Elaboration and Characterization of Whey Protein Beads by an

Whey protein beads were successfully produced using a new emulsification/cold gelation method. The principle of this method is based on an emulsifying...
0 downloads 0 Views 442KB Size
Biomacromolecules 2002, 3, 239-248

239

Elaboration and Characterization of Whey Protein Beads by an Emulsification/Cold Gelation Process: Application for the Protection of Retinol Lucie Beaulieu,† Laurent Savoie,‡ Paul Paquin,† and Muriel Subirade*,† STELA (Dairy Research Centre) and Groupe de recherche en nutrition humaine, Faculte´ des sciences de l’agriculture et de l’alimentation, Universite´ Laval, Que´ bec, Que´ bec, Canada G1K 7P4 Received May 2, 2001; Revised Manuscript Received December 5, 2001

Whey protein beads were successfully produced using a new emulsification/cold gelation method. The principle of this method is based on an emulsifying step followed by a Ca2+-induced gelation of pre-denatured (80 °C/30 min) whey protein. Beads are formed by the dropwise addition of the suspension into a calcium chloride (CaCl2) solution. IR results show that bead formation has a pronounced effect on the secondary structure of whey protein, which leads to the formation of intermolecular hydrogen-bonded β-sheet structures. Their preparation conditions (CaCl2 concentrations of 10, 15, and 20% (w/w)) influence their sphericity and homogeneity: an increase in CaCl2 favors regular-shaped beads. The physicochemical and mechanical characterizations of beads were also carried out. Their properties, such as swelling, elasticity, deformability, and resistance at fracture, change according to pH levels (1.9, 4.5, and 7.5) and preparation conditions. Indeed, protein chain networks exhibit different behavior patterns with respect to their charge. Finally, bead degradation by enzymatic hydrolysis reveals that beads are gastroresistant and form good matrixes to protect fat-soluble bioactive molecules such as retinol, that have in vivo intestinal absorption sites. The experiment demonstrated the potential of whey protein beads to protect molecules sensitive (i.e., vitamins) to oxidation. 1. Introduction The growing interest in effective and selective delivery of bioactive agents to the site of action has led to the development of new encapsulation materials. Despite the successful elaboration of many synthetic polymers as biodegradable microencapsulating media, natural polymers remain attractive agents that are extensively investigated. They have the potential advantages of great availability, low cost, low toxicity, and being easily modifiable.1 Although many wall materials are available for nonfood applications, very few are used in food applications.2 However, among the systems investigated, food proteins have recently received considerable attention because of their excellent functional properties.3 Proteins, such as gelatin,4,5 gliadin,6 human serum albumin,7,8 or egg albumin,9 have been used with success for encapsulating bioactive molecules. Whey proteins, also known as the serum proteins of milk, are widely used in food products because of their high nutritional value and their ability to form gels, emulsions, or foams.10 It has recently been shown, using a spray-drying technique,11,12 that whey proteins form spherical microcapsules. However, this technique involves high temperatures during the drying process and, consequently, limits its use to active, heat-resistant materials.13 Another method, which is based on a water-in-oil (w/o) emulsification with glutar* Corresponding author. Tel.: 418-656-2131-4278. Fax: 418-656-3353. E-mail: Muriel.Subirade@aln.ulaval.ca. † STELA (Dairy Research Centre). ‡ Groupe de recherche en nutrition humaine.

aldehyde cross-linking, has been developed for using whey protein microspheres.14 However, it has the disadvantages of requiring the use of an organic solvent, of being difficult to remove from the finished product, and of using glutaraldehyde, which restricts it to the biomedical field because of its toxic effects. Moreover, both methods exploit the whey protein emulsification properties, which have been extensively studied.15-17 Another important functional property of whey proteins is their ability to form heat-induced gel matrixes, capable of holding large amounts of water.18,19 Depending on the preparation techniques, gels can exhibit different microstructural properties, which are strongly related to the intimate structure of the aggregates.20,21 Recently, it has been shown that cold-induced gelation of whey proteins can be achieved by adding Ca2+ ions to a preheated protein suspension.22 This method requires a heating step during which whey proteins are denaturated and polymerized into soluble aggregates, followed by a cooling step and the subsequent salt addition, which results in the formation of a network via Ca2+mediated interactions of soluble aggregates. Ca2+-induced whey proteins cold gelation may be compared to alginate gelation resulting from a dimeric association of guluronic acid regions with Ca2+ in the “egg box” formation.23 Similarly, a mechanism of cross-linking carboxylate groups with Ca2+ has been suggested for the gelation of predenaturated whey proteins at ambient temperature.24 To overcome the limitations of existing methods (i.e., high temperature, organic solvents, and toxic agents), a new encapsulation method that uses whey proteins had to be

10.1021/bm010082z CCC: $22.00 © 2002 American Chemical Society Published on Web 01/22/2002

240

Biomacromolecules, Vol. 3, No. 2, 2002

developed. First, a two-phase process involving an emulsifying step followed by a Ca2+-induced gelation of predenatured whey protein was designed. Beads were then formed by the dropwise addition of the suspension into a calcium chloride solution according to the method used to produce calcium-alginate beads.25 Second, the molecular changes occurring in the whey protein during bead formation were followed using Fourier transform infrared spectroscopy (FTIR). This technique is versatile and powerful for determining the secondary structure of proteins in aqueous solutions,26 complex systems, such as biological system,27,28 or functional state.20,21 In the third part of the work, the physicochemical and mechanical characterizations of the beads were studied with respect to CaCl2 concentrations (10, 15, 20% w/w) and pH levels (1.9, 4.5, 7.5). The swelling ratio, one of the most important factors affecting the drug release characteristics in drug delivery systems,29 was determined. Indeed, the drug release is dependent on the swelling of the matrix. Thus, the matrix has the ability to release drugs in response to changes in environmental variables such as temperature, pH, ionic strength, etc. As for pH-sensitive drug delivery systems, several studies that address the relationships between the swelling ratio of the vehicle and the drug release characteristics were conducted.30,31 Mechanical properties of the beads were also determined since they are of great importance when they have to be used in a bioreactor,32 implanted in vivo,33 or used in food processes that possibly undergo different treatments such as cutting, slicing, spreading, or mixing.34 In the fourth part of the work, stability assays were carried out with a selected batch of beads using an in vitro protease degradation. Bead susceptibility to proteolytic enzymes has been studied using a two-step proteolysis, which first consisted of the predigestion of beads with pepsin followed by pancreatin.35 Finally, the protector effect of the beads was studied using retinol as a model molecule. Retinol and its derivatives are particularly sensitive to environmental conditions, particularly to the stomach acid pH.36 2. Material and Methods 2.1. Material. Whey protein isolates (WPI) were obtained from Davisco (Food International, Inc, Le Sueur, MN). WPI protein content was 93% (dry matter basis), as determined by the Kjeldahl method (nitrogen X 6.38).37 Soybean oil used to form the emulsions was purchased from a local commercial store (Metro Co., Canada). all-trans-Retinol (vitamin A) was procured from Sigma (Oakville, ON, Canada) while R-tocopherol was acquired from Aldrich Chemical Co., Inc. (Milwaukee, WI). The enzymes used in the study were pepsin 1:60000, from porcine stomach mucosa, crystallized and lyophilized (Sigma Chemical Co., St. Louis, MO), and pancreatin 5X, from hog pancreas (ICN Nutritional Biochemicals Cleveland, OH). Thimerosal (J. T. Baker, Phillipsburg, NJ) was used to prevent bacterial growth, and taurocholic acid, the sodium salt form (Sigma), was used as an emulsifying agent. The chemical products used in high-performance liquid chromatography (HPLC) are dichloromethane (DCM), metha-

Beaulieu et al.

nol (MeOH), acetonitrile (MeCN), and hexane, all HPLC grade (Omnisolv, EM Science, VWR Canlab, Mississauga, Canada), then ammonium acetate (ACS reagent, Sigma) and triethylamine (Sequanal grade, Pierce Chemical Co., Rockford, IL). The HPLC system used is a Hewlett-Packard system, series 1050, manufactured in Germany, which consisted of a system pump (series 1050), an autoinjector (series 1050), and a UV-vis absorbance detector. The system was equipped with the Windows HPChemstation software (Hewlett-Packard Co. HPChemstation System Software Rev. A. 03.03, USA). The HPLC column (Licrosphere 100 RP18e, 5 µm, 250 × 4 mm) was purchased from HewlettPackard Co. (Wilmington, DE). 2.2. Manufacturing Whey Protein Beads. WPI was hydrated in deionized water (8%, w/w). The solution was agitated for 1 h at room temperature and allowed to rest for 2 h before heating in order to permit a good protein hydration. WPI solution was adjusted at pH 8. The solution was heated at 80 °C for 30 min and simultaneously mixed at 300 rpm in a cooker (Stephan U. Sohne Gmbh & Co., Germany). The solution was cooled for 1 h at room temperature (∼23 °C) and then stored overnight at 4 °C. The following day, the solution was equilibrated to room temperature (∼23 °C) and used to produce the emulsion. Protein concentration and oil proportion in the emulsion were 5.6% and 30%, respectively. Prior to preparing the emulsion, the WPI solution and soybean oil were prehomogenized and mixed using an Ultra-Turrax (Janke & Kunkel, IKALabortechnik, Germany). The mixture was then homogenized using an Emulsiflex-C5 high-pressure homogenizer (AVESTIN Inc., Ottawa, Canada). Emulsion preparation was initially performed at 100 MPa pressure and then at 3 MPa. The resulting emulsion was added dropwise into 100 mL of 10, 15, or 20% (w/w) CaCl2 solutions, using a hydraulic pump (Allo Kramer Shear Press, model SP 12, Rockville, MD) equipped with a syringe and needle (Terumo Medical Corp., Elkton, MD). Magnetic stirring was maintained during the gelation. The resulting beads were rinsed with distilled water and dried in P2O5. Retinol-loaded beads were prepared by adding retinol in the soybean oil. First, retinol was solubilized in 95% ethanol and added to soybean oil (0.5% w/w). Retinol concentration in emulsion is 5000 µg/100 g of emulsion. Experimental material used for manipulation of retinol was covered with aluminum foil in order to avoid retinol photodegradation. In addition, nitrogen was injected in surface solutions to prevent oxidation. 2.2.1. Bead Morphology Analyses. Observations of external bead structure were made by macrophotographs using a Minolta camera (35-mm XG-M) with a 55-mm macro lens. 2.2.2. Internal Bead Structure Analyses by Transmission Electron Microscopy (TEM). Beads were fixed with formaldehyde 4% (cacodylate buffer 0.1 M) for 2 h, dehydrated in graded ethanol series, embedded in LRWhite resin, and polymerized under UV. Materials were collected on Formvar-coated nickel grids and stained with uranyl acetate and lead citrate. Observations were carried out under a JEOL 1200X electron microscope.

Characterization of Whey Protein Beads

2.3. Physicochemical and Mechanical Characterizations. 2.3.1. Infrared Measurements. Infrared spectra at a 2 cm-1 resolution were recorded with a Magna 560 Nicolet spectrometer (Madison, WI) equipped with a mercurycadmium-telluride (MCT) detector. The spectrometer was continuously purged with dried and CO2-free air. All experiments were run on a horizontal attenuated total reflectance (ATR) crystal (ZnSe), where solutions and beads were pressed. Each spectrum is the average of 128 scans and apodized with Happ-Genzel function. For the study of the amide I region of the protein, subtractions and Fourier self-deconvolution were performed with the software provided with the spectrometer (Omnic software). The band narrowing was achieved with a full width at half-height of 18 cm-1 and a resolution enhancement factor of 2.20 2.3.2. Swelling Experiment or Water Uptake Ability. Predetermined amounts of dried whey protein beads were placed in a monosodium phosphate buffer solution (0.02 M containing 0.13 M NaCl) at different pH values: pH 1.9, which corresponds to acid stomach pH; pH 4.5, which is near the pI of whey protein; and pH 7.5, which represents the physiological intestinal pH. Temperature was maintained at 37 °C in an incubator. After 6 h, the beads in their equilibrium-swollen state were weighed. Bead swelling ratios were determined from the weight change before and after swelling, expressed in percentages swelling ratio38 or water uptake ability (%)39 ) [(Ww - Wd)/Wd] × 100 where Ww and Wd represent the weight of wet and dry beads, respectively. 2.3.3. Compression Studies. The beads were studied by means of a texture analyzer TA-XT2 version 5.15 (50 N maximum force, precision of 0.001 N; Stable Micro Systems, Haslemere, Surrey, United Kingdom). The apparatus was equipped with a 20 mm diameter cylindrical piston. Each measurement was carried out at room temperature on one bead, which was placed under the piston on a fixed bottom plate. For each CaCl2 concentration (10, 15, and 20% (w/ w)), the measurements were repeated on 2 batches of beads, and on 10 beads per batch. 2.3.2.1. Rupture Study: Stress and Strain at Fracture. The piston went down, keeping contact with the top of the bead, and flattened the bead at a constant rate of 0.2 mm/s, until it reached 90% of its original height. The force exerted by the bead as a function of displacement was recorded. The return speed of the piston to its initial position after compression was 10 mm/s. The force needed for deformation was recorded as a function of time until fracturing of the bead. A force-compression curve was obtained for each sample and stored in a file for calculation of the fracture properties using the “XT.RAD Dimension” software, version 3.7H, from Stable micro System. From each measurement, the stress and strain at fracture were determined. The fracture stress is associated with the first peak on the graphs, representing the force as a function of displacement. Stresses (σ, N m-2) were calculated by dividing the force, registered at every point, by the corre-

Biomacromolecules, Vol. 3, No. 2, 2002 241

sponding bearing area. For gel beads, the stresses were calculated considering the contact area as the area of a sphere and assuming a dissipation of the internal beads force in all directions.40 The fracture strain (), expresses bead deformability, and is calculated as follows41 fracture strain () ) |ln(h0 - ∆h)/h0| where h0 is the initial height and ∆h the change in height. The strain is obtained by relating any strain increase (in an already strained sample) to changes in sample dimension. 2.3.2.2. Stress Relaxation. The piston went down at the rate of 0.2 mm/s until it reached 50% of deformation at bead rupture. The piston then stayed motionless at this position for 30 s and finally returned to its initial position. From the graphs representing the force versus time, the instantaneous resistance strength (F1), which is the force measured when the piston had just reached its maximum displacement, and (F2), the force opposed by the bead after 30 s, are obtained. From these values, sample elasticity was calculated as the ratio of F2 to F1 and expressed as a percentage stress relaxation (%) ) (F1 - F2)/F1 × 100 When the value of stress relaxation is high, the elasticity is low and vice versa.42 2.4. In Vitro Degradation Assays. The enzymatic degradation assay was conducted using a modified version (Lalancette43) of the method of Gauthier et al.:44 Beads (125 mg protein) were suspended in 15 mL of 0.1 N HCl (50 mg/mL thimerosal) in a flat-bottom glass tube and stirred magnetically for 10 min at 37 °C. The volume of the digestion mixture was adjusted to 20 mL, and 0.5 mL of pepsin solution (1 mg/mL 0.1 N HCl) was added to initiate the hydrolysis. The digestion was carried out for 30 min and stopped by raising the pH to 7.5 with NaOH. A concentrated monosodium phosphate solution (1 mL; 0.5 M, contained NaCl 3.25 M, pH 7.5) and taurocholic acid (0.5 mL; 0.25 M) were added, and the reaction mixture was adjusted to 25.5 mL with distilled water. The reaction was initiated by adding 0.5 mL of pancreatic enzymes (10 mg/mL) prepared in monosodium phosphate buffer (0.02 M, contained NaCl 0.13 M, pH 7.5). The final volume is 25 mL because the magnetic bar takes up a volume of 1 mL. The digestion was carried out for 6 h and stopped by placing the tube on ice. The end of lysis was defined as the time it took for all particles to disappear. 2.5. Release Experiments. 2.5.1. In Vitro Retinol Release. The in vitro retinol release assays were conducted as described above for the in vitro degradation assays. Samples (2 mL) were collected at every hour. All assays were done in triplicate and controls were carried out without enzymes. 2.5.2. Retinol HPLC Analysis. The retinol HPLC dosage was adapted from the method described by Sun (1999).45 The HPLC system was operated at 35 °C with a flow rate of 0.75 mL/min. UV detection was at 325 nm for retinol and 290 nm for R-tocopherol. A reversed-phase column was used with an isocratic solvent system (MeCN, 100%) instead

242

Biomacromolecules, Vol. 3, No. 2, 2002

of a step gradient used with three mixtures of solvents: MeCN, MeOH, DCM. The sample preparation was based on extraction methods described by Qian and Sheng (1998)46 and Thompson et al. (1980).47 First, the samples (2 mL) were withdrawn from the digestion mixture, filtered through a 0.5-µm filter (Whatman Inc., Clifton, NJ) and transferred to a 50-mL screw-capped centrifuge tube (Oak Ridge Centrifuge tubes, in polypropylene, Nalgene). Then, 5 mL of absolute ethanol was added. The mixture was shaken on a vortex mixer for 1 min, the sample was allowed to rest for 5 min, and 5 mL of hexane was added. The mixture was shaken vigorously with a vortex mixer for 30 s and allowed to stand for 2 min. These mixing and standing procedures were repeated twice. A quantity of 3 mL of water was added, and the tube was inverted several times. The tube was flushed with a stream of N2 to protect the vitamin from air exposure before its agitation in a rotated cooker (Robbins Scientific Corp., Sunnyvale, CA, model 400) during 1 h. After centrifugation (Sorvall, SS-34) at 5000 rpm for 15 min, 2 mL of hexane (upper phase) was transferred to a tube covered with aluminum foil and evaporated under nitrogen to remove the solvent. The residue was redissolved in mixtures of solvents: DCM-MeCN-MeOH (15:65:20), containing internal standard R-tocopherol at a 10 µg/mL concentration. The sample was filtered through a 0.45-µm filter (Acrodisc 4, HPLC Certified) and put to a 2-mL amber vial before being injected into the HPLC system. Retinol recovery was evaluated by measuring retinol concentration after the extraction method using known concentrations, the same as those used for the calibration curves: 0.05, 0.1, 0.2, 0.4, 0.8, and 1.0 µg/mL. Retinol recovery was 92 ( 8%, and the retinol detection limit was 2.5 ng. 2.7. Statistical Analysis. The combined effects of CaCl2 concentration (10, 15, and 20%) and pH (1.9, 4.5, and 7.5) on swelling, fracture stress, and strain and stress relaxation were studied using a factorial experimental design (3 × 3). Data were analyzed by the Statistical Analysis System (SAS Institute, Inc., Cary, NC) using the general linear model (GLM) procedure for regression analyses, the ANOVA procedure for analysis of variance, and the Levine test to verify variance homogeneity. Analysis of variance was used to determine whether the factors and their interaction had a significant effect on the measured properties. Statistical analyses were performed at R ) 0.05. Error bars on graphs represent standard error obtained from the statistical model.

Beaulieu et al.

Figure 1. Micrographs of whey protein beads: (a) prepared with 10% CaCl2 concentration (w/w); (b) prepared with 15% CaCl2 concentration; (c) prepared with 20% CaCl2 concentration. (Scale bar represents 1 mm.)

3. Results and Discussion 3.1. Bead Formation. 3.1.1. Bead Morphology. Figure 1 shows the macrophotographs of whey protein beads prepared with different calcium chloride (CaCl2) concentrations: 10% (a), 15% (b), and 20% (w/w) (c). The result shows that the CaCl2 concentrations used in the extrusion step have an influence on both the size and appearance of the beads. Indeed, when the CaCl2 concentration increases from 10 to 20%, the size of the beads decreases from 2.1 to 1.8 mm. Moreover, the shape of the beads becomes more

regular and spherical with higher concentrations. At 10% (w/w) concentration, the beads have an irregular shape and aggregate together, while at 15% (w/w) concentration, beads are rounder. Conversely, at 20% (w/w) concentration of CaCl2, the beads are regular and spherical in shape and are characterized by a smooth surface. The increase in sphericity with higher CaCl2 concentrations is interesting since this characteristic is expected in controlled delivery because it allows a constant release.48 The higher sphericity with the elevated CaCl2 concentration may be due to an increase in

Characterization of Whey Protein Beads

the kinetic mechanism of gelation with calcium chloride concentration. Indeed, it has recently been shown that this parameter is likely to be a major determinant in the aggregation process.49 Ca2+ acts as a bridge between protein molecules and favors intermolecular interactions resulting in the aggregation process. Moreover, Ca2+ binding to unfolded protein molecules causes an increase in the reactive sulfhydryl group content thereby participating more easily in the aggregation process.50 Therefore, it is likely that the increase in CaCl2 concentration increases protein-protein interactions and results in further aggregation of the protein to form a gel network.51 3.1.2. Internal Microstructure Analyses of Beads by Transmission Electron Microscopy (TEM). Figure 2 displays microstructures of selected beads prepared with various CaCl2 concentrations: 10% (a), 15% (b), and 20% (w/w) (c). Each image shows a uniform (homogeneous) oil globule distribution in a gel protein network. The micrographs show that increasing CaCl2 concentration from 10% (w/w) to 20% (w/w) resulted in smaller fat globules and in a more homogeneous network. This suggests that increasing CaCl2 concentration prevents coalescence of oil droplets in the protein network. It is known that coalescence is a phenomenon that results from the fusion of individual droplet emulsion into bigger droplets and leads to an increase in average particle size.52 During the emulsification step of bead formation, thermal pre-denatured proteins, acting as an emulsifier, rapidly adsorb to the surface of the oil droplets. The large negative change in free energy associated with protein adsorption creates a stabilizing layer that protects the fine droplets against coalescence and provides physical stability to the emulsion.17 In the second step, addition of Ca2+ reduces the electrostatic stabilization of the emulsion, which could favor the coalescence of some droplets. Increasing Ca2+ enhances the gelation kinetic, which leads to the rapid entrapment and stabilization of the droplets by the protein network. Higher Ca2+ levels also reinforce attractive electrostatic interactions between adsorbed proteins on adjacent droplets and Ca2+. Calcium acts as a bridge between adjacent emulsion droplets and favors their aggregation without disruption of the protective stabilizing protein layer at the interface thereby preventing their coalescence. 3.2. Physicochemical and Mechanical Bead Characterization. 3.2.1. Molecular Changes of Whey Proteins Induced by Bead Formation. Figure 3 displays the deconvoluted spectra of whey proteins in solution (A), after thermal denaturation (B) and in beads (C) in the amide I region (1600-1700 cm-1) recorded at room temperature (20 °C). This region, which is mainly due to CdO stretching vibration and to a small extent C-N stretching vibration of the peptide bonds, is sensitive to the secondary structure of proteins.53 In solution (Figure 3A), the amide I band is composed of eight individual components at 1615, 1629, 1639, 1652, 1662, 1672, 1683, and 1693 cm-1, which correspond to particular secondary structures. The assignment of these components is well-known.53-56 The strongest band at 1629 cm-1 is characteristic of amide groups involved in strongly bonded β-strands, whereas the component at 1639 cm-1 is attributed to β-sheets. The presence of the component at 1683

Biomacromolecules, Vol. 3, No. 2, 2002 243

Figure 2. Representative TEM image of the internal structure of whey protein beads: (a) prepared with 10% CaCl2 concentration (w/w); (b) prepared with 15% CaCl2 concentration; (c) prepared with 20% CaCl2 concentration. (Scale bar represents 1 µm.)

cm-1 suggests that it consists of antiparallel β-sheets. The components at 1652 and 1662 cm-1 are assigned to a-helix/ unordered structures and turns, respectively. The ones at 1672 and 1693 cm-1 are both related to β-sheet and turn, while

244

Biomacromolecules, Vol. 3, No. 2, 2002

Beaulieu et al.

Figure 3. Deconvoluted infrared spectra of whey protein in aqueous solution (8% w/v) (A), after thermal denaturation (B), and in a bead (C).

the small component at 1615 cm-1 is due to the vibration of amino acid residues. Since bead formation requires a preheated step, we recorded the spectrum of whey proteins after the thermal treatment at 20 °C. As expected and according to previous studies,20,21 the spectrum (Figure 3B) reveals that major conformational changes occur upon thermal treatment. First, a new component arises at 1619 cm-1, whereas the one located at 1694 cm-1 increases in intensity. The former is well-known and is assigned to intermolecular β-sheets resulting from aggregation,57 whereas the presence of the latter indicates that the β-sheets are antiparallel.58 At the same time, the amide I band broadens as a result of a further loss of secondary structures. The spectrum contains a featureless broad band between 1680 and 1625 cm-1, with a maximum near 1645 cm-1. Such a profile is indicative of unfolded polypeptide segments. When beads are formed, the spectrum (Figure 3C) keeps the same profile showing that the protein remains aggregated. However, there is a redistribution of broad band intensities in the 1680-1625 cm-1 range compared to the thermal denatured spectrum. At 20 °C and before beads are formed, there is a maximum located at about 1645 cm-1, while in beads, two maximums can be noticed at 1634 and 1652 cm-1, suggesting a reorganization of the polypeptide chain within the network on bead formation. From these results, it is clear that the molecular structure of whey proteins in solution is quite different from that obtained in beads. In solution, whey proteins are mainly composed of β sheets structures due to the predominance of β-lactoglobulin, a β-sheet protein (i.e., as confirmed by crystallographic59,60 and IR data20,21). Upon bead formation, the IR spectrum exhibits significant changes showing that whey proteins are mainly composed of intermolecular β-sheet network, even though residual secondary structures are observed. 3.2.2. Swelling Ratio. Figure 4 displays the equilibriumswelling ratio of the beads as a function of CaCl2 concentration as well as the pH of the swelling medium. The statistical analysis shows that the effect of pH levels on the bead swelling ratio is influenced by the CaCl2 concentration (p < 0.05). The figure reveals that the medium pH has a striking effect on the swelling of the beads. It is at a minimum at pH 4.5, near the pI (5.2) of the whey protein, and increases with changes in pH values (increased, intestinal pH (7.5), or decreased, gastric pH (1.9)). These results suggest that bead

Figure 4. Swelling ratio (%) of beads as a function of CaCl2 concentration (10, 15, 20% w/w) and pH (1.9, 4.5, 7.5).

swelling is mainly governed by the net charge of the protein molecules. At pI, the net charge of the whey protein molecule is at a minimum, which translates into low electrostatic repulsions between chains and results in a low swelling ratio. Protein-protein interactions are favored by protein-solvent interactions. However, as the pH differs from pI, the net charge of the whey protein molecule increases (positive below pI, negative above pI), leading to high electrostatic repulsive forces and an increase in the swelling ratio. The beads are highly swollen at intestinal pH (7.5). This high equilibrium-swelling ratio can be attributed to the electrostatic repulsive force originating from the negative charge of the ionized carboxyl groups, suggesting that these groups are mainly involved in the pH-sensitive swelling property. At the gastric pH (1.90), the beads are less swollen. This suggests that the low repulsion electrostatic interactions, between positive charges, caused by the protonation of the amine groups on the protein chain, resulted in a low network swelling, but their contribution cannot be ruled out of the pH-sensitive swelling mechanism. It can therefore be concluded that the ionizable and/or ionized groups are the major factors that govern the pH-sensitive swelling mechanism of the beads. Although CaCl2 does not have a significant effect on the swelling ratio, we can note a higher swelling trend at 20% CaCl2. This may be due to a more homogeneous protein network at this concentration, as indicated by the internal bead structure, which improves the water-trapping capacity of the gel.61 3.2.3. Rupture Study. 3.2.3.1. Rupture Strength. Figure 5 shows the results of stress measurements at fracture (N m-2) as a function of CaCl2 concentration and pH. The statistical analysis shows that the effect of environment pH on shear stress at bead failure is dependent on the CaCl2 concentration (p < 0.05). The figure shows that higher pH values increase the shear stress of the beads. The shear stress is smaller at pH 1.9 and is relatively constant at pH 4.5 and pH 7.5. Consequently, the resistance at bead failure is higher at both these pH values (4.5, 7.5) compared to pH 1.9. It is interesting to note that at pH 7.5 and 4.5, the beads exhibit similar rupture strengths, despite their different swelling

Characterization of Whey Protein Beads

Biomacromolecules, Vol. 3, No. 2, 2002 245

Figure 5. Fracture stress (N m-2) of beads as a function of CaCl2 concentration (10, 15, 20% w/w) and pH (1.9, 4.5, 7.5).

properties. This unexpected result could be explained by interactions in the protein network. As seen before, at pH 4.5, i.e., near the isoelectric point of β-lactoglobulin, the protein-protein interactions (aggregates) are favored leading to a high shear stress. The rigid structure of the beads in these conditions increases their hardness. At pH 7.5, repulsive electrostatic interactions, between negative charges, prevented the formation of protein-protein interactions and favored the swelling of the bead’s internal network. Thus, the resulting elasticity improves the fracture strength of the beads, which adopt a rubber-like behavior. At pH 1.9, the low repulsion electrostatic interactions, between positive charges, caused a low network swelling and allowed a weak shear stress. The fracture stress is also affected by calcium concentration. Higher calcium concentrations result in lower rupture strength of the beads. Similar results have previously been reported;52 the authors showed that increasing CaCl2 concentration at low protein concentrations ( 0.05), between pH and CaCl2 concentration. The figure shows that bead deformability is relatively constant at pH 1.9 and 4.5 and increases at pH 7.5. As expected, the high swelling ratio obtained at pH 7.5 allows for a greater deformability compared to other pH values. As seen in the figure, the concentration of CaCl2 does not significantly affect shear strain at failure even though lower values were observed at 20% CaCl2 concentration. 3.2.3.3. Stress Relaxation. Figure 7 shows the results of the measurements of stress relaxation as a function of CaCl2 concentration and pH. The effect of environment pH on stress relaxation of the beads was influenced by the CaCl2

Figure 6. Fracture strain of beads as a function of CaCl2 concentration (10, 15, 20% w/w) and pH (1.9, 4.5, 7.5).

Figure 7. Stress relaxation (%) of beads as a function of CaCl2 concentration (10, 15, 20% w/w) and pH (1.9, 4.5, 7.5).

concentration (p < 0.001). The beads stress relaxation increases with pH, up to a maximal value obtained with pH 4.5. Then the stress relaxation considerably decreased at pH 7.5. This means that beads exhibit a higher elasticity at pH 7.5 and a lower one at pH 4.5. These results concur with those previously obtained for swelling properties and might be explained by the effect of the net charge of the protein molecules that favors, depending on its value, either proteinprotein or protein-solvent interactions. As shown earlier, the type of interactions in the protein networks influences the swelling properties of beads and, therefore, their elasticity, which is favored by the swelling of protein network at pH 7.5. Increasing CaCl2 concentration decreases stress relaxation. Consequently, beads have a better elasticity at 20% CaCl2 concentration, possibly because of the internal bead structure, and this confirms the previous explanation. Globule distribution homogeneity in the protein network is conducive to improved flexibility and favors bead elasticity. 3.3. Enzymatic Degradation. Beads prepared with CaCl2 20% (w/w) were degraded using a method that consisted in

246

Biomacromolecules, Vol. 3, No. 2, 2002

Beaulieu et al.

Figure 9. Release of retinol from beads in the presence and in the absence of digestive enzyme.

Figure 8. Micrographs of beads (a) prepared with 20% CaCl2 concentration (w/w), (b) after a 30-min gastric incubation, and (c) after a 6-h pancreatic incubation.

a two-step proteolysis performed at 37 °C and included a pepsin predigestion at pH 1.9, followed by a hydrolysis with pancreatic enzymes at close to neutral pH. Figure 8 shows the macrophotographs of beads during in vitro digestion: intact bead (a), after gastric incubation (b), and after pancreatic incubation (c). This figure reveals that the beads exhibit a resistance to the hydrolytic action of pepsin but that they are totally degraded in pancreatic media. Indeed, macroscopic bead examination, before and after gastric incubation, shows a very slight degradation suggesting that the beads are gastroresistant. As for enzymatic specificity,

pepsin is known to preferentially attack peptide bonds involving hydrophobic aromatic amino acids. In its native structure, the major protein of whey, β-lactoglobulin (β-lg), is known to be resistant to pepsin62 since its hydrophobic amino acids are located in the internal core of its calyx-like structure.59 In the initial step of bead formation, the protein molecules are heated above their denaturation temperature leading to a disruption of both their tertiary and the H-bonded secondary structures. The primary importance of the denaturation process is to expose functional groups, such as CO and NH of peptide bonds, side-chain amide groups, and hydrophobic amino acids. The thermal denaturation of whey proteins was therefore expected to cause a significant increase in the susceptibility of proteins to proteolysis degradation, particularly as far as peptic digestion is concerned. However, in the emulsification step of bead formation, the hydrophobic amino acids that are adsorbed at the surface of the oil droplets are trapped in the protein network by adding Ca 2+. The hydrophobic amino acids are thus masked, which prevents the action of pepsin. As for degradation by pancreatin, beads were completely destroyed within 6 h. After this incubation time, only fat globules remained in the solution. The degradation by pancreatin would then be attributed to the combined effect of the proteases, mainly trypsin, chymotrypsin, and elastase, which catalyze the hydrolysis of the peptide bonds, but with different specificities. The action of trypsin is known to be restricted to the peptide links that involve the carboxylic groups of lysine and arginine, chymotrypsin is specific to bulky hydrophobic residues preceding the scissile peptide bond, and elastase is specific to small neutral residues. It can therefore be concluded that bead degradation is mainly enteric and that these beads can be used as matrix to protect fat-soluble bioactive molecules that are sensitive to stomach pH. 3.4. Release Study of the Retinol. Figure 9 displays the release of retinol in the presence or in the absence of digestive enzymes. This figure shows the plot of the data expressed as the cumulative amounts of retinol released (%) by the whey protein beads as a function of the time. In the

Characterization of Whey Protein Beads

absence of enzymes, retinol is weakly released and its rate is relatively constant. Only 5-10% of the incorporated retinol is released even after 6.5 h, which could be due to a diffusion phenomenon associated to the swelling of the beads. In the presence of enzymes, retinol is little released in the first 30 min, due to gastric digestion. Its release was mostly the same as the release in the absence of enzymes (about 5-10%) and suggests a weak diffusion phenomenon in the stomach. Then, a higher rate of retinol release is seen from beads during pancreatic digestion. Retinol is released progressively in this phase and suggests an in vivo release along its intestinal absorption sites. This suggests that retinol release is mainly due to the biodegradation of the matrixes. The release of retinol from whey protein beads was 75-95% in about 4.5 h and remained constant thereafter. After a certain time, the decrease of the reaction velocity could be due to matrix degradation. This suggests a progressive release of retinol in accordance with the matrix degradation. HPLC shows no apparent sign of oxidation of the retinol after its release suggesting a potential protective effect of the beads (results not shown). 4. Conclusion This work has allowed the development of a new encapsulation method that exploits whey protein emulsification and gelation properties. The emulsification/cold gelation procedure outlined in the present study illustrates an innovative technique for producing whey protein beads. Their physicochemical, mechanical, and degradation properties are interesting. First, bead formation has a pronounced effect on the secondary structure of whey protein, which leads to the formation of intermolecular hydrogen-bonded β-sheet structures as revealed by IR results. Second, Ca2+ modulated the spherical shape of the beads as well as their characteristics: at a high calcium chloride concentration, beads have a lower shear stress and a better elasticity. The gel aggregation is affected by the conditions of the gelation process. Third, bead hydration is dependent on pH medium and involves an improvement of elasticity. At high water content, resistance at fracture could be elevated. Bead protein chains reorganize their interactions according to environmental conditions. Last, bead degradation is mostly enteric. It thus, seems likely that beads are not susceptible to enzymatic attack during a rapid transit in the stomach; the action is prevented by the bead structure. The results of this research demonstrate that beads at a 20% (w/w) CaCl2 concentration have an excellent capacity to encapsulate bioactive molecules that are hydrophobic and sensitive to stomach pH (i.e., retinol). These spherical and elastic beads are composed of a homogeneous distribution of globules in a protein network. These beads therefore appear as promising matrixes with potential applications in various fields, such as food, pharmaceutics, and cosmetics. Acknowledgment. The authors thank the Natural Sciences and Engineering Research Council of Canada (NSERC)sIndustrial chair as well as industrial partners (Agropur, Novalait, and Parmalat-Canada) for their financial

Biomacromolecules, Vol. 3, No. 2, 2002 247

support. They wish to thank Dr. M. Britten for helpful discussions and A. F. Allain, J. Marin, and L. Tremblay for their technical assistance. Muriel Subirade acknowledges the NSERC-Canada Research Chairs Program for its financial contribution. References and Notes (1) Chasin, M.; Langer, R.; Swarbrick J. Biodegradable Polymers as Drug DeliVery Systems; Marcel Dekker Inc: New York, 1990. (2) Dziezak, J. D. Food Technol. 1988, 42, 136. (3) Shahidi, F.; Han, X. Q. Crit. ReV. Food Sci. Nutr. 1993, 33, 501. (4) Draye, J. P.; Delay, B.; Van de Voorde, A.; Van Den Bulcke, A.; Bogdanov, B. Biomaterials 1998, 19. (5) Digenis, G. A.; Gold, T. B.; Shah, V. P. J. Pharm Sci. 1994, 83, 915. (6) Duclairoir, C.; Irache, J. M.; Nakache, E.; Orecchioni, A. M.; Chabenat, C.; Popineau, Y. Polym. Int. 1999, 79, 327. (7) Chen, G. Q.; Lin, W.; Coombes, A. G. A.; Davis, S. S.; Illum, L. J Microencapsulation 1994, 11, 395. (8) Edwards-Le´vy, F.; Le´vy, M. C. Biomaterials 1999, 20, 2069. (9) Torrado, S.; Torrado, S. Drug DeV. Ind. Pharm. 1996, 22, 451. (10) Fox, P. F. DeVelopments in Dairy Chemistry; Applied Sci: London, 1982. (11) Rosenberg, M.; Young, S. L. Food Struct. 1993, 12, 31. (12) Young, S. L.; Sarda, X.; Rosenberg, M. J. Dairy Sci. 1993, 76, 2868. (13) Risch, S. J.; Reineccius, G. A. Encapsulation and Controlled Release of Food Ingredients; ACS Symposium Series 590; American Chemical Society: Washington, DC, 1995. (14) Heelan, B. A.; Corrigan, O. I. Microencapsulation 1998, 15, 93. (15) Damodaran, S.; Paraf, A. Food proteins and their applications; Marcel Dekker, Inc.: New York, 1997. (16) Dickinson, E.; Hong, S. T. J. Agric. Food Chem. 1994, 42, 1602. (17) Dickinson, E. J. Dairy Sci. 1997, 80, 2607. (18) Schmidt, R. H.; Illingworth, B. L.; Ahmed, E. M.; Richter, R. L. J. Food Process. PreserV. 1978, 2, 11. (19) Aguilera, J. M. Food Technol. 1995, 10, 83. (20) Lefe`vre, T.; Subirade, S. Biopolymers 2000, 54, 578. (21) Allain, A. F.; Paquin, P.; Subirade, M. Int. J. Biol. Macromol. 1999, 26, 337. (22) Barbut, S.; Foegeding, E. A. J. Food Sci. 1993, 58, 867. (23) Dziezak, J. D. Food Technol. 1991, 45, 115. (24) Roff, C.; Foegeding, E. A. Food Hydrocolloids 1996, 10, 193. (25) Hackel, U.; Klein, J.; Megent, R.; Wagner, F. W. Eur. J. Appl. Microbiol. 1975, 1, 291. (26) Subirade, M.; Gueguen, J.; Pe´zolet, M. Biochim. Biophys. Acta 1994, 1205, 239. (27) Subirade, M.; Salesse, C.; Marion, D.; Pe´zolet, M. Biophys. J. 1995, 69, 974. (28) Lefe`vre, T.; Subirade, M. Biochim. Biophys. Acta 2001, 15, 37. (29) Rosoff, M. Controlled Release of Drugs: Polymers and Aggregate System; VCH Publishers: New York, 1989. (30) Park, H. Y.; Choi, C. R.; Kim, J. H.; Kim, W. S. Drug DeliVery 1998, 5, 13. (31) Peppas, N. A.; Klier, J. J. Controlled Release 1991, 16, 203. (32) Martins dos Santos, V. A. P.; Leenen, E. J. T. M.; Ripoll, M. M.; van der Sluis, C.; van Vliet, T.; Tramper, J.; Wijffels, R. H. 1997, 56, 517. (33) Joly, A.; Desjardins, J. F.; Fremond, B.; Desile, M.; Campion, J. P.; Malledant, Y.; Lebreton, Y.; Semana, G.; Edwards-Le´vy, F.; Le´vy, M. C.; Cle´ment, B. Transplantation 1997, 63, 795. (34) Moskowitz, H. R. Food Texture, Instrumental and Sensory Measurement; Marcel Dekker, Inc.: New York, 1987. (35) Savoie, L.; Gauthier, S. F. J. Food Sci. 1986, 51, 494. (36) Toursel, P. PROCESS 1999, 1147, 44. (37) NORM IDF, 20B: 1993. Kjeldahl (nitrogen) methods. (38) Choi, Y. S.; Hong, S. R.; Lee, Y. M.; Song, K. W.; Park, M. H.; Nam, Y. S. Biomaterials 1999, 20, 409. (39) Park, H. Y.; Song, I. H.; Kim, J. H.; Kim, W. S. Int. J. Pharm. 1998, 75, 231. (40) Oudet, C. Polyme` res: structure et proprie´ te´ s; Masson: Paris, 1994. (41) Peleg, M.; Bagley, E. B. Physical properties of foods; AVI Publishing, Inc.: Wesport, 1983. (42) Mitchell, J. R. J. Texture Stud. 1980, 11, 315. (43) Lalancette, C. De´veloppement d'une me´thode pour mesurer l′impact de la nature des lipides sur la cine´tique d′hydrolyze des triglyce´rides. Me´moire maıˆtrise, Universite´ Laval, Que´bec, Canada, 1998.

248

Biomacromolecules, Vol. 3, No. 2, 2002

(44) Gauthier, S. F.; Vachon, C.; Savoie, L. J. Food Sci. 1986, 51 (4), 960. (45) Sun, J. J. AOAC Int. 1999, 82, 68. (46) Qian, H.; Sheng, M. J. Chromatogr., A 1998, 825, 127. (47) Thompson, J. N.; Hatina, G.; Maxwell, W. B. Assoc. Off. Anal. Chem. 1980, 63, 894. (48) Tyle, P. Specialized Drug DeliVery Systems. Manufacturing and Production Technology; Marcel Dekker, Inc.: New York, 1990. (49) Hongsprabhas, P.; Barbut, S. Int. Dairy J. 1998, 7, 827. (50) Jeyarajah, S.; Allen, J. C. J. Agric. Food Chem. 1994, 42, 80. (51) Hongspabhas, P.; Barbut, S. J. Food Sci. 1997, 62, 382. (52) Mangino, M. E. J. Dairy Sci. 1984, 67, 2711. (53) Byler, D. M.; Susi, H. Biopolymers 1986, 25, 469-87. (54) Casal, H. L.; Ko¨hler, U.; Mantsch, H. H. Biochim. Biophys. Acta 1988, 957, 11-20. (55) Dong, A.; Matsuura, J.; Allison, S. D.; Chrisman, E.; Manning, M.

Beaulieu et al. C.; Carpenter, J. F. Biochemistry 1996, 35, 1450-7. (56) Jackson, M.; Mantsch, H. H. Biochim. Biophys. Acta 1992, 1118, 139-143. (57) Clark, A. H.; Saunderson, D. H. P.; Suggett, A. Int. J. Peptide Protein Res. 1981, 17, 353-364. (58) Bandekar, J. Biochim. Biophys. Acta 1992, 1120, 123-143. (59) Papiz, M. Z.; Sawyer, L.; Eliopoulos, E. E.; North, A. C. T.; Findlay, J. B. C.; Sivaprasadarao, R.; Jones, T. A.; Newcomer, M. E.; Kraulis, P. J. Nature 1986, 324, 383 (60) Brownlow, S.; Cabral, J. H. M.; Cooper, R.; Flower, D. R.; Yewdall, S. J.; Polikarpov, I.; North, A. CT.; Sawyer, L. Structure 1997, 5, 481. (61) Hongspabhas, P.; Barbut, S. Food Res. Int. 1998, 30, 523. (62) Morr, C. V.; Ha, E. Y. W. Food Sci. Nutr. 1993, 33, 431.

BM010082Z