsion droplet size distribution

×7

(3)

M AN U

164

* 5*  6

RI PT

157

165

Droplet size distribution of emulsions was measured by a Mastersizer 2000 (Malvern Instruments,

166

Worcestershire, U.K.) with a Hydro 2000S sample unit. Emulsions were diluted 1:2 (v:v) with 1 % SDS and

167

volume-weighted mean diameter (D4,3) and surface-weighted mean diameter (D3,2) were determined from

168

each emulsion, resulting in a total of six measurements per sample.

170

TE D

169

2.12. Dynamic rheological analysis of emulsions

Emulsion was transferred to the rheometer (AR G2, TA instruments, New Castle, DE, USA) fitted with a

172

parallel plate-plate geometry with a gap of 1 mm (plate radius 40 mm). After a 10 min equilibration

173

oscillatory frequency sweeps (0.01–50 Hz) at 0.1 % strain and strain sweeps (0.01-10 %) at 0.5 Hz were

174

performed at 20 °C. Two frequency sweeps were performed per emulsion resulting in four measurements,

175

except for the PI HIC 1 fraction where only one sample was available.

177

AC C

176

EP

171

2.13. Statistical analysis

178

The results are presented as means with standard deviations. Significant differences between treatments were

179

determined by one or two-way ANOVA analysis with parameters of protein fraction and pH and interactions

180

hereof using GraphPad Prism 6 (GraphPad Software Inc, La Jolla, USA, version 6.01). Differences were

181

regarded to be significant at minimum 95 %-level (P < 0.05).

8

ACCEPTED MANUSCRIPT 182

AC C

EP

TE D

M AN U

SC

RI PT

183

9

ACCEPTED MANUSCRIPT 184

3. Results and Discussion

185

3.1. Protein composition The protein composition of the samples at c = 1 g/L by SDS-PAGE are visualized in Figure 1. Here, it is

187

obvious that the protein bands with masses of ~ 40-43 kDa dominate the patatin HIC 1 and HIC 2 fractions

188

with patatin HIC 2 being enriched with a higher molecular weight isoform of patatin (see arrow). The PI

189

fractions are dominated by proteins with molecular weight of 10-24 kDa for PI HIC 1 and 15-18 kDa for PI

190

HIC 2, respectively.

RI PT

186

SC

191 192

194

M AN U

193 3.2. Zeta-potential

The zeta-potential was measured of the spray dried powder, the patatin fractions and the PI fractions at three

196

different pH values (Figure 2). The zeta-potential gives the surface charge of the respective proteins in the

197

sample. The ionization of the protein, which depends on the actual pH, ionic strength and the isolectric point

198

of the proteins in concern, is reflected in the zeta-potential measured.

199

All protein samples had positive charges at pH 3, and negative charges at pH 7, which is in line with the

200

literature value of isoelectric point of the patatin protein at ~pH 4.5-5.1 (Barta et al., 2012) and of the

201

protease inhibitors (PI) at ~pH 5.1-9.0 (Povreau et al., 2001). Furthermore, at pH 5 the zeta-potential of the

202

PI fractions was close to 0 mV, while the further fractionation by HIC resulted in the 1st fraction of PI having

203

a higher positive zeta-potential and PI HIC 2 having a slightly negative value. For the patatin fraction sample

204

at pH 5, the zeta-potential showed a negative value, meaning that this fraction carried a negative charge, i.e.

205

is above its isoelectric point, which was also seen to remain in the further fractionated patatin HIC 1 and 2

206

sample.

AC C

EP

TE D

195

207

10

ACCEPTED MANUSCRIPT 208 209

All samples at pH 7 were negative, with the powder and patatin fractions resulting in the lowest values. The

210

zeta-potential can also relate to the stability of the proteins in solution, as a numerically higher zeta-potential

211

indicates a higher ability to stay in solution and resist to aggregation due to repulsive electrostatic forces.

213

RI PT

212 3.3. Turbidity

Turbidity of solutions was measured to indicate solubility and visual appearance of the different protein

215

fractions (Figure 3). The powder and patatin fraction had low turbidity at pH 3 and 7 but high turbidity at pH

216

5, indicating low solubility at this pH due to the pH being close to pI of the patatins as also shown in the

217

zeta-potential measurements. For the PI fractions, low turbidity was observed at pH 3 with slightly

218

increasing turbidity at pH 5 and pH 7, hence not in alignment with the zeta-potential measurements. The

219

patatin HIC 1 fraction showed significantly lower turbidity than patatin and patatin HIC 2 at pH 5. The

220

difference between patatin and PI fractions in turbidity relation to zeta-potential is so far unclear.

M AN U

SC

214

223

TE D

221 222 3.4. Foam studies – capacity and stability

The foaming properties of powder, patatin and PI fractions were evaluated by shaken foam capacity (Figure

225

4) and foam stability as relative volume after 2 hours (Figure 5).

227 228

AC C

226

EP

224

229

In general, all potato protein fractions and the whole spray dried powder had foaming capacities ranging

230

from ~0.8-1.8 L/L. The patatin HIC fraction 2 (most hydrophobic on basis of elution in HIC) was superior at

231

pH 3 to the other protein samples, while the PI-fractions were much lower at pH 3 compared to the two other

232

samples. At pH 5, less difference between the samples appeared with patatin and PI HIC 2 yielding highest

233

values, whereas at pH 7, the patatin HIC fraction 1 was inferior to all other samples, and here the PI HIC

234

fraction 2 (most hydrophobic PI fraction) revealed the highest foam overrun. To conclude, the patatin

11

ACCEPTED MANUSCRIPT fractions showed highest overrun at pH 3, intermediate at pH 5 and lowest at pH 7. The high foam overrun at

236

pH 3 is believed to be related to unfolding of patatin at low pH < 4.5 (Pots, de Jongh, Gruppen, Hessing, &

237

Voragen, 1998), (van Koningsveld, et al., 2002). Studies in which patatin was adjusted to pH 3 and

238

readjusted to pH 7 also showed increased foam formation due to irreversible denaturation (van Koningsveld,

239

et al., 2002). The intermediate foam volume at pH 5 and low foam volume at pH 7 could be related to the

240

zeta-potential, with pH 5 being closer to pI and pH 7 above. The PI fractions had the highest overrun at pH 5,

241

at which the zeta-potential was closest to zero compared to pH 3 and 7, except for the PI HIC 1 fraction at

242

pH 5 that had a higher zeta-potential value (Figure 2). In contrast to our results van Koningsveld, et al.

243

(2002) found the highest foam volume by whipping patatin solutions at pH 5, with a small reduction at pH 3

244

and a large reduction in foam volume at pH 7. Furthermore, the PI fraction had higher foam volume than

245

patatin at pH 5 and 7, while foam stability was highest for the patatin fractions showing significantly lower

246

liquid drainage at pH 5 and 7 than the PI fraction. These differences may be partly explained by the method

247

of foam productions e.g. has it been shown that sparging results in very low overrun and stability of the PI

248

fraction compared to patatin (Ralet, et al., 2001), (van Koningsveld, et al., 2002).

251

SC

M AN U

TE D

250

EP

249

RI PT

235

Foam stability of whole potato protein powder, the patatin fractions and the PI sample, displayed lowest

253

stability at pH 3, this was especially pronounced for the patatin fraction with less than 20 % foam remaining

254

after 2 hours (Figure 5). The two patatin HIC fractions performed very differently at pH 3, with Patatin HIC

255

1 having a significantly higher stability compared to Patatin HIC 2 and Patatin. The reason for the stability

256

difference between patatin fractions is not known, but it has been shown that phenolic compounds can alter

257

the hydrophobic character of proteins (Arntfield, 1996), (Rawel, Kroll, & Rohn, 2001) which results in

258

decreased surface activity, increased foam height stability and increased liquid drainage in model systems

259

(Rodriguez, von Staszewski, & Pilosof, 2015), furthermore HIC purification of the patatin fraction indicates

260

a lowering of phenolic substances (Schmidt, et al., 2017). The protein fractions of PI HIC 1 and PI HIC 2

AC C

252

12

ACCEPTED MANUSCRIPT showed the highest foam stability among all samples at the pH 3. It has been suggested (Partsia &

262

Kiosseoglou, 2001) that oxidation of sulfhydryl groups of small potato proteins with Mw < 12 kDa plays a

263

role in both foam overrun and stability, this observation can however not explain the high stability of the PI

264

fraction since sulfhydryl oxidation rate is minimal at pH 3 (Monahan, German, & Kinsella, 1995).

265

At pH 5 and pH 7, the protein samples had much more comparable foam stability levels. Overall, the PI

266

fraction was least susceptible for the change in pH, and even at pH 5, where the zeta-potential for this

267

fraction was ~0 mV, the foam was stable with 75% foam volume retained 2 hours after foam production. At

268

pH 7, the whole protein powder and the PI HIC 2 fractions performed highest in foam stability with the

269

overall retention of ~80 % foam volume after 2 hours.

270

Looking closer into the foam decay profiles of the foam volume, these exhibited clearly a two-step behaviour

271

at pH 3 of the patatin and patatin HIC 2 foams and to a minor degree of the whole protein powder sample

272

foam, whereas at the other conditions of sample and pH, the foam volume showed a one-step decrease

273

(Figure 6). The two-step decrease is suggested to be due to liquid drainage from the foam in the first step

274

followed by foam bubble collapse due to coalescence and Ostwald ripening in the second step. Most likely,

275

this takes place in all the analysed foams as being common destabilising mechanisms of foams

276

(Hammershoj, Prins, & Qvist, 1999), however the change is more drastic at pH 3.

M AN U

TE D

3.5. Surface tension

EP

277 278

SC

RI PT

261

The surface tensions of the whole protein powder at the different pH-values are illustrated in Figure 7 as

280

function of time until reaching a steady state. For both whole protein powder, the patatin fraction, and the PI

281

fraction, the final value of surface tension reached at steady state and the half-life time to reach 50% decrease

282

in surface tension from the initial γ of the various buffers applied are shown in Table 1. The buffers initial γ

283

were measured to be in the range of 67.8-71.7 mN/m.

AC C

279

284 285 286

13

ACCEPTED MANUSCRIPT For the whole protein powder solution, there was only a small difference in the steady state surface tension

288

as function of pH. However, the half-life time was much lower at pH 3 in comparison with that at pH 7,

289

which took almost twice as long time to reach the half-life. This could indicate that the potato protein powder

290

solution had highest surface activity at the pH 3, less at pH 5 and lowest at pH 7, which correlates well with

291

the foaming capacity results in Figure 4. The patatin fraction showed a similar trend, lowest half-life at pH 3

292

and highest at pH 7, but a significantly higher steady state surface tension at pH 5 compared to pH 3 and pH

293

7, indicating less surface activity close to pI of patatin. The PI fraction showed a different trend with an

294

increase in steady state surface tension going from pH 3 to pH 7, and the highest half-life at pH 5 with lower

295

values at pH 3 and pH 7. For the PI fraction no correlations were found between surface tension

296

measurements and foam overrun.

297

It should be noted that the calculated half-life times depend on the final steady state value, i.e. a high steady

298

state value of 52 mN/m will potentially take less time to reach than a low steady state value of 42 mN/m.

299

Hence, this dependency had an effect on the calculated values of the PI fraction, where a large difference in

300

steady state was observed, but not on the whole protein powder where the difference was minimal.

M AN U

SC

RI PT

287

TE D

301 302

pH 5

pH 7

44.01 ± 0.06A

43.98 ± 0.39A

42.60 ± 0.10A

1342 ± 67Z

1866 ± 161XY

2722 ± 146W

43.38 ± 0.38A

47.21 ± 0.25B

42.49 ± 0.49A

1124 ± 49Z

1533 ± 122YZ

2299 ± 172WX

Steady state γ, mN/m

46.51 ± 0.14B

49.29 ± 0.97C

52.6 ± 0.20D

Half-life time, s

1539 ± 64YZ

2101 ± 126X

1574 ± 24YZ

Steady state γ, mN/m

AC C

Half-life time, s

pH 3

EP

Whole protein powder

Patatin fraction

Steady state γ,mN/m Half-life time, s

Protease inhibitor fraction

303

A-D, Different letters indicate that steady state means significantly differ (P<0.05),

14

ACCEPTED MANUSCRIPT 304

W-Z, Different letters indicate that half-life time means significantly differ (P<0.05).

305 A relationship between a fast air-water adsorption and foamability of whey proteins have previously been

307

reported (Marinova, et al., 2009), but here the fastest adsorption is near pI in contrast to our results for the

308

whole powder and patatin fraction. Both hydrophobicity and charge play a role in adsorption as shown by

309

studies of modified ovalbumin where increased hydrophobicity resulted in faster initial adsorption and

310

increased electrostatic charge in a lower initial adsorption (Wierenga, et al., 2010), but we suggest that

311

increased surface hydrophobicity due to partial unfolding at pH 3 resulting in a structural change is the

312

driving force, overcoming the increase in charge at this pH compared with pH 5 and pH 7.

M AN U

SC

RI PT

306

313 314

3.6. Emulsion studies – interfacial tension and emulsification

The interfacial tension of the various protein fractions was analysed at pH 3 (Figure 8). The interfacial

316

tension at steady-state describes, when proteins have adsorbed and rearranged at the oil-water interface

317

(Romero, et al., 2011). When comparing whole protein powder to patatin and PI, patatin showed the lowest

318

value, and PI the highest, with the powder giving an intermediate value. Further fractionation of patatin by

319

HIC resulted in the least hydrophobic fraction (HIC 1) to yield the highest interfacial tension, while patatin

320

HIC 2 was not being significantly different from patatin. Fractionation of the PI fraction lead to the most

321

hydrophobic fraction (HIC 2) yielding a lower value than unfractionated PI.

322

It has previously been found that potato protein isolate reach equilibrium interfacial tensions between 9-12

323

mN/m after 5 hours dependent on pH and protein concentration (Romero, et al., 2011), which is comparable

324

to the levels in our present results (Figure 8).

AC C

EP

TE D

315

325 326 327

15

ACCEPTED MANUSCRIPT Emulsion activity index (EAI) results are shown in Figure 9. The patatin HIC 2 fraction had a significantly

329

higher emulsifying capacity given as interfacial area stabilized per mass of protein, which was ~20% higher

330

than the second-best performing samples, i.e. patatin. The PI fraction had lower EAI performance. The two

331

HIC fractions of PI performed differently with PI HIC 1 yielding lower values than PI, and PI HIC 2 higher

332

values than PI, where PI HIC 2 was at a level comparable to the patatin fractions.

RI PT

328

333 334

SC

335

Furthermore, the emulsion stability index (ESI), which is a measure of the diluted emulsion’s stability over

337

time is given in Figure 10. Again, the patatin HIC 2 fraction had the highest value, however, insignificantly

338

different from the patatin and patatin HIC 1 fraction. PI HIC 1 showed lower stability than PI, and PI HIC 2

339

higher stability than PI. The PI HIC 2 had values comparable to the patatin fractions.

M AN U

336

342

EP

341

TE D

340

The emulsion droplet diameters analysed as mean of size distributions is given both as volume weighted

344

(Figure 11A) and as surface weighted (Figure 11B) diameters. Regardless of which weighted diameter, the

345

PI fraction produced an emulsion with much larger oil droplets compared with all other samples, and the

346

patatin fraction emulsion appeared to be composed of oil droplets of smaller mean diameter. A significant

347

difference was seen with the HIC fractions of PI, with PI HIC 1 yielding larger, and PI HIC 2 smaller

348

droplets than the PI fraction. These data could partly explain the difference in emulsion activity and stability

349

of the PI fractions, which correlated with oil droplet size. In a previous study, the PI fraction also has the

350

largest oil droplets (D 4,3) at pH 3, compared to patatin and a total protein isolate, with values of 1.23, 0.84

351

and 0.73 µm, respectively (van Koningsveld, et al., 2006). In our present experiments, larger droplet sizes

AC C

343

16

ACCEPTED MANUSCRIPT were measured than previously reported e.g. 0.58-0.84 µm (D 4,3) (Romero, et al., 2011) and 0.55-5.79 µm

353

(D 4,3) (van Koningsveld, et al., 2006), which can depend on different apparatus being applied in the

354

emulsion preparation. Droplet size have been reported to correlate with patatin amount and more specifically

355

the lipid acyl hydrolase activity, resulting in release of surface active molecules lowering surface tension and

356

giving smaller oil droplets (van Koningsveld, et al., 2006). In our studies, we believe this to be of minor

357

contribution since the enzyme activity should be diminished at pH 3, and furthermore, the use of sunflower

358

oil with the majority of the fatty acids being C18 (Orsavova, Misurcova, Ambrozova, Vicha, & Mlcek, 2015)

359

conflicts with patatin having higher affinity for shorter chain fatty acids (Anderson, Pinsirodom, & Parkin,

360

2002).

361 362

3.7. Emulsion rheological properties

M AN U

SC

RI PT

352

Dynamic rheology of emulsions was performed as frequency sweeps, and the recordings of storage moduli

364

(G’) are shown in Figure 12 . There was a significant difference between the potato protein fractions in their

365

rheological properties, with the patatin HIC 2 fraction having the overall highest G’ values, i.e. the emulsion

366

is thicker in consistency. The PI fraction exhibited very low G’ values being only one third the level of all

367

other emulsion samples, i.e. this emulsion was less thick in consistency. The PI HIC 2 fraction showed a

368

significantly higher G’ than the PI fraction and the response in G’ on frequency was different than the other

369

PI samples. The replicate measurements led to consecutive higher G’ values, indicating that the time length

370

after emulsification could be of importance for the emulsion consistency.

EP

AC C

371

TE D

363

372

In summary, did the patatin and especially patatin HIC 2 show the best emulsion properties, i.e. highest EAI

373

and the highest G’-values compared to the total protein powder. The PI fraction and PI HIC 1 displayed the

374

opposite performance i.e. lower ESI, EAI and larger droplet size than the whole protein powder. PI HIC 2

375

did, however, show a performance equal to the whole powder and patatin fractions. Previous studies at low

376

pH (pH 4) find a significant difference between creaming rate of whole protein isolates being ~4 times 17

ACCEPTED MANUSCRIPT higher compared to patatin and ~6 times higher compared to the PI fraction, and furthermore, does the whole

378

protein isolate emulsions have excessive coalescence (Ralet, et al., 2000). However, negligible differences in

379

emulsion aggregation at pH 3 between PI, patatin and a total protein fraction are reported (van Koningsveld,

380

et al., 2006).

RI PT

377

381 382

4. Conclusions

In this study, patatin and protease inhibitor rich fractions obtained by ion exchange chromatography of spray

384

dried potato protein powder, were further purified into low and high hydrophobic fractions based on

385

hydrophobic interaction chromatography, and in total 7 differentsamples and fractions were analysed for

386

interfacial and rheological properties in relation to foam and emulsion studies.

387

Turbidity of total protein powder and patatin was low at pH 3 and pH 7 but high at pH 5. The least

388

hydrophobic patatin fraction (patatin HIC 1) showed decreased turbidity at pH 5 i. The PI fractions displayed

389

low turbidity at all pH values.

390

Foam overrun for the total protein powder and the all patatin rich fractions were highest at pH 3, with

391

successive lower values at pH 5 and 7. HIC fractionation of patatin led to increased foam overrun and

392

stability when compared to patatin at pH 3.

393

The protease inhibitor fractions had highest foam overrun at pH 5 with lower values at pH 3 and 7. Highest

394

overrun and stability at pH 5 and 7 was from the most hydrophobic PI fraction.

395

The spray dried powder and all patatin fractions had better emulsion stability, smaller oil droplets and higher

396

texture than PI and PI HIC 1, with the best performing fraction of all being patatin HIC 2 . Surprisingly, the

397

most hydrophobic PI fraction (PI HIC 2) displayed emulsion characteristics on par with the spray dried

398

powder and patatin.

399

This study shows that improved functionality can be obtained when potato protein fractions are further

400

purified by hydrophobic interaction chromatography and the generally increased performance was associated

401

with the most hydrophobic fractions of either patatin or protease inhibitor.

AC C

EP

TE D

M AN U

SC

383

18

ACCEPTED MANUSCRIPT 402 Acknowledgements

404

The authors thank the Future Food Innovation of region Mid-Jutland, Denmark, KMC, AKV Langholt and

405

Aarhus University for financial support of the present work.

RI PT

403

406 References

408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442

Anderson, C., Pinsirodom, P., & Parkin, K. L. (2002). Hydrolytic selectivity of patatin (lipid acyl hydrolase) from potato (Solanum tuberosum L.) tubers toward various lipids. Journal of Food Biochemistry, 26(1), 63-74. Arntfield, S. D. (1996). Effects of divalent cations, phytic acid, and phenolic compounds on the gelation of ovalbumin and canola proteins. In N. Parris, A. Kato, L. K. Creamer & J. Pearce (Eds.), Macromolecular Interactions in Food Technology (Vol. 650, pp. 82-92). Washington: Amer Chemical Soc. Barta, J., Bartova, V., Zdrahal, Z., & Sedo, O. (2012). Cultivar variability of patatin biochemical characteristics: table versus processing potatoes (Solanum tuberosum L.). J Agric Food Chem, 60(17), 4369-4378. Cheung, L., Wanasundara, J., & Nickerson, M. T. (2014). The Effect of pH and NaCl Levels on the Physicochemical and Emulsifying Properties of a Cruciferin Protein Isolate. Food Biophysics, 9(2), 105-113. Hammershoj, M., Peters, L. V., & Andersen, H. J. (2004). The significance of critical processing steps in the production of dried egg albumen powder on gel textural and foaming properties. J Sci Food Agric, 84(9), 1039-1048. Hammershoj, M., Prins, A., & Qvist, K. B. (1999). Influence of pH on surface properties of aqueous egg albumen solutions in relation to foaming behaviour. Journal of the Science of Food and Agriculture, 79, 859-868. Jorgensen, M., Stensballe, A., & Welinder, K. G. (2011). Extensive post-translational processing of potato tuber storage proteins and vacuolar targeting. FEBS J, 278(21), 4070-4087. Knorr, D., Kohler, G. O., & Betschart, A. A. (1977). Potato protein concentrates: the influence of various methods of recovery upon yield, compositional and functional characteristics. Journal of Food Processing and Preservation, 1(3), 235-247. Laemmli, U. K. (1970). Clevage of structural proteins during assembly of head of bacteriophage-T4. Nature, 227(5259), 680-685. Lokra, S., Helland, M. H., Claussen, I. C., Straetkvern, K. O., & Egelandsdal, B. (2008). Chemical characterization and functional properties of a potato protein concentrate prepared by large-scale expanded bed adsorption chromatography. Lwt-Food Science and Technology, 41(6), 1089-1099. Marinova, K. G., Basheva, E. S., Nenova, B., Temelska, M., Mirarefi, A. Y., Campbell, B., & Ivanov, I. B. (2009). Physico-chemical factors controlling the foamability and foam stability of milk proteins: Sodium caseinate and whey protein concentrates. Food Hydrocolloids, 23(7), 1864-1876. Monahan, F. J., German, J. B., & Kinsella, J. E. (1995). Effect of pH and temperature on protein unfolding and thiol-disulfide interchange reactions during heat-induced gelation of whey proteins J Agric Food Chem, 43(1), 46-52.

AC C

EP

TE D

M AN U

SC

407

19

ACCEPTED MANUSCRIPT

EP

TE D

M AN U

SC

RI PT

Orsavova, J., Misurcova, L., Ambrozova, J., Vicha, R., & Mlcek, J. (2015). Fatty Acids Composition of Vegetable Oils and Its Contribution to Dietary Energy Intake and Dependence of Cardiovascular Mortality on Dietary Intake of Fatty Acids. Int J Mol Sci, 16(6), 12871-12890. Partsia, Z., & Kiosseoglou, V. (2001). Foaming properties of potato proteins recovered by complexation with carboxymethylcellulose. Colloids and Surfaces B-Biointerfaces, 21(1-3), 69-74. Pearce, K. N., & Kinsella, J. E. (1978). Emulsifying properties of proteins: evaluation of a turbidimetric technique. Journal of Agricultural and Food Chemistry, 26(3), 716-723. Pots, A. M., de Jongh, H. H. J., Gruppen, H., Hessing, M., & Voragen, A. G. J. (1998). The pH dependence of the structural stability of patatin. J Agric Food Chem, 46(7), 2546-2553. Pouvreau, L., Gruppen, H., Piersma, S. R., van den Broek, L. A. M., van Koningsveld, G. A., & Voragen, A. G. J. (2001). Relative abundance and inhibitory distribution of protease inhibitors in potato juice from cv. Elkana. J Agric Food Chem, 49(6), 2864-2874. Ralet, M. C., & Gueguen, J. (2000). Fractionation of potato proteins: Solubility, thermal coagulation and emulsifying properties. Lebensmittel-Wissenschaft Und-Technologie-Food Science and Technology, 33(5), 380-387. Ralet, M. C., & Gueguen, J. (2001). Foaming properties of potato raw proteins and isolated fractions. Lebensmittel-Wissenschaft Und-Technologie-Food Science and Technology, 34(4), 266-269. Rawel, H. M., Kroll, J., & Rohn, S. (2001). Reactions of phenolic substances with lysozyme - physicochemical characterisation and proteolytic digestion of the derivatives. Food Chem, 72(1), 59-71. Rodriguez, S. D., von Staszewski, M., & Pilosof, A. M. R. (2015). Green tea polyphenols-whey proteins nanoparticles: Bulk, interfacial and foaming behavior. Food Hydrocolloids, 50, 108-115. Romero, A., Beaumal, V., David-Briand, E., Cordobes, F., Guerrero, A., & Anton, M. (2011). Interfacial and Oil/Water Emulsions Characterization of Potato Protein Isolates. Journal of Agricultural and Food Chemistry, 59(17), 9466-9474. Schmidt, J. M., Greve-Poulsen, M., Damgaard, H., Sunds, A. V., Zdráhal, Z., Hammershoj, M., & Larsen, L. B. (2017). A new two-step chromatographic procedure for fractionation of potato proteins with potato fruit juice and spray dried protein as source materials. Food and Bioprocess Technology, (Unpublished results) Straetkvern, K. O., & Schwarz, J. G. (2012). Recovery of Native Potato Protein Comparing Expanded Bed Adsorption and Ultrafiltration. Food and Bioprocess Technology, 5(5), 1939-1949. van Koningsveld, G. A., Gruppen, H., de Jongh, H. H. J., Wijngaards, G., van Boekel, M., Walstra, P., & Voragen, A. G. J. (2001). Effects of pH and heat treatments on the structure and solubility of potato proteins in different preparations. J Agric Food Chem, 49(10), 4889-4897. van Koningsveld, G. A., Walstra, P., Gruppen, H., Wijngaards, G., van Boekel, M., & Voragen, A. G. J. (2002). Formation and stability of foam made with various potato protein preparations. J Agric Food Chem, 50(26), 7651-7659. van Koningsveld, G. A., Walstra, P., Voragen, A. G. J., Kuijpers, I. J., Van Boekel, M., & Gruppen, H. (2006). Effects of protein composition and enzymatic activity on formation and properties of potato protein stabilized emulsions. J Agric Food Chem, 54(17), 6419-6427. Wierenga, P. A., & Gruppen, H. (2010). New views on foams from protein solutions. Current Opinion in Colloid & Interface Science, 15(5), 365-373.

AC C

443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486

20

ACCEPTED MANUSCRIPT 487 488 489

RI PT

490 491

SC

492 Figure captions

494

Figure 1. SDS-PAGE gel of protein samples. Lane 1, molecular weight marker with protein masses (Mw) in

495

kDa to the left of the gel; Lane 2, potato protein powder; Lane 3, Patatin HIC 1; Lane 4, Patatin HIC 2; Lane

496

5, PI HIC 1; Lane 6, PI HIC 2.

M AN U

493

497

Figure 2. Zeta-potential (mV) of potato protein samples (spray dried powder, patatin fraction, patatin HIC

499

fraction 1 and 2, PI fraction and PI HIC fraction 1 and 2) at pH values of pH 3, pH 5 and pH 7, n = 4. a-j bars

500

with different letters are significantly different (P<0.05).

EP

501

TE D

498

Figure 3. Turbidity analysed as absorbance units (AU) at λ = 500 nm of potato protein powder and fraction

503

solutions of c = 1 g/L as function of pH value. Bars represent means ± SD, n = 6-9. a-i bars with different

504

letters are significantly different (P<0.05).

505

AC C

502

506

Figure 4. Foaming capacity as overrun (FO) of potato protein powder and fractions of c = 1 g/L as function

507

of pH values. Bars are means ± SD, n = 3. a-j, bars with different letters are significantly different (P<0.05).

508

21

ACCEPTED MANUSCRIPT 509

Figure 5. Foam stability 2 hours after foam production as remaining relative foam volume (FV) of potato

510

protein powder and fractions of c = 1 g/L as function of pH values. Bars are means ± SD, n = 3. a-j, bars with

511

different letters are significantly different (P<0.05).

RI PT

512 Figure 6. Foam volume as function of time after foam production and of pH 3 (black), pH 5 (dashed) and pH

514

7 (grey) for A) patatin fraction, B) whole protein powder, C) PI fraction, D) patatin HIC 1 fraction, E) patatin

515

HIC 2 fraction, F) PI HIC 1 fraction, and G) PI HIC2 fraction. Vertical bars indicate ± SD.

SC

513

516

Figure 7. Examples of surface tension curves of potato protein powder solutions at pH 3 (black), pH 5

518

(dashed) and pH 7 (grey) as function of time (s).

M AN U

517

519

Figure 8. Interfacial tension at pH 3 after 9 hours of equilibrium of different fractions of potato proteins, PI

521

= protease inhibitor fraction, HIC = hydrophobic interaction chromatography, n = 3. a-e bars with different

522

letters are significantly different (P<0.05).

523

TE D

520

Figure 9. Emulsifying activity index (EAI) of various potato protein samples and fractions, n = 4. a-d, bars

525

with different letters are significantly different (P<0.05).

AC C

526

EP

524

527

Figure 10. Emulsion stability index (ESI) for various potato protein sample and fractions hereof at pH 3, n =

528

4. a-d, bars with different letters are significantly different (P<0.05).

529 530

Figure 11. Mean emulsion droplet diameters given as A) D4,3 and B) D3,2 for samples of whole potato

531

protein powder, patatin, PI and HIC fractions hereof, n = 6. a-e, bars within each panel A and B with

532

different letters are significantly different (P<0.05).

22

ACCEPTED MANUSCRIPT 533 534

Figure 12. Storage modulus (G’) from oscillatory frequency sweep in plate-plate geometry of emulsions

535

prepared with potato protein powder and fractions hereof. Strain = 0.1% and T = 20°C, n = 4.

RI PT

536 Table captions

538

Table 1 Mean ± S.D. of surface tension as final steady state value and half-life time for potato protein

539

powder and patatin fraction at c = 0.5 g/L as function of pH, n = 2-3

SC

537

AC C

EP

TE D

M AN U

540

23

PT

ED M AN U

TE D

M AN U

TE D

M AN U

TE D

M AN U

TE D

M AN U

AC C EP TE D

M AN US C

RI PT

ED

M AN

TE D

M AN US

TE D

M AN US

TE D

M AN US

TE D

M AN U

TE D

M AN US

ACCEPTED MANUSCRIPT Foam and emulsions of 7 different potato protein fractions were compared

•

Proteins were separated by ion exchange and hydrophobic interaction chromatography

•

Patatin had superior foam performance at pH 3, protease inhibitors at pH 5 and pH 7

•

Patatin had higher emulsion stability and viscosity than protease inhibitors

•

The most hydrophobic protein fractions had enhanced emulsion properties

AC C

EP

TE D

M AN U

SC

RI PT

•