Vous êtes sur la page 1sur 10

ARTICLE IN PRESS

HYDROCOLLOIDS
Food Hydrocolloids 20 (2006) 11141123 www.elsevier.com/locate/foodhyd

FOOD

Impact of a thermal treatment on the emulsifying properties of egg yolk. Part 2: Effect of the environmental conditions
Fabien Guilmineau, Ulrich Kulozik
Technische Universita t Mu nchen, Chair for Food Process Engineering and Dairy Technology, Weihenstephaner Berg 1, 85354 Freising, Germany Received 25 July 2005; accepted 8 December 2005

Abstract The protein solubility and emulsifying properties of native and heat-treated egg yolk (EY) suspensions were investigated in various environmental conditions. Four distinct conditions were tested by combining two levels of pH, namely pH 4.0 and 6.5, and two levels of ionic strength, namely 0.15 and 0.52 M NaCl, in a model oil-in-water (O/W) emulsion containing 30% oil (v/v). Although the protein solubility was greatly reduced by the thermal denaturation in all tested environmental conditions, the average size of oil droplets obtained in emulsions made with heated EY was observed to be either similar or slightly smaller than that obtained with native EY, depending on the environmental conditions. Using heat-treated EY rather than native EY led to a signicant increase of the interfacial protein concentration in all environmental conditions. This increased interfacial protein concentration was shown to have a major impact on the occulation behaviour of the emulsions, as well as on their rheological properties and stability to creaming. Hypotheses regarding the mechanisms by which insoluble protein aggregates stabilise O/W emulsions at various pH and ionic strengths are discussed. r 2006 Elsevier Ltd. All rights reserved.
Keywords: Egg yolk; Protein denaturation; Emulsifying properties; Environment

1. Introduction Egg yolk (EY) is a reference food emulsier used in many applications ranging from bakery products to salad dressings. Commercial EY is routinely pasteurised at temperatures between 60 and 68 1C in order to ensure its microbiological safety (Cunningham, 1995). Applying more severe thermal treatment conditions would allow to extend the shelf life of EY and further reduce the risk of microbial spoilage. But the impact of an increased degree of thermal denaturation of the heat sensitive EY proteins on the emulsifying properties of this key functional ingredient is not well understood. In the rst part of this work we have used dilute EY suspensions to study the impact of the heat treatment intensity on the emulsifying properties of EY (Guilmineau & Kulozik, submitted). This work has notably shown that the interfacial concentration

Corresponding author. Tel.: +49 (0)8161 71 3535; fax: +49 (0)8161 71 4384. E-mail address: ulrich.kulozik@wzw.tum.de (U. Kulozik).

of proteins at the O/W interface was signicantly increased with an increasing degree of protein denaturation. This was shown to allow a signicant decrease of the occulation in O/W emulsions made with denatured EY, and signicantly impacted the rheological properties and stability of the emulsions obtained. This work was, however, carried out on a model emulsion using constant environmental conditions, namely a pH of 6.5 and an ionic strength (IS) corresponding to 0.52 M NaCl. But the many applications of EY in industrial food products cover a much wider range of pH and ionic strengths, varying with the formulation of the products. The environmental conditions have been shown to have a signicant impact on the physico-chemical and emulsifying properties of EY and its components (Aluko & Mine, 1998; Anton & Gandemer, 1999; Le Denmat, Anton, & Beaumal, 2000; Mine, 1998). Fresh EY has a pH of about 6.2 and an IS of about 0.17 M NaCl (Anton, 1998). In these natural conditions, EY is constituted of a continuous aqueous phase containing about 80% of the total dry matter and referred to as plasma, and insoluble denser structures with a size ranging from 0.3 to 2 mm referred to as granules. EY plasma

0268-005X/$ - see front matter r 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.foodhyd.2005.12.006

ARTICLE IN PRESS
F. Guilmineau, U. Kulozik / Food Hydrocolloids 20 (2006) 11141123 1115

contains globular glycoproteins known as a-, b-, and glivetins, as well as the low-density lipoproteins (LDL), which apoproteins are called lipovitellenins. Granules contain a phosphoprotein known as phosvitin, as well as the high-density lipoproteins (HDL), which apoproteins are called lipovitellins (Burley & Vadehra, 1989). Upon increase of the IS to values above 0.3 M, granules dissociate because of the rupture of the phosphocalcic bridges between lipovitellins and phosvitins (Causeret, Matringe, & Lorient, 1991). This is a major change of the EY microstructure driven by the environmental conditions. Partial disruption of EY granules by mechanical shear forces has also been observed, leading to the formation of granule fragments, notably during high pressure homogenisation (Anton, Beaumal, & Gandemer, 2000). Most salad dressings are prepared at a pH around 4.0 by addition of acetic and/or citric acid, which contributes to the products characteristic taste and inhibits growth of most microorganisms (Ford, Borwankar, Pechak, & Schwimmer, 2004). This makes this value of low pH a particularly relevant one when it gets to studying the behaviour of EY in model food emulsions. In this work, model oil-in-water (O/W) emulsions containing 30% (v/v) sunower oil were prepared within a range of pH and IS most commonly occurring in food emulsions. The impact of two levels of pH, namely pH 4.0 and 6.5, and two levels of IS, namely 0.15 and 0.55 M NaCl, on the emulsifying properties of EY was studied. The objective of this work is to compare the behaviour of non-heated (native) EY to that of heat-treated EY in the various environmental conditions dened above. The level of protein denaturation used for the heat-treated EY was higher than that normally reached with industrial EY pasteurisation, and was chosen based on the results obtained in the rst part of this study (Guilmineau & Kulozik, submitted). The solubility of EY proteins as well as the key characteristics of the emulsions such as particle size distribution, occulation and interfacial protein concentration were investigated. Hypothesis regarding the colloidal interactions taking place between oil droplets in the various environmental conditions studied are discussed. 2. Material and methods 2.1. Preparation and heating of an EY suspension Freshly laid eggs from Lohman Tradition hens were collected from the Universitys research farm (Thalhausen) and used within 48 h after collection. A suspension containing 20% (w/w) pure EY in an isotonic NaCl solution (i.e. 0.17 M NaCl) is prepared as described by Guilmineau and Kulozik (submitted): it is referred to as the native EY20. 2.1.1. Thermal treatment of the EY20 suspension Native EY20 was heated as described by Guilmineau and Kulozik (submitted) with a holding time of 12 min at

74 1C: we obtained the heated EY20. Because of the dilution of the native EY prior to heating, the heated suspension remained uid despite the intense heat treatment. Heated EY could therefore be used just as easily as native EY to prepare O/W emulsions. 2.2. Preparation of emulsions 2.2.1. Preparation of the continuous phase The continuous phase was prepared by diluting EY20 (native or heated) so as to obtain a concentration of 1.12% (w/w) EY dry matter. Dilutions were made either in an aqueous NaCl solution (not buffered) or in a 0.01 M acetate buffer (Sodium acetate trihydrate/acetic acid) at pH 4.0, containing NaCl so as to obtain a concentration of either 0.15 or 0.52 M in the nal continuous phase. The continuous phases prepared without buffer were measured to have an average pH of 6.4970.04 and the ones made with the acetate buffer had an average pH of 4.0370.05. All continuous phases were allowed to equilibrate under mild agitation (magnetic stirrer) for 30 min before the production of the emulsion. Commercial rened sunower oil was used to prepare the emulsions (Cereol Deutschland GmbH, Mannheim, Germany). A pre-emulsion with a dispersed phase volume fraction j 0:3 was prepared by pouring 180 ml of sunower oil in 420 ml continuous phase while mixing with a dispersing system model Ultra Turrax T25 (IKA Werke, Staufen, Germany) equipped with an 18 mm diameter toothed-disc dispersing tool (model S25KR-18G) at 8000 rpm. The oil was added over the course of 1 min and the agitation was maintained for a further minute in order to obtain a homogeneous preemulsion. The median oil droplet diameter in the preemulsion was measured to be about d50,31316 mm throughout the trial period. The pre-emulsion was then further emulsied by passing it once through the rst stage of a high-pressure homogeniser model APV 1000 (Invensys APV, Albertslund, Denmark) at a pressure of 200 bars and a temperature of 25 1C. 2.3. Characterisation of the EY suspension and emulsions 2.3.1. Protein solubility The solubility of proteins in the EY samples was assessed using the method published by Morr et al. (1985) with the following modication. Samples of each of the four continuous phases prepared as described in Section 2.2 (two pH and IS levels) were equilibrated under mild agitation at 20 1C for 1 h. An aliquote of each suspension was kept for total protein content measurement (Pt), and the rest was centrifuged twice at 19,000g for 20 min in order to separate the insoluble proteins. An aliquote of the supernatant containing the soluble proteins was also analysed for protein content (Ps). The protein content was determined using the colorimetric method of Markwell, Hass, Bieber, and Tolbert (1978). The percentage of

ARTICLE IN PRESS
1116 F. Guilmineau, U. Kulozik / Food Hydrocolloids 20 (2006) 11141123

solubility was determined as given in Eq. (1) below. Solubility% Ps 100. Pt (1)

determine which means are signicantly different from others.

3. Results 2.3.2. Particle size distribution and occulation measurement All particle size measurements were carried out using a laser diffraction spectroscope model LS230 (BeckmanCoulter, Germany). Details of the method used are given by Guilmineau and Kulozik (submitted). 2.3.3. Stability to creaming The stability to creaming is measured by following the formation of a clear droplet-free phase at the bottom of a sample of emulsion, using a light scattering optical analyser model Turbiscan MA1000 (Formulaction, France). All details regarding the determination of the initial creaming rate and relative height of the cream layer formed during 1 week of storage at 10 1C are given by Guilmineau and Kulozik (submitted). 2.3.4. Rheological characterisation of the emulsions A controlled shear rate rheometer model Rheomat 115 equipped with a double-gap geometry model MS 0/115 (Contraves GmbH, Stuttgart, Germany) was used. The details of the method used are given by Guilmineau and Kulozik (submitted). 2.3.5. Interfacial protein concentration The method used in this study to separate the oil droplets from the continuous phase of the emulsions was adapted from Patton and Huston (1986). All details of the method used are described by Guilmineau and Kulozik (submitted). 2.3.6. Zeta-potential measurements Zeta potential was measured by a laser light-scattering technique using a Zetasizer Nano-Z model ZEN 2600 (Malvern Instruments, Herrenberg, Germany). The drift velocity of oil droplets placed in an electric eld is accurately determined from the Doppler shift of the light scattered by the moving droplets. The emulsions were diluted 10,000 fold (v/v) in a buffer solution corresponding to their continuous phase (same pH and NaCl concentration). All measurements were carried out at a constant temperature of 25 1C. 2.3.7. Statistical analysis Three replicates were carried out on three different weeks. Oil droplet size distribution, occulation, rheological properties, stability, interfacial protein concentration and zeta-potential were measured each time twice. These parameters were subjected to a one-way analysis of variance using Statgraphics software (Statistical Graphics Corporation, Rockville, MD, USA) with a condence level of 95% (po0:05). A multiple range test was used to The solubility of native EY protein is signicantly impacted by the pH and IS (Fig. 1). The protein solubility is lowest at pH 4, independently from the IS. At pH 6.5, virtually all proteins are soluble at an IS of 0.52 M NaCl, whereas only about two thirds are soluble at low IS (0.15 M NaCl). The protein solubility in heated EY is always much lower than in native EY. The inuence of the environmental conditions follows the same trend with heated EY as with the native one. Thus, protein solubility in heated EY is also maximum at pH 6.5 and high IS, and minimal at pH 4. Le Denmat et al. (2000) have shown that in native EY, plasma proteins have a very high solubility over a wide range of pH and IS. On the other hand, granules were shown to be insoluble at a pH of 3, and only partially soluble in the physicochemical conditions occuring in native EY (i.e. pH of 6.5 and IS of 0.15 M NaCl). Our results are consistent with those presented by Le Denmat et al. (2000) regarding the solubility of EY proteins. The low solubility measured at pH 4 is explained by the insolubility of granules at this low pH (Fig. 1). At a pH of 6.5 and 0.15 M NaCl, granules are mostly insoluble because their constituents are aggregated via Ca2+ bridging between negatively charged phosphoserine residues of HDL and phosvitin (Causeret et al., 1991). Increasing the IS from 0.15 to 0.55 M NaCl leads to the disruption of phosphocalcic bridges resulting from the displacement of divalent

120 Native Heated

100

Protein solubility (%)

80

60

40

20

0 pH [NaCl]

4 0.15 M

4 0.52 M

6.5 0.15 M

6.5 0.52 M

Fig. 1. Impact of pH and NaCl concentration on the solubility of proteins in native (non-heated) and heated (i.e. 74 1C for 12 min.) EY suspensions.

ARTICLE IN PRESS
F. Guilmineau, U. Kulozik / Food Hydrocolloids 20 (2006) 11141123 1117

Ca2+ cations by an excess of monovalent Na+. This induces the dissociation of granules, and therefore an increased level of protein solubility. The lower levels of protein solubility measured in heated EY suspensions reect the denaturation of some EY proteins which associate to form insoluble protein aggregates (see Guilmineau & Kulozik, submitted). In heated samples at a pH of 6.5, an increase of the IS from 0.15 to 0.55 M also leads to an increase of the level of solubility from 25% to 37%. This indicates that some granule proteins are still native after the heat treatment, and can therefore still be dissociated by addition of NaCl. This is consistent with observations that granule proteins are more resistant to heat than plasma proteins (Anton, Le Denmat, & Gandemer, 2000; Le Denmat, Anton, & Gandemer, 1999). The median oil droplet diameter obtained in O/W emulsions prepared with native and heated EY is signicantly inuenced by the environmental conditions (Fig. 2). The smallest droplets are formed at a pH of 6.5 and an IS of 0.52 M NaCl, whereas the largest are formed at a pH of 4 and an IS of 0.15 M NaCl. The heat treatment seems to have a rather small impact on the size of oil droplets achieved, except at a pH of 4 and a low IS where heated EY allows to obtain much smaller droplets as native EY. The interfacial protein concentration in emulsions prepared with native EY is signicantly higher at a pH of 6.5 than at a pH of 4 (Fig. 3). The IS does not seem to inuence the interfacial protein load at pH 6.5, but at pH 4 the interfacial load is lower at an IS of 0.52 M NaCl than at
8

4.0 Native Heated

3.5

Interfacial protein load [mg.m-2]

3.0

2.5

2.0

1.5

1.0

0.5

0.0 pH [NaCl]

4 0.15 M

4 0.52 M

6.5 0.15 M

6.5 0.52 M

Fig. 3. Impact of pH and NaCl concentration on the interfacial protein concentration measured in O/W emulsions containing native (i.e. nonheated) and heated (i.e. 74 1C for 12 min.) egg yolk.

7 median droplet diameter d50.3 (m)

Native Heated

6 5

2 1

0 pH [NaCl]

4 0.15 M

4 0.52 M

6.5 0.15 M

6.5 0.52 M

Fig. 2. Impact of pH and NaCl concentration on the median oil droplet diameter (d50.3) achieved in O/W emulsions containing native (i.e. nonheated) and heated (i.e. 74 1C for 12 min.) egg yolk.

an IS of 0.15 M NaCl. On the other hand, when heated EY is used, the interfacial protein concentration is much higher than when native EY is used, under all conditions tested. Moreover, there does not seem to be any negative impact of the environmental conditions on the interfacial protein load, since the differences measured between the various conditions tested are not signicant. The zeta-potential of oil droplets in the emulsion is signicantly impacted by the environmental conditions (Fig. 4). Droplets carry a positive charge at a pH of 4 and a negative one at a pH of 6.5. Increasing the concentration of NaCl at a pH of 4 leads to a decrease of the zeta-potential, which is due to the enhanced screening of the adsorbed proteins positive charges by negative counter ions. However, at a pH of 6.5, increasing the IS leads to an increase of the negative zeta-potential, despite the increased screening by positive ions from the dissolved NaCl. This suggests that the negative charge density of the interface is greatly increased by the addition of NaCl, in a way, which more than compensates the screening effect. This may be due to an increased adsorption of the strongly negative phosvitin, following the dissociation of granules taking place at high pH and IS. A heat treatment of the EY did not have any signicant impact on the zeta-potential obtained at the oil droplet surface, regardless of the environmental conditions. The values of zeta-potential obtained with native EY are consistent with measurements from Le Denmat et al. (2000). The zeta-potential does not exceed 10 mV in all environmental conditions tested, which is always well

ARTICLE IN PRESS
1118 F. Guilmineau, U. Kulozik / Food Hydrocolloids 20 (2006) 11141123

15 Native Heated

5 Native Heated 4

10

Zeta potential (mV)

Flocculation factor (-) 4 0.52 M 6.5 0.15 M 6.5 0.52 M

-5

-10 1 -15 4 pH [NaCl] 0.15 M

Fig. 4. Impact of pH and NaCl concentration on the zeta potential of oi droplets taken from O/W emulsions containing native (i.e. non-heated) and heated (i.e. 74 1C for 12 min.) egg yolk.

0 4 pH [NaCl] 0.15 M

4 0.52 M

6.5 0.15 M

6.5 0.52 M

Fig. 5. Impact of pH and NaCl concentration on the occulation factor measured in O/W emulsions containing native (i.e. non-heated) and heated (i.e. 74 1C for 12 min.) egg yolk.

below the 20 mV evaluated by Friberg (1997) to produce effective electrostatic stabilisation. The charge of the interface changes from a slightly positive charge at pH 4.0 to a slightly negative one at pH 6.5, which is consistent with the expected variation of EY protein charges when predicted from their isoelectric points. Adsorption of heated EY proteins did not affect the charge of the interfacial lm. Similar observations were made with caseins, for which it was reported that a change in the geometry of the protein had relatively small effects upon the zeta-potential of the emulsion droplets (Dalgleish, 1999). This suggests that despite the increased proportion of denatured proteins present at the interface, the adsorbed layer adopts a conformation, which maintains the same charge density as the proteins in the native form. Both the environmental conditions and a heat treatment were measured to have a very signicant impact on the occulation of oil droplets (Fig. 5). The occulation factor of emulsions prepared with native EY is always higher than that of emulsions prepared with heated EY, at all environmental conditions tested. With native EY, the IS seems to drive the occulation, whereby a high IS leads to the largest level of occulation, regardless of the pH. At low IS, a pH of 6.5 leads to a signicantly higher level of occulation than that obtained at a pH of 4. When heated EY is used as opposed to native one, the level of occulation in emulsions containing 0.52 M NaCl is greatly reduced, down to a level similar to that obtained at low IS. The occulation factor seems to be less dependant on the environmental conditions, although emulsions prepared at a pH of 6.5 tend to be slightly more occulated than those prepared at a pH of 4.

25 Native Heated 20 Consistency index K (mPa.sn)

15

10

0 4 pH [NaCl] 0.15 M

4 0.52 M

6.5 0.15 M

6.5 0.52 M

Fig. 6. Impact of pH and NaCl concentration on the consistency index (K) of O/W emulsions containing native (i.e. non-heated) and heated (i.e. 74 1C for 12 min.) egg yolk.

The consistency index (K, according to the HerschellBulkley model) of the emulsions made with native EY is driven by the IS, whereby a high IS promotes a high value of K, regardless of the pH (Fig. 6). When heated EY is

ARTICLE IN PRESS
F. Guilmineau, U. Kulozik / Food Hydrocolloids 20 (2006) 11141123 1119

1.1 Native Heated

7 Native Heated

1.0

0.8

Initial creaming rate (%/h) 4 0.15 M 4 0.52 M 6.5 0.15 M 6.5 0.52 M

0.9 Flow index (-)

0.7

0.6

0.5

0.4 pH [NaCl]

0 4 pH [NaCl] 0.15 M

4 0.52 M

6.5 0.15 M

6.5 0.52 M

Fig. 7. Impact of pH and NaCl concentration on the ow index of O/W emulsions containing native (i.e. non-heated) and heated (i.e. 74 1C for 12 min.) egg yolk.

Fig. 8. Impact of pH and NaCl concentration on the initial creaming rate (rst 611 h after emulsication) measured in O/W emulsions containing native (i.e. non-heated) and heated (i.e. 74 1C for 12 min.) egg yolk.

used, the consistency of the product is less impacted by the environmental conditions as when native EY is used. All emulsions prepared have a ow index comprised between 0.9 and 1.0, which reects the near-newtonian ow behaviour of these uid emulsions. The ow index of emulsions made with native EY is lowest at high IS (Fig. 7), which is when the consistency index is the highest (Fig. 6). At low IS, we observe that the ow index is signicantly higher at a pH of 6.5 than at a pH of 4.0. However, when heated EY is used, the ow index of the emulsion is much less impacted by the environmental conditions than when native EY is used, and averages around a value of 0.95 in all tested environments. For the emulsion prepared with native EY, we measured that the creaming rate of oil droplets is much higher at pH 4 and 0.15 M NaCl than in any other environmental conditions (Fig. 8). However, in emulsions prepared with heated EY, there is no signicant impact of the environmental conditions on the creaming rate. The creaming rate is always slightly lower when heated EY is used as emulsier rather than native EY. The relative cream height reported here after 1-week storage, when the level of the cream is stabilised, reects the ability of oil droplets to pack in an efcient manner under normal gravity (Fig. 9). When native EY is used, the tightest packing of droplets is achieved at pH 4 with an IS of 0.15 M, and the most lose packing at a pH of 6.5 and an IS of 0.52 M. However, when heated EY is used, the nal level of creaming is not inuenced by the environmental conditions. The cream obtained is rather more compact than that obtained with native EY, except at pH 4 and 0.15 M NaCl.

80 70 RCH after 1 week storage (%) 60 50 40 30 20 10 0 4 pH [NaCl] 0.15 M Native Heated

4 0.52 M

6.5 0.15 M

6.5 0.52 M

Fig. 9. Impact of pH and NaCl concentration on the relative cream height (RCH) reached 1 week after emulsication measured in O/W emulsions containing native (i.e. non-heated) and heated (i.e. 74 1C for 12 min.) egg yolk.

4. Discussion The discussion aims at comparing the results of this work to those obtained in similar conditions with native EY by other authors and focuses on explaining the results obtained in the present study regarding the impact of heat denaturation of EY.

ARTICLE IN PRESS
1120 F. Guilmineau, U. Kulozik / Food Hydrocolloids 20 (2006) 11141123

4.1. Thermally aggregated EY proteins can adsorb at the O/W interface independently from the environmental conditions 4.1.1. Emulsifying activity (EA) of native and heat-treated EY Le Denmat et al. (2000) found that the size of oil droplets formed in emulsions containing native EY was independent from the environmental conditions. The energy used by Le Denmat et al. (2000) to disperse the oil by high-pressure homogenisation was much higher than the one used in the present work, and the amount of EY proteins in the continuous phase was 25 mg/ml, which is about 8 times more than the one we used (i.e. 3.2 mg/ml). As could be expected, Le Denmat et al. (2000) obtained an average oil droplet size about 10 times smaller than the one we achieved in the present work. They, however, clearly demonstrated a decreased EA of granules at a pH of 3, which was attributed to the poor solubility of granule proteins at this pH. The authors found that this did not affect the EA of complete EY, since the same mean droplet diameter d 32 0:42 mm was achieved in all environmental conditions. It was concluded that the emulsifying properties of EY were driven by plasma proteins. However, in our study, where we dispersed the oil using a comparatively lower energy (i.e. single pass through one stage homogeniser at 200 bar), we observed an increased d50.3 in emulsions prepared at a pH of 4 and 0.15 M NaCl, indicating a decreased EA (Fig. 2). This could be due to the fact that the dispersing forces generated during emulsication were insufcient to allow a shear-driven dissociation of granule proteins at this low pH and IS. Anton, Beaumal, et al. (2000) showed that insoluble EY granules do adsorb at the O/W interface, and suggested that fragments of granules obtained during homogenisation can also play the same role. As the intensity of disruption forces increases, the partial dissociation of granules could lead to a more efcient adsorption of granule fragments at the O/W interface, and an overall better surface activity of EY. We observed that the EA of EY in our study was high at a pH of 4 and 0.55 M NaCl (Fig. 2), despite the low protein solubility (Fig. 1). This seems to indicate that the presence of high concentrations of NaCl facilitates the shearinduced dissociation of insoluble EY granules at pH 4, even when relatively low shear forces are used. When an EY suspension is heated, part of the granule proteins are denatured and form heterogeneous aggregates together with denatured EY plasma proteins. Previous work has shown that the heat treatment applied in the present study leads to the thermal aggregation of just over 60% of the total EY proteins (Guilmineau & Kulozik, submitted). The forces bonding proteins within heatcoagulated EY has been shown to be mostly hydrophobic interactions (Kiosseoglou & Paraskevopoulou, 2005). These interactions are quite weak and can probably be dissociated by the shear forces present in the conditions used for emulsication (i.e. 200 bar, one stage homogenisa-

tion). The interfacial lm in emulsions prepared with heattreated EY could therefore be formed by the adsorption of a multitude of minute fragments of thermally aggregated proteins, obtained by the shear-driven disruption of much larger protein aggregates during high pressure homogenisation. This hypothesis is supported in the present work by the fact that denatured EY proteins are adsorbed at the O/ W interface, even though they form insoluble protein aggregates in the continuous phase. This change in the nature of the interfacial lm when using heated EY can explain the fact that the EA of EY at a pH of 4 and 0.15 M NaCl is better when heated EY is used as opposed to native EY (Fig. 2). This result suggests that the shear intensity obtained during homogenisation in the conditions of this study was sufcient to allow a thorough dissociation of thermally aggregated EY proteins, although not sufcient to dissociate insoluble native granules at low pH and IS. 4.1.2. Interfacial protein concentration under various environmental conditions The isoelectric region (pI) of apo-LDL has been measured between pH 6.5 and 7.3 (Kojima & Nakamura, 1985; Nakamura, Hayakawa, & Sato, 1977), and that of apo-HDL, although not characterised precisely, is assumed to be in the enlarged neutral pH region as well (Le Denmat et al., 2000). The increased interfacial protein concentration observed with native EY at pH 6.5 (Fig. 3) can be explained by the proximity of this pH to the pI of apo-LDL and apo-HDL. The decreased charge density of proteins around their pI favours their tight re-arrangement leading to the formation of dense interfacial lms, which was notably observed with bovine serum albumin (Das & Chattoraj, 1980). This work shows that the interfacial lm formed with heated EY contains a greater concentration of protein than the lm formed with native proteins (Fig. 3). This reects the formation of a thick lm of aggregated proteins in emulsions made with heated EY. During high-pressure homogenisation, the protein aggregates formed during heating could be broken down by the shear forces into a multitude of smaller aggregates, which seem to adsorb favourably at the O/W interface. It has indeed been reported that the turbulences obtained in high-pressure homogenisation favour the adsorption of large protein aggregates because the convective mass transport rate increases with the size of molecules (Walstra, 1983). Furthermore, the mechanical disruption of protein aggregates held together by hydrophobic interactions probably leads to the exposure of many hydrophobic sites. The increased surface hydrophobicity of the protein fragments could greatly contribute to their rapid adsorption at the O/W interface. After adsorption, interactions between neighbouring aggregates could lead to the formation of a thick and cohesive interfacial lm, possibly forming multiple layers. The adsorption of protein aggregates occurs at all environmental conditions tested, without being negatively impacted by the environment.

ARTICLE IN PRESS
F. Guilmineau, U. Kulozik / Food Hydrocolloids 20 (2006) 11141123 1121

4.2. The adsorption of denatured EY protein aggregates reduces the impact of the environmental conditions on the emulsifying properties of EY 4.2.1. Decreased level of occulation in emulsions made with heated EY The EY proteins forming the interfacial lm are charged polymers, and therefore tend to stabilise emulsion droplets against aggregation through a combination of electrostatic and steric repulsions (Claesson, Blomberg, Fro berg, Nylander, & Arnebrant, 1995). The principal difference between these two types of interactions is their sensitivity to pH and IS (Hunter, 2000). The electrostatic repulsion between emulsion droplets is dramatically decreased when the electrical charge on the droplet surface is reduced (e.g. by altering the pH) or screened (e.g. by increasing the concentration of electrolyte in the aqueous phase). In contrast, steric repulsion is fairly insensitive to both electrolyte concentration and pH (McClements, 1999). The stability of the emulsion, and particularly the tendency for oil droplets to occulate, depends on the intensity of the steric and electrostatic repulsive forces in relation to the attractive forces (e.g. van der Waals, hydrophobic interactions). The magnitude and range of electrostatic repulsion between droplets decrease as the IS of the solution separating them increases because of electrostatic screening. This explains the susceptibility of protein-stabilised emulsions to occulation when electrolyte concentration is increased above a critical level (Demetriades, Coupland, & McClements, 1997). Electrostatic screening can explain the increased occulation at a high IS (0.55 M NaCl) when native EY is used (Fig. 5). At low IS, the lower level of occulation observed at pH 4.0 compared to pH 6.5 is probably due to an increased electrostatic repulsion obtained at pH 4.0 because of the increased interfacial charge density. As already discussed, it appears that the pI of the proteins, which are most active at the interface is close to pH 6, so that their charge density is higher at pH 4 than at pH 6.5. This difference is clearly shown at low IS by the measurement of the zeta-potential of oil droplets (Fig. 4). The decreased level of occulation obtained at high IS when heated EY is used as emulsier indicates a change in the stabilisation mechanism. Indeed, the absence of modication of the zeta-potential when a heat-treatment is applied to the EY, rules out any explanation of the occulation behaviour based on electrostatic interactions. However, a change of the steric interaction is supported by the increase of interfacial protein concentration measured when heated EY is used instead of native EY (Fig. 3). Since the strength of steric repulsions imposed by a polymer when adsorbed at the interface partially depends on the amount of polymer adsorbed (Hunter, 2000), it is expected that the contribution of the steric repulsion to the overall repulsion forces is increased when heated EY is used. The impact of the electrostatic screening which appears to drive the occulation when native EY is used seems to be overridden by the increased steric repulsion when heated

EY is used. These results illustrate the fact that an increased steric contribution over the electrostatic interactions leads to a decreased sensitivity of an emulsions stability to changes of environmental conditions. 4.2.2. Levelling of the rheological properties of the emulsions It appears that the consistency index of the emulsion containing native EY is high when the emulsion is occulated, which is the case at an IS of 0.55 M NaCl (Fig. 6). This can be explained by an increase of the effective volume of the dispersed phase in the occulated emulsion which is due to the continuous phase trapped between oil droplets within the ocs (Dickinson & Stainsby, 1982). The emulsions having a high degree of occulation at high IS also display the lowest values of ow index, which indicates a higher level of pseudoplasticity (Fig. 7). This reects the deocculation of the droplets under the effect of increased shear rate (Bower, Washington, & Peruwal, 1997), and indicates that the ocs formed at high IS are sensitive to shear forces. When heated EY is used, the deocculation of the emulsion at high IS leads to a decrease of its consistency index. We also measure a slight pseudoplasticity in emulsions made with heated EY. This could be due to the presence of EY protein aggregates in the continuous phase, which orientation in the shear eld (with possible dissociation) leads to a decreased resistance to ow at increased shear rate. 4.2.3. Improved creaming stability and formation of a compact cream In emulsions made with native EY, the initial creaming rate seems to be correlated to the median oil droplet diameter (Figs. 8 and 2, respectively), which is consistent with the expected impact of the droplet size on the creaming velocity, as described for example in Stokes law. Given the intermediate oil droplet concentration in our model system (j 0:3), droplet occulation is expected to increase the creaming velocity because the ocs have a larger effective size than the individual droplets. However, we did not observe any correlation between the degree of occulation and the creaming rate of the emulsions. It should be noted that the emulsions were stirred vigorously prior to the beginning of the creaming rate measurement, which probably led to some degree of droplet deocculation. The initial creaming rate in emulsions prepared with heated EY does not seem to be impacted by the oil droplet size as much as those prepared with native EY. Despite signicant differences in oil droplet size with a median oil droplet diameter varying between 3.4 and 4.3 mm, the initial creaming rate of oil droplets in emulsions made with heated EY was constant at about 1% creaming per hour. This result shows that the modication of the interfacial lm composition and density-taking place when heated EY is used has a positive effect on the stability of emulsions to creaming by

ARTICLE IN PRESS
1122 F. Guilmineau, U. Kulozik / Food Hydrocolloids 20 (2006) 11141123

decreasing the creaming rate. This could be due to an increased density of the oil droplets under the effect of the increased density of the interfacial lm (Fig. 3). The adsorption of dense material at the O/W interface can signicantly increase the density of oil droplets, and therefore slow down the creaming (Tan, 1990). This effect is particularly signicant for small oil droplets. In this work, the creaming rate was estimated by observing the kinetic displacement of a front line separating the oating fat phase from the sedimenting fat-free phase in a standing sample of the emulsion. The position of the front moving from the bottom of the tube upwards reects the displacement of the small droplets of the distribution, since they are the slowest to migrate. It is therefore expected that this method would allow to detect changes in the density of oil droplets, which is assumed to be the case in this work. As highlighted by McClements (1999), the thickness of the creamed layer formed after complete creaming of the dispersed phase depends on the effectiveness of droplet packing and is impacted by the same factors as those determining the structure of ocs. When native EY is used, there seems to be a correlation between the median oil droplet size and the compactness of the cream layer (Fig. 9). This can be due to the fact that when the droplet size decreases at constant oil phase volume ratio, the number of droplets and therefore the interfacial area increases as well, so that more continuous phase remains trapped between the oil droplets, giving a less compact cream. Notice that emulsions made with native EY have a tendency to occulate more than the ones made with heated EY, which reects the presence of greater attraction forces between oil droplets. This can lead to oil droplets sticking together when coming in contact during creaming, and therefore forming a cream layer with a relatively open structure. This could explain why emulsions containing denatured EY, which tend to be less occulated, also tend to form a more compact cream layer than those made with native EY (Fig. 9). The only environmental condition in which the cream made with native EY is more compact than that obtained with heated EY (i.e pH 4 and 0.15 M NaCl) is also that for which the occulation factor is minimal, which is consistent with the hypothesis formulated above. It appears that when attraction forces between droplets are weak, the droplets are able to roll around each other more freely and can therefore pack more closely together. Visual observations revealed that the cream formed in emulsions prepared with native EY was cohesive and formed a lump, which did not re-disperse easily under mild agitation. On the other hand, the cream formed in emulsions prepared with heated EY could easily be evenly re-dispersed by simple manual agitation, which can be an advantage in products, which have to be shaken just before use. 5. Conclusions This work has shown that the emulsifying activity of thermally denatured EY is similar or even better than that

of native EY, depending on the environmental conditions. Using heated EY rather than native one led to a signicant increase of the interfacial protein concentration under all environmental conditions. This demonstrates the ability for thermally denatured insoluble EY protein aggregates to adsorb efciently at the O/W interface. EY protein aggregates formed during heating are thought to be disrupted by the high shear forces occurring in highpressure homogenisation. The resulting micro-particles of aggregated proteins would constitute the interfacial lm when heat-treated EY is used as emulsier. The adsorption of denatured EY proteins was shown to greatly decrease the occulation of oil droplets, particularly at high ionic strength. This is thought to be due to an increased contribution of the steric repulsions between droplets when protein aggregates are adsorbed at the interface. The rheological properties as well as the stability of the emulsions against creaming were shown to be much less sensitive to variations of the environmental conditions when heated EY was used rather than native EY. Further work is required to study the impact of the environmental conditions on the composition of the interfacial lm formed with both native and denatured EY. Acknowledgement The authors thank Jan Peter Luh and Mario Henrique Martinez Marins for their valued contribution in the experimental phase of this work. References
Aluko, R. E., & Mine, Y. (1998). Characterization of oil in water emulsions stabilized by hens egg yolk granule. Food Hydrocolloid, 12, 203210. Anton, M. (1998). Structure and functional properties of hen egg yolk constituents. In S. G. Pandalai (Ed.), Recent research developments in agricultural and food chemistry, Vol. 2 (pp. 839864). Trivandrum, India: Research Signpost. Anton, M., Beaumal, V., & Gandemer, G. (2000). Adsorption at the oilwater interface and emulsifying properties of native granules from egg yolk: Effect of aggregated state. Food Hydrocolloids, 14, 327335. Anton, M., & Gandemer, G. (1999). Effect of pH on interface composition and on quality of oil-in-water emulsions made with hen egg yolk. Colloids and Surfaces B, 12, 351358. Anton, M., Le Denmat, M., & Gandemer, G. (2000). Thermostability of hen egg yolk granules: Contribution of native structure of granules. Journal of Food Science, 65(4), 581584. Bower, C., Washington, C., & Peruwal, T. S. (1997). The use of image analysis to characterize aggregates in a shear eld. Colloids and Surfaces A, 127, 105112. Burley, R. W., & Vadehra, D. V. (1989). The Avian egg: Chemistry and biology. New York: Wiley. Causeret, D., Matringe, E., & Lorient, D. (1991). Ionic strength and pH effect on composition and microstructure of granules. Journal of Food Science, 56, 15321536. Claesson, P. M., Blomberg, E., Fro berg, J. C., Nylander, T., & Arnebrant, T. (1995). Protein interactions at solid surfaces. Advances in Colloid and Interface Science, 57, 161227. Cunningham, F. E. (1995). Chapter 12: Egg-product pasteurization. In W. J. Stadelman, & O. J. Cotterill (Eds.), Egg science and technology (pp. 289321). New York: Haworth Food Products Press.

ARTICLE IN PRESS
F. Guilmineau, U. Kulozik / Food Hydrocolloids 20 (2006) 11141123 Dalgleish, D. G. (1999). Interfacial structures and colloidal interactions in protein-stabilised emulsions. In E. Dickinson, & J. M. Rodriguez Patino (Eds.), Food emulsions and foamsInterfaces, interactions and stability (pp. 116). Cambridge: The Royal Society of Chemistry. Das, K. P., & Chattoraj, D. K. (1980). Adsorption of proteins at the polar oilwater interface. Journal of Colloid and Interface Science, 78(2), 422429. Demetriades, K., Coupland, J. N., & McClements, D. J. (1997). Physicochemical properties of whey protein-stabilized emulsions as affected by heating and ionic strength. Journal of Food Science, 62(3), 462467. Dickinson, E., & Stainsby, G. (1982). Colloids in food. London: Applied Science Publishers. Ford, L. D., Borwankar, R. P., Pechak, D., & Schwimmer, B. (2004). Chapter 13: Dressings and sauces. In S. E. Friberg, K. Larsson, & J. Sjo blom (Eds.), Food emulsions (pp. 525572). New York: Marcel Dekker, Inc. Friberg, S. E. (1997). Chapter 1: Emulsion stability. In S. E. Friberg, & K. Larsson (Eds.), Food emulsions (pp. 155). New York: Marcel Dekker, Inc. Guilmineau, F., Kulozik, U. (submitted). Impact of e thermal treatment on the emulsifying properties of egg yolk. Part 1: Effect of the treatment intensity. Food Hydrocolloids, submitted for peer-review. Hunter, R. J. (2000). Introduction to modern colloid science. Oxford, New York: Oxford University Press. Kiosseoglou, V. D., & Paraskevopoulou, A. (2005). Molecular interactions in gels prepared with egg yolk and its fractions. Food Hydrocolloids, 19, 527532. Kojima, E., & Nakamura, R. (1985). Heat gelling properties of hens egg yolk LDL in the presence of other proteins. Journal of Food Science, 50, 6366. 1123 Le Denmat, M., Anton, M., & Beaumal, V. (2000). Characterisation of emulsion properties and of interface composition in O/W emulsions prepared with hen egg yolk, plasma and granules. Food Hydrocolloids, 14, 539549. Le Denmat, M., Anton, M., & Gandemer, G. (1999). Protein denaturation and emulsifying properties of plasma and granules of egg yolk as related to heat treatment. Journal of Food Science, 64, 194197. Markwell, M. A., Hass, S. M., Bieber, L. L., & Tolbert, N. E. (1978). A modication of the lowry procedure to simplify protein determination in membrane and lipoprotein samples. Analytical Biochemistry, 87, 206210. McClements, D. J. (1999). Food emulsionsPrinciples, practice, and techniques. London, New York: CRC Press. Mine, Y. (1998). Emulsifying characterization of hens egg yolk proteins in oil-in-water emulsions. Food Hydrocolloids, 12, 409415. Morr, C. V., German, B., Kinsella, J. E., Regenstein, J. M., Van Buren, J. P., Kilara, A., et al. (1985). A collaborative study to develop a standardized food protein solubility procedure. Journal of Food Science, 50(6), 17151718. Nakamura, R., Hayakawa, R., & Sato, Y. (1977). Isolation and fractionation of the protein moiety of egg yolk low density lipoprotein. Poultry Science, 56, 11481152. Patton, S., & Huston, G. E. (1986). A method for isolation of milk fat globules. Lipids, 21(2), 170174. Tan, C. T. (1990). Beverage emulsions. In K. Larsson, & S. E. Friberg (Eds.), Food emulsions (pp. 445478). New York: Marcel Dekker, Inc. Walstra, P. (1983). Formation of emulsions. In P. Becher (Ed.), Encyclopedia of emulsion technology: Basic theory, Vol. 1 (pp. 57127). New York: Marcel Dekker Inc.

Vous aimerez peut-être aussi