pubs.acs.org/Langmuir © 2009 American Chemical Society
Morphology of Molecular Soy Protein Fractions in Binary Composite Gels† Stefan Kasapis*,‡ and Sok Li Tay§ ‡
School of Applied Sciences, RMIT University, Bundoora West Campus, Melbourne, Vic 3083, Australia, and §Department of Chemistry, National University of Singapore, Block S8, Level 5, Science Drive 3, Singapore 117543 Received October 7, 2008. Revised Manuscript Received November 27, 2008
We investigate the structural properties of gels of binary mixtures of the three major soy protein fractions: 11S, 7S, and 2S. Gels are formed at 25 °C in the presence of glucono-δ-lactone and studied using a combination of dynamic rheology and scanning electron microscopy. The rheological data was then modeled using a blending-law approach that yields insights into the solvent distribution between the gelled protein fractions and first-order reaction kinetics that follow the gelation process of the single fractions and their mixtures. Gelled mixtures of 11S and 7S yielded enhanced network strength with increasing solid content; in these gels, 50% more solvent partitioned into the 11S phase as compared to that in the 7S phase. In contrast, the addition of small-molecular-weight counterpart 2S to either 11S or 7S results in a catastrophic drop in the values of the overall strength of the mixture. The unexpected phase behavior has been rationalized on the basis of the high water-holding capacity of 2S; 450% more solvent partitions preferentially into the 2S phase as compared to that in the 11S phase. As the concentration of 2S is increased relative to that of 11S or 7S, it becomes the dominant phase and entraps the polymeric segments of 11S (or 7S), thus preventing them from becoming the structural knots of the gel. In addition to the solvent distribution in the gel, the rates of gelation differ markedly between 11S and 2S (with the 11S rate of gelation being up to 2 orders of magnitude greater); at fixed 11S concentration, the rate of gelation decreases with increasing amounts of 2S, further confirming that the latter essentially becomes the dominant phase in the composite gel.
Introduction Increasingly, tailor-made isolates of soy protein are being used in a wide range of industrial applications from functional ingredients to imitate desirable organoleptic properties in processed food products to bioactive systems with physiological properties that may prevent age-related chronic disease.1,2 Soybeans have a high protein content and contain three major molecular fractions (11S, 7S, and 2S). The molecular weights of these three fractions in their native forms are approximately 300, 180, and 17 kDa, with the percentage contents being about 42, 34, and 15%, respectively.3 Cellulose or Sephadex chromatography reveals that the centrifugation pattern is a complex mixture of two major storage globulins (glycinin and conglycinin) and “biologically active” proteins (trypsin inhibitors, cytochrome c, lipoxygenases, hemagglutinins, β-amylase and allantoinase).4 Both 11S (glycinin) and 7S (mainly β-conglycinin) dissociate easily into their structural components in the presence of urea or anionic detergents and at extremes of pH,5 suggesting that they are mainly stabilized by noncovalent interactions (hydrophobic, electrostatic, and hydrogen bonds). The quaternary structure of an intact 11S protein molecule is believed to be a dimer with six basic and six acidic subunits alternating in two parallel layers stacked † Part of the Gels and Fibrillar Networks: Molecular and Polymer Gels and Materials with Self-Assembled Fibrillar Networks special issue. *Corresponding author. E-mail:
[email protected]. Fax: + 61 3 9925 7110. Tel: + 61 3 9925 7144.
(1) Nunes, M. C.; Batista, P.; Raymundo, A.; Alves, M. M.; Sousa, I. Colloids Surf., B 2003, 1, 1. (2) Wang, W.; de Mejia, E. G. Compr. Rev. Food Sci. Food Saf. 2005, 4, 63. (3) Heras, J. M.; Marina, M. L.; Garcia, M. C. J. Chromatogr., A 2007, 1153, 97. (4) Kuipers, B. J. H.; van Koningsveld, G. A.; Alting, A. C.; Driehuis, F.; Gruppen, H.; Voragen, A. G. J. J. Agric. Food Chem. 2005, 53, 1031. (5) Tang, C.-H.; Choi, S.-M.; Ma, C.-Y. Int. J. Biol. Macromol. 2007, 40, 96. (6) Garcia, M. C.; Torre, M.; Marina, M. L. J. Chromatogr., A 2000, 881, 47.
8538 DOI: 10.1021/la803290j
on top of each other, whereas the 7S fraction is a complex glycoprotein.6 Through the years, research in structure-function relationships has been carried out mainly on 11S and 7S globulins. One type of structure formation of these two protein fractions has been studied by Kohyama and Nishinari: cold-induced gelation in the presence of glucono-δ-lactone (GDL).7 These workers reported that under large-deformation compression testing the breaking stress and rate of gelation of the 7S gel were lower than the corresponding properties of 11S. This was attributed to the low isoelectric point of 7S, which maintained a relatively high pH value in the gel following GDL addition. Scanning electron microscopy showed that 11S produced a coarse network with a pore size of 2 to 3 μm whereas 7S exhibited a finer structure of about 0.5 μm pore size.8 By comparison, studies dealing with the relationship between structural and functional properties of 2S soy protein are rather scarce. This is probably due to the well-known allergenicity reported for the albumin family of 2S found in Brazilian nuts and to 2S being the smallest of the three major fractions in plant proteins.9 Nevertheless, any extent of allergenicity of the 2S fraction from soybean has not been documented,10 and putting the above considerations aside, a recent investigation attempted to understand better the fundamentals of 2S functionality. Thus, it has been documented that the fraction exhibits better foaming and emulsification properties than do 11S and 7S.11 Work was (7) Kohyama, K.; Yoshida, M.; Nishinari, K. J. Agric. Food Chem. 1992, 40, 740. (8) Tay, S. L.; Perera, C. O. J. Food Sci. 2004, 6, 139. (9) Nordlee, J. A.; Taylor, S. L.; Townsend, J. A.; Thomas, L. A.; Bush, R. K. N. Engl. J. Med. 1996, 334, 688. (10) Lin, J.; Fido, R.; Shewry, P.; Archer, D. B.; Alcocer, M. J. C. Biochimica et Biophysica Acta 2004, 1698, 203. (11) Tay, S. L.; Kasapis, S.; Perera, C. O.; Barlow, P. J. J. Agric. Food Chem. 2006, 54, 6046.
Published on Web 02/09/2009
Langmuir 2009, 25(15), 8538–8547
Kasapis and Tay
extended to structural properties, and it was found that in the initial stages of structure formation 2S fared better than 7S, with 11S exhibiting the highest rates of aggregation. Noncovalent interactions, as opposed to disulphide bridging, were largely responsible for the changing functionality of the molecular fractions throughout the experimentation from the formation of a vestigial structure to that of a mature gel. As far as we are aware, there have not been fundamental studies on the phase behavior of binary composite gels made of molecular soy protein fractions. The present study aims to address this in an effort to facilitate the use of these materials in tailor-made food applications and drug design.
Experimental Procedures Source and Biochemical Characterization. Soybeans be-
longs to the family Leguminosae (genus Glycine L.).12 The material (product code 063-130; graded as 7-B defatted soy flour) used in this investigation to extract the protein fractions of 11S, 7S, and 2S was purchased from Archer Daniels Midland Company (Decatur, IL). The literature includes extensive data on the characterization of these fractions in terms of primary to quaternary structures, sulfydryl content, composition and crystal structure of the molecular subunits, biological activity, and nutritional quality.13-15 Isolation of the 11S Molecular Soy Protein Fraction. This was obtained using the method of Nagano and co-workers with some modifications.16 In doing so, defatted soy flour was mixed with a 15-fold volume of deionized water, and the pH was adjusted to 7.5 with 1 M NaOH. Water-extractable soybean protein was recovered by centrifugation at 9000g for 30 min at 20 °C. Sodium metabisulfite (0.98 g/L) was added to the supernatant, the pH was adjusted to 6.4 with 1 M HCl, and the preparation was kept in an ice bath overnight. This was then centrifuged at 6500g for 20 min at 4 °C, and the 11S obtained in the precipitate was dissolved in deionized water with pH adjustment to 7.5. Dialysis of the solution ensued against deionized water for 24 h at 4 °C. The resultant salt-free dialysate was freeze dried and stored at 5 °C for subsequent experimentation. Eight replicates of the freeze-dried material were analyzed using the semimicro Kjeldahl method and a conversion factor of 6.25 to identify the protein content, which was found to be 95.0 ( 1.0%. Isolation of the 7S Molecular Soy Protein Fraction. As for 11S, this was obtained using the method of Nagano and coworkers with some modifications.16 Defatted soy flour was mixed with a 15-fold volume of deionized water, and the pH was adjusted to 7.5 with 1 M NaOH. Water-extractable soybean protein was recovered by centrifugation at 9000g for 30 min at 20 °C. Sodium metabisulfite (0.98 g/L) was added to the supernatant, the pH was adjusted to 6.4 with 1 M HCl, and the preparation was kept in an ice bath overnight. This was then centrifuged at 6500g for 20 min at 4 °C, and the insoluble 11S was removed in the precipitate. The supernatant was adjusted to contain 0.25 M NaCl and to pH 5.0 with HCl. Stirring for 1 h at 4 °C ensued, and insoluble materials were removed by centrifugation (9000g for 30 min). The supernatant was diluted 2-fold with ice water, adjusted to pH 4.8 with HCl, and centrifuged once more at 6500g for 20 min. The 7S fraction was obtained in the precipitate, washed, redissolved in distilled water, and (12) Clarke, E. J.; Wiseman, J. J. Agric. Sci. 2000, 134, 111. (13) Garcıa, M. C.; Torre, M.; Marina, M. L.; Laborda, F. Crit. Rev. Food Sci. Nutr. 1997, 37, 361. (14) Maruyama, N.; Adachi, M.; Takahashi, K.; Yagasaki, K.; Kohno, M.; Takenaka, Y.; Okuda, E.; Nakagawa, S.; Mikami, B.; Utsumi, S. Eur. J. Biochem. 2001, 268, 3595. (15) Adachi, M.; Kanamori, J.; Masuda, T.; Yagasaki, K.; Kitamura, K.; Mikami, B.; Utsumi, S. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 7395. (16) Nagano, T.; Hirotsuka, M.; Mori, H.; Kohyama, K.; Nishinari, K. J. Agric. Food Chem. 1992, 40, 941.
Langmuir 2009, 25(15), 8538–8547
Article adjusted to pH 7.5 with NaOH; it underwent dialysis against deionized water for 24 h at 4 °C. The resultant salt-free dialysate was freeze dried and stored at 5 °C for subsequent experimentation. Eight replicates of the freeze-dried material were analyzed using the semimicro Kjeldahl method and a conversion factor of 6.25 to identify the protein content, which was found to be 93.0 ( 1.0%. Isolation of the 2S Molecular Soy Protein Fraction. This was obtained by the method of Rao and Rao with some modifications.17 Defatted soybean flour was mixed with a 10fold volume of deionized water containing 0.1% β-mercaptoethanol, and the preparation was stirred for 30 min. Waterextractable soybean protein was recovered by centrifugation at 6000g for 20 min at 20 °C. Magnesium chloride was added to the supernatant to produce a 0.1 M final solution, and this was held at 4 °C for 6 h. The resultant suspension was centrifuged at 10000g for 30 min at 4 °C. Magnesium chloride was added to the supernatant to produce a 0.4 M final solution, and this was held as before. The resultant fine suspension was centrifuged as for the first addition of MgCl2. Ammonium sulfate was added to the supernatant to produce a 32% (w/v) final solution, thus inducing the formation of a protein precipitate, which was obtained by centrifugation as described previously. The precipitate was dissolved in 1 M NaCl solution and dialyzed against the same solution for 24 h at 4 °C. Following this, ethyl alcohol was added to the dialysate in a proportion of 1:1 (v/v), and the resultant suspension was stirred for 6 h at 4 °C. Centrifugation recovered the supernatant, which was adjusted to pH 7.5, dialyzed against deionized water for 24 h at 4 °C, and freeze dried. Samples were dissolved in a phosphate buffer (pH 7.5) and introduced onto Sephacryl S300 and Sephacryl S200 gel filtration columns (60 1.6 cm i.d. for both; GE Healthcare Bio-Sciences AB, Uppsala, Sweden) connected in tandem and equilibrated earlier with the same buffer solution. The columns were eluted with this buffer at a flow rate of 0.6 mL/ min until the absorbance of the eluted fractions at 280 nm returned to the baseline. High absorption fractions were collected, freeze dried, and stored at 5 °C for subsequent use. The protein content was about 92% by the Kjeldahl method (N 6.25) and with that of 11S and 7S was used to prepare the specific concentrations required in single and mixed systems of the present investigation.
Sodium Dodecylsuphate Polyacrylamide Gel Electrophoresis. SDS-PAGE is an assay of protein purity and was employed in a mini-Protean 3 cell (Bio-Rad, Hercules, CA, USA). A sample (100 μL) with a protein composition of 4 mg/ mL was mixed with 95 μL of buffer made of deionized water, Tris-HCl (pH 6.8), glycerol, SDS, and bromophenol blue. Five microliters of mercaptoethanol were added to that, and the preparation was heated to 95 °C for 4 min. Then, 10 μL of the sample mixture was applied to the gel. Stacking and resolving gels were based on distinct formulations of deionized water, separating buffer, and 30% acrylamide/bis, with their concentrations being 4 and 12%, respectively. Precision Plus protein standards from Bio-Rad were used as the molecular weight markers. Appropriate stacking, resolving, and electrolysis buffers were also prepared, and the voltage used for the migration was 125 V. Following migration, the protein fractions were stained with 0.1% Coomassie brilliant blue in water-methanol-acetic acid (4: 1: 5) and distained with water-methanolacetic acid (85:7.5:7.5). Figure 1 reproduces well-resolved SDS-PAGE patterns of the three soy protein fractions (11S, 7S, and 2S). In lane A, the molecular weight of the 2S protein was about 17 kDa, which is identified to be the Bowman-Birk trypsin inhibitor. In lane B, the 7S protein fraction is made up of three subunits: R, R0 , and β with molecular weights of about 85-45 kDa. And in lane C, the 11S protein fraction is made up of acidic and basic subunits (17) Rao, A. G. A.; Rao, M. S. N. Prep. Biochem. 1977, 7, 89.
DOI: 10.1021/la803290j
8539
Article
Kasapis and Tay preparations, with the process of gel formation being reproducible within a 3% margin of error as a function of the time scale of measurement. Scanning Electron Microscopy (SEM). Single 11S, 7S, and 2S gels at various protein concentrations and in the presence of 0.4% (w/v) GDL were prepared. These and their mixtures made at a fixed 11S concentration (4.0%) with increasing amounts of 7S or 2S (from 0 to 4.0%) and a mixture of 7S (3.0%) with 2S (from 0 to 3.0%) were left to age for 24 h at 25 °C prior to examination in order to imitate the experimental protocol utilized in the preceding section. Next, samples were deep frozen rapidly by submerging them in liquid nitrogen, followed by freeze-drying induced dehydration. Specimens were cut into thin slices before being mounted onto a microscopy stub with black double-sided tape and were coated with platinum by means of a sputter coater. Networks were observed using SEM, which operated at an accelerating voltage of 15 kV (JEOL model JSM-5600LV, Tokyo, Japan).19
Results and Discussion Figure 1. SDS-PAGE of soy protein: lane A is the 2S protein fraction; lane B is the 7S protein fraction; and lane C is the 11S protein fraction.
of about 35 and 18 kDa, respectively. Molecular fractions of 11S, 7S, and 2S were used for subsequent experimentation in this article.
Small Deformation Mechanical Spectroscopy on Shear. To evaluate the structural properties of our materials, 4.0 mL solutions were prepared as follows: For each solution, 3.68 mL was drawn from a separate 4.40 mL soy protein preparation made in deionized water, the pH was adjusted to 7.5, and the sample underwent heat treatment at 100 °C for 10 min. The remaining 0.32 mL was drawn from a freshly made 5% (w/v) GDL sample in order to keep the cosolute content constant at 0.4% (w/v) in the final solution. The amount of freeze-dried protein added to the 4.40 mL preparation was adjusted in order to yield the required protein concentration in the 4.0 mL solution used for mechanical analysis. Three milliliters was drawn from the latter and loaded onto the temperaturecontrolled plate of the rheometer (25 °C), with the exposed edges of the sample being covered with a silicone fluid from BDH (100 cs) to minimize water loss. The measuring geometry was a parallel plate with 40 mm diameter and a sample holding gap of 1.5 mm. Measurements were performed with the Advanced Rheometrics Expansion System (ARES), which is a controlled-strain rheometer (TA Instruments, New Castle, DE). ARES has an air-lubricated and essentially noncompliant force rebalance transducer with a torque range between 0.02 and 2000 g cm. For the precise control of sample temperature, an air convection oven was used that has a dual element heater/cooler with counter-rotating air flow covering a wide temperature range (between 130 and -70 °C). The measuring geometry of parallel plates was placed in the oven, which supplies ultraclean, dry air at a controlled airflow through the ARES oven heaters. Samples were subjected to a long isothermal run (25 °C) for up to 24 h at a frequency of 1 rad/s. In doing so, a strain within the linear viscoelastic region was used that varied from 0.005% for 11S to 0.1% for the 7S and 2S gels. The experimental protocol provides readings of the storage modulus (G0 , elastic component of the network) and loss modulus (G00 , viscous component) and a measure of the “phase lag” δ (tan δ = G00 /G0 ) of the relative liquidlike and solidlike character of the material.18 Three replicates were analyzed for selected experimental concentrations of single and mixed (18) Kasapis, S. Biomacromolecules 2006, 7, 1671.
8540 DOI: 10.1021/la803290j
Small Deformation Studies on the Gelation Properties of the Individual Components. Although a scientific understanding of phase separation phenomena in binary biopolymer systems has made considerable advances in the last 20 years or so,20,21 fundamental studies on the phase behavior of molecular soy fractions in composite gels are nonexistent. The design of meaningful experiments to unveil features of the topology of 11S, 7S, and 2S in binary mixtures requires a certain prior knowledge of the gelation behavior of the individual components. The transition from the sol to the gel state and the development of “mature” networks were monitored in the present investigation as a function of the time scale of observation. Particular care was taken to ensure that results were obtained under the appropriate conditions of frequency of oscillation, strain amplitude and controlled temperature in order to reflect true small-deformation properties of a welldeveloped molecular structure. Figure 2a illustrates the time course of gelation for representative samples of glycinin in the presence of 0.4% GDL. Clearly, the polymeric segments of 11S are capable of rapidly developing a 3D structure at 25 °C that reaches constant values for the “pseudoequilibrium” storage modulus on shear within 500 min of the experimental routine. No further changes in the values of G0 were observed at longer recording times (up to 1500 min), and the first part of the rheological characterization is depicted in Figure 2a to contrast the rapid kinetics of gelation with those of 7S and 2S in the following discussion in this section. It should be noted at this juncture that GDL containing 11S solutions undergoes an aggregation process that produces contracted brittle gels with a notable degree of syneresis. This may lead to catastrophic slippage, but the present investigation was able to avoid the breakdown of adhesion between the sample and the flat surface of the measuring plate by minimizing the disturbance of the gel during experimentation. This was achieved by using an applied strain of 0.005% and parallel plates of 40 mm diameter. Besides 3% 11S, which is the smallest and most difficult concentration to work with, no artificial drop in the magnitude of the solidlike response was observed at such a small level of deformation. (19) Al-Amri, I. S.; Al-Adawi, K. M.; Al-Marhoobi, I. M.; Kasapis, S. Carbohydr. Polym. 2005, 61, 379. (20) Mathew, S.; Brahmakumar, M.; Abraham, T. E. Biopolymers 2006, 82, 176. (21) Butler, M. F.; Clark, A. H.; Adams, S. Biomacromolecules 2006, 7, 2961.
Langmuir 2009, 25(15), 8538–8547
Kasapis and Tay
Article
Figure 2. Time sweeps of the storage modulus (G0 ) obtained for (a) 11S, (b) 7S, and (c) 2S soy protein at the concentrations shown beside the experimental traces. Some experimental data are not shown to avoid clutter (0.4% GDL added; strain = 0.005% 11S and 0.1% 7S and 2S; frequency = 1 rad/s; temperature = 25 °C).
overall, smooth traces are obtained in comparison to the 11S counterparts. A characteristic of the isothermal profile is that the onset of gelation is delayed with increasing concentrations of conglycinin at 0.4% GDL. Finally, we have documented the slow gelation profile of 2S where structural development occurs beyond the time frame of 500 min plotted in Figure 2c. Further holding the temperature at 25 °C yields a gradual rise in the magnitude of the storage modulus that asymptotically approaches an equilibrium value, hence it is feasible to monitor this slow gelling process (and that of 7S) via the application of 0.1% strain to the system. The distinct pictorial rheology of the three soy molecular fractions is also manifest in the isothermal properties of their mixture and will be utilized later on in the framework of arguing for the dominant polymeric phase in the topology of the composite gel. It is not within the scope of this investigation to elucidate the type of interactions responsible for the stabilization of the glycinin/conglycinin or 2S protein gel; rather, we establish the mechanical properties of the individual components as a reference state for the subsequent mixed-system analysis. Nevertheless, the utilization of urea and N-ethylmaleimide in soy preparations demonstrated that 2S is largely stabilized via noncovalent interactions, whereas 11S will also draw on disulfide bridging as a means of achieving maximum network strength. Thus, the extent of noncovalent interactions in 2S (i.e., the extent of delayed onset of gelation) is dictated by the isoelectric point/buffering capacity of the macromolecule in relation to the pH of the system.11 As alluded to earlier, data on single soy fractions are needed to relate the gel modulus to polymer concentrations for the modeling of the phase behavior in these blends. Figure 3 reproduces good linear or exponential relationships obtained for this type of double-logarithmic plot incorporating data from the completion of the time sweeps in Figure 2a-c. It appears that the high-molecular-weight distribution of 11S chains facilitates the formation of a functional network with dominant values of the storage modulus in comparison to the gel strength of 2S at a given polymer concentration. A notable exception to the “conventional” concentration dependence of the storage modulus is the pattern of 7S that reaches a maximum at preparations with 4.5% protein. Given that the isoelectric point (pI) of 11S is ∼6.4 and that of 7S is ∼4.8, it seems that at the upper range of conglycinin concentration the high buffering capacity prevents sufficient neutralization of charges in polymeric segments.11 This is needed for the formation of effective cross links leading to a 3D structure. Figure 3 suffices as the step of converting real (effective) polymer concentrations of their respective phases in the composite gel to shear moduli for “blending law” modeling. In studies of the physicochemical aspects of network structure, these data can be fit by an analytical expression (the “cascade” approach), which is a modification of the protocol proposed by Hermans, thus obtaining a set of parameters that relate to the theory of gelation.22,23 Observations on the Structural Behavior of Binary Soy Protein Gels. Following the analysis of single-component gels, a series of binary mixtures was prepared by keeping the concentration of glycinin constant at 4.0% (i.e., at a level leading to a stable gel in Figure 2a) and varying the amount of 7S or 2S from 0 to 4.0% in several concentration steps. A further series of
Figure 2b shows the gelation profile of 7S preparations upon addition of 0.4% GDL at 25 °C for 1500 min. Variation in the isothermal development of G0 values is rather limited, and
(22) Clark, A. H.; Richardson, R. K.; Ross-Murphy, S. B.; Stubbs, J. M. Macromolecules 1983, 16, 1367. (23) Clark, A. H.; Ross-Murphy, S. B. Adv. Polym. Sci. 1987, 83, 57.
Langmuir 2009, 25(15), 8538–8547
DOI: 10.1021/la803290j
8541
Article
Kasapis and Tay
Figure 3. Concentration dependence of the values of the storage modulus (G0 ) obtained for the three soy molecular fractions at the end of the time sweeps illustrated in Figure 2a-c.
preparations kept the amount of 7S constant at 3.0% and varied that of 2S from 0 to 3.0%. Upon mixing the two molecular fractions under conditions (temperature and polymer/GDL addition) at which each protein constituent forms a stable solution in isolation, turbidity is observed because of microphase separation in the binary system. Phase separation phenomena are very common between, for example, two plant polysaccharides (pectin and locust bean gum), two seaweed polysaccharides (alginate and κ-carrageenan) or two globular proteins (BSA and β-lactoglobulin), and the soy proteins are not an exception to that. The three protein fractions of this investigation possess distinct sequences of amino acids and contents of thiol groups that result in particular tertiary and quaternary (11S) structures. This is confirmed for the well-resolved SDS-PAGE patterns in Figure 1, which exhibit the acidic and basic subunits of 11S, the three subunits (R, R0 , and β) of 7S, and the Bowman-Birk trypsin inhibitor of 2S. The implementation of the measuring routine described in Experimental Procedures generates data for the dependence of the storage modulus (25 °C for 1500 min) for the binary combinations, and these are reproduced in Figure 4a-c. Plotting incorporates the final G0 readings of 4% 11S and 3% 7S from Figure 2a,b, respectively. It is clear that these experimental conditions allow mixed gels to mature, approaching asymptotically constant values of network strength. Furthermore, the dominance of the glycinin phase at low additions of 7S or 2S to the mixture is discernible in the moderately spiky gelling processes recorded in Figure 4a,b. The time course of observation in these binary gels unveils a dramatic variation in the gel-like consistency of 11S in the presence of increasing concentrations of 7S or 2S, and Figure 5a,b focuses on this rather unexpected result. Previous explorations of the 3D topology of biopolymer composite gels carried out in mixtures of hydrolyzed starch with gelatin, κ-carrageenan with gelatin, and sodium caseinate with β-glucan unveiled microphase-separation phenomena between the two constituents.24,25 A clear phase-inversion point from one (24) Kontogiorgos, V.; Ritzoulis, C.; Biliaderis, C. G.; Kasapis, S. Food Hydrocolloids 2006, 20, 749. (25) Kasapis, S.; Al-Marhoobi, I. M. Biomacromolecules 2005, 6, 14.
8542 DOI: 10.1021/la803290j
Figure 4. Time sweeps of the storage modulus (G0 ) obtained for mixtures of (a) 4.0% 11S in the presence of increasing concentrations of 7S shown beside the experimental traces, (b) 4.0% 11S in the presence of increasing concentrations of 2S, and (c) 3.0% 7S in the presence of increasing concentrations of 2S. To avoid clutter, some experimental data are not shown. Dashed lines indicate the pseudo-equilibrium G0 value of single preparations of 4.0% 11S and 3.0% 7S (0.4% GDL added; strain = 0.1%; frequency = 1 rad/s; temperature = 25 °C).
dominant phase to another was also detected, leading to reinforced networks with increasing levels of the second component (and total solids) in the system. Langmuir 2009, 25(15), 8538–8547
Kasapis and Tay
Article
In other studies, interactions between soluble-fiber polysaccharides were interpreted either on the basis of “molecularly” interpenetrating networks for agarose/deacylated gellan and agarose/κ-carrageenan gels26,27 or being segregative for appropriate additions of counterions, calcium ions in the case of gelatin/low-methoxy pectin, and sodium ions for deacylated/ high acyl gellan.28,29 In the aforementioned mixtures, the pattern of modulus development was system-specific, but it was found to follow a simple additivity rule or possess “synergistic” character that preserved or enhanced the contributions of the individual polymeric components in the blend. In the event of a prolonged isothermal run on our molecular soy fractions, the familiar increase in the overall modulus is observed in Figure 5a where the addition of conglycinin considerably increases the strength of the composite gel throughout the experimentally accessible concentration range. Storage modulus values of single 11S and 7S gels at 4.0% solids are also shown as dashed lines for comparison. In contrast, gel reinforcement is recorded only in the lower range of 2S additions to the glycinin sample made at the aforementioned concentration of 4.0% (Figure 5b). At levels of 2S higher than 0.5%, there is an immediate diminishing effect on the composite strength, which at 1.5% 2S falls to a value comparable to that of the fixedconcentration glycinin gel. The catastrophic effect continues unabated at the upper range of 2S concentrations where it appears to be virtually complete. At this stage, the experimental storage modulus of the composite gel is close to the network strength recorded for 4.0% 2S. Strikingly, the addition of 2S to a conglycin gel accelerates the loss in mechanical strength, as shown in Figure 5c. Even at the lower range of 2S concentrations, there is a sharp decrease in the G0 values of the mixture, with 7S being unable to maintain structural cohesion. As for Figure 5b, the network strength of the material in the presence of high additions of 2S is effectively identical to that expected from the single 2S gel at equal levels of solids. The above findings indicate that high levels of 2S increasingly entrap polymeric segments of glycinin in the 2S phase, thus “freezing” them in the system and preventing them from contributing to functional connections within the phase/polymeric network of the high-molecularweight fraction. The effect is more pronounced in the case of 7S owing to the shorter chain segments that are unable to withstand the adverse effect on structure formation caused by the addition of 2S. The severe disruption of soy globulin aggregation in the presence of 2S discussed on the basis of mechanical observations can be further documented using scanning electron microscopy (SEM). Figure 6a,b illustrates SEM micrographs obtained for single and mixed preparations at 1500 magnification, thus making the contrast between the phase behavior of 11S/2S and 11S/7S mixtures. At 4% solids and a processing temperature of 25 °C, the high-molecular-weight 11S fraction forms welldeveloped “honeycomb” structures following the addition of GDL (Figure 6a1). Networks appear to have an approximate pore size of 10 μm. By comparison, extensive structural defects are noted in the 3D image of the low-molecular-weight 2S Figure 5. Experimental storage modulus (G0 ) values of (a) 4.0% 11S with varying concentrations of 7S at the completion of the time sweeps illustrated in Figure 4a, (b) 4.0% 11S with varying concentrations of 2S at the completion of the time sweeps illustrated in Figure 4b, and (c) 3.0% 7S with varying concentrations of 2S at the completion of the time sweeps illustrated in Figure 4c. Dashed lines indicate the pseudo-equilibrium G0 value of single preparations of 11S, 7S, and 2S. Remaining conditions are the same as in Figure 4. Langmuir 2009, 25(15), 8538–8547
(26) Amici, E.; Clark, A. H.; Normand, V.; Johnson, N. B. Carbohydr. Polym. 2001, 46, 383. (27) Amici, E.; Clark, A. H.; Normand, V.; Johnson, N. B. Biomacromolecules 2002, 3, 466–474. (28) Gilsenan, P. M.; Richardson, R. K.; Morris, E. R. Food Hydrocolloids 2003, 17, 739. (29) Kasapis, S.; Giannouli, P.; Hember, M. W. N.; Evageliou, V.; Poulard, C.; Tort-Bourgeois, B.; Sworn, G. Carbohydr. Polym. 1999, 38, 145.
DOI: 10.1021/la803290j
8543
Article
Figure 6. Scanning electron microscopy of a(1-4) single and mixed gels of 11S and 2S and b(1-4) single and mixed gels of 11S and 7S at concentrations indicated at the bottom of each micrograph. The magnification is 1500; each scale bar is 10 μm long.
fraction, which includes several pores and cavities with a broad size distribution (Figure 6a4). The inclusion of 2S in the system does not bring about the formation of a highly organized blend where the constituents maintain their macromolecular definition within the corresponding phase-separated domain. Such an image would have been according to experience from conventional studies of binary biopolymer gels.30,31 In the present study, the 3D form of 11S is progressively lost, but the organization of the resultant composite gel does not transform to something else with increasing 2S parentage. Instead, a new type of network is observed at this level of magnification, being almost ruptured in the inner structure (Figure 6a2,3). This exhibits an open surface with large cavities that are 30 μm or more in size (e.g., 4.0% 11S plus 4.0% 2S). Figure 6b(1-4) shows the scanning electron micrographs of the 11S/7S system. The 3D form of single 7S gels is well-defined and supported by a continuum of thin films as compared to the 11S compartments. The pore size and cavities are smaller and regularly organized in comparison to those of 2S (Figure 6a4). In (30) Petersson, L.; Oksman, K. Compos. Sci. Technol. 2006, 66, 2187. (31) Zhang, J.; Jiang, L.; Zhu, L.; Jane, J.-L.; Mungara, P. Biomacromolecules 2006, 7, 1551.
8544 DOI: 10.1021/la803290j
Kasapis and Tay
contrast to the observations of the preceding paragraph, the addition of 2.0 and 4.0% conglycinin to 4.0% glycinin results in the isotropic formation of structures with a pore size of about 5 μm. The composite gel appears to be solidlike, exhibiting closely packed and elongated “cell” walls that should impart rigidity in accordance to the reinforcement of the network strength in Figure 5a. These images are in qualitative agreement with the notion of the chain segments of 2S being in close proximity to those of glycinin, thus preventing the latter from forming a cohesive morphology of high network strength and a characteristic molecular arrangement. Blending-Law-Based Quantitative Analysis in Support of the Molecular Soy Fraction Phase Model. The unexpected phase behavior of our mixture in the presence of 2S poses the question as to its origin. Previous investigations of the topology of binary gels made it evident that the ability to hold solvent within a phase is the ultimate determinant of the relative predominance of this constituent over its counterpart. Because at the moment it is rather difficult to identify an experimental technique for reliable measurements of phase volume in the gel state, a computerized algorithm was devised to consider all possible distributions of solvent between the two polymeric components.32 The methodology was put to the test in composite systems of direct relevance to industrial applications, and it was able to rationalize phase-separation phenomena based on the combined framework of gelation and steric exclusion of the two macromolecules.25,33 Nevertheless, its utility is confined in regimes of a positive relationship between “performance” characteristics (e.g., enhanced strength of the composite gel) and increasing content of solids in the system. A hybrid approach was also developed to address the case where the overall modulus of the composite gel remains close or falls moderately below the modulus of the two constituents at their nominal (original) concentrations.34 This pattern of behavior was treated on the basis of deswelling ideas, assuming that the faster gelling species does so at its original concentration throughout the system, followed by ordering of the slower gelling species that claims part of the solvent from the network formed earlier. An application of this line of attack to soy is illustrated in Figure 7a for the mixture of 4% 11S and 0.5% 2S. Blending laws in the following mathematical format were used to relate the storage modulus values of the two-component phases to the overall network strength of the binary gel23 G0U ¼ jG011S þ ð1 -jÞG02S
ð1Þ
1=G0L ¼ φ=G011S þ ð1 -φÞ=G02S
ð2Þ
where G0 U and G0 L are the shear moduli of the upper and lower bounds of the composite, G0 11S and G0 2S are the shear moduli of the 11S and 2S phases, and φ is the volume of the 11S phase, which in this case is approximated to the amount of solvent held in the phase. The analysis affords a straightforward calculation of the relative amount of solvent held in each polymeric phase using the so-called p factor22 p ¼ ðS11S =X11S Þ=ðS2S =X2S Þ
ð3Þ
(32) Morris, E. R. Carbohydr. Polym. 1992, 17, 65. (33) Gilsenan, P. M.; Richardson, R. K.; Morris, E. R. Food Hydrocolloids 2003, 17, 751. (34) Kasapis, S.; Morris, E. R.; Norton, I. T.; Clark, A. H. Carbohydr. Polym. 1993, 21, 269.
Langmuir 2009, 25(15), 8538–8547
Kasapis and Tay
Figure 7. Analysis of the effect of solvent partition on (a) the calculated storage modulus of single 11S and 2S preparations and the calculated upper/lower bounds of a 4.0% 11S plus 0.5% 2S mixed gel and (b) the calculated storage modulus of single 11S and 7S preparations and the calculated bounds of a 4.0% 11S plus 4.0% 7S mixed gel, with the experimental modulus of the two mixtures (indicated as G0 exp) shown as vertical dashed lines on the y axis.
where S11S and S2S refer to the amount of solvent in the 11S and 2S phases (S11S + S2S) = 1, respectively, with X11S and X2S being the original concentrations of the two soy fractions (i.e., 4% 11S and 0.5% 2S in our example). Interested readers are referred to a recent discussion on the ins and outs of this semitheoretical framework, including the requirement for complete phase separation in the gel.35 Briefly, the assumption that each polymer is confined entirely to its own phase can be rationalized by the fact that in the initial stages of gelation there will be a massive increase in the effective molecular weight, which will promote the phase separation of biphasic solutions. Furthermore, this process is driven toward completion by the progressive migration of nongelling concentrations of the minor component in each phase to join the
(35) Kasapis, S. Crit. Rev. Food Sci. Nutr. 2008, 48, 341.
Langmuir 2009, 25(15), 8538–8547
Article
network being formed by the same material in the other phase, where it is dominant.33 Regarding the choice of mixtures, both approaches of constant composition or increasing concentration of the second component can be found in the literature. Treatment of the former remains qualitative whereas the latter has been utilized to predict the relative solvent avidity of the two polymeric components. The reason is that by choosing a fixed concentration of polymer A alongside increasing concentrations of polymer B and applying the blending-law analysis one can assert whether the estimates of the p factor (eq 3) are due to the chemical/ conformational characteristics of the two polymers or have been affected by the increasing concentrations of polymer B, leading to changing patterns of gelation, phase inversion in the mixture, and so forth. Figure 7a depicts the variation in modulus values (G0 11S and G0 2S) as a function of the hypothetical fraction of solvent in the glycinin phase (0.5 < S11S < 1.0). These were calculated for every possible distribution of solvent, each time yielding the effective (final) concentrations of the two soy proteins in the calibration curves of Figure 3. Moduli and phase volumes of the individual phases are then estimated and utilized in eqs 1 and 2 to produce corresponding values of network strength for the blend. As the value of S11S increases from 0.5 to 1.0, the calculated values of G0 11S decrease whereas those of G0 2S increase. For G0 11S > G0 2S and glycinin forming the supporting matrix, the above approach gives an upper bound behavior [G0 U(11S)], whereas G0 11S < G0 2S with glycinin maintaining phase continuity results in the lower bound model [G0 L(11S)]. During isothermal setting, 11S develops its continuous network prior to structuring of 2S owing to its rapid kinetics of gelation, as contrasted in Figure 2a,c and indicated for the mixture in Figure 4b. A similar picture in relation to the kinetic interplay of the state of gelation and phase-separation phenomena has emerged for other protein-protein and protein-polysaccharide mixtures.36,37 Therefore, the continuous-phase prediction for 11S descending from the top left corner (upper bound) to the bottom right corner (lower bound) of the diagram appears to possess physical meaning. Gratifyingly, the experimental value of the composite gel crosses the upper bound prediction of the glycinin continuous matrix. This relates to a solvent content in the 11S phase of about 0.64. Using eq 3, the p estimate on that is 0.22, making the solvent affinity of 2S several times higher than for 11S, with the effective concentration of the polymeric components in each phase being 1.4 and 6.3%, respectively. This result is in direct contrast to the patterns of solvent partitioning observed for mixtures of glycinin and conglycinin in Figure 7b. The experimental protocol in Figure 5a, which held the overall concentration of 11S constant (4.0%) and varied that of 7S, yields reinforced composite gels and allows the quantitative analysis of the mixed-gel moduli at high levels of 7S addition (4.0%). Owing to the concentration dependence of the storage modulus of 7S (Figure 3), the computerized algorithm in Figure 7b produces mechanical functions for the hypothetical solvent distributions in the 11S phase that exceed those of 7S. The experimental modulus of the 4.0% 11S-4.0% 7S mixture is shown as a dashed line, which intercepts the 11S bound. This is in agreement with the gel-time evidence presented in the isothermal runs of Figure 2a,b, which should yield a continuous (36) Tolstoguzov, V. Food Hydrocolloids 2003, 17, 1. (37) Richardson, R. K.; Kasapis, S. In Instrumental Methods in Food and Beverage Analysis; Wetzel, D. L. B., Charalambous, G., Eds.; Elsevier, Amsterdam, 1998; pp 1-48.
DOI: 10.1021/la803290j
8545
Article
Kasapis and Tay
glycinin matrix. When the data presented in Figure 7b are recast in terms of eq 3, a p value of 1.50 is obtained (S11S = 0.60). As for the treatment of conventional composite gels, the dominant polymeric phase (11S) possesses higher solvent affinity than the molecular sequences of 7S, an outcome that further emphasizes the adverse effect of 2S addition on the glycinin network (Figure 7a). Utilization of Reaction Kinetics to Pinpoint PhaseSeparation Phenomena in 11S/2S Mixtures. In general, the validation of phase-separation phenomena with the blendinglaw protocol can be further substantiated by cross checking with tangible evidence from several experimental techniques. These include the melting endotherm of gel networks utilizing heating runs of differential scanning calorimetry (DSC), the temperature course of gel melting under small-deformation dynamic oscillation on shear, and light/electron microscopy images usually taken between 100 and 2000 magnification.38-40 In the gelatin-maltodextrin system, for example, steric exclusion becomes apparent from the distinct molecular events of the protein and polysaccharide networks, being at least 40 °C apart, which are readily recorded in gel-melting DSC thermograms and rheological heating runs.34 Replacing gelatin and/or the cold gelling polysaccharide with globular proteins creates a composite system where thermal processing causes extensive denaturation of the molecules.41 These form aggregated spheroidal clusters that remain largely impervious to subsequent thermal treatments within the normal constraints of a laboratory investigation. Thus, thermomechanical characterization of order T disorder transitions is no longer a viable proposition, an outcome that may impair the identification of the phase-separated microstructure in the binary gel state. In addition, the comparable (amino acid-based) composition of the three soy protein fractions is not conducive to staining as a means of differentiation of polymeric networks. This, of course, is relevant to the 11S/2S composite gel where the researcher is unable to utilize the experimental characterization discussed above. The successful application of confocal laser scanning microscopy on composite gels that allows for optical sectioning on the z axis may change the current status quo in the future. Presently, however, we have taken advantage of the disparate gelation rates of the two constituents shown in Figure 2a,c in order to circumvent this problem and offer a helpful device for the rationalization of the phase behavior in the glycinin/2S mixture. Modeling was performed on the isothermal time course of gelation of single 11S and 2S gels and the mixtures of the two components at which the 11S content was fixed at 4.0% (Figure 4b), with the outcome being reproduced in Figure 8. It was verified that the gelation curves fitted first-order reaction kinetics, a reasonable result considering that network development is based on “collisions” of functional groups of the same reactant (i.e., polymeric segments of 11S or 2S). The integrated form of the first order rate law is as follows Kt ¼ ln
a a -x
ð4Þ
(38) Monteiro, S. R.; Tavares, C.; Evtuguin, D. V.; Moreno, N.; Lopes da Silva, J. A. Biomacromolecules 2005, 6, 3291. (39) Yan, H.; Saiani, A.; Gough, J. E.; Miller, A. F. Biomacromolecules 2006, 7, 2776. (40) Mano, J. F.; Viana, J. C. Poly. Test. 2006, 25, 953. (41) Batista, A, P.; Portugal, C. A. M.; Sousa, I.; Crespo, J. G.; Raymundo, A. Int. J. Biol. Macromol. 2005, 36, 135.
8546 DOI: 10.1021/la803290j
Figure 8. First-order reaction rate constants plotted as a function of protein concentration for single 11S (]) and 2S (0) preparations and their mixtures of 4.0% 11S with increasing concentrations of 2S (•).
where a and x represent the initial concentration of the reactant and the concentration of the product at time t, respectively. It is seen from eq 4 that the value of the rate constant, κ, depends only on the ratio or the relative values of two concentrations, thus obtaining the dimensions of reciprocal time (e.g., s-1). Because the original concentration of the reactant is not needed in order to find κ, eq 4 can be expressed as the ratio of two experimental parameters a Xf ¼ a -x Xf -X
ð5Þ
where Xf and X are the parameters at the end of the experiment and after time t, respectively. Combining eqs 4 and 5 and implementing a certain rearrangement gives the operational form for modeling the experimental observations of this work, which is G0 ðtÞ ¼ G0f ½1 - exp½-Kðt -to Þ
ð6Þ
where G0 f is the final storage modulus on shear, with to denoting the onset of gelation in the single and binary systems. Table 1 constitutes a concise account of the focal experimental data and parameters derived from eq 6, which was solved by nonlinear analysis using Statistica v. 8.0. It appears that the latent time of gelation, to, is not affected by the polymer concentration in single preparations of glycinin and 2S. The value of G0 f does, though, exhibiting rises of 7- and 16fold, respectively, with increasing solids content in the gel. The rate constant was thus calculated, and it is plotted in Figure 8 as a function of protein concentration. A considerable difference of 1 to 2 orders of magnitude was observed in the values of κ at equal contents of the two molecular fractions of soy, with those of 11S achieving, as expected, the highest rates of gelation. It takes quite a while for 2S to form a coherent network morphology, an observation that is further verified in the time required for the onset of gelation to occur in the presence of 0.4% added GDL (Table 1). Langmuir 2009, 25(15), 8538–8547
Kasapis and Tay
Article
Table 1. Summary of the Main Parameters (Experimental and Derived) Following First-Order Reaction Modeling of Single 11S and 2S Gels and Their Mixtures of 4.0% 11S with Increasing Concentrations of 2Sa 11S conc (%)
G0 final (Pa)
to (min)
2S κ 104 (s-1)
G0 final (Pa)
to (min)
11S + 2S κ 104(s-1)
G0 final (Pa)
to (min)
3.0 1000 36.1 4.067 18.7 538 0.50 3.5 1887 34.7 2.50 46.2 513 0.567 4.0 2551 37.6 2.183 55.8 524 0.417 2551 37.6 4.13 2789 16.2 4.25 2236 26.2 4.5 3153 38.2 1.90 70.6 502 0.183 2164 19.1 4.75 2025 17.6 5.0 4098 37.2 1.567 72.2 517 0.140 1736 22.7 5.5 4219 39.6 1.583 108.8 536 0.150 1109 29.2 6.0 4734 39.1 1.433 147.7 594 0.137 792.4 38.2 6.5 6269 31.1 1.383 685.5 46.1 7.0 7480 34.2 1.250 299.0 550 0.122 181.5 87.3 7.5 73.8 141 8.0 26.8 221 a The regression coefficient of the non-linear analysis for the gelation kinetics of single and mixed systems varied from 0.966 to 0.983.
Normally, gelation would proceed faster at higher concentrations of a given reactant, but the opposite occurrence is evaluated graphically in Figure 8. The drop in pH due to GDL hydrolysis to gluconic acid is very much the determinant of gelation kinetics in these systems, but the increasing buffering capacity of high protein concentrations diminishes the acidification and hence coagulation of the macromolecules. Clearly, two distinct regions are noted that can be attributed to the gelation kinetics of 11S and 2S at rate constant values above 1.25 10-4 s-1 and below 0.64 10-4 s-1, respectively. Changes in the molecular organization of the mixture are now evident with increasing additions of 2S to 4.0% 11S. Thus there is a transformation from a glycinin continuous matrix at 2S levels below 1.0% to a system where the small-molecularweight counterpart is capable of forming the dominant phase at concentrations higher than 3.0%. In conclusion, 2S dictates the structural characteristics of 11S at comparable concentrations of the two materials.
Langmuir 2009, 25(15), 8538–8547
κ 104 (s-1)
2.183 1.767 1.550 1.483 1.452 1.322 1.218 0.908 0.833 0.637 0.423 0.368
Finally, a quantitative assessment of the solvent partition between two polymeric constituents using blending laws is confined in regimes of a positive relationship between “performance” characteristics (e.g., enhanced strength of the composite gel) and increasing solid content in the system. This is the case for the 11S/7S and 11S/2S mixtures at 2S concentrations of up to 0.5% w/w. In the 7S/2S system, however, the addition of 2S results in an immediate decrease in the values of G0 , an outcome that does not allow the application of the blending-law model. Nevertheless, the loss in mechanical strength of the 7S/2S mixtures is similar to that recorded for 11S/2S at 2S concentrations above 0.5%, hence arguing that in both cases 2S is capable of sustaining a disproportionate phase of high solvent volume. Rationalization of the phase-separation phenomena along the lines advocated in this work may prove to be useful in tailor-made applications of the molecular soy protein fractions employed as functional components in processed biomaterials.
DOI: 10.1021/la803290j
8547
