Food Hydrocolloids 112 (2021) 106270

Contents lists available at ScienceDirect

Food Hydrocolloids
journal homepage: http://www.elsevier.com/locate/foodhyd

Foams and air-water interfaces stabilised by mildly purified rapeseed

proteins after defatting
Jack Yang a, b, Iris Faber b, Claire C. Berton-Carabin c, d, Constantinos V. Nikiforidis e,
Erik van der Linden a, b, Leonard M.C. Sagis b, **
a
TiFN, Nieuwe Kanaal 9A, 6709, PA, Wageningen, the Netherlands
b
Laboratory of Physics and Physical Chemistry of Foods, Wageningen University, Bornse Weilanden 9, 6708WG, Wageningen, the Netherlands
c
Laboratory of Food Process Engineering, Wageningen University, Bornse Weilanden 9, 6708WG, Wageningen, the Netherlands
d
INRAE, UR BIA, F-44316, Nantes, France
e
Laboratory of Biobased Chemistry and Technology, Wageningen University, Bornse Weilanden 9, 6708WG, Wageningen, the Netherlands

A R T I C L E I N F O A B S T R A C T

Keywords: Rapeseed protein isolate has promising functional properties (e.g. emulsifying and foaming), but is often
Rapeseed proteins extracted with intensive purification steps. This requires a considerable use of resources and damages protein
Oleosomes functionality regarding, for instance, foam stabilization. We studied the interfacial and foaming properties of a
Surface rheology
mildly obtained rapeseed protein concentrate that contained oleosomes, and of its derived defatted rapeseed
Lissajous plots
Atomic force microscopy
protein concentrate after solvent-based defatting. The air-water interfaces were deformed with large amplitude
Foam dilatational and shear deformations, which were analysed with Lissajous plots. At low bulk concentrations
(0.01% w/w), the rapeseed protein-stabilised interfaces behaved as viscoelastic solids. The interfacial films
became weaker and more stretchable at higher concentrations, suggesting that more non-protein components
interfere with the intermolecular interactions between the adsorbed proteins at higher bulk concentrations. We
confirmed the presence of such non-protein components at the interface by analysing Langmuir-Blodgett films
with atomic force microscopy. The stability and air bubble size of foams prepared with either rapeseed protein
concentrate or defatted rapeseed protein concentrate were similar. Mild purification of rapeseed resulted in a
protein concentrate containing lipids in their native oleosome form, which have a minor destabilizing effect on
foams. We conclude that mild purification is a suitable method to obtain sustainably produced protein con­
centrates with promising foaming properties.

1. Introduction studies dealing with rapeseed protein functionality involve extensive

purification steps, which include defatting with organic solvents, alka­
Rapeseed (Brassica napus) is cultivated worldwide for canola oil line extraction of proteins, isoelectric point precipitation, and some­
extraction, which yields a protein-rich side-product with 35–40% (w/w) times further purification using preparative chromatography. This
protein (Wanasundara, 2011). Currently, the side-product is utilised as results in materials with high protein purity (>80%), sometimes further
animal feed, while it can have a larger added value to human nutrition separated into albumins and globulins (Bérot, Compoint, Larré, Malabat,
(Lamminen, Halmemies-Beauchet-Filleau, Kokkonen, Vanhatalo, & & Guéguen, 2005; Malabat, nchez-Vioque, Rabiller, & Gu guen, 2001).
Jaakkola, 2019). Rapeseed proteins are a potential food ingredient, as: Production of such fractions requires extensive use of energy, water, and
(i) the proteins are of high nutritional quality, with a balanced amino chemicals, and on top of this, the additional processing might cause loss
acid composition (Aider & Barbana, 2011); and (ii) they have promising in functionality. For example, defatted oilseed protein mixtures
functional properties in forming gels, emulsions and foams (Cheung, (43–49% w/w protein) have shown better emulsion and foam stability
Wanasundara, & Nickerson, 2014a, 2014b; Krause & Schwenke, 2001; compared to the purified isolates (70–84% w/w protein) (Aluko &
Larré et al., 2006; Tan, Mailer, Blanchard, & Agboola, 2014; Tan, Mailer, McIntosh, 2004; Aluko, McIntosh, & Katepa-Mupondwa, 2005). Addi­
Blanchard, Agboola, & Day, 2014; Wu & Muir, 2008). The majority of tional processing can change the protein conformation and

** Corresponding author.
E-mail addresses: [email protected] (J. Yang), [email protected] (L.M.C. Sagis).

https://doi.org/10.1016/j.foodhyd.2020.106270
Received 20 June 2020; Received in revised form 19 August 2020; Accepted 19 August 2020
Available online 22 August 2020
0268-005X/© 2020 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
J. Yang et al. Food Hydrocolloids 112 (2021) 106270

hydrophobicity, and affect their functionality in interface and emulsion and other solutes, and a pellet with solids. The subnatant was separated
stabilization (Geerts, Nikiforidis, van der Goot, & van der Padt, 2017; and diafiltrated over a 5 kDa membrane to remove salts and free phe­
Jiang, Chen, & Xiong, 2009; L’Hocine, Boye, & Arcand, 2006). For pea nols. Finally, the material was freeze-dried and stored at − 20 ◦ C.
proteins, it was found that heating upon spray-drying resulted in
chemical and structural changes, and decreased emulsifying properties 2.2.2. Defatting of mildly purified rapeseed protein concentrate
(Estrada et al., 2018; Geerts et al., 2017). Shifts in pH for soy protein The mRPC was defatted four times by dispersing the concentrate in
aqueous dispersions also affected the protein hydrophobicity and hexane with a 1:10 (w/v) ratio, followed by 2 h stirring at room tem­
structure, which resulted in decreased protein solubility and emulsifying perature. The hexane was filtrated using filtration paper and the final
properties (Jiang et al., 2009). With the above in mind, the use of plant retentate was dried overnight under a continuous flow of nitrogen.
protein concentrates obtained via mild processing has gained increasing
interest (Geerts et al., 2017; Karefyllakis, Octaviana, van der Goot, & 2.2.3. Preparation of the soluble fraction
Nikiforidis, 2019; Ntone, Bitter, & Nikiforidis, 2020; Pelgrom, Boom, & The mRPC and defatted mRPC were further processed by dissolving
Schutyser, 2014). in a 20 mM sodium phosphate buffer (20 mM, pH 7.0) in a solid-to-water
Mildly purified plant protein concentrates require fewer purification weight ratio of 1:10. The pH was adjusted to 7.0 and the sample was
steps and fewer resources compared to extensive purification, but stirred overnight at 4 ◦ C. Subsequently, the samples were centrifuged
contain substantial amounts of non-proteinaceous components, such as twice at 16,000×g for 30 min at room temperature to obtain the soluble
lipids, phenols, and carbohydrates. One concern is the possible detri­ fraction (supernatant). The supernatant was freeze-dried for further
mental effect of such components on the functional properties of the analysis. The soluble fraction of mRPC is called rapeseed protein
concentrate, such as interfacial, emulsifying, and foaming properties. concentrate (RPC), and the soluble fraction of defatted mRPC is called
This has not been studied extensively yet. In a yellow pea concentrate defatted rapeseed protein concentrate (DRPC).
produced by aqueous dispersion and centrifugation only, increased
emulsion stability was found comparable to commercially available 2.2.4. Preparation of protein solutions
protein isolates (Geerts et al., 2017). In another example, sunflower All RPC/DRPC solutions in this work were prepared by dissolving
seeds, which have been cold-pressed to obtain a sunflower cake, con­ RPC/DRPC in a sodium phosphate buffer (20 mM, pH 7.0), unless stated
taining proteins, fibres, and residual oil in the form of oleosomes, have otherwise. The samples were stirred at room temperature for 4 h before
successfully been used to prepare emulsions (Karefyllakis, Octaviana, further studies.
et al., 2019). Nonetheless, little is known about the effect of the
non-proteinaceous components in foams. The presence of lipid struc­ 2.3. Composition analysis
tures and fibres may have a limited effect on emulsifying properties, but
this may differ immensely in foams, where the presence of lipids can be Protein content was determined by measuring the nitrogen content
disastrous for foam stability. For example, fatty acids and oil/fat drop­ of triplicates using a Flash EA 1112 Series Dumas (Interscience, The
lets reduce the surface elasticity of protein-stabilised interfaces and Netherlands). A nitrogen conversion factor of 5.7 was used to calculate
thereby increase the chance of bubble coalescence (Denkov, 2004; the protein content (Fetzer, Herfellner, & Eisner, 2019). The oil content
Denkov & Marinova, 2006, pp. 383–444; Wilde et al., 2003). of the samples was studied by performing Soxhlet extraction with the
In this study, we produced a rapeseed protein concentrate by a mild solvent petroleum ether for 6 h and the oil content was determined by
process involving twin-screw pressing and centrifugation, which only weighing the starting material and the oil in the collection flasks. All
partially removes lipids and insoluble materials. The obtained protein samples were prepared in independent duplicates.
concentrate thus had a considerable fraction of lipids (8.3% w/w). We
examined the colloidal state of the lipids and their effect on foaming 2.4. Protein composition by SDS-PAGE
properties by completely defatting the concentrate to obtain a defatted
rapeseed protein concentrate. Both samples were analysed for their A solution of 0.1% protein (w/w) in water was prepared and 45 μL
physicochemical, interfacial, and foaming properties to evaluate the were mixed with 6 μL of 500 mM DTT and 7 μL of NuPAGE LDS sample
necessity of lipids removal in such mildly obtained concentrates. buffer. These samples were heated 10 min for 70 ◦ C and loaded on a
4–12% (w/w) BisTris gel, next to a molecular weight marker in the range
2. Experimental section of 2.5–200 kDa. The electrophoresis was performed for 30 min at 200 V.
Finally, the proteins on the gel were stained with SimplyBlue Safestain
2.1. Materials and the gels were scanned using a gel scanner.

Untreated Alizze rapeseeds obtained from a seed producer were used 2.5. Determination of the zeta-potential and the particle size
as starting material. All chemicals (Sigma-Aldrich, USA) and the mate­
rials for SDS-PAGE (Invitrogen Novex, ThermoFisher Scientific, USA) The zeta-potential and the particle size of the proteins in solution
were used as received. Ultrapure water (MilliQ Purelab Ultra, Germany) (filtered over 0.45 μm syringe filter) were determined using dynamic
was used for all experiments. light scattering in a Zetasizer Nano ZS (Malvern Instruments, UK).
Refractive indices used were 1.45 for the proteins and 1.33 for the
2.2. Sample preparation continuous phase. All measurements were performed in triplicates at
20 ◦ C.
2.2.1. Production of mildly purified rapeseed protein concentrate
Mildly purified rapeseed protein concentrate (mRPC) was produced 2.6. Alteration of tertiary structure of proteins
using a method described by Ntone et al. (Ntone et al., 2020). In short,
rapeseed kernels were dehulled and soaked at pH 9.0 for 4 h in a Possible alterations of the tertiary structure of the proteins were
rapeseed-to-water weight ratio of 1:8. Subsequently, the mixture was probed using intrinsic fluorescence of the tryptophan residues. Protein
blended for 2 min at max speed in a Vita-Prep blender (Vitamix, USA), solutions of 0.01% protein (w/w) were analysed on a LS 50 B lumines­
and solids and supernatant were separated using a twin-screw press. The cence spectrometer (PerkinElmer, USA). The excitation and emission
pH before blending and after screw-pressing was adjusted to 9.0. The slits were set at 5 nm. The excitation wavelength was 295 nm and the
supernatant was centrifuged at 10,000×g for 30 min (4 ◦ C), which emission spectra were recorded from 300 to 450 nm with steps of 0.5 nm
resulted in a top cream layer, a subnatant containing soluble proteins at a scan speed of 120 nm/min. Each sample was measured in duplicate.

2
J. Yang et al. Food Hydrocolloids 112 (2021) 106270

2.7. Determination of protein thermal stability by DSC 2.12. Determination of interfacial layer thickness

The protein thermal stability was studied by differential scanning The thickness of the air-water interfacial film formed by RPC and
calorimetry using a Q100 DSC (TA Instruments, USA). Aliquots of about DRPC was studied using an imaging nulling ellipsometer EP4 (Accurion,
50 μL of 10% protein (w/w) solution were weighed in a stainless steel Germany). A 783 cm2 Langmuir film balance (Langmuir-Blodgett
high volume pan. An empty stainless steel pan was used as a reference trough, KSV NIMA/Biolin Scientific Oy, Finland) filled with a sodium
and nitrogen was used as carrier gas during the measurements. Samples phosphate buffer (20 mM, pH 7.0) was positioned under the ellips­
were equilibrated at 20 ◦ C for 5 min, followed by a heating step to 140 ◦ C ometer. Protein solution (200 μL of 0.12% (w/w) protein) was spread on
at a rate of 2 ◦ C/min, and finally cooled down to 20 ◦ C with a rate of the surface using a syringe, and the protein layer was equilibrated for 30
10 ◦ C/min. Each sample was prepared in duplicate and each of these min. Afterwards, it was compressed with barriers moving at 5 mm/min
duplicates were measured in duplicate. to target surface pressures, as measured by a Wilhelmy plate (platinum,
perimeter 20 mm, height 10 mm). At these surface pressures, the
2.8. Confocal laser scanning microscopy (CLSM) interfacial layers were measured over a wavelength range from 499.8 to
793.8 nm of two zones at an angle of incidence of 50◦ . The ellipsometric
The RPC and DRPC were dissolved at 15% protein content (w/w) in parameters δ and ψ were determined for the air-water interface with and
water for 1 day at 4 ◦ C. The samples were analysed by confocal laser without the protein layer. All measurements were performed at room
scanning microscopy (CLSM) using a Leica TSC SP8x confocal laser temperature, and at least in duplicate. The output was analysed using
microscope (Leica Microsystems Inc., Germany), which was fitted with a the software from the supplier (EP4Model v.3.6.1.). The model was
white laser and HyD detector. Lipids and proteins were fluorescently created by combining: the ambient medium air, the protein layer, and
labelled by adding 4 μL 0.1% (w/v) Nile Red and 7 μL 0.1% (w/v) Fast the substrate buffer. The parameters of the protein layer in the model
Green per mL of RPC sample. The mixtures were placed on microscope were fitted using a Cauchy model (Eq. (1)):
slides and examined with a 63x magnification and a water immersion
B C
lens (refractive index = 1.20). Nile Red and Fast Green were excited n(λ) = A + + 4 (1)
λ2 λ
using a wavelength of 488 and 633 nm, respectively. Images were
analysed using the Leica Application Suite X software. where n is the refractive index, λ is the wavelength of the polarised light,
and A, B, and C are fitting parameters.
2.9. Determination of surface shear properties
2.13. Preparation of Langmuir-Blodgett films
Surface shear properties were determined using a magnetic air
bearing stress-controlled rheometer (AR-G2, TA Instruments, USA) with
Langmuir-Blodgett (LB) films of protein-stabilised interfaces were
a double-wall-ring (DWR) geometry. The diamond-edged ring was made
prepared with a 243 cm2 Langmuir film balance (Langmuir-Blodgett
of iron and platinum and was positioned at the air-water interface of a
Trough, KSV NIMA/Biolin Scientific Oy, Finland). The balance was filled
protein sample in a double wall trough. A vapour cap was installed to
with sodium phosphate buffer (20 mM, pH 7.0). Protein solution (200
avoid evaporation, and the sample was pre-sheared for 5 min at a shear
μL, 0.02% (w/w) protein) was spread on the surface using a syringe and
rate of 10/s. Afterwards, the interface was equilibrated for 10,800 s,
the protein layer was equilibrated for 30 min. Subsequently, the layer
before frequency sweeps were performed with frequencies varying from
was compressed with barriers moving at 5 mm/min to reach the target
0.01 to 10 Hz at a constant amplitude of 1%. Strain sweeps were per­
surface pressure, as monitored with a Wilhelmy plate (platinum,
formed with strains varying from 0.01% to 100% at a constant frequency
perimeter 20 mm, height 10 mm). The films were deposited on a freshly
of 0.1 Hz. All experiments were performed at least in triplicate at 20 ◦ C.
cleaved mica sheet (Highest Grade V1 Mica, Ted Pella, USA) at 1 mm/
The Boussinesq number, which is the ratio between surface and bulk
min withdraw speed. The films were produced in duplicate and dried in
stress, was >1 for all systems studied in our work. This implies that the
a desiccator at room temperature.
contribution of the subphase to the stress signal is negligible.

2.10. Determination of surface tension and surface dilatational properties 2.14. Determination of the interfacial structure by AFM

The air-water surface dilatational properties were determined with a The interfacial microstructure was studied using an atomic force
drop tensiometer PAT-1M (Sinterface Technologies, Germany). A microscope (AFM, MultiMode 8-HR, Bruker, USA). The AFM images of
hanging drop of protein solution was formed at the tip of the needle, the LB films were recorded in tapping mode using Scanasyst-air model
with a surface area of 20 mm2. The surface tension of the drop was non-conductive pyramidal silicon nitride probes (Bruker, USA) with a
calculated by fitting the drop contour with the Young-Laplace equation. normal spring constant of 0.40 N/m. A lateral scan frequency of 0.977
The surface tension was monitored for 10,800 s, which was followed by Hz was employed for all topographical images. The lateral resolution
oscillatory dilatational deformations of the interface. The deformation was 512 × 512 pixels2 in a scan area of 2 × 2 μm2. At least two locations
amplitude was varied between 3 and 30% at a fixed frequency of 0.02 Hz on the film were scanned to ensure good representativeness. Images
in so-called amplitude sweeps. Five sinusoidal oscillations were per­ were analysed using Nanoscope Analysis 1.5 software (Bruker, USA).
formed at each amplitude. Step dilatation of the interface was done by a
sudden extension or compression (step time of 2 s) of the area by 10%. 2.15. Determination of foam properties
All measurements were performed at least in triplicate at 20 ◦ C.
The foams were produced using a Foamscan foaming device (Teclis
2.11. Rheology data analysis IT-Concept, France) by sparging nitrogen through a metal frit (27 μm
pore size, 100 μm distance between centres of pores, square lattice). A
The oscillatory sweep results (in dilatation) were studied by plotting total volume of 40 mL was sparged in a glass cylinder with a diameter of
Lissajous curves of the surface pressure (Π = γ-γ 0) versus the deforma­ 60 mm at a gas flow rate of 400 mL/min to a maximum foam volume of
tion ((A-A0)/A0). Here, γ and A are the surface tension and area of the 400 cm3. The foam decay was monitored by a camera and Foamscan
deformed interface, γ 0, and A0 are the surface tension and area of the software. From this, the foam half-life time was measured, which is the
non-deformed interface. The middle three oscillations of each amplitude time at which 50% of the initial foam volume had collapsed. Images of
cycle were processed into plots. the bubbles were taken with a camera and analysed using a custom

3
J. Yang et al. Food Hydrocolloids 112 (2021) 106270

Matlab script that ran a DIPlip and DIPimage image analysis software.
From this, an average bubble size was obtained.

3. Results and discussion

3.1. Composition

Mild purification resulted in a rapeseed protein concentrate (RPC)

with a protein content of 71.7% (w/w) and a lipid content of 8.3% (w/
w) (Table 1). These lipids were likely the natural oil storage organelles of
plant seeds, namely the oleosomes (OS), also known as oil bodies. OS
store lipids, mainly triglycerides, surrounded by a membrane consisting
of phospholipids and specific proteins called oleosins. These OS exist in
sizes ranging from 0.1 to 3 μm depending on the source of seed and Fig. 1. CLSM images of rapeseed protein concentrate (RPC) and defatted
environmental factors (Nikiforidis et al., 2013). The presence of OS was rapeseed protein concentrate (DRPC). Red indicates the labelled lipids and
green indicates the labelled proteins. (For interpretation of the references to
confirmed with CLSM imaging, which revealed small spherical lipid
colour in this figure legend, the reader is referred to the Web version of
structures (Fig. 1, red colour). The OS in the images were relatively large
this article.)
(a few microns), which is attributed to coalescence during the
freeze-drying step (data not shown). Defatting of RPC was effective, as it
resulted in a defatted rapeseed protein concentrate (DRPC) with a pro­
tein content of 81.8%, and no lipids were found in the lipid content
analysis. Despite the defatting step, minor residual fluorescing material
(red colour) was observed in the CLSM images (Fig. 1), which could be
membrane polar lipids (phospholipids) or protein hydrophobic domains
(Karefyllakis, Van Der Goot, & Nikiforidis, 2019). SDS-PAGE (Fig. 2)
showed a comparable protein composition for RPC and DRPC with the
presence of the two major rapeseed proteins: the water-soluble albumin,
called napin, and the salt-soluble globulin, called cruciferin.

3.2. Protein structure

Intense processing may alter the protein structure at different scales,

and was previously demonstrated for rapeseed proteins after treatments
with organic solvents (Das Purkayastha et al., 2014). We will use the
same techniques here to evaluate whether changes occurred in the
protein structure. First, intrinsic fluorescence was analysed to evaluate
the tertiary protein structure (graph S1 in SI). The emission spectra were
similar for both samples, which indicates no alteration of the chemical
environment of the fluorescent amino acids, mainly tryptophan. Addi­
tionally, proteins were analysed by DSC to probe protein denaturation Fig. 2. SDS-PAGE profile under reducing conditions containing M (Marker; the
temperatures and enthalpy (Table 2). The thermogram showed two corresponding molecular weight are indicated on the left), 1. Rapeseed protein
concentrate, and 2. Defatted rapeseed protein concentrate (DRPC).
peaks at 89.3–90.0 and 106.7–107.3 ◦ C, which can be attributed to the
denaturation temperature of cruciferin and napin, respectively (Wu &
Muir, 2008). The comparable denaturation enthalpies and fluorescence
Table 2
spectra for both protein fractions suggest that the protein structure is Protein denaturation temperature and enthalpy of rapeseed protein concentrate
retained under the defatting procedure. The combination of these (RPC) and defatted rapeseed protein concentrate (DRPC). The average and
measurements allows us to conclude that the protein structure is overall standard error were obtained from four replicates.
not affected by the additional defatting step.
Denaturation temperature (◦ C) Enthalpy (J/g protein)

Cruciferin Napin Total Cruciferin Napin

3.3. Adsorption behaviour
RPC 90.0 ± 0.7 107.3 ± 0.5 14.6 ± 0.1 6.9 ± 0.1 7.8 ± 0.2
DRPC 89.3 ± 0.0 106.7 ± 0.9 14.1 ± 0.0 6.1 ± 0.1 8.0 ± 0.1
The adsorption behaviour of the RPC and DRPC at an air-water
interface was studied (Fig. 3), for which the protein bulk concentra­
tions were varied between 0.01 and 1% (w/w). The surface activity was to DRPC. More specifically, at bulk concentrations of 0.01 and 1% (w/
concentration-dependent, i.e., higher bulk concentrations resulted in w), the surface pressure for RPC was between 2 and 3 mN/m higher than
higher surface pressures even after a substantial adsorption time of 3 h. for DRPC. We observed a larger difference at 0.1% (w/w), where RPC
At all concentrations, the RPC gave higher surface pressures compared had a surface pressure of up to 7 mN/m higher compared to DRPC. The
fact that the presence of OS in the protein concentrate caused a higher
Table 1 surface pressure compared to the defatted protein concentrate suggests
Protein and lipid content in rapeseed protein concentrate (RPC) and defatted that the OS are surface-active. This was also observed in other types of
rapeseed protein concentrate (DRPC). The averages and standard deviation were OS. Soy OS can lower the surface tension by adsorbing and unfolding at
obtained from two replicates. the air-water interface, forming a mixed phospholipid- and oleosin-
RPC DRPC stabilised layer (Waschatko, Junghans, & Vilgis, 2012; Waschatko,
Protein 71.7% ± 1.8 81.8% ± 3.5 Schiedt, Vilgis, & Junghans, 2012).
Oil 8.3% ± 0.2 0.0% In all protein systems at all concentrations, we observed a relatively

4
J. Yang et al. Food Hydrocolloids 112 (2021) 106270

the rapeseed protein-stabilised interfaces were composed of disordered

solid structures, very similar to whey protein-stabilised interfaces (Yang
et al., 2020).
In the frequency sweeps, the RPC- and DRPC-stabilised interfaces
showed comparable moduli at 0.01% (w/w). An increase in bulk con­
centration to 1% (w/w) gave lower moduli for both samples, of which
the decrease was more pronounced for RPC than for DRPC. This may
seem surprising, as in pure protein systems, such an increase in bulk
concentration results typically in similar or higher moduli, which we
showed for interfaces stabilised by whey protein isolate with a bulk
concentration of 0.01 or 2% (w/w) (see graph S2 in SI). For RPC, a
higher protein concentration also resulted in a higher concentration of
non-proteinaceous material, such as OS. A higher concentration of the
latter would result in more competition with the proteins for adsorption
at the interface, and adsorbed OS could reduce the connectivity between
adsorbed proteins, thereby reducing the surface shear moduli.
We also observed a modulus decrease for the DRPC-stabilised
Fig. 3. Surface pressure as a function of time at an air-water interfaces in the
interface. Oil extraction using hexane removed the oil inside the OS,
presence of rapeseed protein concentrate (RPC, black) and defatted rapeseed
but some membrane material most likely remained in the DRPC. For a
protein concentrate (DRPC, grey). The material concentration (w/w) in the
aqueous phase was 0.01% (dotted line), 0.1% (dashed line), or 1% (solid line). milk fat membrane, it is known that fragments of the membrane are
For clarity one representative isotherm is shown for each sample, but compa­ surface-active and can compete with the proteins for adsorption at the
rable isotherms were obtained on at least three replicates. interface (Chen & Sagis, 2019; Karefyllakis, Van Der Goot, & Nikiforidis,
2019; Waninge et al., 2005). Therefore, the decrease in moduli for the
fast increase in surface pressure until roughly 1000 s, which is often DRPC may have been caused by the adsorption of membrane material of
described as a regime where proteins adsorb at the interface (Beverung, the OS, thereby, again decreasing the connectivity between the adsorbed
Radke, & Blanch, 1999). Afterwards, the surface pressure increase was proteins.
slower, which is often related to rearrangements (e.g., conformational
changes or clustering) of the adsorbed proteins (Yang, Thielen, 3.4.2. Strain sweeps
Berton-Carabin, van der Linden, & Sagis, 2020). The surface rheology In strain sweeps (Fig. 4B), the Gi’ and Gi” of both protein-based
tests described in the next sections should ideally be performed in interfacial films at 0.01% (w/w) were also relatively close. At this
equilibrium conditions, to limit the influence of the aging processes. As concentration, the Gi’ in the linear viscoelastic (LVE) regime were 0.032
we do not reach equilibrium conditions, even after 50,000 s, we decided and 0.041 Pa m for RPC- and DRPC-stabilised interfaces, respectively.
to let all interfacial films age for 3 h before subjecting them to Increasing the protein bulk concentration to 1% (w/w) resulted in a
deformations. lower Gi’ in the LVE regime, which were found to be 0.0015 and 0.0059
Pa m for RPC- and DRPC-stabilised interfaces, respectively. All four
curves in Fig. 4B showed a LVE regime, followed by a nonlinear visco­
3.4. Surface shear rheology elastic (NLVE) regime, where the materials soften at higher de­
formations. In the whole strain range from 0.1 to 100%, the Gi’ was
3.4.1. Frequency sweeps higher than Gi”, suggesting solid-like behaviour, which was also
The rapeseed protein-stabilised interfaces were subjected to oscilla­ demonstrated by a tanδ (Gi”/Gi’) <1 in the LVE regime (0.29–0.37 for all
tory shear deformations in frequency and strain sweeps (Fig. 4A), at two interfaces).
different protein bulk concentrations, of 0.01% and 1% (w/w). All in­ The moduli shown in Fig. 4 are calculated from the intensity and
terfaces showed a slight frequency dependence in the storage (Gi’) and phase shift of the first harmonic of the Fourier transform of the stress
loss moduli (Gi’‘). In all cases, Gi’ scaled as Gi’ ~ ωn and this resulted in signal, while contributions from higher harmonics are neglected.
n-values of between 0.19 ± 0.02 and 0.23 ± 0.02. This suggests a weak Higher- order harmonics do exist in the NLVE regime and are the results
frequency dependency and power-law behaviour, which was also found of changes in the interfacial microstructure upon higher deformation. To
for gels and soft glassy materials (Jaishankar & McKinley, 2012). A include these non-linearities into the analysis, we have plotted the tor­
combination of power-law behaviour and Gi’ > Gi” modulus implies that que against deformation in Lissajous plots (Sagis, Humblet-Hua, & van

Fig. 4. The storage (closed symbols) and loss moduli (open symbols) as a function of frequency (A) and strain (B) of air-water interfacial films stabilised by rapeseed
protein concentrate (RPC, squares) or defatted rapeseed protein concentrate (DRPC, triangles). The different bulk protein concentrations (w/w) are depicted in the
legend. The graphs represent averages obtained from at least three replicates, and the standard deviation for each sample was below 5%.

5
J. Yang et al. Food Hydrocolloids 112 (2021) 106270

Kempen, 2014). increasingly disrupted at higher deformations, which suggests the for­
mation of a solid viscoelastic layer, as observed earlier in the surface
3.4.3. Lissajous plots shear rheology tests. Just as for these tests, the moduli for the RPC-based
The Lissajous plots (Fig. 5) showed a narrow ellipsoidal shape at layers show a decrease with increasing bulk concentration. The Ed’ of a
strains in the LVE, which suggested viscoelastic behaviour where the RPC-stabilised interface with 1% (w/w) bulk concentration was around
elastic component dominated. The plots became wider at higher defor­ 29 mN/m, and independent of the applied amplitude, which suggests a
mation amplitudes, which indicates an increase of the viscous contri­ less stiff and more stretchable layer compared to an interface stabilised
bution in the film response. Further into the NLVE regime, the plots by the same material at lower bulk concentrations. The same behaviour
started to exhibit a rhomboidal shape, which was clearly present for was observed for the DRPC-stabilised interfaces (Fig. 6B). At a protein
both rapeseed protein-stabilised interfaces formed from 0.01% (w/w) bulk concentration of 1% (w/w), the defatted protein concentrate
bulk concentration, at 100% strain. In the rhomboidal shape, we resulted in a slightly stiffer interface compared to the protein concen­
observed initially a highly elastic response, at the start of the cycle (the trate. The Ed’ of the DRPC exhibited a minor degree of amplitude
lower left corner of the plot). This was evident from the steepness of the dependence at 1% (w/w) bulk concentration, which was non-existent in
curve in this regime, which indicated high values for the tangential the case of RPC. We also studied the non-linearities in these responses by
shear modulus (the local slope of the plot). This was followed by a plotting Lissajous curves.
gradual but significant change in the slope. This is typical for (partial)
intra-cycle yielding, where a considerable part of the elasticity is lost, 3.5.2. Lissajous plots
and the response becomes dominated by the viscous contribution. The The Lissajous plots (Fig. 7) of RPC- and DRPC-stabilised interfaces at
critical stress where this happens is the yield stress of the interfacial a bulk concentration of 0.01% (w/w) had a symmetric shape at 5%
microstructure. It is important to note that intra-cycle yielding is not the deformation. The plots were becoming less symmetric at higher de­
same as macroscopic fracture of the interface. When the deformation formations, which was most pronounced at 30% deformation. In those
continued in the opposite direction, we again observed a highly elastic plots, we observed a relatively steep surface pressure increase from the
response, followed by intra-cycle yielding, and a predominantly viscous left-bottom corner (start of extension of interface), which represents an
response of the interface. This suggests that the microstructure of the elastic response, followed by a gradual softening of the interfacial
interface (partially) recovers at both ends of the deformation cycle. structure. In this part of the cycle, the elastic component of the response
The Lissajous plots at 100% strain for 0.01% (w/w) bulk protein diminished and the viscous contribution increased. The surface pressure
concentration were comparable for RPC and DRPC, and indicated an reached higher values in maximum compression than in maximum
interface that behaved like a viscoelastic solid. Increasing the bulk extension, which is an indication for intra-cycle strain hardening. The
protein concentration to 1% (w/w) resulted in narrower Lissajous plots strain hardening upon compression (at high deformations) most likely
that were more tilted towards the horizontal axis, which indicates a resulted from jamming of densely clustered protein regions. These in­
more dominant elastic contribution in the response, and a less stiff terfaces show the rheological behaviour of viscoelastic solids, and their
interface. Even at 100% deformation, the plots are still nearly ellip­ behaviour is comparable to that of whey protein-stabilised interfaces
soidal, which indicates that response is still close to linear, and the (Yang et al., 2020).
interfacial film has a considerably larger maximum linear strain Increasing the protein bulk concentration to 1% (w/w) resulted in
compared to 0.01% (w/w) (Fig. 4B). This concentration dependency was narrower Lissajous plots, and also in plots more tilted towards the hor­
also more pronounced in the RPC-stabilised interface compared to the izontal axis. These findings, again, suggest the formation of a less stiff
DRPC-stabilised one. The presence of non-protein components (OS or OS and much more stretchable interfacial layer, where the elastic response
membranes) resulted in the formation of a less stiff and more stretchable is more dominant over the whole range of deformations, compared to
interface. the interface formed at lower bulk concentrations. The dilatational
rheology results thus confirm the shear rheology results, for which we
proposed that non-protein components, such as OS and OS-membranes,
3.5. Surface oscillatory dilatational rheology could hamper the formation of a viscoelastic protein network at the air-
water interface.
3.5.1. Amplitude sweeps Understanding the air-water interface stabilising properties of the
The same air-water interfacial films stabilised with rapeseed protein proteins can contribute to a better insight on the foaming-stabilising
ingredients were studied by performing dilatational deformations in a properties. Interfacial rheology can be used to evaluate the type of
drop tensiometer. In amplitude sweeps, the RPC-stabilised interface at interfacial layer formed, and its strength. It is beneficial to study both
0.01% (w/w) bulk concentration showed an amplitude-dependent dilatational and shear rheology, as both deformation could occur in a
modulus (Fig. 6A), as the Ed’ decreased from 70 to 34 mN/m when foam. For instance, the extension of the air bubble upon
increasing the amplitude from 3 to 30%. The interfacial layer was

Fig. 5. Lissajous plots of torque as a function of

0.5% 6% 18% 32% 100%
deformation obtained during strain sweeps of air-water
interfacial films stabilised by rapeseed protein concen­
trate (RPC) and defatted rapeseed protein concentrate
(DRPC). The vertical axes were normalised by the
RPC

maximum torque. The dark solid lines and light dashed

0.01%
lines indicate 0.01 and 1% (w/w) bulk protein con­
1% centration, respectively. For clarity, one representative
distribution is shown for each sample, but comparable
distributions were obtained on at least three replicates.
DRPC

0.01%
1%

6
J. Yang et al. Food Hydrocolloids 112 (2021) 106270

Fig. 6. Surface elastic (closed symbols) and loss (open symbols) dilatational moduli as a function of deformation amplitude of air-water interfacial films stabilised by
rapeseed protein concentrate (RPC, A) or defatted rapeseed protein concentrate (DRPC, B). The different bulk protein concentrations tested (%, w/w) are given in the
legend. The graphs represent averages and standard deviations from at least three replicates.

Fig. 7. Lissajous plots of surface pressure as a func­

tion of the applied dilatational deformation, obtained
during amplitude sweeps of air-water interfacial films
stabilised by rapeseed protein concentrate (RPC) or
defatted rapeseed protein concentrate (DRPC). The
plots were normalised to the same maximum surface
pressure value. The dark lines and light lines indicate
0.01 and 1% (w/w) bulk protein concentrations,
respectively. For clarity one representative distribu­
tion is shown for each sample, but comparable dis­
tributions were obtained on at least three replicates.

disproportionation is a dilatational deformation of the surface, whereas

Table 3
the collision of two bubbles and a possible rupture could have a stronger
β and τ1 for interfacial films stabilised by rapeseed protein concentrate (RPC) or
shear component. Also in-plane interactions are analysed to a larger
defatted rapeseed protein concentrate (DRPC) with various bulk concentrations.
extent in shear than in dilatational deformations. We should keep in The averages and standard deviation were obtained from at least three
mind that the interfacial properties alone do not dictate the foaming replicates.
properties, as for instance the bulk properties, such as viscosity and
Extension Compression
presence/size of particles, can also largely influence foam formation and
stability. β τ1 β τ1
RPC 0.01% 0.61 ± 0.03 26.2 ± 4.4 0.58 ± 0.03 24.1 ± 2.3
3.6. Step-dilatation 0.1% 0.56 ± 0.04 28.0 ± 4.4 0.60 ± 0.05 22.3 ± 3.8
1% 0.59 ± 0.04 29.4 ± 5.2 0.67 ± 0.04 23.0 ± 1.1
DRPC 0.01% 0.59 ± 0.04 31.7 ± 5.2 0.63 ± 0.02 22.4 ± 4.1
The interfaces were also subjected to sudden step extensions/com­ 0.1% 0.58 ± 0.06 28.8 ± 4.1 0.60 ± 0.02 23.2 ± 2.7
pressions, and the resulting relaxation response was fitted with a 1% 0.54 ± 0.06 28.3 ± 6.6 0.66 ± 0.06 25.1 ± 4.9
Kohlrausch-William-Watts (KWW) stretch exponential term and a reg­
ular exponential (Equation (2)) (Watts & Davies, 1969). The regular
observed in the Lissajous plots in Fig. 7. β-values between 0.42 and 0.73
exponential is added to decouple the continuous aging of the interface
have been reported for various interfacial layers stabilised by plant,
from the actual relaxation process.
insect, and milk proteins (Cicuta, 2007; Felix et al., 2019; Sagis et al.,
γ(t) = ae− (t/τ1 )β
+ be− t/τ2
+c (2) 2019). Such values were also observed in relaxation processes in 3D
disordered solids (Klafter & Shlesinger, 1986). The relaxation time τ1
In Table 3, we show the relaxation time τ1 and stretch exponent β was between 22.3 and 31.7 s for all rapeseed protein-stabilised in­
determined for the different systems tested. Parameters τ2 and constants terfaces. This suggests comparable relaxation processes for interfaces
a, b, and c can be found in Table S1 in the SI. stabilised by both rapeseed protein samples, at bulk concentrations be­
The KWW is a phenomenological model for describing relaxation tween 0.01 and 1% (w/w).
processes in disordered systems, which has also been applied for protein-
stabilised interfaces (Felix, Yang, Guerrero, & Sagis, 2019; Sagis et al.,
2019). The stretch component β < 1 reveals dynamic heterogeneity, as it 3.7. Ellipsometry
is related to local variations in the relaxation kinetics, and indicates the
presence of a wide range of relaxation times. The β′ s of both RPC- and The thickness of the interfacial layer was studied using ellipsometry
DRPC-stabilised interfaces were 0.54–0.61 in extension and 0.58 to 0.67 in combination with a Langmuir trough, which allowed us to perform
in compression. In general, compression of the interface resulted in a measurements at various surface pressures (Fig. 8). Both rapeseed pro­
higher β than extension of the interface, in line with the asymmetry also tein samples showed a similar interfacial thickness, ranging from 2.1 nm

7
J. Yang et al. Food Hydrocolloids 112 (2021) 106270

at a surface pressure of 13 mN/m to 3.5 nm at 21 mN/m. At the highest long strands that were prominently present at 13 mN/m and slightly
tested surface pressure of 25 mN/m, the DRPC-stabilised interface had a visible at 17 mN/m. Conversely, at the highest compression of 21 mN/
higher thickness (5.1 nm) compared to that formed with RPC (4.3 nm). m, where the films are denser, the strands seem to be barely visible.
This could be related to the stronger interfacial layer formed by DRPC, as These strands have a thickness of <2 nm, and lengths varying from 0.2 to
stronger layers could retain more material at higher compressions, while 2 μm. The strands seem to be close to protein-dense regions, and also
a weaker and more stretchable layer, such as formed with RPC, might related to the clustering of these regions. The presence of such strands
result in loss of interfacial material towards the bulk phase at such high could have reduced the overall connectivity between the proteins at the
compressions. In general, both protein samples gave layers with com­ interface, which would be in line with the lower moduli at higher con­
parable average thicknesses, while we may have expected an increase in centrations that we observed in shear and dilatation deformation ex­
layer thickness for the RPC-stabilised interfaces, as OS were present. Yet, periments. The exact nature of the strands cannot be established based
the possible contribution of OS to the interfacial composition and on AFM images alone, but we can speculate that the strands could be
structure may not be captured with this method, as ellipsometry mea­ polysaccharides that were not removed during centrifugation. Another
sures the average properties over an area of several square microns. explanation for the structures could be membrane fragments of OS, if
Additionally, large OS that were subjected to coalescence during these would unfold at the air-water interface. We also cannot exclude
extraction might also unfold and disintegrate at the air-water interface, the possibility that OS is present in a sublayer beneath the scanned
which already has been demonstrated for soybean and sunflower oleo­ interface, as we only studied the topography of the interfaces using AFM.
somes (Karefyllakis, Van Der Goot, & Nikiforidis, 2019; Waschatko, As we also suggested in previous sections, the OS might disintegrate at
Schiedt, et al., 2012). the interface after adsorption.

3.9. Foams
3.8. Interfacial microstructure
Rapeseed protein-stabilised foams were studied for their average
We also studied the interfacial microstructure by visualization of the bubble diameter, and the foam half-life time, which is the duration for
topography using AFM on Langmuir-Blodgett films (Fig. 9), which were the foam volume to decrease down to half of its original value (Fig. 10).
produced by transferring a protein film at different surface pressures The RPC-based samples did not reach the set initial volume of 500 cm3 at
onto a solid substrate. Surface pressure isotherms obtained upon film a bulk concentration below 0.02% (w/w). Increasing the protein bulk
compression are shown in graph S3 in the SI. For both RPC- and DRPC- concentration resulted in a higher half-life time of the foam, which can
based films at all surface pressures tested, a highly heterogeneous reach up to 361 ± 84 and 288 ± 63 min at a concentration of 1% (w/w)
structure was observed with many thicker regions (white dots). The for RPC- and DRPC-based foams, respectively. Both protein samples
thickness of these regions was between 6 and 8 nm, which is comparable showed comparable foam half-life times over the entire range of protein
to the protein size measured by dynamic light scattering (8.7 nm, graph concentrations between 0.02 and 1% (w/w). The foam stability is
S4 in SI), suggesting that these are protein-dense regions. An increase in closely related to the air bubble size, which decreased with increased
surface pressure (and thus higher compression of the film) resulted in protein concentrations. The smallest bubbles can be achieved at con­
more protein-dense regions on the AFM images. Comparable heteroge­ centrations higher than 0.2% (w/w) and are 0.052 mm in diameter for
neity in the interfacial microstructure has also been observed in many both rapeseed protein samples.
other protein-based interfaces (Berton-Carabin, Genot, Gaillard, Gui­ So, the foam stability increased with a higher protein content, while
bert, & Ropers, 2013; Sah & Kundu, 2017; Yang et al., 2020). These the interfacial properties were affected by the bulk concentration of the
protein-stabilised interfaces exhibit solid-like behaviour and can form different components in RPC or DRPC in the opposite way. A low bulk
highly disordered structures at the interface, which is also found in concentration (0.01% w/w) gave a strong viscoelastic solid-like inter­
three-dimensional gel systems. The dynamic heterogeneity in the step face, while a high concentration (1% w/w) resulted in a less stiff and
dilatation experiments was suggested to be associated with structural more stretchable interface. The mechanical strength of the interfacial
heterogeneity in the interfacial microstructure, which we herein confirm layer is often related to the foam stability, where an increased layer
via the AFM images. strength would result in higher foam stability. A stronger viscoelastic
Another interesting observation related to these AFM images was the layer could potentially slow down the bubble rupture. Yet, we did not
observe such a relationship here for rapeseed protein-based foams. The
increased stability appeared to be more correlated with the smaller
bubble size at higher protein concentrations. Although the moduli of the
interfacial films obtained with bulk concentrations of 0.01% (w/w) were
substantially higher compared to bulk concentrations of 1% (w/w), the
adsorption rate in these samples was considerably slower (Fig. 3), which
can be related to lower foamability and larger bubble sizes. The foams at
higher concentration, because of the faster adsorption, have a smaller
and more homogeneous distributed bubble size, and are therefore more
stable. The higher stretchability of the corresponding interfaces may
also have provided additional protection against coalescence of the
bubbles. The more brittle interfaces formed at low concentration dis­
played yielding at large deformations, possibly making the bubbles more
prone to coalescence.
With regard to foam formation and stability, we also observed very
minor differences between the rapeseed protein concentrate and defat­
ted protein concentrate. The presence of OS in RPC did not appear to
Fig. 8. Thickness of interfacial layer as function of surface pressure for Lang­ negatively influence the foam stability in the tested conditions (pH, ionic
muir films made of rapeseed protein concentrate (RPC, black square) or strength, and lipid content). Conversely, in many studies, protein-
defatted rapeseed protein concentrate (DRPC, grey circle), measured by stabilised foams were destabilised due to the presence of fatty acids
ellipsometry. The graph represents averages and standard deviations from at and oil/fat droplets, as they can decrease surface elasticity, thus
least two replicates. increasing the chance for coalescence of air bubbles (Denkov, 2004;

8
J. Yang et al. Food Hydrocolloids 112 (2021) 106270

Fig. 9. AFM images of Langmuir-Blodgett films made from rapeseed protein concentrate (RPC) and defatted rapeseed protein concentrate (DRPC). The surface
pressure indicates the conditions during film sampling.

Fig. 10. The half-life time (A) and bubble diameter (B) of foams stabilised by rapeseed protein concentrate (RPC) or defatted rapeseed protein concentrate (DRPC).
The graphs represent averages and standard deviations of at least three replicates.

Denkov & Marinova, 2006, pp. 383–444; Maldonado-Valderrama & polar lipids from the concentrate to produce a defatted rapeseed pro­
Patino, 2010; Wilde et al., 2003). This was not observed in our work, as tein concentrate allowed us to study their effects in more detail. The
both RPC and DRPC formed foams with comparable stability. The OS defatted protein concentrate contained only residual oleosome mem­
were not detected in AFM images, which could suggest that the OS are branes, which consists of proteins and polar lipids. The defatted protein
indeed not present at the interface as such. However, the OS could in concentrate led to slightly stronger interfacial films compared to the
principle be attached to the interfacial film as a sublayer, and therefore concentrate; in both cases, the interfacial films became weaker with
invisible for AFM measurements. Based on our data this cannot be increasing bulk concentrations. This suggests that oleosome membranes
excluded, but even if this would happen, it does not affect the stability of influenced the interface stabilization by weakening the interfacial layer.
foams in the tested conditions. On the other hand, both the rapeseed protein concentrate and defatted
rapeseed protein concentrate could form foams with comparable sta­
4. Conclusions bility. The presence of oleosomes in the concentrate (and at the con­
centrations studied in this work) did not influence the foam stability.
We studied the interfacial and foaming properties of a mildly ob­ The fact that non-protein components present after mild purification do
tained rapeseed protein concentrate that contained oleosomes, and of its not negatively influence the macroscopic foam properties, implies
derived defatted protein concentrate after solvent-based defatting. For rapeseed protein mixtures have promising foam stabilising properties.
the concentrate, the air-water interface displayed the behaviour of a The findings presented in this manuscript have important implications,
disordered viscoelastic solid, and a weakening of the interfacial layer in since protein mixtures with high functionality can be produced, while
shear and dilatational rheology was observed for increasing bulk con­ reducing the use of water, chemicals, and energy.
centration. This suggests that the presence of non-protein components
hampered the connectivity of the interfacial layer. In the AFM images, CRediT authorship contribution statement
we observed the presence of strand-like structures that may cause a
weaker interfacial layer at higher concentrations. Removing the non- Jack Yang: Conceptualization, Methodology, Investigation,

9
J. Yang et al. Food Hydrocolloids 112 (2021) 106270

Validation, Visualization, Writing - original draft. Iris Faber: Method­ Denkov, N. D., & Marinova, K. G. (2006). Antifoam effects of solid particles, oil drops and
oil–solid compounds in aqueous foams. In Colloidal particles at liquid interfaces.
ology, Investigation. Claire C. Berton-Carabin: Methodology, Writing -
https://doi.org/10.1017/CBO9780511536670.011.
review & editing. Constantinos V. Nikiforidis: Conceptualization, Estrada, P. D., Berton-Carabin, C. C., Schlangen, M., Haagsma, A., Pierucci, A. P. T. R., &
Methodology, Writing - review & editing. Erik van der Linden: Writing Van Der Goot, A. J. (2018). Protein oxidation in plant protein-based fibrous
- review & editing, Funding acquisition. Leonard M.C. Sagis: Concep­ products: Effects of encapsulated iron and process conditions. Journal of Agricultural
and Food Chemistry, 66(42), 11105–11112. https://doi.org/10.1021/acs.
tualization, Methodology, Supervision, Writing - review & editing. jafc.8b02844.
Felix, M., Yang, J., Guerrero, A., & Sagis, L. M. C. (2019). Effect of cinnamaldehyde on
interfacial rheological properties of proteins adsorbed at O/W interfaces. Food
Declaration of competing interest Hydrocolloids, 97, 105235. https://doi.org/10.1016/j.foodhyd.2019.105235.
Fetzer, A., Herfellner, T., & Eisner, P. (2019). Rapeseed protein concentrates for non-food
applications prepared from pre-pressed and cold-pressed press cake via acidic
The authors have declared that no competing interest exist. This precipitation and ultrafiltration. Industrial Crops and Products, 132, 396–406. https://
manuscript has not been published and is not under consideration for doi.org/10.1016/j.indcrop.2019.02.039.
publication in any other journal. All authors approve this manuscript Geerts, M. E. J., Nikiforidis, C. V., van der Goot, A. J., & van der Padt, A. (2017). Protein
nativity explains emulsifying properties of aqueous extracted protein components
and its submission to Food Hydrocolloids. from yellow pea. Food Structure, 14, 104–111. https://doi.org/10.1016/j.
foostr.2017.09.001.
Acknowledgements Jaishankar, A., & McKinley, G. H. (2012). Power-law rheology in the bulk and at the
interface: Quasi-properties and fractional constitutive equations. Proceedings of the
Royal Society A, 469(2149). https://doi.org/10.1098/rspa.2012.0284,
The authors thank Helene Mocking for her contribution to the pro­ 20120284–20120284.
tein extraction and Eleni Ntone and Simha Lakshminarasimh for their Jiang, J., Chen, J., & Xiong, Y. L. (2009). Structural and emulsifying properties of soy
protein isolate subjected to acid and alkaline pH-shifting processes. Journal of
contribution in the CLSM. The project is organised by and executed Agricultural and Food Chemistry, 57(16), 7576–7583. https://doi.org/10.1021/
under the auspices of TiFN, a public-private partnership on pre­ jf901585n.
competitive research in food and nutrition. The authors have declared Karefyllakis, D., Octaviana, H., van der Goot, A. J., & Nikiforidis, C. V. (2019). The
emulsifying performance of mildly derived mixtures from sunflower seeds. Food
that no competing interests exist in the writing of this publication.
Hydrocolloids, 88, 75–85. https://doi.org/10.1016/j.foodhyd.2018.09.037.
Funding for this research was obtained from Unilever, Groupe Bel, Karefyllakis, D., Van Der Goot, A. J., & Nikiforidis, C. V. (2019). The behaviour of
Danone, PepsiCo, the Netherlands Organisation for Scientific Research sunflower oleosomes at the interfaces. Soft Matter, 15(23), 4639–4646. https://doi.
org/10.1039/c9sm00352e.
(NWO) and the Top-sector Agri&Food (TKI). NWO project number:
Klafter, J., & Shlesinger, M. F. (1986). On the relationship among three theories of
ALWTF. 2016.001. relaxation in disordered. Proceedings of the National Academy of Sciences of the United
States of America, 83(4), 848–851. https://doi.org/10.1073/pnas.83.4.848.
Krause, J. P., & Schwenke, K. D. (2001). Behaviour of a protein isolate from rapeseed
Appendix A. Supplementary data (Brassica napus) and its main protein components - globulin and albumin - at air/
solution and solid interfaces, and in emulsions. Colloids and Surfaces B: Biointerfaces,
Supplementary data to this article can be found online at https://doi. 21(1–3), 29–36. https://doi.org/10.1016/S0927-7765(01)00181-3.
Lamminen, M., Halmemies-Beauchet-Filleau, A., Kokkonen, T., Vanhatalo, A., &
org/10.1016/j.foodhyd.2020.106270.
Jaakkola, S. (2019). The effect of partial substitution of rapeseed meal and faba
beans by Spirulina platensis microalgae on milk production, nitrogen utilization, and
References amino acid metabolism of lactating dairy cows. Journal of Dairy Science, 102(8),
7102–7117. https://doi.org/10.3168/jds.2018-16213.
Larré, C., Mulder, W., Sánchez-Vioque, R., Lazko, J., Bérot, S., Guéguen, J., et al. (2006).
Aider, M., & Barbana, C. (2011). Canola proteins: Composition, extraction, functional
Characterisation and foaming properties of hydrolysates derived from rapeseed
properties, bioactivity, applications as a food ingredient and allergenicity - a
isolate. Colloids and Surfaces B: Biointerfaces, 49(1), 40–48. https://doi.org/10.1016/
practical and critical review. Trends in Food Science & Technology, 22(1), 21–39.
j.colsurfb.2006.02.009.
https://doi.org/10.1016/j.tifs.2010.11.002.
L’Hocine, L., Boye, J. I., & Arcand, Y. (2006). Composition and functional properties of
Aluko, R. E., & McIntosh, T. (2004). Electrophoretic and functional properties of mustard
soy protein isolates prepared using alternative defatting and extraction procedures.
seed meals and protein concentrates. JAOCS, Journal of the American Oil Chemists’
Journal of Food Science, 71(3). https://doi.org/10.1111/j.1365-2621.2006.tb15609.
Society, 81(7), 679–683. https://doi.org/10.1007/s11746-004-961-0.
x.
Aluko, R. E., McIntosh, T., & Katepa-Mupondwa, F. (2005). Comparative study of the
Malabat, C., nchez-Vioque, R. I. S., Rabiller, C., & Gu guen, J. (2001). Emulsifying and
polypeptide profiles and functional properties of Sinapis alba and Brassica juncea
foaming properties of native and chemically modified peptides from the 2S and 12S
seed meals and protein concentrates. Journal of the Science of Food and Agriculture, 85
proteins of rapeseed (Brassica napus L.). Journal of the American Oil Chemists’ Society,
(11), 1931–1937. https://doi.org/10.1002/jsfa.2199.
78(3), 235–242. https://doi.org/10.1007/s11746-001-0251-x.
Bérot, S., Compoint, J. P., Larré, C., Malabat, C., & Guéguen, J. (2005). Large scale
Maldonado-Valderrama, J., & Patino, J. M. R. (2010). Interfacial rheology of protein-
purification of rapeseed proteins (Brassica napus L.). Journal of Chromatography B,
surfactant mixtures. Current Opinion in Colloid & Interface Science, 15(4), 271–282.
818, 35–42. https://doi.org/10.1016/j.jchromb.2004.08.001.
https://doi.org/10.1016/j.cocis.2009.12.004.
Berton-Carabin, C., Genot, C., Gaillard, C., Guibert, D., & Ropers, M. H. (2013). Design of
Nikiforidis, C. V., Ampatzidis, C., Lalou, S., Scholten, E., Karapantsios, T. D., &
interfacial films to control lipid oxidation in oil-in-water emulsions. Food
Kiosseoglou, V. (2013). Purified oleosins at air-water interfaces. Soft Matter, 9(4),
Hydrocolloids, 33(1), 99–105. https://doi.org/10.1016/j.foodhyd.2013.02.021.
1354–1363. https://doi.org/10.1039/c2sm27118d.
Beverung, C. J., Radke, C. J., & Blanch, H. W. (1999). Protein adsorption at the oil/water
Ntone, E., Bitter, J. H., & Nikiforidis, C. V. (2020). Not sequentially but simultaneously:
interface: Characterization of adsorption kinetics by dynamic interfacial tension
Facile extraction of proteins and oleosomes from oilseeds. Food Hydrocolloids, 102,
measurements. Biophysical Chemistry, 81(1), 59–80. https://doi.org/10.1016/S0301-
105598. https://doi.org/10.1016/j.foodhyd.2019.105598.
4622(99)00082-4.
Pelgrom, P. J. M., Boom, R. M., & Schutyser, M. A. I. (2014). Functional analysis of
Chen, M., & Sagis, L. M. C. (2019). The influence of protein/phospholipid ratio on the
mildly refined fractions from yellow pea. Food Hydrocolloids, 44, 12–22. https://doi.
physicochemical and interfacial properties of biomimetic milk fat globules. Food
org/10.1016/j.foodhyd.2014.09.001.
Hydrocolloids, 97, 105179. https://doi.org/10.1016/j.foodhyd.2019.105179.
Sagis, L. M. C., Humblet-Hua, K. N. P., & van Kempen, S. E. H. J. (2014). Nonlinear stress
Cheung, L., Wanasundara, J., & Nickerson, M. T. (2014a). Effect of pH and NaCl on the
deformation behavior of interfaces stabilized by food-based ingredients. Journal of
emulsifying properties of a napin protein isolate. Food Biophysics, 10(1), 30–38.
Physics: Condensed Matter, 26(46), 9. https://doi.org/10.1088/0953-8984/26/46/
https://doi.org/10.1007/s11483-014-9350-7.
464105.
Cheung, L., Wanasundara, J., & Nickerson, M. T. (2014b). The effect of pH and NaCl
Sagis, L. M. C., Liu, B., Li, Y., Essers, J., Yang, J., Moghimikheirabadi, A., et al. (2019).
levels on the physicochemical and emulsifying properties of a cruciferin protein
Dynamic heterogeneity in complex interfaces of soft interface-dominated materials.
isolate. Food Biophysics, 9(2), 105–113. https://doi.org/10.1007/s11483-013-9323-
Scientific Reports, 9(1), 1–12. https://doi.org/10.1038/s41598-019-39761-7.
2.
Sah, B. K., & Kundu, S. (2017). Modification of hysteresis behaviors of protein monolayer
Cicuta, P. (2007). Compression and shear surface rheology in spread layers of β-casein
and the corresponding structures with the variation of protein surface charges.
and β-lactoglobulin. Journal of Colloid and Interface Science, 308(1), 93–99. https://
Colloids and Surfaces B: Biointerfaces, 159, 696–704. https://doi.org/10.1016/j.
doi.org/10.1016/j.jcis.2006.12.056.
colsurfb.2017.08.032.
Das Purkayastha, M., Gogoi, J., Kalita, D., Chattopadhyay, P., Nakhuru, K. S., Goyary, D.,
Tan, S. H., Mailer, R. J., Blanchard, C. L., & Agboola, S. O. (2014a). Emulsifying
et al. (2014). Physicochemical and functional properties of rapeseed protein isolate:
properties of proteins extracted from Australian canola meal. Lebensmittel-
Influence of antinutrient removal with acidified organic solvents from rapeseed
Wissenschaft und -Technologie- Food Science and Technology, 57(1), 376–382. https://
meal. Journal of Agricultural and Food Chemistry, 62(31), 7903–7914. https://doi.
doi.org/10.1016/j.lwt.2013.12.040.
org/10.1021/jf5023803.
Tan, S. H., Mailer, R. J., Blanchard, C. L., Agboola, S. O., & Day, L. (2014b). Gelling
Denkov, N. D. (2004). Mechanisms of foam destruction by oil-based antifoams. Langmuir,
properties of protein fractions and protein isolate extracted from Australian canola
20(22), 9463–9505. https://doi.org/10.1021/la049676o.

10
J. Yang et al. Food Hydrocolloids 112 (2021) 106270

meal. Food Research International, 62(August), 819–828. https://doi.org/10.1016/j. Watts, C., & Davies, E. (1969). Non-symmetrical dielectric relaxation behaviour arising
foodres.2014.04.055. from a simple empirical decay function. Transactions of the Faraday Society, 66(1),
Wanasundara, J. P. D. (2011). Proteins of brassicaceae oilseeds and their potential as a 80–85. https://doi.org/10.1039/TF9706600080.
plant protein source. Critical Reviews in Food Science and Nutrition, 51(7), 635–677. Wilde, P. J., Wilde, P. J., Husband, F. a, Husband, F. a, Cooper, D., Cooper, D., et al.
https://doi.org/10.1080/10408391003749942. (2003). Destabilization of beer foam by lipids: Structural and interfacial effects.
Waninge, R., Walstra, P., Bastiaans, J., Nieuwenhuijse, H., Nylander, T., Paulsson, M., Journal of the American Society of Brewing Chemists, 61, 196–202. https://doi.org/
et al. (2005). Competitive adsorption between β-casein or β-lactoglobulin and model 10.1094/ASBCJ-61-0196.
milk membrane lipids at oil-water interface. Journal of Agricultural and Food Wu, J., & Muir, A. D. (2008). Comparative structural, emulsifying, and biological
Chemistry, 53(3), 716–724. https://doi.org/10.1021/jf049267y. properties of 2 major canola proteins, cruciferin and napin. Journal of Food Science,
Waschatko, G., Junghans, A., & Vilgis, T. A. (2012a). Soy milk oleosome behaviour at the 73(3), 210–216. https://doi.org/10.1111/j.1750-3841.2008.00675.x.
air-water interface. Faraday Discussions, 158, 157–169. https://doi.org/10.1039/ Yang, J., Thielen, I., Berton-Carabin, C. C., van der Linden, E., & Sagis, L. M. C. (2020).
c2fd20036h. Nonlinear interfacial rheology and atomic force microscopy of air-water interfaces
Waschatko, G., Schiedt, B., Vilgis, T. A., & Junghans, A. (2012b). Soybean oleosomes stabilized by whey protein beads and their constituents. Food Hydrocolloids, 101.
behavior at the air − water interface. The Journal of Physical Chemistry B, 116, https://doi.org/10.1016/j.foodhyd.2019.105466, 105466.
10832–10841. https://doi.org/10.1021/jp211871v.

11