Soft Matter

View Article Online

PAPER View Journal | View Issue

Behaviour of whey protein emulsion gel during oral

and gastric digestion: effect of droplet size
Cite this: Soft Matter, 2014, 10, 4173
Published on 14 April 2014. Downloaded by Temple University on 26/10/2014 11:51:36.

Qing Guo,a Aiqian Ye,*a Mita Lad,a Douglas Dalgleishb and Harjinder Singha

Received 19th March 2014
Accepted 10th April 2014
DOI: 10.1039/c4sm00598h
www.rsc.org/softmatter
A set of whey protein stabilized-emulsion gels with different droplet size distributions (D4,3 ¼ 1, 6 and 12
mm) was produced, and the mechanical properties of the gels in the linear viscoelastic region and at large
deformation were measured, along with the physicochemical and structural changes of the gels during oral
mastication and gastric digestion. The gels containing 1 mm oil droplets had an aggregated particle structure
with proteins coating at oil droplets whereas the gels containing 12 mm oil droplets had a particle-filled
structure with spatially continuous matrix. During oral processing, the release of oil droplets from the
gels increased as the droplet size increased, with coalescence being seen in gels containing oil droplets
of 6 and 12 mm diameter. Under gastric digestion, high degrees of coalescence and phase separation of
oil droplets occurred in the gels containing 6 and 12 mm oil droplets because of oil droplet release from
the gel matrix; this led to slow gastric emptying. The gels were finally broken down into peptide
aggregates and oil droplets (or free oil). The gels, containing 1 mm oil droplets disintegrated into various
particles of several to several tens of microns with a low degree of oil droplet release and coalescence.
Protein breakdown was slower in these gels, suggesting that the protein structures of the gel matrices
were affected by the sizes of the incorporated oil droplets.

result of the grinding and pyloric discrimination process, the

1 Introduction
Nearly all solid or semi-solid foods are subjected to the process
of chewing in the oral cavity. From a physiological point of view,
there are two main roles of chewing: to ensure fragmentation of
food into particles small enough to be properly lubricated by
saliva and form a cohesive bolus for swallowing;1,2 and to have
an enhanced release of avour and aroma from food structure.3
The result of chewing is mainly determined by oral factors (i.e.
dental factors, jaw muscle activity, bite force, masticatory
performance, saliva and swallowing of food) and the charac-
teristics of the food.3 The former is about the individuality of
human beings which leads to variations of chewing. The latter is
the key factor inuencing how a food is orally processed and
sensually perceived.4 The stomach is the key organ which
performs four main functions: storing food, mixing food with
gastric juice, reducing the size of food particles and emptying
the digested food. The bolus of solid food resulting from
mastication is disintegrated further under both mechanical
shearing (antral contraction wave) and biochemical digestion
(pepsin, low pH, temperature, etc.) in the human stomach.5 For
all forms of food, gastric digestion is greatly inuenced by the
initial characteristics of the food material (e.g. food structures,
food composition, etc.) and the action of oral processing.6,7 As a
a
Riddet Institute, Massey University, Private Bag 11 222, Palmerston North 4442, New
Zealand. E-mail: A.M.Ye@massey.ac.nz; Fax: +64-6-350-5655; Tel: +64-6-350-5072
b
Department of Food Science, University of Guelph, Canada
majority of digestible solids empty from the stomach as small
particles (<2 mm in diameter, and most actually are <1 mm).5,8,9
Subsequently, the physicochemical characteristics of chyme
play an important role in the digestion and adsorption of foods
in the human intestine.
Nutritionally, lipid is an important source of energy and has
shown to have important nutraceutical properties, such as
carrying vitamins A, D, E and K, u-3 fatty acids, phytosterols and
carotenoids.10 However, ample research from animal or clinical
studies, and epidemiologic or ecologic analyses provides strong
evidence that dietary fats, especially saturated fats, trans fatty
acids, and cholesterol play a role in the development of obesity
which has been linked to increased rates of heart disease, dia-
betes, certain cancers, gall bladder disease, osteoarthritis and
other obesity related disorders.11–14 Therefore, unravelling the
digestion behaviour in the digestive tract of lipids in different
food systems is critical to lipid regulation and human health.15
The digestion of lipids is an interfacial process which is
controlled by interfacial properties of emulsied oil drop-
lets.16,17 Thus, the states of oil/fat in solid food (e.g. released or
trapped within the food matrix) during oral and gastric pro-
cessing are important to the lipid digestion which mainly
occurs in the human intestine.18 Emulsion gels are semi-solid
materials containing dispersed oil droplets whose formation
and mechanical properties have been a subject of study over
several decades.19 The category of protein-based gel-like mate-
rials containing emulsied lipid includes a wide range of

This journal is © The Royal Society of Chemistry 2014 Soft Matter, 2014, 10, 4173–4183 | 4173

Soft Matter
traditional and processed food products like cheese, sausage,
set yogurt, dairy dessert, soybean curd and etc.19,20 The structure
of an emulsion gel can be understood as a combination of the
structure of the matrix, the size distribution of oil droplets, the
crystallinity of the oil and the extent of oil droplet–matrix
interactions. Because the structures of emulsion gels can be
precisely manipulated by changing protein/oil ratios, oil droplet
sizes and salt content, the emulsion gel model represents an
ideal food material for investigating the mechanistic relation-
ships between food structure and food digestion (especially the
digestion of emulsied lipid in solid foods with varying
Published on 14 April 2014. Downloaded by Temple University on 26/10/2014 11:51:36.
structures).
Recent research has shown that the release of well-bonded
small oil droplets (0.45 and 1.0 mm) in whey protein oil-in-
water emulsion gels is small during both in vitro and in vivo oral
processing, as a result of the strong interactions (hydrophobic
interactions, disulde bridges, etc.) between oil droplets and
protein matrix.21,22 Moreover, the structure of whey protein
emulsion gels can be modied by using different ionic strengths
during the gelation process, and this plays a key role in their
gastric digestion. Only a few oil droplets are released from
particulate whey protein emulsion gels (formed at high ionic
strength) in the course of 5 hours gastric digestion, whereas
large quantities of oil droplets are released from ne-stranded
whey protein emulsion gel (formed at low ionic strength) aer 3
hours gastric digestion.23 These differences in the behavior of oil
droplets during gastric digestion were attributed to the varied
crosslinking originally present in the gel structures which may
lead to the generation of different sets of peptides.
The initial droplet size distribution is an important factor,
which strongly inuences both the structure and sensory
characteristics of food emulsions.24 In the present study, heat-
set whey protein emulsion gels with different structures were
formed by varying oil droplet size distributions. The behavior of
emulsion gels during oral and gastric processing was investi-
gated. The main objective was to examine the effect of gel
structure varied by oil droplet size on the disintegration of gels
during oral and gastric processing with a focus on the behavior
of oil droplets bound to the gel matrix during digestion. This
study may help to develop a rational means of controlling lipid
digestion and target release of oil-soluble nutrients in the
human body.
2 Materials and methods
2.1 Materials
Whey protein isolate (WPI) with a protein content of $90% was
purchased from Fonterra Co-operative Group Limited, Auck-
land, New Zealand. Soybean oil was purchased from a local
supermarket. The activity of the pepsin used (Merck, Darm-
stadt, Germany) was 0.7 FIP units per mg. Milli-Q water (Milli-
pore Corporation, Bedford, MA, USA) was used for all
experiments. HCl (37%) was purchased from EMD Millipore
Corporation (Billerica, MA, USA). All other chemicals were
analytical grade. Articial saliva and simulated gastric uid
(SGF) (pH 1.5, 3 g L1 pepsin and 150 mM NaCl) were prepared
respectively.23
4174 | Soft Matter, 2014, 10, 4173–4183
View Article Online
Paper
2.2 Methods
2.2.1 Emulsion gel preparation. WPI-stabilized emulsions
(10 wt% protein and 20 wt% oil) with particle size (D4,3) of 1, 6
and 12 mm were prepared separately. A coarse emulsion was
initially prepared using a high speed mechanical mixer (L5M,
Silverson, Massachusetts, USA) at 9000 rpm for 3 min which
gave an emulsion that had a D4,3 diameter of 12 mm. The
emulsions with D4,3 of 1 and 6 mm were prepared by homog-
enizing coarse emulsions using a high-pressure two-stage valve
homogenizer (APV 2000, Albertslund, Denmark) through 4
passes at the pressure of 100 and 0 bars for the rst stage and 20
and 10 bars for the second stage respectively. The required
quantities of salt to give a nal concentration of 100 mM were
added to the emulsions, with gentle stirring to allow the salt to
completely dissolve, before heating. The emulsions were then
put in sealed cylindrical containers (inner diameter: 25 mm)
and heated at 90  C for 30 min in a water bath. Aer heating, the
gels were cooled at 4  C for 24 hours before further use.
2.2.2 Measurement of mechanical properties. Dynamic
rheological measurements were performed by a rheometer
(ARG2, TA instruments, Delaware, USA) with cup (diameter 30
mm) and rotor (28 mm) geometry. Gelation was induced in situ
by (1) heating the emulsion from 30 to 90  C at a rate of 3  C
min1; (2) holding at 90  C for 30 min; (3) cooling back down to
30  C at a rate of 1  C min1; and (4) holding at 30  C for 20 min.
The shear storage moduli (G0 ) were monitored as functions of
time. All measurements were performed in triplicate and were
made in the linear viscoelastic region (0.5% strain) at a constant
frequency of 1 Hz. The fracture test was carried out using a
texture analyser (TA instruments, Delaware, USA), with a frac-
ture wedge set with the angle of the wedges at 30 . The width,
height and length for the samples for this test were 16  16 
25 mm, respectively. The test was performed to a strain of 95%
with a speed of 2 mm s1 and a trigger force of 0.05 N. The
fracture force and strain were recorded. All tests were carried
out at room temperature with at least 12 replicates.
2.2.3 In vivo oral processing. A human mastication study
was approved by Massey University Human Ethics Committee:
Southern A (Application 11/60). Eight human subjects (4 males
and 4 females) aged from 18–49 were selected for this study. The
subjects were pre-trained to enable them to become familiar
with the textures of the samples. Prior to chewing samples,
subjects were asked to rinse their mouth using water (3  30
mL). The subjects were given samples (5 g for each sample, 25
mm and 10 mm in the diameter and height respectively)
randomly and instructed to chew samples for as long as
necessary and expectorate into a container provided before they
felt the need to swallow. Once the subject had expectorated the
sample (bolus) they were asked to rinse their mouth with a
further 30 mL of water and the debris was collected into another
container.
2.2.4 Creation of simulated gel bolus (in vitro oral pro-
cessing). A mechanical grinder (MultiGrinder II, EM0405,
Sunbeam, Australia) was used to mimic mastication and to
prepare simulated gel boluses with similar size distributions to
the boluses of human subjects. The gel samples (diameter
This journal is © The Royal Society of Chemistry 2014

Paper
25 mm and height 20 mm) with 12 mm oil droplets were ground
for 1 s, whereas the gel samples containing 1 and 6 mm droplets
were ground for 1 s, then the gel pieces were mixed and ground
for another 1 s. Simulated gel bolus was prepared by mixing 200
g of ground gel and 40 mL of articial saliva.
2.2.5 Measurement of particle size distribution of gel
bolus. The masticated gel (bolus and debris) from each subject
or simulated gel bolus was poured through a stack of sieves
(mesh size: 0.038, 0.425, 0.85, 1.40, 2.00 and 3.15 mm). Gel
particles retained in each sieve were washed carefully using
mild running water with gentle shaking. The retained gel
Published on 14 April 2014. Downloaded by Temple University on 26/10/2014 11:51:36.
particles in each sieve were washed off onto pre-dried and pre-
weighed lter papers and dried in an oven (105  C) for 24 h
before reweighing. The dry matter retained in each sieve was
calculated from the difference in weights. It is difficult to collect
and weigh the particles passing through the sieve of 0.0380 mm.
Therefore, the weight of gel particles passing through the 0.038
mm sieve was estimated from the ratio of area of particles
<0.038 mm and particles >0.038 mm in the particle size distri-
butions measured by a Mastersizer 2000 (Malvern Instruments,
Worcestershire, UK) of gel fragments passing through a sieve
with the aperture of 0.85 mm (weight of gel particles >0.038 mm
known by sieving analysis). The weight of dry matter retained in
each sieve and that smaller than 0.038 mm were expressed as a
percentage of the whole dry weight. The particle size distribu-
tions were expressed as the cumulative weight (%) of gel parti-
cles passing through a sieve as a function of the sieve mesh size
and the median size was extracted from the curves.
2.2.6 Quantication of oil droplet release. Oil droplet
release during in vivo and in vitro oral processing was measured
by centrifugation.21 The masticated gels including gel bolus and
debris from subjects or simulated gel bolus (40 g as gel) were
put in tubes and centrifuged at 2000g for 20 min. Aer centri-
fugation, the tubes were frozen at 20  C and the top oil layer
was cut off. The oil phase was thawed and put back into a tube
(diameter: 15 mm) before further centrifugation aer which the
height of oil layer was measured aer centrifugation (4500g)
which gives the extent of oil droplet release.
2.2.7 In vitro gastric digestion. A human gastric simulator
(HGS) was used for in vitro gastric digestion, the apparatus of
which was described in our previous paper.23 First, a mesh bag
(pore size 1 mm) was placed inside the stomach (latex
chamber) to mimic stomach pyloric sieving. Then the simulated
gel bolus (200 g mixed with 40 mL of articial saliva) was fed to
the HGS. Before digestion, 70 mL of SGF was loaded to mimic
the fasting condition of the stomach. Then the simulated
digestion process (removal of digesta and addition of SGF) was
started immediately. The secretion rate of the SGF was 2.5 mL
min1. The digesta began to be emptied out from 30 min. For
accurate control of the gastric emptying, the digesta (45 mL
aliquots) were removed from the bottom of the stomach every
15 min, equalling the gastric emptying rate of 3.0 mL min1.
The contraction frequency was 3 times per min, simulating the
actual stomach contraction frequency of 3 cycles per min. The
temperature of the HGS was maintained at 37  C.
2.2.8 Gastric emptying. The digesta, which were emptied at
different digestion times (45 mL each time), were dried in an
This journal is © The Royal Society of Chemistry 2014
View Article Online
Soft Matter
oven (105  C) for 24 h and the mass of dried digesta (X) was
determined. In addition, a control experiment using 200 g of
water as sample was done to calculate the dry matter of SGF
retained in each digesta (Y). The actual dry weight of gel parti-
cles in the digesta emptied at different times was determined by
subtracting Y from X. Then the cumulative dry weight of gel
emptied out of the stomach at different times was calculated.
The gastric retention percentage was calculated as: y(t) ¼ (W0 
Wt)/W0, where W0 is the initial dry weight (g) of gel and Wt (g) is
the cumulative dry weight of emptied gel at time t. In addition,
the digesta in the stomach aer the 300 min digestion period
was immediately put into a beaker to test whether an oil layer
formed on the top.
2.2.9 Analysis of protein hydrolysis. The protein composi-
tions of the emptied digesta were determined by sodium
dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE)
under reducing conditions. Samples (0.2 g) were treated with
1.8 mL of water and 2 mL of tricine sample buffer and were
heated in a boiling water bath for 10 min. They were centrifuged
for 20 min at 4200g aer they had been cooled to room
temperature to remove separated oil, and an amount of sub-
natant with the same protein weight for each sample of digesta
was loaded on to gels previously prepared on a Mini-PROTEAN
II system (Bio-Rad Laboratories, Richmond, CA, USA). The
resolving gel contained 16.0% acrylamide and the stacking gel
was made up of 4.0% acrylamide. Aer running, the gel was
stained for 60 min with a Coomassie Brilliant Blue R-250 solu-
tion and 20% isopropanol. The gel was destained with a solu-
tion of 10% acetic acid and 10% isopropanol and scanned using
a Molecular Imager Gel Doc XR system (Bio-Rad Laboratories).
2.2.10 Measurement of particle size distribution of
emptied digesta. Particle size distributions of digesta emptied
through the mesh bag at 30, 60, 90, 120, 150, 180, 210, 240, 270
and 300 min were measured immediately by the Mastersizer
(Malvern Instruments, Worcestershire, UK).23 The refractive
index of the gel particles of a whey protein emulsion gel was set
as 1.47 and that of the dispersion medium (i.e. water) was set as
1.33. The average volume-weighed diameter (D4,3) was reported
for the particle size distributions.
2.2.11 Measurement of size distribution of oil droplets in
the gel, bolus and digesta. The procedures for measuring size
distribution of oil droplets followed our previous study.21 The
addition of SDS and b-mercaptoethanal to the pieces of original
gels, the gel bolus (including debris for human gel bolus) and
gastric emptied digesta liberated and stabilized oil droplets.
The protein matrix was dissolved during this process. The
particle size distribution of the oil droplets (refractive index:
1.47) was measured using the Mastersizer (Malvern Instru-
ments, Worcestershire, UK). A large volume manual wet sample
dispersion unit designed for use with standard laboratory
beakers (Hydro 2000MU) was used to disperse samples with de-
ionized water as dispersant. Individual particles were sus-
pended in water by stirring. The average volume-weighed
diameter (D4,3) was reported for the particle size distributions.
2.2.12 Confocal scanning laser microscopy (CLSM). The
microstructure of original gel, gel bolus (mixed gel bolus by
different human subjects) and gastric digesta was studied using
Soft Matter, 2014, 10, 4173–4183 | 4175
 View Article Online

Soft Matter Paper

confocal laser scanning microscopy (Leica, Heidelberg, Ger-

many). Nile Red (1%) was used to stain for oil (argon laser with
an excitation line at 488 nm) and Fast Green (1%) was used to
stain for protein (He–Ne laser with an excitation line at 633 nm).
The small pieces of original gels were stained for 2 hours in
the mixture of dyes (1 : 1). A small quantity of gel bolus or
gastric digesta was placed on a concave confocal microscope
slide immediately aer the oral or gastric processing, mixed
with 12 mL of Nile Red and 12 mL of Fast Green, stained for 15
min and then covered with a cover slip.
2.2.13 Statistical analysis. The data of mechanical proper-
Published on 14 April 2014. Downloaded by Temple University on 26/10/2014 11:51:36.

ties, oil droplet size in the original gel, gel bolus and gastric
emptied digesta, oil droplet release, particle size of gastric
emptied digesta during gastric digestion and gastric emptying
were analyzed using one-way ANOVA. Means were compared by
Duncan multiple-range test at P < 0.05 (using SPSS Statistics 21
soware).

3 Results and discussion

3.1 Gel structure and mechanical properties
As shown in Fig. 1 (LM), oil droplets appeared to be homoge-
neously distributed within the protein networks in the gels
containing 1 and 6 mm oil droplets. For the gels containing oil
droplets of 12 mm, the droplets did not appear as evenly
distributed or as rmly bound to the protein network as did the
smaller droplets. CLSM image of the gels containing 1 mm oil
droplets with high magnication shows the porous network
structure (Fig. 1 (HM)). Proteins adsorbed and surrounded the
oil droplets to form a thick layer and the gel appeared to be
formed by the aggregation of protein-coated oil droplets via Fig. 1 CLSM images of whey protein emulsion gels (A, B and C: gels
surface–surface cross-links or interactions. However, in the gel containing 1, 6 and 12 mm oil droplets respectively). Green colour
represents protein, red colour represents the oil phase, and black
containing 12 mm oil droplets, large oil droplets were just colour represents air or water. LM and HM represent the low and high
embedded or bound to a typical heat-induced whey protein magnification respectively.
network formed at high ionic strength (i.e. syneresis and
spatially continuous protein matrix) (Fig. 1 (HM)).25 The struc-
ture of protein matrix in the gels containing 1 mm droplets was spatially continuous protein matrix.19 The structure of the gels
totally modied by the adsorption and surrounding of whey containing 6 mm oil droplets is between these two gel types.
protein strands or aggregates at oil–water interfaces during gel The mechanical properties of whey protein emulsion gels are
formation because of a much higher surface area of the oil presented in Table 1. The shear storage modulus (i.e. mechan-
phase.23,26,27 Whey protein strands (several tens of nanometres) ical property in linear viscoelastic region) of the emulsion gels
assembled from whey protein molecules are the primary decreased signicantly with increasing droplet size which is in
structural units for protein gelation (heating).28 At high ionic accordance with the study of Ye and Taylor.30 Effective stress
strength, these strands randomly aggregate into larger dense transfer is the most important factor which contributes to the
aggregates (several microns) which are cross-linked into the 3D strength of two-phase composite gels.31 For a given content,
gel network.29 Moreover, for an oil content of 20%, and even smaller oil droplets induce strong bonding between oil droplets
distribution of droplets, it can be calculated that each droplet
resides inside a cube with a side 2.75R, where R is the radius of
4 Table 1 Mechanical properties of whey protein emulsion gelsa
droplet volume n pR3
the droplet (i.e. ¼ 3 z 20%, n is the
cube volume nð2:75RÞ3 Storage Fracture Fracture
Gel type modulus (kPa) force (N) strain (%)
number of droplets). The distance between the droplets'
surfaces is 0.75R (i.e. 0.38, 2.25 and 4.50 mm for the gels A 84.0  3.5* 12.0  2.7* 50.6  4.8*
containing 1, 6 and 12 mm oil droplets respectively). Therefore, B 71.4  1.7** 7.5  1.2** 46.2  5.7**
the gels containing 1 mm oil droplets can be regarded as a type C 63.7  3.9*** 4.0  0.2*** 31.3  2.4***
of aggregated particle gel whereas the gels containing 12 mm oil a
A, B and C represent the gel with oil droplet size (D4,3) of 1, 6 and 12
droplets can be regarded as a type of particle-lled gel with mm respectively. * to *** represent signicant difference between
different gels (P < 0.05).

4176 | Soft Matter, 2014, 10, 4173–4183 This journal is © The Royal Society of Chemistry 2014

Paper
and protein matrix because of the higher interfacial surface area
to which protein is adsorbed compared to large oil droplets.
This strong bonding can effectively transfer the applied stress to
the emulsion droplets from the protein matrix and improve gel
strength,32 which explains the decrease of storage modulus with
increasing oil droplet size. Similarly, the fracture force and
strain signicantly decreased with the increase of droplet size,
indicating that large oil droplets may act as defects in the
fracture test.33
Published on 14 April 2014. Downloaded by Temple University on 26/10/2014 11:51:36.
3.2 Breakdown properties in the human mouth
CLSM images of gel boluses are presented in Fig. 2. Aer
mastication, only a few oil droplets were released from the gels
Fig. 2 CLSM images of boluses of whey protein emulsion gel after human mastication and oil droplet distributions in the gel boluses (A, B and C:
gels containing 1, 6 and 12 mm oil droplets respectively). Green colour represents protein, red colour represents the oil phase, and black colour
represents air or water.
This journal is © The Royal Society of Chemistry 2014
View Article Online
Soft Matter
containing smaller oil droplets (1 mm) whereas large quantities
of oil droplets were released from the protein matrices of gels
containing the larger oil droplets. The difference in oil droplet
release is attributed to the differences in gel structure caused by
the oil droplet size. Oil droplet release is difficult in the oral
processing of aggregated particle gel because of the protection
of thick protein coating around them and strong interactions
between protein-coated oil droplets under low electrostatic
repulsion. However, for the particle-lled gel, the oil droplets
released easily upon the deformation or cutting because of the
low stress transfer capacity across oil–protein interface thereby
leading to cracking of the interface, which is supported by the
decrease of the fracture force and strain with increasing oil
droplet size. In the CLSM images there appeared to be larger oil
Soft Matter, 2014, 10, 4173–4183 | 4177
 View Article Online

Soft Matter Paper

droplets in the bolus of the gels containing 12 mm oil droplets

than in the original gel, suggesting that coalescence occurred
during mastication (Fig. 2). The value of D4,3 of oil droplets in
the original gels (1, 6 and 12 mm) were 1.03, 6.01 and 11.84 mm,
respectively. Aer mastication, droplet size distributions had no
signicant changes with those before mastication for the rst
two gels (Fig. 2). The value of D4,3 of oil droplets in the bolus
showed only a slight increase in the gels containing 12 mm oil
droplets. A similar phenomenon has also been observed in the
oral behaviour of liquid emulsions.34 This could be explained by
the limitation of particle size measuring technique (i.e. large oil
Published on 14 April 2014. Downloaded by Temple University on 26/10/2014 11:51:36.

droplets and free oil are difficult to disperse in water) and

creaming or phase separation of large oil droplets before
particle size measurement.
The mean particle size distributions of the gel boluses are
presented in Fig. 3. The shapes of the curves of three gels were
similar. The approximate median size d50 extracted from these
curves was around 1 mm for different gels, and there was not
much difference between the gels, indicating that heat-set whey
protein emulsion gels with different droplet size distributions
had similar degrees of fragmentation in the mouth.

3.3 Characterization of the simulated gel bolus

The simulated gel boluses which were created for the gastric
digestion in the HGS had similar particle size distributions
(D50: 1 mm) with those of human subjects (data not shown).
The behaviour of oil droplets in gels during the creation of
simulated gel boluses is presented in Fig. 4. During in vitro oral
processing, the gels containing 1 mm oil droplets showed
similar behaviour to that by in vivo oral processing. By
contrast, the value of D4,3 of oil droplets signicantly increased
from 5.94 and 11.84 to 7.74 and 24.92 mm for the gels con-
taining 6 and 12 mm oil droplets respectively (Fig. 4A) indi-
cating that signicant coalescence of oil droplets occurred
during the creation of simulated boluses of the gels containing
6 and 12 mm oil droplets. In addition, the oil droplet release
during both in vivo and in vitro oral processing increased with
Fig. 4 Particle size distributions of oil droplets in the simulated gel
bolus (A) and oil droplet release during the oral processing (B).
increasing droplet size (Fig. 4B). Oil droplet release from the
simulated bolus of gels containing 6 and 12 mm oil droplets
was higher than that from the gel bolus produced by human
subjects. The different behaviour between the in vitro and in
vivo oral processing for the gels containing large oil droplets is
attributed to a higher retention of coalesced oil droplets in in
vivo processing. Dresselhuis et al.35 found that over 15% WPI-
stabilized emulsion (0.3 wt% protein and 7 wt% oil) was
retained in the mouth aer the oral processing (i.e. processing
10 mL emulsion by moving the tongue against the palate for 15
s and rinsing the mouth with water using the same procedure
as for the emulsion immediately aer expectoration). In these
studies the fat retention increased with the increase of droplet
size, which explains a higher oil droplet release during the in
vitro oral processing. Furthermore, ex vivo oral processing
(tribology measurement using pig's tongue) of WPI-stabilized
emulsions showed that coalescence of emulsion droplets
occurred due to the surface friction. Therefore, the retention of
coalesced oil droplets in the mouth during oral processing is
likely to be the key factor leading to a lower apparent coales-
cence of oil droplets in the gels containing large oil droplets (6
and 12 mm) shown in the CLSM images of gel boluses of
human subjects.

3.4 In vitro gastric digestion

Fig. 3 Mean particle size distribution of fragments of whey protein

3.4.1 Gastric emptying. The acidic environment, pepsin-
emulsion gels with different droplet size after chewing (obtained from digestion and mechanical shearing led to further disintegra-
8 human subjects). tion of emulsion gels in the HGS. The gastric emptying

4178 | Soft Matter, 2014, 10, 4173–4183 This journal is © The Royal Society of Chemistry 2014
 View Article Online

Paper Soft Matter

proles of emulsion gels which are represented by the reten-
tion of solids in the HGS as a function of time are illustrated
in Fig. 5. The gastric emptying curve of the gels containing 1
mm oil droplets was more rapid than those of the gels con-
taining larger oil droplets and there were 20.1, 27.9 and
33.6% of total gel particles retained in the HGS for the gels
containing 1, 6 and 12 mm oil droplets respectively aer 300
min digestion. This indicates the emptying rate of the gels
decreased with the increase of droplet size, which can be
attributed to the creaming of large oil droplets during gastric
digestion. Aer 300 min digestion, an oil layer was observed in
Published on 14 April 2014. Downloaded by Temple University on 26/10/2014 11:51:36.
gel structure (Fig. 1 and Table 1). This strong effect of the oil
droplets on the protein matrix is likely to modify the cleavage
sites of the whey proteins or to restrict the access of pepsin to
the cleavage sites, which will hinder the whey protein
hydrolysis. The protein in the gels containing the largest oil
droplets was easily attacked because of the limited effect of
oil droplets on the protein matrix structure (i.e. small effects
of the oil droplets on the protein matrix). The protein
hydrolysis rate in the gels containing 6 mm oil droplets is
between those two gels (i.e. aggregated particle and particle-
lled gels).

the gels (6 and 12 mm) because of the creaming of released oil

droplets (Fig. 5), which is consistent with previous studies
showing acid-unstable emulsions rapidly separated into lipid-
depleted ‘aqueous’ and lipid layers in the human stomach, as
observed by magnetic resonance imaging36 and whey protein
stabilized-emulsions showed distinct creamed layers during in
vitro gastric digestion.37 However, creaming of emulsion
droplets was not observed in the gels containing small oil
droplets (1 mm).
3.4.2. Protein hydrolysis. The gastric proteolysis of whey
proteins within the emptied digesta including bovine serum
albumin (BSA) (66 kDa), a-lactalbumin (a-la) (14 kDa) and
b-lactoglobulin (b-lg) (18 kDa) was followed by tricine-SDS-
PAGE in reducing conditions (Fig. 6). BSA disappeared rapidly
during hydrolysis; a-la and b-lg were gradually hydrolysed
with digesting time. For all three emulsion gels, the whey
proteins were gradually hydrolysed into peptides <10 kD as
the digestion proceeded. Although all three gels had the same
protein and oil contents, the intensity of a-la and b-lg bands
decreased more rapidly in the samples with larger oil droplets
and especially these bands disappeared at an early time in
the digestion of gels containing 12 mm oil droplets (i.e. the
hydrolysis rate of whey proteins increases with increasing
droplet size). This may be attributed to the effect of droplet
size on the protein network. In the gels with the smallest oil
droplets, the droplets themselves have a strong effect on the

Fig. 5 Gastric emptying profiles of whey protein emulsion gels with

varied droplet size distribution in the HGS and images of top oil layer Fig. 6 Tricine SDS-PAGE patterns under reducing conditions of
after 300 min of in vitro gastric digestion (A, B and C: gels containing 1, proteins in emptied digesta (A, B and C: gels containing 1, 6 and 12 mm
6 and 12 mm oil droplets respectively). oil droplets respectively).

This journal is © The Royal Society of Chemistry 2014 Soft Matter, 2014, 10, 4173–4183 | 4179

Soft Matter
3.4.3 Breakdown of gel particles and behaviour of oil
droplets. The evolution of particle size of the gel fragments in
emptied digesta during gastric digestion was followed by light
scattering (Fig. 7). For the gels containing 1 mm oil droplets, the
value of D4,3 gradually increased followed by a rapid decrease
(Fig. 7A-1). At 30 min, the particle size distribution was bimodal,
with peaks at about 100 and 1000 mm. With further digestion,
the area of the peak at 1000 mm increased to a maximum at 180
min because of the breakdown of large gel particles and then
decreased because of strong pepsin-digestion. The peak at 100
mm gradually shied towards smaller size (10 mm) (Fig. 7A-2).
Published on 14 April 2014. Downloaded by Temple University on 26/10/2014 11:51:36.
The gels containing 6 mm oil droplets showed a similar trend to
the gels containing 1 mm oil droplets (Fig. 7B). For the gels
containing 12 mm oil droplets, the value of D4,3 gradually
decreased from 219.3 to 79.6 mm aer 300 min digestion
(Fig. 7C-1). At 30 min, the particle size distribution was mono-
modal (Fig. 7C-2), and this pattern remained over the rst 120
min. With further digestion a bimodal pattern appeared, and
the peak slowly shied to smaller size.
Coalescence of oil droplets was observed in all three gels
during gastric digestion (Fig. 8). Coalescence of emulsion
droplets during gastric digestion can be induced by the
Fig. 7 Evolution of the volume-weighted diameter (D4,3) (-1) and size distributions (-2) of emptied digesta (i.e. gel fragments) of gels with varied
droplet size (A, B and C: gels containing 1, 6 and 12 mm oil droplets respectively).
4180 | Soft Matter, 2014, 10, 4173–4183
View Article Online
Paper
decreased electrostatic repulsion at isoelectric point, mechan-
ical shearing and pepsin-digestion of the adsorbed protein
layer.23,38,39 The degree of coalescence of oil droplets in the gels
containing 1 mm oil droplets was low because of limited oil
droplet release, as shown by the appearance of only small
amounts of material above about 5 mm diameter (Fig. 8A). van
Aken et al.37 found that WPI-stabilized emulsions had a signi-
cant coalescence during gastric digestion, suggesting that the
gel matrix in our experiments provides a good protection for oil
droplets in the gel containing 1 mm oil droplets. In contrast, a
high degree of coalescence occurred in the gels containing 6 and
12 mm oil droplets during gastric digestion because of oil droplet
release and large droplet size (Fig. 8B and C). However, the value
of D4,3 of oil droplets decreased quickly aer 60–90 min due to
the creaming of large coalescenced oil droplets. Compared with
Fig. 7, a small overlap of particle size distributions between
emptied digesta and oil droplets in the emptied digesta at
different times in the gels containing 1 mm oil droplets sug-
gested most oil droplets were still retained in the gel particles
during gastric digestion. However, the overlaps of the gels con-
taining larger oil droplets (6 and 12 mm) were high suggesting
large quantities of oil droplets released from gel particles.
This journal is © The Royal Society of Chemistry 2014

Paper
Published on 14 April 2014. Downloaded by Temple University on 26/10/2014 11:51:36.
Fig. 8 Evolution of the volume-weighted diameter (D4,3) (-1) and size distributions (-2) of oil droplets in emptied digesta of gels with varied
droplet size during gastric digestion (A, B and C: gels containing 1, 6 and 12 mm oil droplets respectively).
Microstructure of emptied digesta at different times is
shown in Fig. 9 which is consistent with the results of light
scattering. A few of coalesced oil droplets were observed in the
gels containing 1 mm oil droplets (Fig. 9A (LM)). 10 mm gel
particles were still retained aer 300 min digestion and oil
droplets were within the protein matrix (Fig. 9A (HM)). By
contrast, there were signicant oil droplet release and coales-
cence of oil droplets at 60 and 120 min observed for the gels
containing 6 and 12 mm oil droplets (Fig. 9B and C). A few oil
droplets appeared to be still linked by the proteins aer the
300 min digestion for the gels containing 6 mm oil droplets
whereas the gel particles were broken down into free oil
droplets and peptide aggregates in the gels containing 12 mm
oil droplets.
The gel structure plays a key role in the breakdown of gel
particles during gastric digestion. Slow protein hydrolysis
hindered the breakdown of gel particles of the gel containing
small oil droplets (1 mm). Although very few intact whey
proteins survived aer 300 min digestion, the peptide aggre-
gates formed mainly via disulde bridges and hydrophobic
interactions can link oil droplets together because of the
aggregated particle structure.23,40–42 However, for the gels
This journal is © The Royal Society of Chemistry 2014
View Article Online
Soft Matter
containing large oil droplets (12 mm), the protein hydrolysis
was much higher, which accelerated the breakdown of
gel particles. The continuous protein matrix of particle lled
gel was easily disintegrated by pepsin-digestion, mechanical
shearing and acidic effect and peptide aggregates were
difficult to link large oil droplets over long distances (Fig. 1
and 9).
In summary, the structure of whey protein emulsion gels
was determined by the oil droplet distribution. The solid
protein matrix of the gels containing large oil droplets (6 and
12 mm) cannot protect oil droplets during the oral and gastric
digestion because of the particle-lled structure. When the oil
droplets were released from the gel matrix, they behaved
similarly to a protein-stabilized oil-in-water emulsion (i.e.
destabilization of oil droplets will occur). Meanwhile, small
well-bonded oil droplets (1 mm) can pass the mouth and
stomach with limited destruction because of the aggregated
particle structure with thick coating at oil droplets. Moreover,
the aggregated particle structure can hinder the hydrolysis of
whey proteins. To unravel the overall mechanism of food
emulsion gel behaviour during digestion, further work will
focus on intestinal digestion.
Soft Matter, 2014, 10, 4173–4183 | 4181
 View Article Online

Soft Matter Paper

Published on 14 April 2014. Downloaded by Temple University on 26/10/2014 11:51:36.

Fig. 9 Microstructure of emptied digesta at different times (A, B and C: gels containing 1, 6 and 12 mm oil droplets respectively). LM and HM
represent low and high magnification respectively. Green colour represents protein, red colour represents the oil phase, and black colour
represents air or water.

12 P. T. James, R. Leach, E. Kalamara and M. Shayeghi, Obesity,

Acknowledgements
2001, 9, 228S–233S.
The authors gratefully acknowledge Miss Janiene Gilliland from 13 J. O. Hill, E. L. Melanson and H. T. Wyatt, J. Nutr., 2000, 130,
the Riddet Institute for the help with the gastric digestion 284S–288S.
experiments and Dr Jianyu Chen from the Manawatu Micros- 14 S. Gesta, Y.-H. Tseng and C. R. Kahn, Cell, 2007, 131, 242–
copy and Imaging Centre for her help with confocal microscopy. 256.
15 P. J. Wilde and B. S. Chu, Adv. Colloid Interface Sci., 2011,
165, 14–22.
References 16 H. Singh and A. Ye, Curr. Opin. Colloid Interface Sci., 2013, 18,
360–370.
1 J. B. Hutchings and P. J. Lillford, J. Texture Stud., 1988, 19,
17 J. Maldonado-Valderrama, P. Wilde, A. Macierzanka and
103–115.
A. Mackie, Adv. Colloid Interface Sci., 2011, 165, 36–46.
2 J. F. Prinz and P. W. Lucas, Arch. Oral Biol., 1995, 40, 401–403.
18 H. Singh, A. Ye and D. Horne, Prog. Lipid Res., 2009, 48, 92–
3 A. van der Bilt, in Food Oral Processing, Wiley-Blackwell,
100.
Oxford, 1st edn, 2012, ch. 4, pp. 61–93.
19 E. Dickinson, Food Hydrocolloids, 2012, 28, 224–241.
4 M. L. Jalabert-Malbos, A. Mishellany-Dutour, A. Woda and
20 Y. Matsumura, I.-J. Kang, H. Sakamoto, M. Motoki and
M. A. Peyron, Food Qual. Prefer., 2007, 18, 803–812.
T. Mori, Food Hydrocolloids, 1993, 7, 227–240.
5 K. Schulze, Neurogastroenterol. Motil., 2006, 18, 172–183.
21 Q. Guo, A. Ye, M. Lad, D. Dalgleish and H. Singh, Food
6 M. Koziolek, G. Garbacz, M. Neumann and W. Weitschies,
Hydrocolloids, 2013, 33, 215–224.
Mol. Pharm., 2013, 10, 1610–1622.
22 G. Sala, F. van de Velde, M. A. Cohen Stuart and G. A. van
7 F. Kong and R. P. Singh, J. Food Sci., 2008, 73, R67–R80.
Aken, Food Hydrocolloids, 2007, 21, 977–985.
8 J.-R. Malagelada and F. Azpiroz, in Comprehensive Physiology,
23 Q. Guo, A. Ye, M. Lad, D. Dalgleish and H. Singh, So Matter,
John Wiley & Sons, Inc., 2010.
2014, 10, 1214–1223.
9 J. H. Meyer, H. Ohashi, D. Jehn and J. B. Thomson,
24 D. J. McClements and J. Weiss, in Bailey's Industrial Oil and
Gastroenterology, 1981, 80, 1489–1496.
Fat Products, John Wiley & Sons, Inc., 2005.
10 M. Golding and T. J. Wooster, Curr. Opin. Colloid Interface
25 K. Ako, T. Nicolai and D. Durand, Biomacromolecules, 2010,
Sci., 2010, 15, 90–101.
11, 864–871.
11 G. Bray and B. Popkin, Am. J. Clin. Nutr., 1998, 68, 1157–1173.

4182 | Soft Matter, 2014, 10, 4173–4183 This journal is © The Royal Society of Chemistry 2014

Paper
26 D. G. Dalgleish, Trends Food Sci. Technol., 1997, 8, 1–6.
27 T. Phan-Xuan, D. Durand, T. Nicolai, L. Donato,
C. Schmitt and L. Bovetto, Biomacromolecules, 2013, 14,
1980–1989.
28 S. Ikeda and V. J. Morris, Biomacromolecules, 2002, 3, 382–
389.
29 S. Ikeda, Food Hydrocolloids, 2003, 17, 399–406.
30 A. Ye and S. Taylor, Int. Dairy J., 2009, 19, 721–727.
31 S.-Y. Fu, X.-Q. Feng, B. Lauke and Y.-W. Mai, Composites, Part
B, 2008, 39, 933–961.
32 D. Qian, E. C. Dickey, R. Andrews and T. Rantell, Appl. Phys.
Published on 14 April 2014. Downloaded by Temple University on 26/10/2014 11:51:36.
Lett., 2000, 76, 2868–2870.
33 D. J. McClements, F. J. Monahan and J. E. Kinsella, J. Texture
Stud., 1993, 24, 411–422.
34 D. M. Dresselhuis, E. H. A. de Hoog, M. A. Cohen Stuart,
M. H. Vingerhoeds and G. A. van Aken, Food Hydrocolloids,
2008, 22, 1170–1183.
This journal is © The Royal Society of Chemistry 2014
View Article Online
Soft Matter
35 D. M. Dresselhuis, M. A. C. Stuart, G. A. van Aken,
R. G. Schipper and E. H. A. de Hoog, J. Colloid Interface
Sci., 2008, 321, 21–29.
36 L. Marciani, M. S. J. Wickham, D. Bush, R. Faulks, J. Wright,
A. J. Fillery-Travis, R. C. Spiller and P. A. Gowland, Br. J. Nutr.,
2006, 95, 331–339.
37 G. A. van Aken, E. Bomhof, F. D. Zoet, M. Verbeek and
A. Oosterveld, Food Hydrocolloids, 2011, 25, 781–788.
38 A. Sarkar, K. K. T. Goh, R. P. Singh and H. Singh, Food
Hydrocolloids, 2009, 23, 1563–1569.
39 M. Golding, T. J. Wooster, L. Day, M. Xu, L. Lundin, J. Keogh
and P. Clion, So Matter, 2011, 7, 3513–3523.
40 N. Creusot, H. Gruppen, G. A. van Koningsveld, C. G. de
Kruif and A. G. J. Voragen, Int. Dairy J., 2006, 16, 840–849.
41 M. R. Peram, S. M. Loveday, A. Ye and H. Singh, J. Dairy Sci.,
2013, 96, 63–74.
42 D. G. Dalgleish, V. Senaratne and S. Francois, J. Agric. Food
Chem., 1997, 45, 3459–3464.
Soft Matter, 2014, 10, 4173–4183 | 4183