Morphology of Soy Protein Isolate at Oil / Water and Oil / Air Interfaces

Neste artigo, as propriedades emulsificantes da proteína isolada de soja (SPI) foram evidenciadas mostrando que estas macromoléculas sofrem mudanças conformacionais quando adsorvidas em interfaces. Investigou-se a conformação das cadeias proteicas ancoradas nas regiões interfaciais de emulsões de óleo em água (o/a) através de técnicas de espalhamento de raios X (SAXS) e de imagem (microscopia eletrônica de varredura (SEM)). O valor médio do raio de giro (Rg) da SPI (aq) é 20 nm e aumenta para 30 nm em emulsões o/a; as proteínas atuam como moléculas anfifílicas expondo seus núcleos hidrofóbicos ao óleo e os resíduos hidrofílicos à fase aquosa. Este valor ainda é maior após o spray drying das emulsões, na interface o/ar das respectivas microcápsulas. As paredes das microcápsulas são fractais de objetos agregados com superfícies rugosas, que são alisadas pela presença de um agente de reticulação.


Introduction
The benefits of using biopolymers as emulsifiers are several; for instance, they are usually cheaper than synthetic polymers, natural polymers are more environmentally friendly (biocompatible and biodegradable) and they can usually be obtained through simple extraction/ precipitation techniques, with no need for hard chemistry or organic solvents.The fact that Brazil is one of the top world producers of soybean is another motivation for the development of new products based on the biopolymers that can be obtained from this product. 1e of the peptide fractions of the dried extract of soybeans is known as soy protein isolate (SPI), which is a mixture of several different proteins, in which the major constituents are globulins. 2These globular proteins may self-associate to form spherical aggregates in water and the resulting system is an aqueous suspension of these aggregates. 3he emulsifying properties of SPI are already known and it has been found to be a very good foam stabilizer. 2,4[7][8][9][10] The cysteine residues in the protein backbone allow cross-linking of the polypeptide chains by simply heating, Vol. 24, No. 6, 2013   with no need for cross-linking agents due to the formation of inter and intramolecular disulfide bridges. 11his research forms part of a strategy to encapsulate the oil from the Copaiba tree, which has been used as an anti-inflammatory and anti-septic and has shown activity against the parasite Leishmania amazonensis. 12,13In this study, oil in water (o/w) emulsions using SPI as the emulsifier were prepared and from the obtained samples microcapsules of SPI with oil were prepared through spray drying the emulsions.Herein, the emulsifying properties of SPI and the morphological changes that SPI undergoes when nested at oil/water and oil/air interfaces are reported.Scattering techniques, such as SAXS, can provide a good insight into the conformation of protein chains both in solution and at interfaces. 14,15The size distribution of droplets of both emulsions, the storage stability of the oil/water emulsions and the zeta potential (ξ) were also measured by dynamic light scattering (DLS).Images of the dried microcapsules obtained by scanning electron microscopy (SEM) are provided, with the size and morphology.

Materials
The soy protein isolate (SPI-Supro 500E) was donated by Solae (São Paulo State, Brazil).It was dissolved in water, filtered through a 0.45 mm cellulose acetate membrane and freeze-dried prior to use.The copaiba oil was extracted from Amazon Forest Copaiba trees and kindly donated by Dr. Luiz A. Kanis (Universidade do Sul de Santa Catarina (UNISUL), Santa Catarina State, Brazil).Sodium hydroxide, acetic acid, hydrochloric acid and sodium acetate were purchased from Nuclear (São Paulo State, Brazil).The cross-linking agent 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide (EDC) was purchased from Sigma-Aldrich (USA).

Emulsion preparation
The emulsions were prepared by the following method: aqueous suspensions of SPI with the concentrations listed in Table 1 were prepared, with pH adjusted to 10 by NaOH addition.The suspensions were kept under stirring overnight, at room temperature.After this period, the pH was adjusted to 7 through HCl addition.For the samples with cross-linking agent, EDC 5% (m/m) was added at room temperature under stirring for 10 min.The suspensions were then placed in an UltraTurrax mechanical disperser (Quimis, Brazil) operated at 12,000 rpm.The copaiba oil was slowly added to the mixture, in the proportion of 0.04:1 (v/v), always under stirring.

Emulsion size distribution and zeta potential
The mean diameter of the oil drops from the emulsions was measured on a Malvern NanoZetasizer, which uses back light scattering (at 173°) from a 4 mW He Ne laser at 632.8 nm to evaluate the size, by correlating the scattering fluctuation and using the Stokes-Einstein approximation to calculate the size from the mean diffusion coefficient.The equipment is also able to simultaneously measure the zeta potential of the sample by applying an electric potential and measuring the laser Doppler effect.

Microcapsule preparation
The resulting emulsions were kept under stirring at 12,000 rpm, and then pumped to a Spray Dryer (B-290, Büchi, Germany) in which the influx air temperature was 200 °C and outflux 70 °C, at a feeding speed of 8 mL min -1 and influx air speed of 35 m 3 h -1 .At this temperature, all the water was removed and the microcapsules were collected at the bottom of the equipment.

Imaging technique (SEM)
The morphology of the microcapsules was studied by field-emission gun scanning electron microscopy (FEG-SEM) (JEOL JSM-6701F, Japan), at a maximum amplification of 37,000 times.The samples were fixed onto metallic supports with silver glue and then sputter-coated with gold film.

Scattering techniques (SAXS and DLS)
The average size of the oil drops of the emulsions was measured through the angular dependency of the dynamic and static light scattering (DLS and SLS, respectively) using an ALV-7002 goniometer (Germany) operated with a 22 mV laser at a wavelength of 638.2 nm and able to collect data at scattering angles from 32° up to 150°.All measurements were carried out by collecting a 1.5 mL sample of each emulsion under stirring and immediately adding it to a cylindrical glass cell without dilution.The measurements were taken immediately after that at room temperature and the data were always collected for 600 s.
The morphology of the protein chains was analyzed by SAXS at the SAX2 beam line at LNLS (Campinas-SP, Brazil), at l = 0.176 nm and q range of 0.05 < q < 1.15.The detection was performed with a CCD (charge-coupled device) camera with two frames of 150 s exposure and the I(q) curves were obtained after azimuthal integration of the images.The liquid samples (emulsions) were injected into a sample holder with a mica window and the solid samples were fixed with Kapton tape.

Oil/water emulsions
The size distributions of all prepared emulsions were measured immediately after preparation and after storage periods of up to ten weeks.The mean diameter of the oil drops is shown in Figure 1.As indicated in the figure, the size of the drops is close to 1 µm and varies only slightly even after 10 weeks of storage: first, an increase in size was observed, probably due to the coalescence of the oil drops, followed by a decrease, which was due to the observed creaming of larger drops, leaving only the smaller drops in suspension.Sample S4, which contained the cross-linking agent and 4% of SPI, had smaller oil drops, whereas sample S1, with less protein (3%) and without the cross-linking agent, had the largest oil drops.The obtained results also indicated that the size distributions were broad, with the polydispersity index ranging from 0.5 to 0.8.
In all cases, the zeta potential of the samples was negative and close to -50 mV, as shown in Table 2.
The negative values for zeta potential arise from the deprotonation of carboxylic acid groups of the protein chains since the pH (7) is higher than the usual pKa values of these groups in these proteins (3.5 to 5). 15 The data also show four features: (i) the absolute value of the zeta potential seems to decrease with time of storage; (ii) this decrease is more pronounced in the samples without cross-linking agent, suggesting that some of the protein chains may leave the interface and migrate to the water phase; (iii) the increase in SPI concentration leads to lower absolute potential values probably due to the accompanying increase in the ionic strength of the aqueous phase; (iv) the absolute values for the samples with EDC are always lower than those without EDC but with the same SPI concentration.This observation might be related to the fact that after cross-linking, less acid groups are available on the protein backbone. 16

SPI microcapsules
All of the emulsions listed in Table 1 were spray dried in order to obtain the microcapsules.Since the mean diameter of the oil drops were observed to be dependent on the storage time, the emulsions were pumped to the spray dryer apparatus immediately after their preparation.SEM images of the dried microcapsules obtained from the emulsions are shown in Figure 2.There is a broad size distribution in which the capsule diameter ranged from 0.5 to 20 mm.The figure also shows that the surfaces of the capsules are not smooth and that some of the capsules are deflated probably due to the spray drying process. 6Some of the images show broken capsules, in which it is possible to see that they are hollow and, therefore, these are in fact capsules and not solid spheres.
The wall thickness of the capsules is not homogeneous: the width varies from 100 to 300 nm, as shown in the FEG-SEM images in Figure 3.Some small holes in the microcapsule walls can be seen, even in the samples containing EDC.These porous structures are desirable since the intended application of these microcapsules involves the controlled-release of the inner oil phase.
1. Mean diameter of oil drops from emulsions as a function of storage time.Protein at oil/water and oil/air interfaces Since the imaging techniques provided information on the microcapsule morphology, it was possible to understand the protein chain conformations at the interfaces using these methods.However, through scattering techniques, the morphology of the SPI polymeric chains was accessed at the interface of the oil drop in water and also at the oil/air interface of the dried microcapsules.
[19] (1) The method is illustrated in Figure 4 for the SPI suspension, the emulsion S2 and microcapsules S2 and the results for all samples are summarized in Table 3.
As shown in Figure 4 and Table 3, the size of the SPI chains in aqueous suspension is around 24 nm.When SPI migrates to the oil-water interface, the size increases to 31 nm, indicating that the hydrophobic residues of the globular proteins are exposed to the oil phase.When the water phase is removed (dried microcapsules), the SPI chains become even larger, showing that more hydrophobic content is being exposed to the solid-air interface.In other words, the removal of water molecules from the SPI neighborhood allows the polymer chains to relax and to assume larger conformations.
Therefore, the thickness of the wall of the microcapsules (100 to 300 nm) seen in Figure 3 originates from the stacking of 4 to 12 layers of SPI.Similar values were obtained when computing the amount of SPI, the concentration of oil drops (number) and the volume fraction of the oil drops in the emulsions.
Another interesting feature observed is that in the presence of EDC (emulsions S2 and S4), the R g values are slightly lower, indicating the contraction of the protein chains after cross-linking.The analysis of the angular dependence of the SAXS intensity is also able to show the appearance of any volume or surface fractal structure, in our case, it comprised of repetitions of SPI aggregates at the interface.This can be achieved by fitting any linear region of the curve ln I(q) vs. ln q, according to equation 2. 19,20 If 1 < a < 3, the linear region represents a volume fractal with a dimension d f = a.If the angular coefficient is 3 < a < 4, then the linear region originates from a surface fractal with dimension d s of 6 -a. 21) The relationship between d f and the topology of the scattering object can be seen in Table 4.
The SAXS data also provided some information regarding the fractal structures created by the assembly of SPI chains at the oil-water and oil-air interface.The method is exemplified in Figure 5, in which the appearance of some linear regions on the curves can be seen.Similar behaviors were observed for all other samples.
The small bump seen at around q = 0.22 nm -1 on the SAXS curve corresponding to the emulsion (Figure 5) originates from the structure factor of the organization of the oil drops. 27It was observed (not shown here) that the peak disappears with the dilution of the emulsion.The process of drying the emulsions also leads to the vanishing of the structure factor peak, as seen when comparing this curve with that for the microcapsules.
The a values are given in Table 5.By comparing these values with the topologies presented in Table 4, the SPI chains assume a random coil configuration when in aqueous suspension, with a d f value close to 1.8.;However, these chains produce a mass fractal corresponding to a sphere with a rough surface with d f = 2.8 when SPI is nested at the o/w or o/air interface, which is in agreement with the information obtained from observation of the SEM images.
All of the curves corresponding to the dried microcapsules presented two different linear regions, at low and at high q values; the first corresponding to the surface fractal (a > 3) and the second arising from volume fractal (a < 3).These surface fractals with d s = 2.3 only appear when SPI is nested at the microcapsule interfaces, indicating that the capsule surfaces are almost spherical.Another feature seen in Table 5 is the effect of the crosslinking agent EDC (emulsions S2 and S4); while there is no difference observed in the d f values for emulsions with and without EDC, larger values are found for the dried samples with the cross-linking agent, in which the d f value is almost 3, indicating increased smoothness of the surface due to the action of EDC.However, the presence of EDC does not affect the smoothness of SPI at the o/w interface, indicating that there is indeed the cross-linking of SPI, which only occurs during the spray-drying process, and is responsible for the increase in d f .

Conclusions
The size distributions of all emulsions and dried microcapsules represented very polydisperse systems and this may be a consequence of the intrinsic polydispersity of SPI, which is in fact a mixture of several protein chains of different sizes and amphiphilic characteristics.However, even in such polydisperse systems, it was possible to observe some interesting features of SPI at the interfaces.Our results show that the biopolymer SPI can be used to  stabilize emulsions and to prepare microcapsules.The stabilization is due to the hydrophilic characteristics of the protein chains.Morphological changes were observed in the SPI chains when nested at o/w and o/air interfaces: the chains increase in size, indicating the rupture of the globular conformation of the protein promoting the exposure of the hydrophobic amino acid residues.Thus, the hydrophilic characteristics of the protein are strongly associated with its emulsifying properties.It was also shown that a mass fractal emerges when SPI is nested at these interfaces; the fractal dimensions indicate that the surface is rougher in the absence of the cross-linking agent and becomes smoother when the chains are cross-linked during the formation of the microcapsules.
A good agreement between the SAXS results and the SEM images was observed, indicating that the microcapsule interfaces are rough.

Figure 5 .
Figure 5. Double logarithmic plot of the angular dependence of the SAXS intensity for the SPI suspension and at the interface of the oil drops in the emulsion and microcapsules.The angular coefficients of the linear regions are indicated.

Table 1 .
Composition of o/w emulsions Morphology of Soy Protein Isolate at Oil/Water and Oil/Air Interfaces J. Braz.Chem.Soc.1014

Table 2 .
Zeta potential (ξ) of emulsions measured after different periods of storage

Table 5 .
a values in the SAXS I(q) m q -a relation at high q for SPI in all dispersions