- Email: [email protected]
Stiffening in gels containing whey protein isolate
International Dairy Journal 28 (2013) 62e69
Contents lists available at SciVerse ScienceDirect
International Dairy Journal journal homepage: www.elsevier.com/locate/idairyj
Stiffening in gels containing whey protein isolate Nanik Purwanti a, b,1, Emma van der Veen b, Atze Jan van der Goot a, b, *, Remko Boom b a b
Top Institute Food and Nutrition, Nieuwe Kanaal 9a, 6709 PA Wageningen, The Netherlands Food Process Engineering Group, Agrotechnology and Food Sciences, Wageningen University, Bomenweg 2, 6703 HD Wageningen, The Netherlands
a r t i c l e i n f o
a b s t r a c t
Article history: Received 18 November 2011 Received in revised form 26 August 2012 Accepted 26 August 2012
Gels made only from whey protein isolate (WPI) stiffened over the first few days of storage, after which the textural properties remained nearly constant. However, protein gels containing WPI microparticles, at the same total protein content, stiffened over a longer period than those without microparticles. This stiffening was suggested to be the result of rearrangement of crosslinks in the gel. Addition of particles induces additional effects leading to water distribution between the protein particles and continuous phase. The stiffness change over time was different for gels made from a mixture of locust bean gum and xanthan gum containing microparticles. The stiffness of matrix gel and of gels containing 20% (w/w) microparticles was rather stable over time; microscopy analysis of these gels showed that particle size was constant after 72 h storage. Nevertheless, changes were observed in small deformation; this might be the consequence of slow rearrangements within the protein particles. Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction Protein-enriched products, for example protein bars, are nowadays formulated for health-conscious consumers (Loveday, Hindmarsh, Creamer, & Singh, 2009). Many studies reveal the positive effects of a higher daily intake of protein for better body weight control (Luhovyy, Akhavan, & Anderson, 2007), higher muscle strength (Hayes & Cribb, 2008), and delay in onset of sarcopenia (Campbell & Leidy, 2007). However, high-protein products, which mostly consist of a mixture of protein, sugars and water, undergo sensorial changes over time that are not appreciated by consumers. The main changes relate to the increase of product stiffness with time (Labuza, Zhou, & Davis, 2007). Several studies have discussed the mechanisms behind the stiffening of these products (Li, Szlachetka, Chen, Lin, & Ruan, 2008; Liu, Zhou, Tran, & Labuza, 2009; Loveday et al., 2009; Loveday, Hindmarsh, Creamer, & Singh, 2010; McMahon, Adams, & McManus, 2009; Purwanti, van der Goot, Boom, & Vereijken, 2010). Li et al. (2008) suggested that an increase in water activity during storage indicates that water becomes less bound to the matrix. Consequently, the effect of
* Corresponding author. Tel.: þ31 317 480852. E-mail address: [email protected] (A.J. van der Goot). 1 Present address: Department of Mechanical and Biosystem Engineering, Faculty of Agricultural Engineering and Technology, Bogor Agricultural University, IPB Darmaga Campus, PO Box 220, Bogor 16002, Indonesia. 0958-6946/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.idairyj.2012.08.011
water as plasticiser might be less and thus the product becomes stiffer. McMahon et al. (2009) proposed that separation between the protein and the carbohydrate in a mixture of protein, carbohydrate and water causes stiffening of the bar and subsequently, increases protein aggregation. Loveday et al. (2010) found no evidence of the formation of additional covalent bonds in protein aggregation, but confirmed that water migration from the protein phase to a glucoseeglycerol phase is correlated with the stiffening of protein bars. The questions remain whether stiffening is relevant in highprotein systems without ingredients such as sugars or other polyols, and to what extent stiffening is a property of the protein matrix itself. A previous study suggested that protein aggregation is responsible for structural and textural changes in a concentrated matrix of whey protein isolate (WPI) (Zhou, Liu, & Labuza, 2008a). However, this study only explored the changes in a model product made from 60% (w/w) native WPI in phosphate buffer. This study hypothesises that stiffening in a high-protein product is not only due to the properties of the protein, but also related to the structure in the product and the resulting dynamics within the product. Therefore, the changes of stiffness in a protein model product were followed over time, without and with a dispersed protein phase. The model product comprises a total protein concentration above the critical gelling concentration of native WPI. We compared protein gels made from native WPI only and protein gels containing a dispersed protein phase. The later gels were made from dispersions of WPI microparticles in native WPI solution and in a polysaccharide continuous phase.
N. Purwanti et al. / International Dairy Journal 28 (2013) 62e69
2. Materials and methods 2.1. Materials WPI (Bipro) lot no. JE 034-70-440 (Davisco Food International Inc., Le Sueur, Minnesota, USA), with a reported protein content of 97.9% on a dry basis, was used to prepare the WPI solutions and WPI microparticles. A mixture of locust bean gum (LBG) (Degussa, #56244, Ghent, Belgium) and xanthan gum (XG) derived from Xanthomonas campestris (Fluka, #95465, Zwijndrecht, Netherlands) was used as a polysaccharide phase. Isopropanol EmpluraÒ (Merck, Darmstadt, Germany) was used as one of the carrier fluids for WPI microparticles during particle size measurements. Sodium azide (SigmaeAldrich, Schnelldorf, Germany) was used at 0.02% (w/w) as a preservative. Paraffin oil (Merck, Darmstadt, Germany) was used as a lubricant during compression tests and as a vapour lock during rheological measurements. Milli-Q water (resistivity of 18.2 MU cm at 25 C, total oxidisable carbon <10 ppb, Millipore, Molsheim, France) was used in all experiments, unless mentioned otherwise. 2.2. Production of whey protein isolate microparticles The WPI microparticles were produced following the method described by Purwanti, Moerkens, van der Goot, and Boom (2012), but on a laboratory scale. A 40% (w/w) WPI suspension was centrifuged (Beckman Coulter B.V., Woerden, Netherlands) in 50 mL centrifuge tubes at 78 g for 10 min (20 C) to remove air bubbles. The suspension was then heated at 90 C for 50 min while mixing in a bowl mixer type W50 (Brabender OHG, Duisburg, Germany) that was connected to a Brabender Do-corder E330 (Brabender OHG). The mixer was operated at 0 rpm for 5 min, 5 rpm for 5 min, and 200 rpm for 40 min. The temperature was controlled by circulating water from a water bath. The mixer was cooled down by circulating water having a temperature of 4 C for about 5 min before taking the material out of the mixer. The resulting gel pieces were dried in an oven at 50 C (Termaks, Gemini B.V, Apeldoorn, Netherlands) for 16 h. The remaining processing steps were the same as described in the previous study. 2.3. Characterisation of whey protein isolate microparticles The WPI microparticles were characterised in terms of the degree of denaturation, the particle size, density, protein concentration, and the water holding capacity (WHC). The measurement procedures were described earlier in the previous study (Purwanti et al., 2012). Differential scanning calorimetry (PerkineElmer Diamond DSC, Norwalk, USA) was performed for the 40% WPI suspension and the pieces of gel obtained after heating and mixing. The measurement was done with 3 batches of production; each was at least in triplicate. The microparticle size was measured in triplicate using static light scattering (Mastersizer 2000, Malvern Instruments, Sysmex Nederland B. V., Etten-Leur, Netherlands) using isopropanol and water as the carrier fluids. Dispersed microparticles in water was measured at different dispersion times (0, 1, and 24 h) and at a condition after it was heated at 90 C for 30 min. The protein concentration in the WPI microparticles was measured using the Dumas method (Nitrogen analyser, FlashEA 1112 series, ThermoQuest, Rodano, Italy). Samples for this measurement were taken from several batches. The protein contents were determined in duplicate. The density of the microparticle powder was measured with a manometric gas expansion pycnometer (Ultrapycnometer 1000, Quantachrome, Odelzhausen, Germany) in triplicate. The WHC of microparticles before and after
63
heating was determined with a centrifugation procedure using 845 g (Centrifuge 5424, Eppendorf AG, Hamburg, Germany). Five samples per condition were measured and each condition was repeated three times. 2.4. Production of high-protein model product Three types of model product were prepared: product A, a gel originating from a solution of WPI only; product B, a gel made from a dispersion of WPI microparticles in a WPI solution; product C, a gel made from a dispersion of WPI microparticles in a polysaccharide continuous phase. The concentrations in all products are listed in Table 1. The samples for the compression tests were prepared as follows. Product A was prepared by dissolving WPI powder in water. The solution was then stirred at room temperature for 2 h and kept at 4 C while stirring overnight. The WPI solution was centrifuged in 50-mL centrifuge tubes at 78 g for 10 min (20 C) to remove air bubbles. The pH of the WPI solutions was about 6.8. Cylindrical teflon tubes (diameter 20 mm, length 100 mm) were used to create gels with a uniform cylindrical shape. The WPI solution was poured into the tubes, after which the tubes were tightly sealed. The tubes were clamped in a rotating device set at 30 rpm and immersed in a water bath set at 90 C for 30 min. The tubes were then cooled in an iceewater mixture for 30 min. The resulting gel was carefully removed from the tubes using a cylindrical plunger. The gel was cut with a wire cutter into sections 20mm long. This resulted in cylindrical gel sections having a diameter/height ratio of 1. These gel sections were tightly wrapped with stretch-plastic foil. In this way, a number of samples for each variant of product A were produced that were then stored at 20 C and evaluated at various times up to 8 days. The first measurement was done after storing the samples for 2 h at 20 C. This was defined as day 0. Product B was prepared by dispersing WPI microparticles in a WPI solution. The dispersion was stirred for 30 min before pouring the dispersion into the cylindrical tubes. The remaining procedures were the same as those for product A. However, in this case, rotation of the tubes during heating was important to prevent sedimentation of the WPI microparticles and to retain an even distribution over the product. The homogeneity of the distribution was monitored visually. Product C was made by preparing the LBGeXG mixture first. LBG was dissolved in water and stirred at room temperature for at least 1 h. The solution was heated at 80 C for 30 min in a water bath, and subsequently cooled to room temperature while stirring. Then, XG was dispersed in the LBG solution while stirring at room temperature for 1 h. Finally, the WPI microparticles were dispersed in the LBGeXG mixture and mechanically stirred for 1 h. The final dispersion was centrifuged at 78 g for 10 min (20 C) to remove the air bubbles. The remaining Table 1 Composition of the model products: gels containing whey protein isolate. Variant codes
WPI (%, w/w)
WPI microparticles (%, w/w)
LGB (%, w/w)
XG (%, w/w)
Total protein concentration (%, w/w)
A1 A2 B1 B2 B3 B4 C1 C2 C3a
20 15 17.5 12.5 12.5 7.5 e e e
e e 2.5 7.5 2.5 7.5 20 15 e
e e e e e e 0.6 0.6 0.75
e e e e e e 0.6 0.6 0.75
20 15 20 20 15 15 20 15 0
a
C3 was the matrix for C1.
64
N. Purwanti et al. / International Dairy Journal 28 (2013) 62e69
steps were the same as those for other products. All variants contained 0.02% (w/w) sodium azide as a preservative. For the rheological measurements, the preparation of the WPI solution, the dispersion of WPI microparticles in a WPI solution, and the dispersion of WPI microparticles in an LBGeXG mixture was the same as described above, except that no preservative was added. 2.5. Large deformation tests All products were subjected to uniaxial compression tests (Instron 5564, Norwood, Massachusetts, USA), unless mentioned otherwise. A sample was placed on a lubricated stainless steel base and compressed with a lubricated stainless steel cylindrical probe with a diameter of 50 mm. A load cell of 2000 N was used. Before compression test was applied, each gel was subjected to a pre-load. The pre-load was to make sure that surface of the sample is entirely covered by the compression probe. The pre-load was 0.06e0.5 N depending on the gel’s stiffness. After pre-load, a relaxation time was applied and then, the compression test was started. The sample was compressed using a speed of 1 mm s1 until 80% compressive (Cauchy) strain. The output data were recorded every 10 ms. The number of samples (n) of each product variant being compressed was 7 for A1, B1, B2, and C3, and at least 2 for the other variants. The true stress and true strain were determined from the forceedistance data provided by BluehillÒ material testing software (Instron). The true stress and true strain of the samples at fracture were determined at the first maximum true stress along the true strain, just before the true stress started to drop. The Young’s modulus of the sample was determined by the slope of true stress at true strains of 0.3e1% for products A and B, and at true strains of 4e8% for product C because of data instabilities below 4% in this product. Young’s modulus was used to represent the stiffness of the gel. 2.6. Small deformation tests Textural changes were also investigated using rheological measurements. The samples were prepared in situ in the rheometer (Anton-Paar Physica MCR301, Gentbrugge, Belgium) using a Couette geometry (CC17). The bob and cup were heated to 90 C and kept at this temperature for at least 15 min. A WPI solution, a dispersion of WPI microparticles in the WPI solution, or a dispersion of WPI microparticles in an LBGeXG mixture was placed inside the cup. The bob was then lowered to 3 mm above the zero position normally defined in the rheometer. This position was chosen to accommodate the movement of the bob during measurement. The solutions or dispersions were equilibrated at 90 C for 10 min, and then kept at this temperature for an additional 30 min. Then, the sample was cooled down to 20 C at a rate of 2 C min1 and kept at this temperature for 30 min. A constant strain of 0.01% at a frequency of 1 Hz was applied during heating and cooling. After cooling, the sample was kept in the geometry at 20 C for 72 h, while a constant strain at 0.1% at a frequency of 1 Hz was applied. The time at which this constant strainefrequency was started was defined as time 0 h. The changes in storage modulus (G0 ), loss modulus (G00 ) and damping factor (tan d) were recorded every minute. They were normalised against each of their value at time 0 h and then plotted every hour. The measurement was repeated twice for a chosen variant of each product. Four repetitions were done for variant B3 because the measurement showed larger difficulties in reproducibility because of particle sedimentation.
2.7. Microscopy observation The behaviour of microparticles in water after heating and in gels was visualised with a light microscope (Zeiss, Axiovert 200 MAT, Oberkochen, Germany). This observation was performed to check the swelling of the microparticles in the model products over time. About 20 mL of 2.5% (w/w) microparticle dispersion in water was heated at 90 C for 30 min, cooled in iceewater mixture for 30 min, and then equilibrated at 20 C for 30 min prior to observation. Product B with a total protein concentration of 15% (w/w) was prepared according to the method described in Section 2.4. Product C was prepared with a final protein concentration of 2.5% (w/w) to ease the observation of the microparticles under microscope. The product with 15% (w/w) microparticles was also observed for the comparison, however, particle analysis was not possible due to the high microparticle concentration. No preservative was added, therefore, the observation was done until 72 h only. Samples were placed on a counting glass chamber (0.1 mm depth, 0.0025 m2, Optik Labor, Friedrichshafen, Germany) and the pictures were taken at 100 magnification. Microparticles were analysed using Image-J (free software) to measure their area. From the area, the diameter of microparticles was calculated assuming that the microparticles were spherical. Around 150e300 microparticles were analysed in each product at time 0 h and 72 h. The volume weighted mean diameter (d4,3) was calculated. 3. Results 3.1. Properties of whey protein isolate microparticles DSC measurements indicated that the protein in dry WPI microparticles was denatured completely. The protein concentration of dry WPI microparticles was 92.7 1.2% (w/w), which was similar to the protein concentration of native WPI determined using the same method. WPI microparticles had a density of 1.3 g cm3, similar as reported in the previous study (Purwanti et al., 2012). The average diameter (d4,3) of dry microparticles (unswollen condition) was around 70 mm. Dispersing the microparticles in to water resulted in a particle size increase. The total increase depended on the contact time with water. Microparticles had a diameter of around 97 mm when freshly dispersed in water, about 105 mm after 1 h in water, and around 114 mm after 24 h. Their diameter increased further up to 125 mm by heating the dispersion at 90 C for 30 min. The WHC measurement confirmed that the microparticles swelled upon heating. The WHC of WPI microparticles without heating was 3.55 g water (g powder)1, which gave 3.83 g water (g protein)1 when taking the measured protein concentration of the microparticles into account. The WHC increased to 5.92 g water g1 powder (6.38 g water g1 protein) after heating the dispersion at 90 C for 30 min. The fact that heating leads to additional swelling suggests that the swelling at lower temperature was not completed yet and might continue upon time, as heat generally only accelerates the effects. 3.2. Large deformation properties of the model product The changes in Young’s moduli of product A over time are depicted in Fig. 1. Both variants became stiffer during a limited period of about 3 days, with a larger stiffening rate in A1. Fig. 2a shows that the fracture stress of A1 increased during the first 3 days and then stabilised or even slightly decreased. This change was in line with the change in Young’s modulus. The fracture stress of A2 remained constant for 8 days of storage. This change was not in line with the change in Young’s modulus. This could be due to a subtle
N. Purwanti et al. / International Dairy Journal 28 (2013) 62e69
Fig. 1. The changes in Young’s moduli for product A over time: -, A1 (20%, w/w, whey protein isolate); ,, A2 (15%, w/w, whey protein isolate). The lines are only plotted to guide the eyes. The vertical bars are the standard deviations.
change in product stiffness, which did not significantly affect the fracture stress. The fracture strains of both variants were relatively constant with the storage time. The results from compression tests show that protein gels made from WPI only became stiffer in the first few days, after which their stiffness became constant. The Young’s moduli of product B are shown in Fig. 3. When native WPI was partly replaced by 2.5% (w/w) WPI microparticles (B1), the value of Young’s modulus was rather similar to the one without microparticles (A1). Lower values for Young’s moduli were obtained when the amount of microparticles was increased while keeping the total protein concentration constant. Product variants that contained 7.5% (w/w) microparticles resulted in a lower Young’s modulus than the variants containing 2.5% (w/w). The lower modulus was also obtained when the total protein concentration was decreased. The total change in Young’s modulus of B1 was similar to that of A1 over time. However, the change occurred over a longer period than 3 days for B2 and B3. Variant B4, which contained 7.5% (w/w) WPI microparticles at 15% w/w overall protein concentration, did not form a gel. Nevertheless, its changes in consistency over time could be clearly observed visually. This variant became more viscous during a 6-day storage period, which may be interpreted as being analogous to stiffening. Overall, it seems that product B showed a different stiffening behaviour compared with A, even though the effects were small. The changes in the fracture stresses and strains over time for product B are depicted in Fig. 4. The changes of the fracture stresses for B1 and B2 were in accordance with the changes of their Young’s moduli. The fracture strains of B1 and B2 slightly decreased over
65
time. This indicates that a stiffer gel behaves more as a brittle material. No data are shown for B3 because most samples of this variant did not fracture. The compression tests of product C did not result in fractured samples in all variants. Therefore, only the result on Young’s modulus can be extracted from the large deformation test. Fig. 5 shows that Young’s moduli of C1 and C2 were larger than the values for the matrix (C3). This increase is typical for a filled gel with a dispersed phase that has a higher modulus than the continuous phase or has a good interaction with the continuous phases (Sala, van Aken, Stuart, & van de Velde, 2007; van Vliet, 1988; Yost & Kinsella, 1992). The change of Young’s modulus of C3 over time was limited. Variation in the first 3 days was hardly significant. The modulus of C2 increased in the first 3 days and then stabilised. C1 showed similar behaviour as C3 over time. If the Young’s modulus of C1 and of C2 was normalised to its value at time 0, it turned out that Young’s modulus of C1 changed in similar way as of C3 over time (Fig. 5, inset). For C2, its Young’s modulus indeed increased in the first 3 days and then, became stable. 3.3. Small deformation properties of model products The nature of the rheological measurement used in this study was good for liquid or semi-solid sample and was rather tricky for materials with sedimenting particles. Therefore, we tested high WPI concentration (A1) and product B (B1 and B2) in replicates at least to check the reproducibility of the measurement solid sample (A1) and samples that contained microparticles. The measurement for C3 was also repeated twice to confirm the reproducibility for semi-solid sample. Fig. 6 shows the result of the normalised G0 , G00 and tan d-values from product A, B and C over time. The deviation of the variants that measured in repetition is also presented. Fig. 6a shows the changes in normalised G0 and G00 -values of product A at 20 C during a 72-h period. The normalised G0 -value of A1 increased until around 60 h, after which it seemed to decrease. For A2, the normalised G0 -value slightly increased throughout the whole time range. The normalised G00 -value for A1 increased in the first 30 h and then it decreased. For A2, this value tended to decrease over time. The changes in G0 or/and G00 -value resulted in a change in damping factor (tan d) over time. The normalised tan d-values of A1 slightly increased in the first 24 h and then it decreased. The standard deviation shows that the increase in the first 24 h was not significant. The value for A2 slightly decreased over the storage time. The change in the normalised G0 -value over time for B1over time was similar to that of for A1 (Fig. 6b). The changes in the normalised
Fig. 2. The changes in (panel a) fracture stresses and (panel b) fracture strains of product A over time: - and C, A1 (20%, w/w, whey protein isolate); , and B, A2 (15%, w/w, whey protein isolate). The lines guide the eyes to the fracture trends over time. The vertical bars are the standard deviations.
66
N. Purwanti et al. / International Dairy Journal 28 (2013) 62e69
3.4. Microscopy observation
Fig. 3. The changes in Young’s moduli of product B over time containing a total protein concentration of (B1 and B2) 20% (w/w) and (B3) 15% (w/w): -, B1 (17.5% WPIe2.5% microparticles); ,, B2 (12.5% WPIe7.5% microparticles); , B3 (12.5% WPIe2.5% microparticles). The lines guide the eyes. The vertical bars are the standard deviations. Inset is the B3 with an enhanced y-axis.
G00 and tan d-values were rather different from those of for A1. The value for B1 tended to increase, while it decreased for A1 during the first 30 h, suggesting the occurrence of another effect influence the dynamic properties. Rather similar changes between A1 and B1 might be attributed to the relatively small amount of WPI microparticles applied (2.5%, w/w) in a concentrated WPI matrix (17.5%, w/w). The normalised G0 -values of B2 and B4 were rather stable over time, while, their normalised G00 -values tended to decrease. Therefore, their normalised damping factor decreased slightly with time. Contrary to the large deformation measurements, the changes in G0 and G00 seem to occur only in the first days of measurements. The reason for this difference is not clear yet, but might be related to the fact that here the sample was deformed continuously. For the large deformation tests, separate samples were used that were not deformed upon storage. Fig. 6c shows that the addition of microparticles to the matrix resulted in lower normalised G0 -values than the matrix itself. The actual G0 -values of C1 and C2 were much higher than the matrix. The normalised G0 -values for all variants of product C increased over time with the matrix showed a larger increase in the normalised G0 -values than C1 and C2 over time. The normalised G00 -values for all variants were the same and they tended to be constant over time. Therefore, their normalised tan d tended to decrease over time. The results from small deformation indicate that all the product variants underwent changes during the storage time, unlike the results from large deformation.
Fig. 7 shows the swelling of microparticles when being dispersed in water and heated. After 72 h storage, some microparticles seem to fall apart, which gave a smaller particle size. Swelling also occurred when the microparticles were dispersed in solution containing either WPI or polysaccharide and heated. Fig. 7 shows the changes of microparticles inside the protein gels and polysaccharide gel after 72 h storage. The ratios of particle size (d4,3) in the gels after 72 h storage and the particle size in the gels after heating (0 h) were 1.3 for B3 and 1.4 for B4. This indicates that microparticles continued to swell over time when they were in gels. The ratio of particle size in the polysaccharide gel after 72 h storage was 1.0, indicating no significant swelling. This is an interesting finding, which is in line with the behaviour of this product over time measured with large deformation. The microparticles in polysaccharide gel at 15% (w/w) were packed in the matrix, therefore, particle analysis could not be performed for this sample. A possible explanation for the lack of swelling could be the fact that the XG is charged naturally, thereby lowering the water activity in that phase. This reduction in water activity will reduce the driving force for water migration and hence swelling of the particles. 4. Discussion Changes in the textural properties of high-protein gels were studied with Young’s modulus (stiffness), fracture properties, normalised G0 , G00 and tan d. Overall, it was concluded that the extent of stiffening of all products was rather limited. The changes of protein gels made from only native WPI took place in the first three days mainly. Previous studies suggested a mechanism of stiffening for a concentrated WPI matrix based on protein aggregation through disulphide bond formation (Zhou & Labuza, 2007; Zhou et al., 2008a; Zhou, Liu, & Labuza, 2008b). These changes in the gel properties caused by additional disulphide bonds could account for the changes in the first few days of storage for WPI gels. The changes in the products containing WPI microparticles were also rather limited. However, it seems that the stiffening occurred throughout the whole storage period tested, meaning that the stiffening continued for a longer period than the WPI-gels. This suggests that an additional mechanism might play a role in those structured products. The microscopy analysis showed changes in particle sizes upon heating and storage. This indicates that water migration between the protein matrix and the protein particles takes place, which effect could be responsible for stiffening over a prolonged period. Water migration occurs as long as there is a difference in water activity between different phases. The water redistribution itself might take place over a time scale of a few days.
Fig. 4. The changes in (panel a) fracture stresses and (panel b) fracture strains of product B over time: - and C, B1 (17.5% WPIe2.5% microparticles); , and B, B2 (12.5% WPIe7.5% microparticles). The lines guide the eyes. The vertical bars are the standard deviations.
N. Purwanti et al. / International Dairy Journal 28 (2013) 62e69
67
Fig. 5. The changes in Young’s moduli of product C over time, at a total protein concentration of: -, C1 (20%, w/w, microparticles); ,, C2 (15%, w/w, microparticles); and , C3 (0%, w/w, microparticles: the polysaccharide matrix). The lines guide the eyes. The vertical bars are the standard deviation. Inset is the normalised Young’s moduli over time.
A low estimate of the water diffusivity, i.e., 1012 m2 s1 results in a typical diffusion-length of 800 mm over 2 days. This is around 11 times the average size of dry microparticles. Therefore, diffusion might not be the rate limiting step. Another effect, such as the internal relaxation of the protein inside the microparticles, could be the limiting step for water redistribution. The swelling of microparticles is limited by the intramolecular disulphide bonds, which are able to slowly rearrange upon time. Once a disulphide bridge opens, the protein matrix will rearrange and it will be able to absorb additional water. This row of events (water absorption, disulphide rearrangements followed by additional swelling) might be an explanation for the changes of the product over a longer period. The importance of the number of disulphide bridges on swelling was shown in similar WPI particles produced in a different manner that allowed controlling the number of bridges present in the system (van Riemsdijk, van der Goot, & Hamer, 2011). No changes were observed in the product containing WPI microparticles in polysaccharide matrix in large deformation, except for C2, while limited changes were observed in small deformation. In large deformation, the change in C2 with 15% (w/w) microparticles (Fig. 5) was similar to that in product A and B1. It was limited in the first few days. By increasing particle concentration (C1), hardly any changes were observed anymore. The change in Young’s modulus was similar to the matrix (C3). The microscopy analysis revealed that the microparticle diameter in polysaccharide matrix is relatively constant over time after the initial swelling. Storage of the dispersion at ambient conditions resulted in relatively stable particle. This behaviour might be suggested by thermodynamic equilibrium between the microparticles and the polysaccharide matrix after heating. Therefore, hardly any changes can occur anymore upon storage at ambient temperature. This was observed especially at 20% (w/w) of microparticles. Nevertheless, all variants of product C showed slight changes in small scale deformation, indicating some kind of change in the
Fig. 6. Normalised G0 , G00 and tan d-values over storage time. Panel a, products A1 and A2: - and :, normalised G0 for A1 and A2, respectively; , and 6, normalised G00 for A1 and A2, respectively; inset is normalised tan d for A1 (-) and A2 (6). Panel b, products B1, B2 and B4: -, : and A, normalised G0 for B1, B2 and B4, respectively; ,, 6 and >, normalised G00 for B1, B2 and B4, respectively; inset is normalised tan d for B1 (-), B2 (6) and B4 (A). Panel c, products C1, C2 and C3: -, : and A, ratio G0 for C1, C2 and C3, respectively; ,, 6 and >, ratio G00 for C1, C2 and C3, respectively; inset is normalised tan d for C1 (-), C2 (6) and C3 (A). Product composition is as defined in Table 1. The vertical bars are the standard deviations.
product. This could be the result of slow rearrangements within the microparticles in which the changes only implied in small deformation. It has already been reported that an increase in hardness based on fracture stress can continue for 50 days for protein bars made
68
N. Purwanti et al. / International Dairy Journal 28 (2013) 62e69
Fig. 7. Microparticles in water after heating and in gels, observed at time 0 (AeD) and 72 h (EeH). Panels a and e: heated microparticle dispersion (2.5%, w/w) in water. Panels B and F: 12.5% (w/w) whey protein isolate e 2.5% (w/w) microparticles: Panels C and G: 7.5% (w/w) whey protein isolate e 7.5% (w/w) microparticles. Panels D and G: 2.5% (w/w) microparticles in locust bean gum e xanthan gum.
with calcium caseinate (CaCas) or milk protein concentrate (MPC) (Loveday et al., 2009, 2010). These bars normally contain sugar or polyol as additional components and do not undergo heat treatment during preparation (Li et al., 2008; Liu et al., 2009; Loveday et al., 2009, 2010; McMahon et al., 2009). The absence of a heating step in the preparation could explain the longer period during which stiffening occurs as potential effects of water migration are not diminished due to a thermal treatment. Based on this and other studies, we conclude that stiffening of high-protein product is not caused by the properties of the protein only. A process consisting of various events with mutual interactions such as difference in water activity between phases, rearrangement in the dispersed phase and the matrix phase, or crystallisation in a sugar-rich phase might cause the stiffening over a prolonged period. When WPI microparticles are used together with WPI matrix, stiffening effect that occurs over a longer time might be related to the slow protein rearrangement in the microparticles, which are caused by and induces water migration. Stopping or shortening the stiffening process might be achieved by applying an extensive heating step to accelerate water redistribution to reach thermodynamic equilibrium inside between the various phases inside the product. Another method could be to stop moisture migration happening. This can be done, for example, by modifying the production of microparticles in such a way that the microparticles cannot absorb water. When polysaccharide is used as the matrix, further study is interesting to understand more about the dynamic between WPI microparticles and the matrix. 5. Conclusions The stiffening over time was investigated in protein gels made from WPI only and in gels in which part of the protein was present as a dispersed phase (microparticles). Overall, it can be
concluded that the extent of hardening is limited in those model products. The use of a dispersed protein phase did not reduce the extent of stiffening. Water redistribution and rearrangements of crosslinks within the protein microparticles were used as possible explanations for the changes of the product over time. Better control over the stiffening process can be achieved by accelerating the re-ordering and transport during processing, for example, by heating. This could stabilise the products over a shorter period of time. When the continuous phase is different from the protein phase, changes in the product might not be avoided as long as there is not thermodynamic equilibrium between different phases. References Campbell, W. W., & Leidy, H. J. (2007). Dietary protein and resistance training effects on muscle and body composition in older persons. Journal of the American College of Nutrition, 26, 696Se703S. Hayes, A., & Cribb, P. J. (2008). Effect of whey protein isolate on strength, body composition and muscle hypertrophy during resistance training. Current Opinion in Clinical Nutrition and Metabolic Care, 11, 40e44. Labuza, T. P., Zhou, P., & Davis, L. (2007). Whey to go: application of whey proteins in nutrition bars. World of Food Ingredients, October/November, 108e112. Li, Y., Szlachetka, K., Chen, P., Lin, X., & Ruan, R. (2008). Ingredient characterization and hardening of high-protein food bars: an NMR state diagram approach. Cereal Chemistry, 85, 780e786. Liu, X., Zhou, P., Tran, A., & Labuza, T. P. (2009). Effects of polyols on the stability of whey proteins in intermediate-moisture food model systems. Journal Agricultural Food and Chemistry, 57, 2339e2345. Loveday, S. M., Hindmarsh, J. P., Creamer, L. K., & Singh, H. (2009). Physicochemical changes in a model protein bar during storage. Food Research International, 42, 798e806. Loveday, S. M., Hindmarsh, J. P., Creamer, L. K., & Singh, H. (2010). Physicochemical changes in intermediate-moisture protein bars made with whey protein or calcium caseinate. Food Research International, 43, 1321e1328. Luhovyy, B. L., Akhavan, T., & Anderson, G. H. (2007). Whey proteins in the regulation of food intake and satiety. Journal of the American College of Nutrition, 26, 704Se712S.
N. Purwanti et al. / International Dairy Journal 28 (2013) 62e69 McMahon, D. J., Adams, S. L., & McManus, W. R. (2009). Hardening of high-protein nutrition bars and sugar/polyol-protein phase separation. Journal of Food Science, 74, E312eE321. Purwanti, N., Moerkens, A., van der Goot, A. J., & Boom, R. (2012). Reducing the stiffness of concentrated Whey Protein Isolate (WPI) gels by using WPI microparticles. Food Hydrocolloids, 26, 240e248. Purwanti, N., van der Goot, A. J., Boom, R., & Vereijken, J. (2010). New directions towards structure formation and stability of protein-rich foods from globular proteins. Trends in Food Science and Technology, 21, 85e94. van Riemsdijk, L. E., van der Goot, A. J., & Hamer, R. (2011). On the stability of whey protein particles during production of gluten-free bread. Food Hydrocolloids, 25, 1744e1750. Sala, G., van Aken, G. A., Stuart, M. C., & van de Velde, F. (2007). Effect of dropletmatrix interactions on large deformation properties of emulsion-filled gels. Journal of Texture Studies, 38, 511e535.
69
van Vliet, T. (1988). Rheological properties of filled gels: influence of filler matrix interaction. Colloid Polymer Science, 266, 518e524. Yost, R. A., & Kinsella, J. E. (1992). Microstructure of whey protein isolate gels containing emulsified butterfat droplets. Journal of Food Science, 57, 892e897. Zhou, P., & Labuza, T. (2007). Effect of water content on glass transition and protein aggregation of whey protein powders during short-term storage. Food Biophysics, 2, 108e116. Zhou, P., Liu, X., & Labuza, T. P. (2008a). Effects of moisture-induced whey protein aggregation on protein conformation, the state of water molecules, and the microstructure and texture of high-protein-containing matrix. Journal of Agricultural and Food Chemistry, 56, 4534e4540. Zhou, P., Liu, X., & Labuza, T. P. (2008b). Moisture-induced aggregation of whey proteins in a protein/buffer model system. Journal of Agricultural and Food Chemistry, 56, 2048e2054.
Contents lists available at SciVerse ScienceDirect
International Dairy Journal journal homepage: www.elsevier.com/locate/idairyj
Stiffening in gels containing whey protein isolate Nanik Purwanti a, b,1, Emma van der Veen b, Atze Jan van der Goot a, b, *, Remko Boom b a b
Top Institute Food and Nutrition, Nieuwe Kanaal 9a, 6709 PA Wageningen, The Netherlands Food Process Engineering Group, Agrotechnology and Food Sciences, Wageningen University, Bomenweg 2, 6703 HD Wageningen, The Netherlands
a r t i c l e i n f o
a b s t r a c t
Article history: Received 18 November 2011 Received in revised form 26 August 2012 Accepted 26 August 2012
Gels made only from whey protein isolate (WPI) stiffened over the first few days of storage, after which the textural properties remained nearly constant. However, protein gels containing WPI microparticles, at the same total protein content, stiffened over a longer period than those without microparticles. This stiffening was suggested to be the result of rearrangement of crosslinks in the gel. Addition of particles induces additional effects leading to water distribution between the protein particles and continuous phase. The stiffness change over time was different for gels made from a mixture of locust bean gum and xanthan gum containing microparticles. The stiffness of matrix gel and of gels containing 20% (w/w) microparticles was rather stable over time; microscopy analysis of these gels showed that particle size was constant after 72 h storage. Nevertheless, changes were observed in small deformation; this might be the consequence of slow rearrangements within the protein particles. Ó 2012 Elsevier Ltd. All rights reserved.
1. Introduction Protein-enriched products, for example protein bars, are nowadays formulated for health-conscious consumers (Loveday, Hindmarsh, Creamer, & Singh, 2009). Many studies reveal the positive effects of a higher daily intake of protein for better body weight control (Luhovyy, Akhavan, & Anderson, 2007), higher muscle strength (Hayes & Cribb, 2008), and delay in onset of sarcopenia (Campbell & Leidy, 2007). However, high-protein products, which mostly consist of a mixture of protein, sugars and water, undergo sensorial changes over time that are not appreciated by consumers. The main changes relate to the increase of product stiffness with time (Labuza, Zhou, & Davis, 2007). Several studies have discussed the mechanisms behind the stiffening of these products (Li, Szlachetka, Chen, Lin, & Ruan, 2008; Liu, Zhou, Tran, & Labuza, 2009; Loveday et al., 2009; Loveday, Hindmarsh, Creamer, & Singh, 2010; McMahon, Adams, & McManus, 2009; Purwanti, van der Goot, Boom, & Vereijken, 2010). Li et al. (2008) suggested that an increase in water activity during storage indicates that water becomes less bound to the matrix. Consequently, the effect of
* Corresponding author. Tel.: þ31 317 480852. E-mail address: [email protected] (A.J. van der Goot). 1 Present address: Department of Mechanical and Biosystem Engineering, Faculty of Agricultural Engineering and Technology, Bogor Agricultural University, IPB Darmaga Campus, PO Box 220, Bogor 16002, Indonesia. 0958-6946/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.idairyj.2012.08.011
water as plasticiser might be less and thus the product becomes stiffer. McMahon et al. (2009) proposed that separation between the protein and the carbohydrate in a mixture of protein, carbohydrate and water causes stiffening of the bar and subsequently, increases protein aggregation. Loveday et al. (2010) found no evidence of the formation of additional covalent bonds in protein aggregation, but confirmed that water migration from the protein phase to a glucoseeglycerol phase is correlated with the stiffening of protein bars. The questions remain whether stiffening is relevant in highprotein systems without ingredients such as sugars or other polyols, and to what extent stiffening is a property of the protein matrix itself. A previous study suggested that protein aggregation is responsible for structural and textural changes in a concentrated matrix of whey protein isolate (WPI) (Zhou, Liu, & Labuza, 2008a). However, this study only explored the changes in a model product made from 60% (w/w) native WPI in phosphate buffer. This study hypothesises that stiffening in a high-protein product is not only due to the properties of the protein, but also related to the structure in the product and the resulting dynamics within the product. Therefore, the changes of stiffness in a protein model product were followed over time, without and with a dispersed protein phase. The model product comprises a total protein concentration above the critical gelling concentration of native WPI. We compared protein gels made from native WPI only and protein gels containing a dispersed protein phase. The later gels were made from dispersions of WPI microparticles in native WPI solution and in a polysaccharide continuous phase.
N. Purwanti et al. / International Dairy Journal 28 (2013) 62e69
2. Materials and methods 2.1. Materials WPI (Bipro) lot no. JE 034-70-440 (Davisco Food International Inc., Le Sueur, Minnesota, USA), with a reported protein content of 97.9% on a dry basis, was used to prepare the WPI solutions and WPI microparticles. A mixture of locust bean gum (LBG) (Degussa, #56244, Ghent, Belgium) and xanthan gum (XG) derived from Xanthomonas campestris (Fluka, #95465, Zwijndrecht, Netherlands) was used as a polysaccharide phase. Isopropanol EmpluraÒ (Merck, Darmstadt, Germany) was used as one of the carrier fluids for WPI microparticles during particle size measurements. Sodium azide (SigmaeAldrich, Schnelldorf, Germany) was used at 0.02% (w/w) as a preservative. Paraffin oil (Merck, Darmstadt, Germany) was used as a lubricant during compression tests and as a vapour lock during rheological measurements. Milli-Q water (resistivity of 18.2 MU cm at 25 C, total oxidisable carbon <10 ppb, Millipore, Molsheim, France) was used in all experiments, unless mentioned otherwise. 2.2. Production of whey protein isolate microparticles The WPI microparticles were produced following the method described by Purwanti, Moerkens, van der Goot, and Boom (2012), but on a laboratory scale. A 40% (w/w) WPI suspension was centrifuged (Beckman Coulter B.V., Woerden, Netherlands) in 50 mL centrifuge tubes at 78 g for 10 min (20 C) to remove air bubbles. The suspension was then heated at 90 C for 50 min while mixing in a bowl mixer type W50 (Brabender OHG, Duisburg, Germany) that was connected to a Brabender Do-corder E330 (Brabender OHG). The mixer was operated at 0 rpm for 5 min, 5 rpm for 5 min, and 200 rpm for 40 min. The temperature was controlled by circulating water from a water bath. The mixer was cooled down by circulating water having a temperature of 4 C for about 5 min before taking the material out of the mixer. The resulting gel pieces were dried in an oven at 50 C (Termaks, Gemini B.V, Apeldoorn, Netherlands) for 16 h. The remaining processing steps were the same as described in the previous study. 2.3. Characterisation of whey protein isolate microparticles The WPI microparticles were characterised in terms of the degree of denaturation, the particle size, density, protein concentration, and the water holding capacity (WHC). The measurement procedures were described earlier in the previous study (Purwanti et al., 2012). Differential scanning calorimetry (PerkineElmer Diamond DSC, Norwalk, USA) was performed for the 40% WPI suspension and the pieces of gel obtained after heating and mixing. The measurement was done with 3 batches of production; each was at least in triplicate. The microparticle size was measured in triplicate using static light scattering (Mastersizer 2000, Malvern Instruments, Sysmex Nederland B. V., Etten-Leur, Netherlands) using isopropanol and water as the carrier fluids. Dispersed microparticles in water was measured at different dispersion times (0, 1, and 24 h) and at a condition after it was heated at 90 C for 30 min. The protein concentration in the WPI microparticles was measured using the Dumas method (Nitrogen analyser, FlashEA 1112 series, ThermoQuest, Rodano, Italy). Samples for this measurement were taken from several batches. The protein contents were determined in duplicate. The density of the microparticle powder was measured with a manometric gas expansion pycnometer (Ultrapycnometer 1000, Quantachrome, Odelzhausen, Germany) in triplicate. The WHC of microparticles before and after
63
heating was determined with a centrifugation procedure using 845 g (Centrifuge 5424, Eppendorf AG, Hamburg, Germany). Five samples per condition were measured and each condition was repeated three times. 2.4. Production of high-protein model product Three types of model product were prepared: product A, a gel originating from a solution of WPI only; product B, a gel made from a dispersion of WPI microparticles in a WPI solution; product C, a gel made from a dispersion of WPI microparticles in a polysaccharide continuous phase. The concentrations in all products are listed in Table 1. The samples for the compression tests were prepared as follows. Product A was prepared by dissolving WPI powder in water. The solution was then stirred at room temperature for 2 h and kept at 4 C while stirring overnight. The WPI solution was centrifuged in 50-mL centrifuge tubes at 78 g for 10 min (20 C) to remove air bubbles. The pH of the WPI solutions was about 6.8. Cylindrical teflon tubes (diameter 20 mm, length 100 mm) were used to create gels with a uniform cylindrical shape. The WPI solution was poured into the tubes, after which the tubes were tightly sealed. The tubes were clamped in a rotating device set at 30 rpm and immersed in a water bath set at 90 C for 30 min. The tubes were then cooled in an iceewater mixture for 30 min. The resulting gel was carefully removed from the tubes using a cylindrical plunger. The gel was cut with a wire cutter into sections 20mm long. This resulted in cylindrical gel sections having a diameter/height ratio of 1. These gel sections were tightly wrapped with stretch-plastic foil. In this way, a number of samples for each variant of product A were produced that were then stored at 20 C and evaluated at various times up to 8 days. The first measurement was done after storing the samples for 2 h at 20 C. This was defined as day 0. Product B was prepared by dispersing WPI microparticles in a WPI solution. The dispersion was stirred for 30 min before pouring the dispersion into the cylindrical tubes. The remaining procedures were the same as those for product A. However, in this case, rotation of the tubes during heating was important to prevent sedimentation of the WPI microparticles and to retain an even distribution over the product. The homogeneity of the distribution was monitored visually. Product C was made by preparing the LBGeXG mixture first. LBG was dissolved in water and stirred at room temperature for at least 1 h. The solution was heated at 80 C for 30 min in a water bath, and subsequently cooled to room temperature while stirring. Then, XG was dispersed in the LBG solution while stirring at room temperature for 1 h. Finally, the WPI microparticles were dispersed in the LBGeXG mixture and mechanically stirred for 1 h. The final dispersion was centrifuged at 78 g for 10 min (20 C) to remove the air bubbles. The remaining Table 1 Composition of the model products: gels containing whey protein isolate. Variant codes
WPI (%, w/w)
WPI microparticles (%, w/w)
LGB (%, w/w)
XG (%, w/w)
Total protein concentration (%, w/w)
A1 A2 B1 B2 B3 B4 C1 C2 C3a
20 15 17.5 12.5 12.5 7.5 e e e
e e 2.5 7.5 2.5 7.5 20 15 e
e e e e e e 0.6 0.6 0.75
e e e e e e 0.6 0.6 0.75
20 15 20 20 15 15 20 15 0
a
C3 was the matrix for C1.
64
N. Purwanti et al. / International Dairy Journal 28 (2013) 62e69
steps were the same as those for other products. All variants contained 0.02% (w/w) sodium azide as a preservative. For the rheological measurements, the preparation of the WPI solution, the dispersion of WPI microparticles in a WPI solution, and the dispersion of WPI microparticles in an LBGeXG mixture was the same as described above, except that no preservative was added. 2.5. Large deformation tests All products were subjected to uniaxial compression tests (Instron 5564, Norwood, Massachusetts, USA), unless mentioned otherwise. A sample was placed on a lubricated stainless steel base and compressed with a lubricated stainless steel cylindrical probe with a diameter of 50 mm. A load cell of 2000 N was used. Before compression test was applied, each gel was subjected to a pre-load. The pre-load was to make sure that surface of the sample is entirely covered by the compression probe. The pre-load was 0.06e0.5 N depending on the gel’s stiffness. After pre-load, a relaxation time was applied and then, the compression test was started. The sample was compressed using a speed of 1 mm s1 until 80% compressive (Cauchy) strain. The output data were recorded every 10 ms. The number of samples (n) of each product variant being compressed was 7 for A1, B1, B2, and C3, and at least 2 for the other variants. The true stress and true strain were determined from the forceedistance data provided by BluehillÒ material testing software (Instron). The true stress and true strain of the samples at fracture were determined at the first maximum true stress along the true strain, just before the true stress started to drop. The Young’s modulus of the sample was determined by the slope of true stress at true strains of 0.3e1% for products A and B, and at true strains of 4e8% for product C because of data instabilities below 4% in this product. Young’s modulus was used to represent the stiffness of the gel. 2.6. Small deformation tests Textural changes were also investigated using rheological measurements. The samples were prepared in situ in the rheometer (Anton-Paar Physica MCR301, Gentbrugge, Belgium) using a Couette geometry (CC17). The bob and cup were heated to 90 C and kept at this temperature for at least 15 min. A WPI solution, a dispersion of WPI microparticles in the WPI solution, or a dispersion of WPI microparticles in an LBGeXG mixture was placed inside the cup. The bob was then lowered to 3 mm above the zero position normally defined in the rheometer. This position was chosen to accommodate the movement of the bob during measurement. The solutions or dispersions were equilibrated at 90 C for 10 min, and then kept at this temperature for an additional 30 min. Then, the sample was cooled down to 20 C at a rate of 2 C min1 and kept at this temperature for 30 min. A constant strain of 0.01% at a frequency of 1 Hz was applied during heating and cooling. After cooling, the sample was kept in the geometry at 20 C for 72 h, while a constant strain at 0.1% at a frequency of 1 Hz was applied. The time at which this constant strainefrequency was started was defined as time 0 h. The changes in storage modulus (G0 ), loss modulus (G00 ) and damping factor (tan d) were recorded every minute. They were normalised against each of their value at time 0 h and then plotted every hour. The measurement was repeated twice for a chosen variant of each product. Four repetitions were done for variant B3 because the measurement showed larger difficulties in reproducibility because of particle sedimentation.
2.7. Microscopy observation The behaviour of microparticles in water after heating and in gels was visualised with a light microscope (Zeiss, Axiovert 200 MAT, Oberkochen, Germany). This observation was performed to check the swelling of the microparticles in the model products over time. About 20 mL of 2.5% (w/w) microparticle dispersion in water was heated at 90 C for 30 min, cooled in iceewater mixture for 30 min, and then equilibrated at 20 C for 30 min prior to observation. Product B with a total protein concentration of 15% (w/w) was prepared according to the method described in Section 2.4. Product C was prepared with a final protein concentration of 2.5% (w/w) to ease the observation of the microparticles under microscope. The product with 15% (w/w) microparticles was also observed for the comparison, however, particle analysis was not possible due to the high microparticle concentration. No preservative was added, therefore, the observation was done until 72 h only. Samples were placed on a counting glass chamber (0.1 mm depth, 0.0025 m2, Optik Labor, Friedrichshafen, Germany) and the pictures were taken at 100 magnification. Microparticles were analysed using Image-J (free software) to measure their area. From the area, the diameter of microparticles was calculated assuming that the microparticles were spherical. Around 150e300 microparticles were analysed in each product at time 0 h and 72 h. The volume weighted mean diameter (d4,3) was calculated. 3. Results 3.1. Properties of whey protein isolate microparticles DSC measurements indicated that the protein in dry WPI microparticles was denatured completely. The protein concentration of dry WPI microparticles was 92.7 1.2% (w/w), which was similar to the protein concentration of native WPI determined using the same method. WPI microparticles had a density of 1.3 g cm3, similar as reported in the previous study (Purwanti et al., 2012). The average diameter (d4,3) of dry microparticles (unswollen condition) was around 70 mm. Dispersing the microparticles in to water resulted in a particle size increase. The total increase depended on the contact time with water. Microparticles had a diameter of around 97 mm when freshly dispersed in water, about 105 mm after 1 h in water, and around 114 mm after 24 h. Their diameter increased further up to 125 mm by heating the dispersion at 90 C for 30 min. The WHC measurement confirmed that the microparticles swelled upon heating. The WHC of WPI microparticles without heating was 3.55 g water (g powder)1, which gave 3.83 g water (g protein)1 when taking the measured protein concentration of the microparticles into account. The WHC increased to 5.92 g water g1 powder (6.38 g water g1 protein) after heating the dispersion at 90 C for 30 min. The fact that heating leads to additional swelling suggests that the swelling at lower temperature was not completed yet and might continue upon time, as heat generally only accelerates the effects. 3.2. Large deformation properties of the model product The changes in Young’s moduli of product A over time are depicted in Fig. 1. Both variants became stiffer during a limited period of about 3 days, with a larger stiffening rate in A1. Fig. 2a shows that the fracture stress of A1 increased during the first 3 days and then stabilised or even slightly decreased. This change was in line with the change in Young’s modulus. The fracture stress of A2 remained constant for 8 days of storage. This change was not in line with the change in Young’s modulus. This could be due to a subtle
N. Purwanti et al. / International Dairy Journal 28 (2013) 62e69
Fig. 1. The changes in Young’s moduli for product A over time: -, A1 (20%, w/w, whey protein isolate); ,, A2 (15%, w/w, whey protein isolate). The lines are only plotted to guide the eyes. The vertical bars are the standard deviations.
change in product stiffness, which did not significantly affect the fracture stress. The fracture strains of both variants were relatively constant with the storage time. The results from compression tests show that protein gels made from WPI only became stiffer in the first few days, after which their stiffness became constant. The Young’s moduli of product B are shown in Fig. 3. When native WPI was partly replaced by 2.5% (w/w) WPI microparticles (B1), the value of Young’s modulus was rather similar to the one without microparticles (A1). Lower values for Young’s moduli were obtained when the amount of microparticles was increased while keeping the total protein concentration constant. Product variants that contained 7.5% (w/w) microparticles resulted in a lower Young’s modulus than the variants containing 2.5% (w/w). The lower modulus was also obtained when the total protein concentration was decreased. The total change in Young’s modulus of B1 was similar to that of A1 over time. However, the change occurred over a longer period than 3 days for B2 and B3. Variant B4, which contained 7.5% (w/w) WPI microparticles at 15% w/w overall protein concentration, did not form a gel. Nevertheless, its changes in consistency over time could be clearly observed visually. This variant became more viscous during a 6-day storage period, which may be interpreted as being analogous to stiffening. Overall, it seems that product B showed a different stiffening behaviour compared with A, even though the effects were small. The changes in the fracture stresses and strains over time for product B are depicted in Fig. 4. The changes of the fracture stresses for B1 and B2 were in accordance with the changes of their Young’s moduli. The fracture strains of B1 and B2 slightly decreased over
65
time. This indicates that a stiffer gel behaves more as a brittle material. No data are shown for B3 because most samples of this variant did not fracture. The compression tests of product C did not result in fractured samples in all variants. Therefore, only the result on Young’s modulus can be extracted from the large deformation test. Fig. 5 shows that Young’s moduli of C1 and C2 were larger than the values for the matrix (C3). This increase is typical for a filled gel with a dispersed phase that has a higher modulus than the continuous phase or has a good interaction with the continuous phases (Sala, van Aken, Stuart, & van de Velde, 2007; van Vliet, 1988; Yost & Kinsella, 1992). The change of Young’s modulus of C3 over time was limited. Variation in the first 3 days was hardly significant. The modulus of C2 increased in the first 3 days and then stabilised. C1 showed similar behaviour as C3 over time. If the Young’s modulus of C1 and of C2 was normalised to its value at time 0, it turned out that Young’s modulus of C1 changed in similar way as of C3 over time (Fig. 5, inset). For C2, its Young’s modulus indeed increased in the first 3 days and then, became stable. 3.3. Small deformation properties of model products The nature of the rheological measurement used in this study was good for liquid or semi-solid sample and was rather tricky for materials with sedimenting particles. Therefore, we tested high WPI concentration (A1) and product B (B1 and B2) in replicates at least to check the reproducibility of the measurement solid sample (A1) and samples that contained microparticles. The measurement for C3 was also repeated twice to confirm the reproducibility for semi-solid sample. Fig. 6 shows the result of the normalised G0 , G00 and tan d-values from product A, B and C over time. The deviation of the variants that measured in repetition is also presented. Fig. 6a shows the changes in normalised G0 and G00 -values of product A at 20 C during a 72-h period. The normalised G0 -value of A1 increased until around 60 h, after which it seemed to decrease. For A2, the normalised G0 -value slightly increased throughout the whole time range. The normalised G00 -value for A1 increased in the first 30 h and then it decreased. For A2, this value tended to decrease over time. The changes in G0 or/and G00 -value resulted in a change in damping factor (tan d) over time. The normalised tan d-values of A1 slightly increased in the first 24 h and then it decreased. The standard deviation shows that the increase in the first 24 h was not significant. The value for A2 slightly decreased over the storage time. The change in the normalised G0 -value over time for B1over time was similar to that of for A1 (Fig. 6b). The changes in the normalised
Fig. 2. The changes in (panel a) fracture stresses and (panel b) fracture strains of product A over time: - and C, A1 (20%, w/w, whey protein isolate); , and B, A2 (15%, w/w, whey protein isolate). The lines guide the eyes to the fracture trends over time. The vertical bars are the standard deviations.
66
N. Purwanti et al. / International Dairy Journal 28 (2013) 62e69
3.4. Microscopy observation
Fig. 3. The changes in Young’s moduli of product B over time containing a total protein concentration of (B1 and B2) 20% (w/w) and (B3) 15% (w/w): -, B1 (17.5% WPIe2.5% microparticles); ,, B2 (12.5% WPIe7.5% microparticles); , B3 (12.5% WPIe2.5% microparticles). The lines guide the eyes. The vertical bars are the standard deviations. Inset is the B3 with an enhanced y-axis.
G00 and tan d-values were rather different from those of for A1. The value for B1 tended to increase, while it decreased for A1 during the first 30 h, suggesting the occurrence of another effect influence the dynamic properties. Rather similar changes between A1 and B1 might be attributed to the relatively small amount of WPI microparticles applied (2.5%, w/w) in a concentrated WPI matrix (17.5%, w/w). The normalised G0 -values of B2 and B4 were rather stable over time, while, their normalised G00 -values tended to decrease. Therefore, their normalised damping factor decreased slightly with time. Contrary to the large deformation measurements, the changes in G0 and G00 seem to occur only in the first days of measurements. The reason for this difference is not clear yet, but might be related to the fact that here the sample was deformed continuously. For the large deformation tests, separate samples were used that were not deformed upon storage. Fig. 6c shows that the addition of microparticles to the matrix resulted in lower normalised G0 -values than the matrix itself. The actual G0 -values of C1 and C2 were much higher than the matrix. The normalised G0 -values for all variants of product C increased over time with the matrix showed a larger increase in the normalised G0 -values than C1 and C2 over time. The normalised G00 -values for all variants were the same and they tended to be constant over time. Therefore, their normalised tan d tended to decrease over time. The results from small deformation indicate that all the product variants underwent changes during the storage time, unlike the results from large deformation.
Fig. 7 shows the swelling of microparticles when being dispersed in water and heated. After 72 h storage, some microparticles seem to fall apart, which gave a smaller particle size. Swelling also occurred when the microparticles were dispersed in solution containing either WPI or polysaccharide and heated. Fig. 7 shows the changes of microparticles inside the protein gels and polysaccharide gel after 72 h storage. The ratios of particle size (d4,3) in the gels after 72 h storage and the particle size in the gels after heating (0 h) were 1.3 for B3 and 1.4 for B4. This indicates that microparticles continued to swell over time when they were in gels. The ratio of particle size in the polysaccharide gel after 72 h storage was 1.0, indicating no significant swelling. This is an interesting finding, which is in line with the behaviour of this product over time measured with large deformation. The microparticles in polysaccharide gel at 15% (w/w) were packed in the matrix, therefore, particle analysis could not be performed for this sample. A possible explanation for the lack of swelling could be the fact that the XG is charged naturally, thereby lowering the water activity in that phase. This reduction in water activity will reduce the driving force for water migration and hence swelling of the particles. 4. Discussion Changes in the textural properties of high-protein gels were studied with Young’s modulus (stiffness), fracture properties, normalised G0 , G00 and tan d. Overall, it was concluded that the extent of stiffening of all products was rather limited. The changes of protein gels made from only native WPI took place in the first three days mainly. Previous studies suggested a mechanism of stiffening for a concentrated WPI matrix based on protein aggregation through disulphide bond formation (Zhou & Labuza, 2007; Zhou et al., 2008a; Zhou, Liu, & Labuza, 2008b). These changes in the gel properties caused by additional disulphide bonds could account for the changes in the first few days of storage for WPI gels. The changes in the products containing WPI microparticles were also rather limited. However, it seems that the stiffening occurred throughout the whole storage period tested, meaning that the stiffening continued for a longer period than the WPI-gels. This suggests that an additional mechanism might play a role in those structured products. The microscopy analysis showed changes in particle sizes upon heating and storage. This indicates that water migration between the protein matrix and the protein particles takes place, which effect could be responsible for stiffening over a prolonged period. Water migration occurs as long as there is a difference in water activity between different phases. The water redistribution itself might take place over a time scale of a few days.
Fig. 4. The changes in (panel a) fracture stresses and (panel b) fracture strains of product B over time: - and C, B1 (17.5% WPIe2.5% microparticles); , and B, B2 (12.5% WPIe7.5% microparticles). The lines guide the eyes. The vertical bars are the standard deviations.
N. Purwanti et al. / International Dairy Journal 28 (2013) 62e69
67
Fig. 5. The changes in Young’s moduli of product C over time, at a total protein concentration of: -, C1 (20%, w/w, microparticles); ,, C2 (15%, w/w, microparticles); and , C3 (0%, w/w, microparticles: the polysaccharide matrix). The lines guide the eyes. The vertical bars are the standard deviation. Inset is the normalised Young’s moduli over time.
A low estimate of the water diffusivity, i.e., 1012 m2 s1 results in a typical diffusion-length of 800 mm over 2 days. This is around 11 times the average size of dry microparticles. Therefore, diffusion might not be the rate limiting step. Another effect, such as the internal relaxation of the protein inside the microparticles, could be the limiting step for water redistribution. The swelling of microparticles is limited by the intramolecular disulphide bonds, which are able to slowly rearrange upon time. Once a disulphide bridge opens, the protein matrix will rearrange and it will be able to absorb additional water. This row of events (water absorption, disulphide rearrangements followed by additional swelling) might be an explanation for the changes of the product over a longer period. The importance of the number of disulphide bridges on swelling was shown in similar WPI particles produced in a different manner that allowed controlling the number of bridges present in the system (van Riemsdijk, van der Goot, & Hamer, 2011). No changes were observed in the product containing WPI microparticles in polysaccharide matrix in large deformation, except for C2, while limited changes were observed in small deformation. In large deformation, the change in C2 with 15% (w/w) microparticles (Fig. 5) was similar to that in product A and B1. It was limited in the first few days. By increasing particle concentration (C1), hardly any changes were observed anymore. The change in Young’s modulus was similar to the matrix (C3). The microscopy analysis revealed that the microparticle diameter in polysaccharide matrix is relatively constant over time after the initial swelling. Storage of the dispersion at ambient conditions resulted in relatively stable particle. This behaviour might be suggested by thermodynamic equilibrium between the microparticles and the polysaccharide matrix after heating. Therefore, hardly any changes can occur anymore upon storage at ambient temperature. This was observed especially at 20% (w/w) of microparticles. Nevertheless, all variants of product C showed slight changes in small scale deformation, indicating some kind of change in the
Fig. 6. Normalised G0 , G00 and tan d-values over storage time. Panel a, products A1 and A2: - and :, normalised G0 for A1 and A2, respectively; , and 6, normalised G00 for A1 and A2, respectively; inset is normalised tan d for A1 (-) and A2 (6). Panel b, products B1, B2 and B4: -, : and A, normalised G0 for B1, B2 and B4, respectively; ,, 6 and >, normalised G00 for B1, B2 and B4, respectively; inset is normalised tan d for B1 (-), B2 (6) and B4 (A). Panel c, products C1, C2 and C3: -, : and A, ratio G0 for C1, C2 and C3, respectively; ,, 6 and >, ratio G00 for C1, C2 and C3, respectively; inset is normalised tan d for C1 (-), C2 (6) and C3 (A). Product composition is as defined in Table 1. The vertical bars are the standard deviations.
product. This could be the result of slow rearrangements within the microparticles in which the changes only implied in small deformation. It has already been reported that an increase in hardness based on fracture stress can continue for 50 days for protein bars made
68
N. Purwanti et al. / International Dairy Journal 28 (2013) 62e69
Fig. 7. Microparticles in water after heating and in gels, observed at time 0 (AeD) and 72 h (EeH). Panels a and e: heated microparticle dispersion (2.5%, w/w) in water. Panels B and F: 12.5% (w/w) whey protein isolate e 2.5% (w/w) microparticles: Panels C and G: 7.5% (w/w) whey protein isolate e 7.5% (w/w) microparticles. Panels D and G: 2.5% (w/w) microparticles in locust bean gum e xanthan gum.
with calcium caseinate (CaCas) or milk protein concentrate (MPC) (Loveday et al., 2009, 2010). These bars normally contain sugar or polyol as additional components and do not undergo heat treatment during preparation (Li et al., 2008; Liu et al., 2009; Loveday et al., 2009, 2010; McMahon et al., 2009). The absence of a heating step in the preparation could explain the longer period during which stiffening occurs as potential effects of water migration are not diminished due to a thermal treatment. Based on this and other studies, we conclude that stiffening of high-protein product is not caused by the properties of the protein only. A process consisting of various events with mutual interactions such as difference in water activity between phases, rearrangement in the dispersed phase and the matrix phase, or crystallisation in a sugar-rich phase might cause the stiffening over a prolonged period. When WPI microparticles are used together with WPI matrix, stiffening effect that occurs over a longer time might be related to the slow protein rearrangement in the microparticles, which are caused by and induces water migration. Stopping or shortening the stiffening process might be achieved by applying an extensive heating step to accelerate water redistribution to reach thermodynamic equilibrium inside between the various phases inside the product. Another method could be to stop moisture migration happening. This can be done, for example, by modifying the production of microparticles in such a way that the microparticles cannot absorb water. When polysaccharide is used as the matrix, further study is interesting to understand more about the dynamic between WPI microparticles and the matrix. 5. Conclusions The stiffening over time was investigated in protein gels made from WPI only and in gels in which part of the protein was present as a dispersed phase (microparticles). Overall, it can be
concluded that the extent of hardening is limited in those model products. The use of a dispersed protein phase did not reduce the extent of stiffening. Water redistribution and rearrangements of crosslinks within the protein microparticles were used as possible explanations for the changes of the product over time. Better control over the stiffening process can be achieved by accelerating the re-ordering and transport during processing, for example, by heating. This could stabilise the products over a shorter period of time. When the continuous phase is different from the protein phase, changes in the product might not be avoided as long as there is not thermodynamic equilibrium between different phases. References Campbell, W. W., & Leidy, H. J. (2007). Dietary protein and resistance training effects on muscle and body composition in older persons. Journal of the American College of Nutrition, 26, 696Se703S. Hayes, A., & Cribb, P. J. (2008). Effect of whey protein isolate on strength, body composition and muscle hypertrophy during resistance training. Current Opinion in Clinical Nutrition and Metabolic Care, 11, 40e44. Labuza, T. P., Zhou, P., & Davis, L. (2007). Whey to go: application of whey proteins in nutrition bars. World of Food Ingredients, October/November, 108e112. Li, Y., Szlachetka, K., Chen, P., Lin, X., & Ruan, R. (2008). Ingredient characterization and hardening of high-protein food bars: an NMR state diagram approach. Cereal Chemistry, 85, 780e786. Liu, X., Zhou, P., Tran, A., & Labuza, T. P. (2009). Effects of polyols on the stability of whey proteins in intermediate-moisture food model systems. Journal Agricultural Food and Chemistry, 57, 2339e2345. Loveday, S. M., Hindmarsh, J. P., Creamer, L. K., & Singh, H. (2009). Physicochemical changes in a model protein bar during storage. Food Research International, 42, 798e806. Loveday, S. M., Hindmarsh, J. P., Creamer, L. K., & Singh, H. (2010). Physicochemical changes in intermediate-moisture protein bars made with whey protein or calcium caseinate. Food Research International, 43, 1321e1328. Luhovyy, B. L., Akhavan, T., & Anderson, G. H. (2007). Whey proteins in the regulation of food intake and satiety. Journal of the American College of Nutrition, 26, 704Se712S.
N. Purwanti et al. / International Dairy Journal 28 (2013) 62e69 McMahon, D. J., Adams, S. L., & McManus, W. R. (2009). Hardening of high-protein nutrition bars and sugar/polyol-protein phase separation. Journal of Food Science, 74, E312eE321. Purwanti, N., Moerkens, A., van der Goot, A. J., & Boom, R. (2012). Reducing the stiffness of concentrated Whey Protein Isolate (WPI) gels by using WPI microparticles. Food Hydrocolloids, 26, 240e248. Purwanti, N., van der Goot, A. J., Boom, R., & Vereijken, J. (2010). New directions towards structure formation and stability of protein-rich foods from globular proteins. Trends in Food Science and Technology, 21, 85e94. van Riemsdijk, L. E., van der Goot, A. J., & Hamer, R. (2011). On the stability of whey protein particles during production of gluten-free bread. Food Hydrocolloids, 25, 1744e1750. Sala, G., van Aken, G. A., Stuart, M. C., & van de Velde, F. (2007). Effect of dropletmatrix interactions on large deformation properties of emulsion-filled gels. Journal of Texture Studies, 38, 511e535.
69
van Vliet, T. (1988). Rheological properties of filled gels: influence of filler matrix interaction. Colloid Polymer Science, 266, 518e524. Yost, R. A., & Kinsella, J. E. (1992). Microstructure of whey protein isolate gels containing emulsified butterfat droplets. Journal of Food Science, 57, 892e897. Zhou, P., & Labuza, T. (2007). Effect of water content on glass transition and protein aggregation of whey protein powders during short-term storage. Food Biophysics, 2, 108e116. Zhou, P., Liu, X., & Labuza, T. P. (2008a). Effects of moisture-induced whey protein aggregation on protein conformation, the state of water molecules, and the microstructure and texture of high-protein-containing matrix. Journal of Agricultural and Food Chemistry, 56, 4534e4540. Zhou, P., Liu, X., & Labuza, T. P. (2008b). Moisture-induced aggregation of whey proteins in a protein/buffer model system. Journal of Agricultural and Food Chemistry, 56, 2048e2054.