Food Hydrocolloids 20 (2006) 35–43

www.elsevier.com/locate/foodhyd

Oil-in-water emulsion properties and interfacial characteristics

of hen egg yolk phosvitin
Oscar Castellania,1, Corinne Belhommea, Elisabeth David-Brianda,
Catherine Guérin-Dubiardb, Marc Antona,*
a
Institut National de la Recherche Agronomique, Laboratoire Biopolymères Interactions Assemblages, BP 71627, 44316 Nantes Cedex 3, France
b
Agrocampus Rennes, 65 rue de Saint-Brieuc, CS 84215-35042 Rennes Cedex, France
Received 22 July 2004; accepted 22 February 2005

Abstract
Phosvitin is an egg yolk protein, constituted by 50% of serines, which are all phosphorylated. This singularity makes of phosvitin one of
the most safety natural iron binding molecules. When hen egg yolk is used as emulsifying agent, the adsorption of phosvitin could alter this
iron binding property. Consequently, this work was performed to better understand emulsifying and interfacial properties of phosvitin in
optimal metal binding conditions. We have studied sunflower oil-in-water emulsions of phosvitin at pH 5 and 6, and ionic strengths of 0.05
and 0.15 M. The oil droplet size, the stability against coalescence and flocculation, the composition of the interfacial protein film and the
interfacial activity at model interfaces were analysed. Finally, the capacity of phosvitin to bind iron when it is anchored at the emulsion
interface was investigated.
Phosvitin showed satisfactory emulsifying capacity in conditions favouring iron fixation. In these conditions, emulsifying activity is
sensitive to pH whereas flocculation is influenced by ionic strength. Coalescence destabilisation is not extended and the interfacial protein
film has better characteristics at pH 5. Phosvitin was not efficient to reduce the oil-in-water interfacial tension, although at 0.1 mg/ml the
interfacial tension was reduced from 31 mN/m to 15 mN/m. The high iron binding capacity of phosvitin in solution is kept by interfacial
adsorbed phosvitin. Finally, to explain the poor adsorption efficiency coupled with the suitable emulsifying properties of phosvitin and the
preservation of the iron binding capacity, it is likely that phosvitin is anchored at the interface only by a tiny terminal portion, presenting the
rest of the molecule solvated on the aqueous phase.
q 2005 Elsevier Ltd. All rights reserved.

Keywords: Phosphoprotein; Phosvitin; Emulsion; Interfacial-protein film

1. Introduction can be represented like an elongated core of negatively

Phosvitin (PVT) is a phosphoprotein that constitutes, in
combination with high-density lipoproteins (HDL), the
granules of hen egg yolk. It represents 25% of granules yolk
proteins, and one of its singularities is that it presents a high
proportion of serines which are almost all phosphorylated
(Clark, 1985). Some phosvitin polypeptides have been
sequenced, and, from the amino acid sequences, phosvitin
* Corresponding author. Tel.: C33 2 40 675080; fax: C33 2 40 675084.
E-mail address: anton@nantes.inra.fr (M. Anton).
1
Present address: Centro de Investigación y Desarrollo en Criotecnologı́a
de Alimentos. CONICET-UNLP (Facultad de Ciencias Exactas). Calle 47
esq. 116 (1900) La Plata, Argentina.
0268-005X/$ - see front matter q 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.foodhyd.2005.02.010
charged phosphoserines, with a C-terminal part of about 15
residues relatively rich in hydrophobic amino acids (Byrne
et al., 1984; Mabuchi et al., 1996). Phosvitin has important
features as metal chelator, and is one of the molecules with
the highest iron and calcium binding capacity (Grizzuti &
Perlman, 1975; Hegenauer, Saltman, & Nace, 1979). The
metal-chelator antioxidant activity of phosvitin in solution
is well established (Lu & Baker, 1987; Yamamoto, Sogo,
Iwao, & Miyamoto, 1990).
Hen egg yolk is commonly used in emulsion industry to
form and stabilise emulsions, and it is a key ingredient in
products such as mayonnaise or salad dressing. During the
past 15 years, emulsifying behaviour of egg yolk has begun
to be studied intensively (Anton & Gandemer, 1997;
Kiosseoglou, 1989; Martinet, Beaumal, Dalgalarrondo,
& Anton, 2002; Martinet, Saulnier, Beaumal, Couthaudon,
36 O. Castellani et al. / Food Hydrocolloids 20 (2006) 35–43
& Anton, 2003; Mine, 1998a,b). The main studies have been
focused on the lipoproteins of egg yolk and it has been
specified that, among egg yolk constituents, low-density
lipoproteins bring a great contribution to the emulsifying
properties of yolk.
Concerning phosvitin, the principal investigations have
been essentially focused on iron fixation in solution
(Castellani, Guérin-Dubiard, David-Briand, & Anton,
2004; Lee, Han, & Decker, 2002), and, on conditions
affecting the emulsifying and adsorption properties (Chung
& Ferrier, 1991, 1992; Damodaran & Xu, 1996), However,
conditions and mechanisms allowing simultaneously good
iron fixation and suitable emulsifying properties have not
been studied yet. When employing egg yolk in emulsion
formation, phosvitin migrates to the interface and its iron
affinity could result modified. If the iron affinity of phosvitin
decreases, it could release iron to the solution, and therefore,
increase the oxidant capacity of solution.
Consequently, our objectives in this study were to
characterise phosvitin emulsions and interfacial activity of
phosvitin in conditions where it can bind iron (pH 5–6). To
increase the knowledge of interfacial properties of this
protein, we have determined the oil–water interfacial
protein composition (qualitatively and quantitatively), and
both the structural modifications followed by phosvitin and
its iron binding capacity when it is placed at the interface.
2. Materials and methods
2.1. Materials
Isabrown eggs (three or four days old) were obtained
from local wholesale distributor. All chemicals (analytical
grade) were purchased from Sigma (Saint Quentin-
Fallavier, France), excepted Hydrochloric acid 37% (pro-
analysis) which was purchased from Carlo Erba Reagenti
(Val de Reuil, France).
2.2. Methods
2.2.1. Phosvitin isolation
Hen eggs were manually broken, and yolks were carefully
freed of adhering white and chalazae by rolling on a filter
paper (Whatman). The vitellin membrane was punctured
with a lancet and the content was collected in a beaker cooled
in iced water. After that, temperature was maintained at 4 8C
all through the process. Granules were extracted from yolk
according to the method of McBee and Cotterill (1979).
Yolk was diluted with an equal mass of a 0.17 M NaCl
solution and mixed with a magnetic stirrer. After 1 h, the
solution was centrifuged at 10,000!g for 45 min in a Jouan
centrifuge (model GR 2022, St Herblain, France) and the
pellet (granules) was collected and dissolved in a 1.74 M
NaCl solution (10% wt/vol %). The mixture was stirred to
complete dissolution keeping the pH adjusted to 7.25.
The solution was then dialysed against several changes of
distilled water for 24 h and centrifuged at 10,000!g for
30 min. The high-density lipoproteins precipitated, and the
supernatant was diluted with 0.9 M MgSO4 solution to
obtain a 0.2 M final concentration of this salt. After
centrifugation (10,000!g for 30 min), a precipitate of
phosvitin was collected at the bottom of the tubes and it
was dialysed against distilled water and freeze-dried (PVT).
2.2.2. Protein determination
The protein concentration was determined by means of
phosphorus content considering 9 wt% of this mineral in
PVT (Castellani, Martinet, David-Briand, Guérin-Dubiard,
& Anton, 2003). Phosphorus was determined by triplicate,
using the colorimetric method described by Bartlett (1959),
with hydrazine sulphate and sodium molybdate as reagents.
2.2.3. Emulsion preparation
Dispersions of 5 mg/ml of lyophilised PVT were made at
pH 6 (0.05 M 2-(N-porpholino) ethanesulfonic acid (MES)
buffer) and at pH 5 (0.05 M acetic acid/sodium acetate
buffer). NaCl was added to obtain either 0.05 or 0.15 M
values of ionic strength. They were gently stirred for 1 h,
then solutions were centrifuged 30 min at 3000!g, and the
supernatant was recovered. PVT final concentration of all
supernatants were measured and fixed at 3.3 mg/ml. Oil-in-
water emulsions were prepared with 3 ml of sunflower oil
and 27 ml of the respective PVT solution (oil volume
fraction 0.1). The system was premixed for 1 min at
20,000 rpm using a 12 mm diameter head attached to a
polytron PT 3000 homogeniser (Kinematica, Switzerland).
Then, homogenisation of the emulsion premix was achieved
with a high-pressure valve homogeniser at 70 bars (Stansted
Fluid Power Ltd. model AO 812 W, Stansted, Essex, UK).
Recirculation of each emulsion was made during 5 min at a
flow rate of 80 ml/min.
2.2.4. Particle size distribution
The volume-surface area average diameter distribution
(d32) and the volume frequency distribution (d43) of the
emulsion droplets were determined by laser light diffraction
using a Saturn DigiSizere 5200 (Micromeritics Instrument
Corporation, USA). The refractive index of the oil was
1.475 and the imaginary part of refractive index (due to
absorption) was fixed at 0.01. After homogenisation, a
1/12.5 emulsion dilution in 0.05 M Tris–HCl buffer (pH 8)
containing 1 wt/vol.% of sodium dodecyl sulphate (SDS)
was made to deflocculate the oil droplets.
2.2.5. Flocculation
Emulsions were kept at ambient temperature (18 8C) for
4 h, and then the volume frequency distribution was
measured as indicated previously. To detect the flocculation
process, emulsions were diluted in buffers with and without
SDS. The flocculation index was evaluated as: FiZ
(d43/d43SDSK1)!100, where d43 is the volume frequency
 O. Castellani et al. / Food Hydrocolloids 20 (2006) 35–43
mean of emulsion, and d43SDS the volume frequency mean
of emulsion in the presence of 1 wt/vol.% of SDS.
2.2.6. Coalescence
Emulsions were kept at room temperature (20 8C)
without stirring, and coalescence was measured 24 h
after emulsion preparation by estimation of the specific
surface area (m2/ml oil, SSZ6/d32) according to Walstra
(1983). The coalescence index was evaluated as follows:
CiZ(1KSS24/SSo)!100, where SSo is the specific surface
area at zero time (fresh emulsion), and SS24 the specific
surface area at 24 h after emulsion preparation. For this type
of experiments emulsions were diluted in a buffer contain-
ing 1 wt% SDS.
2.2.7. Adsorbed proteins and interfacial protein
concentration
To wash off the emulsion droplets from native emulsion
the method of Patton and Huston (1986) was slightly
modified. Native emulsion was diluted one time with a
50 wt/vol.% saccharose solution in the same buffer of the
emulsion aqueous phase. An aliquot of this dilution (5 ml)
was then carefully deposited at the bottom of a centrifuging
tube containing 9 ml of a 5 wt/vol.% saccharose solution in
the corresponding buffer. These tubes were centrifuged
at 3000!g for 2 h at 15 8C and then immediately frozen at
K20 8C. Three phases were distinguishable in the tubes:
creamed oil droplets at the top (adsorbed phase), an
intermediate separating phase, which corresponds to 5 wt/
vol.% saccharose separating solution, and the aqueous phase
of the emulsion at the bottom. The frozen tubes were cut so
as to separate the different phases, and protein content was
determined in each of them like previously indicated (item 2
in this section). Proteins from the upper phase are adsorbed
proteins and those from the bottom phase are unadsorbed
proteins. When the middle phase was turbid, presence of
small oil droplets, its protein content was included in the
adsorbed protein quantity. The amount of protein adsorbed
to the oil droplet surface and interfacial protein concen-
tration (G, mg/m2) was calculated like indicated in Martinet
et al. (2002).
2.2.8. Protein composition of the interfacial film
The three recovered phases were analysed by SDS-
PAGE: samples were diluted (4:5) in sample buffer (0.25 M
Tris–HCl pH 6.8, 0.04% bromophenol blue and 30%
glycerol, 15% 2-mercaptoethanol and 6% SDS solution).
The mixture was then treated for 5 min in boiled water and
then analysed. Stacking gel was constituted by 3.5% of
acrylamide in 0.124 M Tris–HCl buffer at pH 6.8, and the
running gel by 10% of acrylamide in 0.05 M Tris–
HCl/0.36 M glycine at pH 8.8. Electrophoresis buffer was
constituted by 0.05 M Tris–HCl, 0.42 M glycine at pH 8.8
and SDS 0.1%. About 30 mg of total protein were loaded
onto the gels. Electrophoresis was performed at 30 mA
per gel for 1.5 h in an Amersham Biosciences system
37
(Hoefer, Mighty Small II SE250/SE260, San Francisco,
USA). Gels were stained by the ‘Coomassie Brilliant Blue
with aluminium mordant’ method as described by Martinet
et al. (2002). Destaining procedure employed 7% acetic acid
and 40% ethanol solution. Molecular weights were
estimated with the Rf of the following standard proteins:
phosphorylase b (94,000 Da), bovine serum albumin
(67,000 Da), ovalbumin (43,000 Da), carbonic anhydrase
(30,000 Da), trypsin inhibitor (20,100 Da) and a-lactalbu-
min (14,400 Da).
Gels were scanned on an imaging densitometer Biorad
GS710 and both Rf and percentage ratio of each band
were determined with Quantity One 4.1 software (Biorad,
Ivry-sur-Seine, France).
2.2.9. Interfacial tension
The interfacial tension was measured using the pendant
drop technique (tensiometer CAM 200, Helsinki, Finland).
A pendant drop of buffer or protein solution was held in the
oil phase by the tip of a computer-controlled syringe. The
drop was illuminated by a uniform light source and its
profile was imaged and digitised by a camera coupled to a
PC. The software analysed the shape of the drop image and
determination of the interfacial tension was based on true
Young and Laplace equation. Data presented are averages
of at least three measurements. The protein concentration
was 0.01 mg/ml and 0.1 mg/ml in the buffers used for
emulsion preparation, and the oil was purified twice using
silica cartridges (Sep-Pak, Waters, Milford, USA). Tem-
perature was set to 20 8C using a Julabo F12 cryostat
(Seelbach, Germany). To obtain the final interfacial
pressure each measurement was extended through 18 h,
when the interfacial pressure kept at the same value
(G0.05 mN/m) at least 30 min.
2.2.10. Surface film balance
Measurements of the surface pressure (p)-surface area
(A) isotherms were performed by the Wilhelmy plate
method on an isolated and fully automated Langmuir-type
film balance (KSV 3000 V2, Helsinki, Finland). The
maximum and minimum areas of the trough were 7.30!
10K2 m2 and 1.07!10K2 m2, respectively. It was filled
with each of the four buffer systems employed for emulsion
preparation and temperature was kept at 20 8C by water
circulation from a thermostat Bioblock Ministat (Illkirch,
France). The cleanliness of the buffers was checked by a
compression and expansion cycle using the buffer without
protein injection, considering a cleaned surface when
surface pressure rising was negligible (less than
0.05 mN/m). Then, an aliquot of 200 ml of phosvitin
solution (3.3 mg/ml) in the respective buffer was spread on
the surface by means of a micrometric syringe. To allow the
process of protein spreading, adsorption and rearrange-
ments, samples were allowed to stand 40 min before
compression. The compression speed was maintained
constant at 28.4 cm2/min, which is sufficiently low to
38 O. Castellani et al. / Food Hydrocolloids 20 (2006) 35–43
prevent secondary effects dues to the barrier displacement.
At least four isotherms were performed for each sample.
2.2.11. Front-surface fluorescence spectroscopy
Front-surface fluorescence spectra of both phosvitin in
solution and phosvitin adsorbed at the O–W interface were
obtained with an F-4500 FL fluorescence spectrophotometer
(Hitachi, Tokio, Japon) at 20 8C using the method described
by Rampon, Lethuaut, Mouhous-Riou, and Genot (2001).
The excitation wavelength was fixed at 278 nm and the
emission was recorded between 300 and 450 nm. The slit
widths were set at 5 nm in both excitation and emission
pathways. Adsorbed phosvitin concentration was obtained
as expressed in Section 2.2.7, and the cream of emulsion
was measured directly.
2.2.12. Determination of iron binding capacity of interfacial
adsorbed phosvitin
Iron-binding capacity of adsorbed phosvitin was deter-
mined by the same method as Castellani et al. (2004), with a
different sample preparation. Cream (adsorbed proteins)
was obtained by the method described in Section 2.2.7.
After that, the adsorbed protein concentration was
measured, and ferrous iron was added, in three different
reaction tubes, at the following ratios: 0.5, 1, and 1.5 iron
atoms per two phosphate groups. At the same time, protein
solutions in the same conditions and concentration were
tested to compare the iron binding capacity of phosvitin in
solution and adsorbed to the O–W interface.
2.2.13. Statistical analysis
At least three replicates were made for all measurements.
Statistical analyses were made by a 1-way analysis of
variance using STATGRAPHICS software (Statistical
Graphics Corporation, Rockville, USA). Confidence inter-
vals were set at 95% (p!0.05).
3. Results
3.1. Droplet size and stability of phosvitin emulsions
Table 1 shows that 3.3 mg/ml phosvitin solutions are
able to stabilise oil-in-water emulsions, giving droplet sizes
Table 1
Characteristics of phosvitin emulsions
pH Ionic d43 (mm)
strength (M)
5 0.05 0.9a
5 0.15 0.9a
6 0.05 1.0b
6 0.15 1.3c
Droplet size (d43), flocculation index, coalescence index, % adsorbed proteins, and interfacial protein concentration (G) of phosvitin emulsions prepared at two
pH and ionic strength (M) values: pH 5 (0.05 or 0.15 M NaCl) or pH 6 (0.05 or 0.15 M NaCl). Phosvitin concentration was fixed at 3.3 mg/ml. Values within the
same column followed by a different letter are significantly different (p!0.05).
of about 1 mm. Phosvitin oil droplet sizes (volume
frequency diameter, d43) at pH 5 were smaller than at pH
6, indicating that protein emulsifying capacity in the first
condition is better. On the contrary, the influence of ionic
strength at the studied pH values was only evident at pH 6.
The flocculation index was greatly influenced by the
ionic strength. Values at 0.05 M of ionic strength for both
pH values were around 6 times smaller than at 0.15 M. On
the contrary, no influence of pH was detected at the studied
ionic strengths at a 5% confidence level. We can notice that,
for example at 0.15 M ionic strength at both pH values,
flocculation indexes are the same whereas droplet sizes are
very different, implying that this last parameter is not a
relevant factor in the flocculation of phosvitin emulsions.
The correlation of emulsion destabilisation by both,
flocculation and the increasing of ionic strength, was also
observed at higher salt concentrations (0.30 and 0.50 M of
ionic strength). Even though, the emulsions were very
unstable and they could not be used in further analysis
(results not shown).
Concerning the coalescence index in quiescent con-
ditions 24 h after emulsion preparation, Table 1 shows that
the increase in ionic strength from 0.05 to 0.15 M
accelerates the coalescence process. But also, the increase
of pH from 5 to 6 provokes an augmentation of the
coalescence index. We can then remark that the high degree
of flocculation previously observed exerts some influence on
coalescence but other influences exist as we have obtained
the same level of coalescence (0.2) with two samples
presenting different levels of flocculation (6.2 for pH 5 and
0.15 M ionic strength and 1.4 for pH 6 and 0.05 M ionic
strength).
3.2. Adsorbed proteins, quantitative and
qualitative analyses
As shown in Table 1, at pH 5 interfacial adsorbed
proteins were under 40% of total protein whilst at pH 6 this
percentage was higher (nearby 63%). Whatever the pH
value, the percentage of adsorbed protein was statistically
the same at both ionic strength conditions. In addition, like
indicated in Table 1, the concentration of adsorbed protein
at pH 5 (nearby 1.6 mg of protein/m2 of interface) is
sensibly lesser than at pH 6 (3.0 mg/m2). This better
Flocculation Coalescence % Adsorbed G (mg/m2)
index (4 h) index (24 h) proteins
1.1a 0.12a 39a 1.9a
6.2b 0.22b 33a 1.3a
1.4a 0.23b 67b 2.9b
6.7b 0.47c 61b 3.1b
 O. Castellani et al. / Food Hydrocolloids 20 (2006) 35–43
Fig. 1. SDS-PAGE electrophoresis of adsorbed (Ads and Ads2) and non-
adsorbed (Na) phosvitin at the emulsion interface. sd: standard proteins.
Arrow indicates the band of different phosvitine polypeptides (see the text).
Acrylamide concentration 10 wt%. Stain: Coomassie blue with aluminium
mordant.
adsorption is the combination of the elevated percentage of
protein adsorption and of the slightly smaller interfacial
surface generated at pH 6. The electrophoresis analysis of
proteins obtained during the emulsion droplet separation
is shown in Fig. 1 (0.05 M ionic strength). The pattern
shows that adsorbed protein (lines ‘Ads’ in the figure)
is constituted by the two main phosvitin polypeptides,
a-phosvitin (40 kDa) and b-phosvitin (45 kDa). The first
one polypeptide is also present in the separation phase (lines
‘Ads2’ in Fig. 1). The proteins present in this phase were
considered like adsorbed protein because the 5% saccharose
solution was faintly turbid at the end of sample centrifu-
gation, indicating the presence of small oil droplets that are
resistant to the creaming process. In non-adsorbed phosvi-
tin, both polypeptides are present and exhibit a slightly
modified mobility because of the presence of residual
saccharose in the sample. Phosvette polypeptides, with
molecular masses of about 22 kDa, are only present, as pale
bands, in this aqueous phase. Comparable results were
obtained at 0.15 M of ionic strength (results not shown).
3.3. Model interfaces covered by phosvitin (O–W and A–W
interfacial tension)
Fig. 2 shows the equilibrium interfacial tension obtained
at 20 8C, between purified sunflower oil and the four
aqueous phosvitin solutions, at two different protein
concentrations (0.01 and 0.1 mg/ml). The dashed line
indicates the interfacial tension obtained without protein
(30.5 mN/m), which is consistent with values already
39
6 - 0.05 6 - 0.15 5 - 0.05 5 - 0.15
35
Interfacial tension (mN/m)
30
25
20
15
10
0
0.01 0.1
pvt concentration (mg/ml)
Fig. 2. Equilibrium interfacial tension (mN/m) of an oil–water interface at
different concentrations of phosvitin (mg/ml), and physicochemical
conditions (indicated on the top pH—ionic strength). Dashed line indicates
the value obtained without protein for all the studied conditions.
Temperature: 20 8C.
reported for the oil-aqueous solution interface (Dickinson,
1996). At 0.01 mg/ml it was observed a slight reduction of
the interfacial tension to 26–28 mN/m, without a significant
difference between the different conditions. This behaviour
is very different to that one of other proteins like bovine
serum albumin or ovalbumin, which can drop to much
smaller levels the interfacial tension (Bos & van Vliet,
2001). Even when comparing with b-casein, which is a non-
ordered and phosphorylated protein, the interfacial activity
of phosvitin was weak (Bos et al., 2001). Nonetheless, at
higher concentrations (i.e. 0.1 mg/ml), equilibrium inter-
facial tension was significantly reduced. Whatever the pH of
emulsion, interfacial tension was more effectively reduced
at 0.15 M than at 0.05 M of ionic strength, with differences
of 5.5 mN/m (pH 6) and 3 mN/m (pH 5). Between different
pH values the dissimilarity was less pronounced. Phosvitin
exhibited a slow kinetic of adsorption, and in order to obtain
the final interfacial tension of equilibrium it was necessary
to measure throughout 20 h. These results are coincident
with those obtained by Dickinson, Hunt, and Dalgleish
(1991), where phosvitin shown poor surface activity and
long time to reach the steady-state for the n-tetradecane-
water interface.
Like a complement, we have studied the behaviour of
this protein at a planar air–water surface by the Langmuir
trough method. The p-A isotherms (20 8C) at the four
conditions are shown in Fig. 3. Phosvitin monolayer
presents a liquid-expanded-like structure under the exper-
imental conditions, with one inflexion point where the
properties of the film change significantly. The monolayer
collapse was observed at the highest surface pressure. These
characteristics were similar to the results obtained by
Rodrı́guez Niño, Carrera Sánchez, and Rodrı́guez Patino
(1999) with the structurally related b-casein. Besides, and
for the two studied pH values, phosvitin isotherms were
displaced towards the p-axis when ionic strength drops
from 0.15 to 0.05 M. The p-A profiles at 0.05 M present less
surface pressure than 0.15 M profiles for the same specific
40 O. Castellani et al. / Food Hydrocolloids 20 (2006) 35–43

5 -0.05 5 -0.15
sol 5 0.05 ads 5 0.05 sol 5 0.15 ads 5 0.15 pH 5
45 1200 343.8 nm
335.7 nm
surface pressure (mN/m)

40 1000
pH 5
35 345.0 nm
800

FI (AU)
30
25 600

20 400
15
200
10
0
5 300 320 340 360 380 400 420 440
0 emission wavelength (nm)
100 300 500 700
area (cm2) sol 6 0.05 ads 6 0.05 sol 6 0.15 ads 6 0.15 pH 6
1200
344.2 nm
6 -0.05 5 - 0.15 336.8 nm
1000
45

FI (AU)
344.4 nm
40 pH 6 800
surface pressure (mN/m)

35 600
338.1 nm

30
25
20
15
10
5 emission wavelength (nm)
0
100 300 500
area (cm2)
Fig. 3. Isotherms of surface pressure (mN/m) versus specific area (cm2/mg
of proteins) for phosvitin films. Buffer solutions at pH 5 (0.05 or 0.15 M
NaCl) or at pH 6 (0.05 or 0.15 M NaCl). Temperature 20 8C.
surface area, even at the collapse. At mZ0.15 M the protein
film showed a collapse pressure of about 40 mN/m, whilst at
0.05 M collapse pressure was around 30 mN/m. This result
is correlated with those of pendant drop interfacial tension,
indicating a similar behaviour of Pvt at oil–water or air–
water interface. In addition, the concentration employed to
obtain these results are relatively high in comparison with
other proteins. For example, b-casein could behave like Pvt
when spreading in 22 times lesser protein concentration on
the trough (Rodrı́guez Patino, Sánchez, & Rodrı́guez Niño,
1999). One of the reasons is that phosvitin is a highly
charged protein, with only a few percentage of hydrophobic
amino acids. So, it is possible that some phosvitin molecules
were solubilised in solution, although the protein was spread
at the water surface, with the consequent diminution of its
surface concentration.
3.4. Front-surface fluorescence analysis
Fig. 4 shows the spectra of fluorescence emission of PVT
in buffer solution and adsorbed at the emulsion droplet
interface at the four studied conditions. This experiment was
performed in view to test if phosvitin structure was altered
by its adsorption at the oil–water interface. The results at pH
5 and 6 are similar. In both cases, and regardless of the ionic
strength value, the wavelength at maximum emission in
400
200
0
300 320 340 360 380 400 420 440
Fig. 4. Front-surface fluorescence emission spectra (lexcZ278 nm) of
700
phosvitin solutions (sol) and cream of phosvitin emulsions (ads). Buffer
solutions at pH 5 (0.05 or 0.15 M NaCl) or at pH 6 (0.05 or 0.15 M NaCl).
Temperature: 20 8C. Fluorescence intensity is expressed in arbitrary units.
solution (open symbols) was around 345.0 nm. This is
characteristic of systems where the fluorescent chromo-
phores are in a high hydrophilic environment (Permyakov,
1993). On the other hand, the spectra of PVT adsorbed at the
interface (solid symbols) are displaced to shorter wave-
lengths, and the maximum emission occurred between
335.7 and 343.8 nm. The fluorescence intensity of turbid
solutions (adsorbed phosvitin) and translucent solutions
(phosvitin in solution) could be differently quenched, so the
evaluation and comparison of fluorescence intensity is not
so obvious like the wavelength of maximum emission
fluorescence. The changes in spectra features indicate
important modifications in phosvitin structure, and are the
consequence of a more hydrophobic environment of its
tryptophan residues when phosvitin is at the interface. In
addition, at both pH values the displacement of emission
maximum was more important at 0.15 M of ionic strength
than at 0.05 M. This could be related to the fact that in the
presence of salt hydrophobic interactions are favoured, and
phosvitin could be more easily or deeply anchored to the
interface. The blue shift at pH 5 and 0.15 of ionic strength
was the more important one (10.1 nm).
3.5. Iron binding of adsorbed phosvitin.
The iron binding capacity of phosvitin in solution and
adsorbed at the oil–water emulsion interface is shown
in Fig. 5 for the four physicochemical conditions at three
 O. Castellani et al. / Food Hydrocolloids 20 (2006) 35–43
0.5
1.50
iron atoms/2 P of Pvt
1.20
0.90
0.60
0.30
0
5-0.05
0.5
1.50
iron atoms/2 P of Pvt
1.20
0.90
0.60
0.30
0
5-0.05
physicochemical conditions
Fig. 5. Iron fixation (three different levels of iron 0.5, 1 and 1.5 atoms per two phosphate groups) by phosvitin versus pH and ionic strength (M) values: pH 5
(0.05 or 0.15 M NaCl) or pH 6 (0.05 or 0.15 M NaCl), in solution or adsorbed at the oil–water interface.
different levels of iron. These two figures clearly indicate
that the high iron binding capacity of phosvitin in solution is
kept when phosvitin is at the oil-in-water interface. The
same behaviour was obtained at the three ratios of iron/
phosphate analyzed, signifying that the high iron binding
capacity is maintained for a wide range of iron/protein
ratios. So, phosvitin could bind iron ions with the same
capacity regardless of where it is placed in the emulsion
(bulk solution or oil–water interface).
4. Discussion
4.1. Emulsifying activity is sensitive to pH, while
flocculation is affected by ionic strength conditions
The results concerning emulsifying activity of phosvitin
and emulsion stability indicate that phosvitin is an adequate
emulsifier, in conditions favouring iron fixation (pH values
between 5 and 6 and ionic strength values until 0.15 M).
This is in agreement with, and at the same time extends, the
studies made by Dickinson et al. (1991) at pH 7. The higher
emulsifying activity exhibited at pH 5 is probably the effect
of the decrease in negative charge when decreasing the pH
value. If we have in mind that phosvitin is composed of
about 50% of phosphoserines, and that the pKa of its second
proton dissociation is in the proximity of pH 6.0
41
1.0 1.5 solution
5-0.15 6-0.05 6-0.15
physicochemical conditions
1.0 1.5
adsorbed
5-0.15 6-0.05 6-0.15
(Damodaran et al., 1996), this difference of one pH unit
could be more relevant than it could be for other proteins.
Nonetheless, phosvitin molecules remain with a relatively
high negative charge, even at pH 5.
The flocculation process is very affected by the ionic
strength conditions. We have already observed this
influence on emulsions made with aggregated and non-
aggregated phosvitin at pH 6 (results not shown). The effect
of sodium salt reflects the screening of negative charge
density at the surface of the oil droplets by salt ions.
Consequently, electrostatic repulsions have less intensity
and the flocculation results facilitated (McClements, 1999).
Apparently, the charge difference between pH 5 and 6 is not
enough to influence on the flocculation process, although it
has some influence on the emulsion activity. The lack of
strict correlation between emulsion activity and emulsion
stability was already described (Chung et al., 1992;
Dickinson et al., 1991). Even though the influence of oil
droplet size on the flocculation process was established
(McClements, 1999), the results obtained for phosvitin
emulsions clearly indicate that, in this case, the oil droplet
size has no influence on flocculation. The electrostatic
repulsion was the prevailing stabilising force. It is also
interesting to note that at pH 5 the coalescence is minimal,
even at ionic strength of 0.15 M where the flocculation was
elevated. This seems to indicate that the interfacial protein
film formed at this pH is more resistant and has better
42 O. Castellani et al. / Food Hydrocolloids 20 (2006) 35–43
viscoelastic properties than the protein film at pH 6, because
the oil droplets resist to the coalescence process at the same
time that they are in contact as deduced by the high
flocculation index. These findings are in accordance with
the measures of the elasticity of phosvitin interfacial film,
which show higher values at pH 5 than at pH 6 for the
air-water interface (results not shown).
4.2. Phosvitin is a poor surfactant, but a good
emulsion stabiliser
The percentage of adsorbed protein at pH 5 is smaller
than at pH 6, indicating a better efficiency of phosvitin to
form the interfacial film of emulsions in the former
condition. The quantity of adsorbed protein at pH 5 was
1.9 mg (at 0.05 M) and 1.3 mg (at 0.15 M) per m2 of oil
interface, which is in accordance with the emulsion activity
at this pH value. At pH 6 there is a higher quantity of protein
adsorbed at the interface, approximately 63% of the added
protein that correspond to a surface concentration of 3.0 mg
of protein/m2 of oil interface. These values are at the same
level of other more surface-active proteins like b-casein
(Bos et al., 2001; Dickinson et al., 1991). In the case of
phosvitin, and at pH 6, a secondary layer of interfacial
protein would probably be formed. Phosvitin employed in
this work possesses a certain proportion of residual Mg2C
ions (Castellani et al., 2003), which at pH 6 could make
electrostatic bridges between different phosvitin molecules.
So, the generation of a secondary protein layer or a thicker
interfacial film protein is probable. This must result in a
higher surface protein concentration, which is not necessary
accompanied by a better emulsifying activity. We have
observed that the presence of aggregates at the oil droplet
interface has a negative influence on the emulsifying
activity.
The nature of adsorbed proteins is the same for all
conditions: a and b-phosvitin polypeptides are present at the
interface. The presence of phosvettes polypeptides in the
aqueous emulsion phase indicates that this kind of
polypeptides, with smaller molecular mass, have not affinity
for the interface.
Actually, adsorbed phosvitin polypeptides have not a
thermodynamic effect for dropping the oil–water interfacial
tension. On the contrary, we have demonstrated that oil–
water and air–water interfacial tensions are only diminished
at very high phosvitin/interface ratio. These high ratio
values are usually achieved when working with model
interfaces, but not when working with emulsions. In
addition, and as also described by Damodaran et al.
(1996) and Dickinson, Hunt, and Horne (1992), the time
to obtain the steady-state values of interfacial tension were
longs, while the interaction time between the generated
surface and the protein solution is very short. In conse-
quence, this protein would be placed at the interface during
the high turbulent homogenisation occurred during the
emulsion preparation, and once phosvitin is at the interface
it could stabilise the emulsion mainly by electrostatic charge
repulsion of oil droplets.
4.3. Phosvitin changes its structure at the interface but
it keeps its high iron binding capacity
Phosvitin presents high changes in fluorescence emission
spectra when it is adsorbed at the interface. The blue shift of
the maximum when the protein is at the interface indicates
that, in this condition, tryptophan groups are placed in a
more hydrophobic environment. This result is not surpris-
ing, because phosvitin polypeptides are highly charged and
exhibit mainly a non-ordered structure at the experimental
conditions. So, tryptophan is normally exposed to the
solvent. Then, when phosvitin is adsorbed at the oil–water
interface, the C-terminal extremity of the protein, which has
the hydrophobic amino acids responsible for fluorescence
emission (Byrne et al., 1984), could be buried in the oil side
of the interface. The N-terminal region of the molecule (by
15 amino acids) has also hydrophobic features, and even
though it lacks of tryptophan, we could hypothesise that it
could also be anchored at the interface.
This structural change has apparently no influence on the
rest of the molecule, which normally presents a highly
proportion of non-ordered structure. So, phosphoserine
groups keep highly solvated as in phosvitin aqueous
solution, and phosvitin molecules must have a great
proportion of the molecule emerging from the interface to
the aqueous phase. In this way, results could explain the
high iron binding capacity of phosvitin at the interface.
Moreover, this is in agreement with the predicted model of
phosvitin adsorption based on a statistical analysis, where
the protein might interact with the hydrophobic surface by
the two ends of the molecule, forming a loop like structure
(Dickinson, Pinfield, & Horne, 1997).
5. Conclusion
Despite a poor surfactant activity, phosvitin provides
excellent properties of emulsion stabilisation in regions of
pH and ionic strength that favours the iron binding
capacity. Furthermore, the ability of phosvitin to bind iron
is not modified when phosvitin is adsorbed as compared
to phosvitin in solution. Consequently, iron will not be
released when phosvitin emulsion is created. In such
medium conditions, emulsifying activity is sensitive to pH
while flocculation is influenced by ionic strength.
Coalescence destabilisation is not high, and the protein
film has better characteristics at pH 5. The two principal
polypeptides of phosvitin are adsorbed at the interface of
emulsions and they interact by, at least, its C-terminal
portion, with the core of the molecule protruding to the
aqueous solution.
 O. Castellani et al. / Food Hydrocolloids 20 (2006) 35–43
Acknowledgements
We want to thank INRA TPA Department, Pole
Agronomique Ouest, and the Region of Pays de la Loire
for their financial support. We also want to thank CONICET
of Argentina for O. Castellani’s partial financial support.
References
Anton, M., & Gandemer, G. (1997). Composition, solubility and
emulsifying properties of granules and plasma from egg yolk. Journal
of Food Science, 62, 484–487.
Bartlett, G. (1959). Phosphorus assay in column chromatography. Journal
of Biological Chemistry, 234(3), 466–468.
Bos, M. A., & van Vliet, T. (2001). Interfacial rheological properties of
adsorbed protein layers and surfactants: A review. Advances in Colloid
and Interface Science, 91, 437–471.
Byrne, B. M., van het Schip, A. D., van de Klundert, J. A. M., Arnberg,
A. C., Gruber, M., & Geert, A. B. (1984). Amino acid sequence of
phosvitin derived from the nucleotide sequence of part of the chicken
vitellogenin gene. Biochemistry, 23, 4275–4279.
Castellani, O., Guérin-Dubiard, C., David-Briand, E., & Anton, M. (2004).
Influence of physicochemical conditions and technological treatments
on the iron binding capacity of egg yolk phosvitin. Food Chemistry, 85,
569–577.
Castellani, O., Martinet, V., David-Briand, E., Guérin-Dubiard, C., &
Anton, M. (2003). Egg yolk phosvitin: Preparation of metal-free
purified protein by fast protein liquid chromatography using aqueous
solvents. Journal of Chromatography B, 791, 273–284.
Chung, S. L., & Ferrier, L. K. (1991). Conditions affecting emulsifying
properties of egg yolk phosvitin. Journal of Food Science, 56,
1259–1262.
Chung, S. L., & Ferrier, L. K. (1992). pH and sodium chloride effects on
emulsifying properties of egg yolk phosvitin. Journal of Food Science,
57, 40–42.
Clark, R. C. (1985). The primary structure of avian phosvitins
(contributions through the Edman degradation of methylmercaptovitins
prepared from the constituent phosphoproteins). International Journal
of Biochemistry, 17, 983–988.
Damodaran, S., & Xu, S. (1996). The role of electrostatic forces in
anomalous adsorption behavior of phosvitin at the air/water interface.
Journal of Colloid Interface Science, 178, 426–435.
Dickinson, E. (1996). Les colloı̈des alimentaires. Paris: Masson.
Dickinson, E., Hunt, J. A., & Dalgleish, D. G. (1991). Competitive
adsorption of phosvitin with milk proteins in oil-in-water emulsions.
Food Hydrocolloids, 4, 403–414.
Dickinson, E., Hunt, J. A., & Horne, D. S. (1992). Calcium induced
flocculation of emulsions containing adsorbed b-casein or phosvitin.
Food Hydrocolloids, 6, 359–370.
Dickinson, E., Pinfield, V. J., & Horne, D. S. (1997). NOTE. On the
‘anomalous’ adsorption behaviour of phosvitin. Journal of Colloid and
Interface Science, 187, 539–541.
43
Grizzuti, K., & Perlmann, G. E. (1975). Further studies on the binding of
divalent cations to the phosphoglycoprotein phosvitin. Biochemistry,
14, 2171–2175.
Hegenauer, J., Saltman, P., & Nace, G. (1979). Iron (III)-phosphoprotein
chelates: Stoichiometric equilibrium constant for interaction of iron
(III) and phophorylserine residues of phosvitin and casein. Biochem-
istry, 18, 3865–3879.
Kiosseoglou, V. D. (1989). Egg yolk. In G. Charalambous, & G. Doxastakis
(Eds.), Food emulsifiers: Chemistry, technology, functional properties
and applications (pp. 63–85). London: Elsevier, 63–85.
Lee, S. K., Han, J. H., & Decker, E. A. (2002). Antioxidant activity of
phosvitin in phosphatidylcholine liposomes and meat model systems.
Food Chemistry and Toxicology, 67, 30–41.
Lu, CH.-L., & Baker, R. C. (1987). Effect of pH and food ingredients on the
stability of egg yolk phospholipids and the metal-chelator antioxidant
activity of phosvitin. Journal of Food Science, 52, 613–616.
Mabuchi, N., Yamamura, J.-I., Adachi, T., Aoki, N., Nakamura, R., &
Nakamura, T. (Dec-1996). Submitted to the EMBL/GenBank/DDBJ
databases (P87498).
Martinet, V., Beaumal, V., Dalgalarrondo, M., & Anton, M. (2002).
Emulsifying properties and adsorption behaviour of egg yolk
lipoproteins (LDL and HDL) in oil-in-water emulsions. In M. Anton
(Ed.), Food emulsions and dispersions (pp. 103–116). Kerala: Research
Signpost, 103–116.
Martinet, V., Saulnier, P., Beaumal, V., Couthaudon, J. L., & Anton, M.
(2003). Surface properties of hen egg yolk low-density lipoproteins
spread at the air–water interface. Colloids and surfaces B: Biointer-
faces, 31, 185–194.
McBee, L., & Cotterill, O. (1979). Ion exchange chromatography and
electrophoresis of egg yolk. Journal of Food Science, 44, 656–660.
McClements, D. (1999). Food emulsions. Boca Raton: CRC Press.
Mine, Y. (1998a). Adsorption behavior of egg yolk low-density-
lipoproteins in oil-in-water emulsions. Journal of Agricultural and
Food Chemistry, 46, 36–41.
Mine, Y. (1998b). Emulsifying characterization of hens egg yolk proteins in
oil-in-water emulsions. Food Hydrocolloids, 12, 409–415.
Patton, S., & Huston, G. E. (1986). A method for isolation of milk fat
globules. Lipids, 21, 170–174.
Permyakov, E. A. (1993). Luminescent spectroscopy of proteins. Boca
Raton: CRC Press.
Rampon, V., Lethuaut, L., Mouhous-Riou, N., & Genot, C. (2001).
Interface characterization and aging of bovine serum albumin stabilized
oil-in-water emulsions as revealed by front-surface fluorescence.
Journal of Agricultural and Food Chemistry, 49, 4046–4051.
Rodrı́guez Niño, M. R., Carrera Sánchez, C., & Rodrı́guez Patino, J. M.
(1999). Interfacial characteristics of b-casein spread films at the air-
water interface. Colloids and Surfaces B: Biointerfaces, 12, 161–173.
Rodrı́guez Patino, J. M., Sánchez, C. C., & Rodrı́guez Niño, M. R. (1999).
Structural and morphological characteristics of b-casein monolayers at
the air–water interface. Food Hydrocolloids, 13, 401–408.
Walstra, P. (1983). Formation of emulsions. In P. Becher, Encyclopedia of
emulsion technology: Basic theory (Vol. 1) (pp. 57–127). New York:
Marcel Dekker.
Yamamoto, Y., Sogo, N., Iwao, R., & Miyamoto, T. (1990). Antioxidant
effect of egg yolk on linoleate on emulsion. Agricultural and Biological
Chemistry, 54, 3099–3104.