The dilatable membrane of oleosomes (lipid droplets) allows their in vitro resizing and triggered release of lipids.

It has been reported that lipid droplets (LDs), called oleosomes, have an inherent ability to inflate or shrink when absorbing or fueling lipids in the cells, showing that their phospholipid/protein membrane is dilatable. This property is not that common for membranes stabilizing oil droplets and when well understood, it could be exploited for the design of responsive and metastable droplets. To investigate the nature of the dilatable properties of the oleosomes, we extracted them from rapeseeds to obtain an oil-in-water emulsion. Initially, we added an excess of rapeseed oil in the dispersion and applied high-pressure homogenization, resulting in a stable oil-in-water emulsion, showing the ability of the molecules on the oleosome membrane to rearrange and reach a new equilibrium when more surface was available. To confirm the rearrangement of the phospholipids on the droplet surface, we used molecular dynamics simulations and showed that the fatty acids of the phospholipids are solubilized in the oil core and are homogeneously spread on the liquid-like membrane, avoiding clustering with neighbouring phospholipids. The weak lateral interactions on the oleosome membrane were also confirmed experimentally, using interfacial rheology. Finally, to investigate whether the weak lateral interactions on the oleosome membrane can be used to have a triggered change of conformation by an external force, we placed the oleosomes on a solid hydrophobic surface and found that they destabilise, allowing the oil to leak out, probably due to a reorganisation of the membrane phospholipids after their interaction with the hydrophobic surface. The weak lateral interactions on the LD membrane and their triggered destabilisation present a unique property that can be used for a targeted release in foods, pharmaceuticals and cosmetics.

On the Emulsifying Properties of Self-Assembled Pea Protein Particles.

Pea proteins are promising oil-in-water emulsifying agents at both neutral and acidic conditions. In an acidic environment, pea proteins associate to form submicrometer-sized particles. Previous studies suggested that the emulsions at acidic pH were stabilized due to a Pickering mechanism. However, protein particles can be in equilibrium with protein molecules, which could play a significant role in the stabilization of emulsion droplets. Therefore, we revisited the emulsion stabilization mechanism of pea proteins at pH 3 and investigated whether the protein particles or the protein molecules are the major emulsifying agent. The theoretical and experimental surface load of dispersed oil droplets were compared, and we found that protein particles can cover only 3.2% of the total oil droplet surface, which is not enough to stabilize the droplets, whereas protein molecules can cover 47% of the total oil droplet surface. Moreover, through removing protein particles from the mixture and emulsifying with only protein molecules, the contributions of pea protein molecules to the emulsifying properties of pea proteins at pH 3 were evaluated. The results proved that the protein molecules were the primary stabilizers of the oil droplets at pH 3.

Stability properties of surfactant-free thin films at different ionic strengths: measurements and modeling.

Foam lamellae are the smallest structural elements in foam. Such lamellae can experimentally be studied by analysis of thin liquid films in glass cells. These thin liquid films usually have to be stabilized against rupture by surface active substances, such as proteins or low molecular weight surfactants. However, horizontal thin liquid films of pure water with a radius of 100 µm also show remarkable stability when created in closed Sheludko cells. To understand thin film stability of surfactant-free films, the drainage behavior and rupture times of films of water and NaCl solutions were determined. The drainage was modeled with an extended Derjaguin-Landau-Verwey-Overbeek (DLVO) model, which combines DLVO and hydrophobic contributions. Good correspondence between experiment and theory is observed, when hydrophobic interactions are included, with fitted values for surface potential (ψ(0,water)) of -60 ± 5 mV, hydrophobic strength (B(hb,water)) of 0.22 ± 0.02 mJ/m(2), and a range of the hydrophobic interaction (λ(hb, water)) of 15 ± 1 nm in thin liquid films. In addition, Vrij's rupture criterion was successfully applied to model the stability regions and rupture times of the films. The films of pure water are stable over long time scales (hours) and drain to a final thickness >40 nm if the concentration of electrolytes is low (resistivity 18.2 MQ). With increasing amounts of ions (NaCl) the thin films drain to <40 nm thickness and the rupture stability of the films is reduced from hours to seconds.

Disintegration of protein microbubbles in presence of acid and surfactants: a multi-step process.

The stability of protein microbubbles against addition of acid or surfactants was investigated. When these compounds were added, the microbubbles first released the encapsulated air. Subsequently, the protein shell completely disintegrated into nanometer-sized particles. The decrease in the number of intact microbubbles could be well described with the Weibull distribution. This distribution is based on two parameters, which suggests that two phenomena are responsible for the fracture of the microbubble shell. The microbubble shell is first weakened. Subsequently, the weakened protein shell fractures randomly. The probability of fracture turned out to be exponentially proportional to the concentration of acid and surfactant. A higher decay rate and a lower average breaking time were observed at higher acid or surfactant concentrations. For different surfactants, different decay rates were observed. The fact that the microbubble shell was ultimately disintegrated into nanometer-sized particles upon addition of acid or surfactants indicates that the interactions in the shell are non-covalent and most probably hydrophobic. After acid addition, the time at which the complete disintegration of the shell was observed coincided with the time of complete microbubble decay (release of air), while in the case of surfactant addition, there was a significant time gap between complete microbubble decay and complete shell disintegration.

Analysis of light scattered by turbid media in cylindrical geometry.

The angle dependence of the transmitted light through a cylindrical turbid sample (latex suspension, developing milk gel, draining/coarsening milk, and protein foams) in a standard light scattering setup was analyzed in terms of the transport mean free path length or scattering length l* (a measure for the turbidity) and the absorption length labs. By variation of the concentration of an absorbing dye, the independence of l* and labs was demonstrated. The resulting value of the specific extinction coefficient of the dye was found to be in fair agreement with direct spectroscopic determination and practically identical in milk and latex suspensions. The validity of this technique for obtaining l* was demonstrated by monitoring the acid-induced gelation of milk. The possibility to simultaneously determine l* and labs was used to follow the time development of a draining and coarsening protein foam which contained an absorbing dye. It was shown that labs can be used as a measure for the volume fraction of air in the foam. This method of monitoring the transmission of multiple light scattering provides an easy way to determine l* and, specifically for foams, quantitative data dominated by the bulk of the foam.

Application of physics encoded neural networks to improve predictability of properties of complex multi-scale systems.

Predicting physical properties of complex multi-scale systems is a common challenge and demands analysis of various temporal and spatial scales. However, physics alone is often not sufficient due to lack of knowledge on certain details of the system. With sufficient data, however, machine learning techniques may aid. If data are yet relatively cumbersome to obtain, hybrid methods may come to the rescue. We focus in this report on using various types of neural networks (NN) including NN's into which physics information is encoded (PeNN's) and also studied effects of NN's hyperparameters. We apply the networks to predict the viscosity of an emulsion as a function of shear rate. We show that using various network performance metrics as the mean squared error and the coefficient of determination ( R 2 ) that the PeNN's always perform better than the NN's, as also confirmed by a Friedman test with a p-value smaller than 0.0002. The PeNN's capture extrapolation and interpolation very well, contrary to the NN's. In addition, we have found that the NN's hyperparameters including network complexity and optimization methods do not have any effect on the above conclusions. We suggest that encoding NN's with any disciplinary system based information yields promise to better predict properties of complex systems than NN's alone, which will be in particular advantageous for small numbers of data. Such encoding would also be scalable, allowing different properties to be combined, without repetitive training of the NN's.

Effect of Pea Legumin-to-Vicilin Ratio on the Protein Emulsifying Properties: Explanation in Terms of Protein Molecular and Interfacial Properties.

In isolates from different pea cultivars, the legumin-to-vicilin (L:V) ratio is known to vary from 66:33 to 10:90 (w/w). In this study, the effect of variations in the L:V ratio on the pea protein emulsifying properties (emulsion droplet size (d3,2) vs protein concentration (Cp)) at pH 7.0 was investigated using a purified pea legumin (PLFsol) and pea vicilin fraction (PVFsol). Despite a different Γmax,theo, the interfacial properties at the oil-water interface and the emulsifying properties were similar for PLFsol and PVFsol. Hence, the L:V ratio did not affect the pea protein emulsifying properties. Further, PLFsol and PVFsol were less efficient than whey protein isolate (WPIsol) in stabilizing the emulsion droplets against coalescence. This was explained by their larger radius and thus slower diffusion. For this reason, the difference in diffusion rate was added as a parameter to the surface coverage model. With this addition, the surface coverage model described the d3,2 versus Cp of the pea protein samples well.

The emulsifying ability of oleosomes and their interfacial molecules.

Oleosomes are natural oil droplets, present in all organisms and abundant in oilseeds. After their aqueous extraction from oilseeds, they can be directly utilized as oil droplets in food, cosmetics and all types of oil-in-water emulsion systems. However, to expand the potential uses of oleosomes as green ingredients and to valorize oilseeds as efficient as possible, we explored their emulsifying ability. Oleosomes were extracted from rapeseeds, and 10.0 wt% oil-in-water emulsions were created after homogenization with 0.5-6.0 wt% oleosomes, and the droplet size of the emulsions and their structure was measured by laser diffraction and confocal laser scanning microscopy (CLSM), respectively. The emulsion with an oleosome concentration lower than 1.0 wt% gave unstable emulsions with visible free oil. At oleosome concentrations at 1.5 wt% or higher, we obtained stable emulsions with droplet sizes between 2.0 and 12.0 µm. To investigate the role of the oleosome interfacial molecules in stabilizing emulsions we also studied their emulsifying and interfacial properties (using drop tensiometry) after isolating them from the oleosome structure. Both oleosomes and their isolated interfacial molecules exhibited a similar behavior on the oil-water interfaces, forming predominantly elastic interfacial films, and also showed a similar emulsifying ability. Our results show that oleosomes are not stabilizing the oil-in-water emulsions as intact particles, but they provide their interfacial molecules, which are enough to stabilize an oil-water surface up to about 2 times bigger than the initial oleosome surface. The understanding of the behavior of oleosomes as emulsifiers, opens many possibilities to use oleosomes as alternative to synthetic emulsifiers in food and pharma applications.

Interfacial protein-protein displacement at fluid interfaces.

Protein blends are used to stabilise many traditional and emerging emulsion products, resulting in complex, non-equilibrated interfacial structures. The interface composition just after emulsification is dependent on the competitive adsorption between proteins. Over time, non-adsorbed proteins are capable of displacing the initially adsorbed ones. Such rearrangements are important to consider, since the integrity of the interfacial film could be compromised after partial displacement, which may result in the physical destabilisation of emulsions. In the present review, we critically describe various experimental techniques to assess the interfacial composition, properties and mechanisms of protein displacement. The type of information that can be obtained from the different techniques is described, from which we comment on their suitability for displacement studies. Comparative studies between model interfaces and emulsions allow for evaluating the impact of minor components and the different fluid dynamics during interface formation. We extensively discuss available mechanistic physical models that describe interfacial properties and the dynamics of complex mixed systems, with a focus on protein in-plane and bulk-interface interactions. The potential of Brownian dynamic simulations to describe the parameters that govern interfacial displacement is also addressed. This review thus provides ample information for characterising the interfacial properties over time in protein blend-stabilised emulsions, based on both experimental and modelling approaches.

Predictive model to describe water migration in cellular solid foods during storage.

BACKGROUND: Water migration in cellular solid foods during storage causes loss of crispness. To improve crispness retention, physical understanding of this process is needed. Mathematical models are suitable tools to gain this physical knowledge. RESULTS: Water migration in cellular solid foods involves migration through both the air cells and the solid matrix. For systems in which the water migration distance is large compared with the cell wall thickness of the solid matrix, the overall water flux through the system is dominated by the flux through the air. For these systems, water migration can be approximated well by a Fickian diffusion model. The effective diffusion coefficient can be expressed in terms of the material properties of the solid matrix (i.e. the density, sorption isotherm and diffusion coefficient of water in the solid matrix) and the morphological properties of the cellular structure (i.e. water vapour permeability and volume fraction of the solid matrix). The water vapour permeability is estimated from finite element method modelling using a simplified model for the cellular structure. CONCLUSION: It is shown that experimentally observed dynamical water profiles of bread rolls that differ in crust permeability are predicted well by the Fickian diffusion model.

Scaling of sound emission energy and fracture behavior of cellular solid foods.

A detailed study was performed of the fracture behavior of toasted rusk rolls, a cellular solid food, at different water activities and morphologies. We find that the energies of the emitted sound pulses follow Gutenberg-Richter power laws with characteristic exponents b ~ 1.5 . The scaling exponents varied only within a range of 0.2 when the method of fracture, humidity, or morphology was changed. However, differences in b were observed, indicating nonuniversal behavior, that seems to be related to morphology and water activity. Also, power law scaling behavior was observed for the waiting time distributions with an exponent a ~ 1.9.

Interfacial properties, thin film stability and foam stability of casein micelle dispersions.

Foam stability of casein micelle dispersions (CMDs) strongly depends on aggregate size. To elucidate the underlying mechanism, the role of interfacial and thin film properties was investigated. CMDs were prepared at 4°C and 20°C, designated as CMD4°C and CMD20°C. At equal protein concentrations, foam stability of CMD4°C (with casein micelle aggregates) was markedly higher than CMD20°C (without aggregates). Although the elastic modulus of CMD4°C was twice as that of CMD20°C at 0.005Hz, the protein adsorbed amount was slightly higher for CMD20°C than for CMD4°C, which indicated a slight difference in interfacial composition of the air/water interface. Non-linear surface dilatational rheology showed minor differences between mechanical properties of air/water interfaces stabilized by two CMDs. These differences in interfacial properties could not explain the large difference in foam stability between two CMDs. Thin film analysis showed that films made with CMD20°C drained to a more homogeneous film compared to films stabilized by CMD4°C. Large casein micelle aggregates trapped in the thin film of CMD4°C made the film more heterogeneous. The rupture time of thin films was significantly longer for CMD4°C (>1h) than for CMD20°C (<600s) at equal protein concentration. After homogenization, which broke down the aggregates, the thin films of CMD4°C became much more homogeneous, and both the rupture time of thin films and foam stability decreased significantly. In conclusion, the increased stability of foam prepared with CMD4°C appears to be the result of entrapment of casein micelle aggregates in the liquid films of the foam.

Effect of Temperature and Pressure on the Stability of Protein Microbubbles.

Protein microbubbles are air bubbles with a network of interacting proteins at the air-water interface. Protein microbubbles are commonly used in medical diagnostic and therapeutic research. They have also recently gained interest in the research area of food as they can be used as structural elements to control texture, allowing for the manufacture of healthier foods with increased consumer perception. For the application of microbubbles in the food industry, it is important to gain insights into their stability under food processing conditions. In this study, we tested the stability of protein microbubbles against heating and pressurization. Microbubbles could be heated to 50 °C for 2 min or pressurized to 100 kPa overpressure for 15 s without significantly affecting their stability. At higher pressures and temperatures, the microbubbles became unstable and buckled. Buckling was observed above a critical pressure and was influenced by the shell modulus. The addition of cross-linkers like glutaraldehyde and tannic acid resulted in microbubbles that were stable against all tested temperatures and overpressures, more specifically, up to 120 °C and 470 kPa, respectively. We found a relation between the storage temperatures of microbubble dispersions (4, 10, 15, and 21 °C) and a decrease in the number of microbubbles with the highest decrease at the highest storage temperature. The average rupture time of microbubbles stored at different storage temperatures followed an Arrhenius relation with an activation energy for rupture of the shell of approximately 27 kT. This strength ensures applicability of microbubbles in food processes only at moderate temperatures and storage for a moderate period of time. After the proteins in the shell are cross-linked, the microbubbles can withstand pressures and temperatures that are representative of food processes.

Quantitative description of the relation between protein net charge and protein adsorption to air-water interfaces.

In this study a set of chemically engineered variants of ovalbumin was produced to study the effects of electrostatic charge on the adsorption kinetics and resulting surface pressure at the air-water interface. The modification itself was based on the coupling of succinic anhydride to lysine residues on the protein surface. After purification of the modified proteins, five homogeneous batches were obtained with increasing degrees of modification and zeta-potentials ranging from -19 to -26 mV (-17 mV for native ovalbumin). These batches showed no changes in secondary, tertiary, or quaternary structure compared to the native protein. However, the rate of adsorption as measured with ellipsometry was found to decrease with increasing net charge, even at the initial stages of adsorption. This indicates an energy barrier to adsorption. With the use of a model based on the random sequential adsorption model, the energy barrier for adsorption was calculated and found to increase from 4.7 kT to 6.1 kT when the protein net charge was increased from -12 to -26. A second effect was that the increased electrostatic repulsion resulted in a larger apparent size of the adsorbed proteins, which went from 19 to 31 nm2 (native and highest modification, respectively), corresponding to similar interaction energies at saturation. The interaction energy was found to determine not only the saturation surface load but also the surface pressure as a function of the surface load. This work shows that, in order to describe the functionality of proteins at interfaces, they can be described as hard colloidal particles. Further, it is shown that the build-up of protein surface layers can be described by the coulombic interactions, exposed protein hydrophobicity, and size.

Identification of pitfalls in the analysis of heat capacity changes of beta-lactoglobulin A.

Information on changes in heat capacity (DeltaCp) of proteins upon unfolding is used frequently in literature to understand possible follow-up reactions of protein denaturation, like their aggregation propensity. This thermodynamic property is intrinsic to the protein's architecture and unfolding and should be independent of the approach used to evaluate it. However, for many proteins, the reported values for DeltaCp vary considerably. To identify whether the origin of these discrepancies lies within the experimental approach chosen and/or in the too simplified unfolding models used in the analysis of the data, we choose beta-lactoglobulin A, a relatively small protein, but disputed for its two-state unfolding, and established its DeltaCp from tryptophan fluorescence, near-UV circular dichroism and differential scanning calorimetric measurements. In view of the large variation for the obtained DeltaCp (between 3.2 and 10.1+/-0.8 kJ/(mol K)), it is evident that: (1) the sensitivity of different approaches to the structural changes; (2) irreversibility of unfolding; (3) non-ideal two-state unfolding behaviour need to be considered prior to interpretation. While the first two points can be addressed by using multiple approaches, the applicability of the selected unfolding behaviour for the analysis is often less easy to establish. In this work, we illustrate that by checking the wavelength-dependence used to detect protein conformational changes a tool is provided that gives a direct insight in the validity of the interpretation in these studies. An experimentally validated determination of DeltaCp allows a more proper use for the mechanistic understanding of protein denaturation and its follow-up reactions, avoiding pitfalls in the interpretation.

The role of interfacial rheological properties on Ostwald ripening in emulsions.

The coarsening of emulsion droplets by Ostwald ripening is studied by means of numerical simulations in which time-dependent (elastic) interfacial behaviour is taken into account. Theoretical calculations on the dissolution of a single emulsion droplet in an infinite medium at saturated conditions show that the dissolution process can be stopped only when the interfacial tension goes to zero. When interfacial stress relaxation is included, which prevents a continuous zero interfacial tension, no stabilisation of the dissolution process is observed and the droplet dissolves completely. In the case of an ensemble of droplets, numerical calculations on the coarsening of emulsion droplets with finite interfacial elasticity show that a stable situation occurs at finite interfacial tensions of the droplets. This applies for a closed system with the same assumptions as those made in the Lifshitz-Slyozov-Wagner (LSW) theory. The coarsening behaviour strongly depends on the saturation of the dispersed phase in the continuous phase. If the system is in contact with atmosphere, saturation will finally go to unity and stabilisation will only occur for zero interfacial tension of the droplets. For an ensemble of droplets in a closed system, the calculations show that stress-relaxation of the interface causes the Ostwald-ripening process to continue, so no stable situation is reached. Stabilisation can only be accomplished by adding insoluble species to the dispersed phase, by using particles as stabilisers or by micro-encapsulation of the emulsion droplets by thick insoluble interfacial layers, which have a thickness that is in the order of the radius of the droplet.

Proteins at air-water interfaces studied using external reflection circular dichroism.

In this report we describe the first attempts to record external reflection circular dichroism (ERCD) spectra of beta-lactoglobulin solutions. It is shown that the accumulated proteins at and near the air-water interface can be detected using ERCD and that the signals obtained contain information on the conformational properties and concentration of the proteins residing at the interface. The local protein concentration and its conformation are in full agreement with previous observations using external reflection infrared spectroscopy. The ERCD signals are dominated by linear dichroism (LD) due to non-ideal behavior of the instrumental optics, but can be explained for using the theoretical description of chiral reflection. This allows the analysis of ERCD spectra of protein solutions. The measured ERCD signals are described accurately in the region between 190 and 220 nm, but poor resemblance is obtained at higher wavelengths. We are however confident that improvement of experimental conditions and theoretical description will allow that in the near future, external reflection circular dichroism (CD) can be a valuable tool that complements the application of external reflection infrared spectroscopy to study interfacial systems.

Water uptake mechanism in crispy bread crust.

Crispness is an important quality characteristic of dry solid food products such as crispy rolls. Its retention is directly related to the kinetics of water uptake by the crust. In this study, a method for the evaluation of the water sorption kinetics in bread crust is proposed. Two different sorption experiments were used: an oscillatory sorption test and a sorption test in which the air relative humidity (RH) was increased stepwise. These two experiments had different time scales, which made it possible to get a better understanding of the mechanisms involved. Results show that the adsorption and desorption dynamics of the oscillatory sorption test could be described by a single exponential in time. The water uptake rate ( k) was one of the fitting parameters. A maximum in the water uptake rate was found for a RH value between 50 and 70%. The rate parameters of the experiment where RH was increased stepwise were around a factor 10 lower than those derived from oscillatory sorption experiments. This is an important factor when designing experiments for the determination of water uptake rates. In addition, also a parameter describing the time dependence of the rate parameters of the oscillatory sorption experiment was calculated (C), again by fitting a single exponential to the rate parameters. C was in the same range as the rate parameter of the isotherm experiment. This indicates that different (relaxation) processes are acting at the same time in the bread crust during water uptake.

Limited conformational change of beta-lactoglobulin when adsorbed at the air-water interface.

Detailed insight can be obtained from proteins at and near the air-water interface using external reflection IR and circular dichroism techniques. Besides information on local protein concentrations and surface layer thickness, it is shown that beta-lactoglobulin displays a limited unfolding at the interface. The conformational change is comparable to that observed upon heat-induced aggregation of the protein and can be understood in view of the high surface concentration of the protein (approximately 40% volume fraction). The layer thickness and the conformational properties of the protein do not depend on the bulk concentration. After adsorption of beta-lactoglobulin to a preformed lipid monomolecular layer a similar conformational change is induced, suggesting that the folding properties of the protein itself determine the extent of conformational changes at the interfaces.

Protein adsorption at air-water interfaces: a combination of details.

Using a variety of spectroscopic techniques, a number of molecular functionalities have been studied in relation to the adsorption process of proteins to air-water interfaces. While ellipsometry and drop tensiometry are used to derive information on adsorbed amount and exerted surface pressure, external reflection circular dichroism, infrared, and fluorescence spectroscopy provide, next to insight in layer thickness and surface layer concentration, molecular details like structural (un)folding, local mobility, and degree of protonation of carboxylates. It is shown that the exposed hydrophobicity of the protein or chemical reactivity of solvent-exposed groups may accelerate adsorption, while increased electrostatic repulsion slows down the process. Also aggregate formation enhances the fast development of a surface pressure. A more bulky appearance of proteins lowers the collision intensity in the surface layer, and thereby the surface pressure, while it is shown to be difficult to affect protein interactions within the surface layer on basis of electrostatic interactions. This work illustrates that the adsorption properties of a protein are a combination of molecular details, rather than determined by a single one.