Pektin & Gelatin

Food Hydrocolloids 15 (2001) 643653
www.elsevier.com/locate/foodhyd
Texture and structure of gelatin/pectin-based gummy confections

Laura L. DeMars 1, Gregory R. Ziegler*
Department of Food Science, The Pennsylvania State University, 116 Borland Laboratory, University Park, PA 16802, USA Received 1 July 2000; accepted 5 March 2001
Abstract The texture of gelatin:high-methoxyl pectin gummy gels was quantied by instrumental and sensory techniques and their microstrucuture examined by light and transmission electron microscopy. Gelatin:HM pectin confectionery gels (33.4% sucrose and 29.8% 42 DE corn syrup solids) with 3.0, 4.5, or 6.0% gelatin and 0.0, 0.5, 1.0, or 1.5% HM pectin were formed into oval-shaped samples and fractured in tension. Descriptive sensory evaluation was done on seven of these gels in duplicate by 10 experienced panelists using free choice proling. The addition of pectin reduced the strain at fracture of gelatin gels. Stress at fracture could be described by upper and lower bound behavior. Microstructural analysis suggested that at high total polymer or pectin concentration, increased phase viscosity and rate of gelation inuenced structure by preventing coalescence of the dispersed gelatin-rich phase. Micrographs suggested that gelatin in the pectin-rich phase was concentrated enough to gel and contributed to mixed gel properties. Gels were described variously as soft to rm and brittle to rubbery. Mixed gels were more fruity, sweet, and tart than pure gelatin gels. Gels with a high degree of coalescence of the dispersed phase were described as pulpy. Sensory texture rst principal component values correlated with strain at fracture (r 0.90), log [stress at fracture] (r 0.87), and avor rst principal component values (r 0.83). q 2001 Elsevier Science Ltd. All rights reserved.
Keywords: Gelatin; Pectin; Mixed gels; Texture; Flavor; Free choice proling
1. Introduction Structure-forming polysaccharides and proteins provide desired functional properties to a wide range of foods (Kinsella, Rector, & Phillips, 1994), and blends of such hydrocolloids have lately been the subject of an increasing number of investigations because of the prospect of discovering useful synergistic effects (Ipsen, 1995). This research has begun to demonstrate how physical properties of blends can be related to phase morphology (Owen & Jones, 1998). A number of different analytical approaches have been employed to elucidate structure-function relationships (BeMiller, 1996; Dickinson & McClements, 1995; Kalab, Allan-Wojtas, & Miller, 1995; Ross-Murphy, 1994). Lillford (2000) suggested that food science should borrow from the science of structural mechanics and concentrate on failure properties, since these show the greatest relationship to the perception of texture in the mouth. Confectionery products are ideal for such study. They are simple enough so as to avoid some of the complexity inherent in biological systems, but are real food products. Non* Corresponding author. Tel.: 11-814-863-2960; fax: 11-814-863-6132. E-mail address: grzl@psu.edu (G.R. Ziegler). 1 Current address: Kraft Inc., Glenview, IL.
chocolate confectionery products are distinguished from one another, to a large extent, on the basis of their texture, being largely composed of similar ingredients. Gelled confections are one of the fastest growing candy markets in the country. Opportunities exist for improving or modifying their texture, e.g. pectin is believed to shorten the elastic texture of gelatin-based gummy confections (Poppe, 1995); however, the resulting textures have never been adequately dened or quantied. Unlike simpler mixed-gel model systems, gummy gels contain little moisture (,20%) and a signicant proportion of low molecular weight carbohydrates (sucrose and corn syrups). Tschumak, Wajnermann, and Tolstoguzov (1976) studied complexes of gelatin and low-methoxyl pectin at low pH and found that the complex precipitate could be used to form thermostable gels. Tolstoguzov (1990) studied the compatibility of dilute aqueous mixtures of gelatin and pectin and developed phase diagrams for varying conditions of NaCl, pH, temperature, and pectin degree of esterication. Clewlow, Rowe, and Tombs (1995a); Clewlow, Clark, Rowe, and Tombs (1995b) have also investigated the phase behavior of gelatin-pectin mixtures. Poppe (1995) studied high sugar solids and low pH gelatin:high-methoxyl (HM) pectin gels and found that the instrumental hardness increased with: increasing concentration of gelling agents,
0268-005X/01/$ - see front matter q 2001 Elsevier Science Ltd. All rights reserved. PII: S 0268-005 X(01)00 044-3
644
L.L. DeMars, G.R. Ziegler / Food Hydrocolloids 15 (2001) 643653
-vis decreasing pH between 3.35 and 3.25, 60 DE vis-a 40 DE corn syrup, and decreasing sugar/corn syrup ratio below 40/60 when 40 DE corn syrup was used. Texture of mixed gels can be quantied by measuring fracture properties. Highly elastic gels may not fracture in compression or torsion tests, so tension tests may be the most appropriate (Hamann & Foegeding, 1994; Langley, Green, & Brooker, 1991). McEvoy, Ross-Murphy, and Clark (1985a) fractured agar:gelatin gels in tension using circular ring-shaped samples (Rodgers, 1973) stretched between dowel pins. Oval-shaped samples were used by Meyers and Wenrick (1974). The method of free-choice proling (FCP) (Williams & Langron, 1984) has been used to describe numerous complex foods whose sensory quality is an integrated response to multiple stimuli, including texture-avor interactions in gelatin desserts (Jaime, Mela, & Bratchell, 1993). FCP requires panelists, either experienced in sensory scaling or inexperienced, to develop their own list of descriptors and range of scaling on an unstructured line scale. This method alleviates problems arising from forced agreement between panelists and reduces training time and cost involved in developing a consensus. Generalized Procrustes Analysis (Gower, 1975) is used to analyze the data. The present study was conducted to quantify the texture of a range of gelatin/HM pectin-based gummy gels by instrumental and sensory analysis, and to relate this qualitatively to gel microstructure with the goal of establishing structure-property-perception relationships.
Fig. 1. Flow diagram for preparation of mixed gels.
2. Materials and methods 2.1. Materials The following ingredients were used in the formulation of gummy gels: Rousselot 250 A 30 gelatin (Systems BioIndustries, Inc., Waukesha, WI), Unipectin Yellow Ribbon high-methoxyl pectin (Systems Bio-Industries, Inc., Waukesha, WI), sucrose (Rened Sugars, Inc., Yonkers, NY), 42 DE corn syrup (ADM Corn Processing, Decatur, IL), food grade citric acid monohydrate (ADM Food Additives Division, Decatur, IL), Red Liquid Color 0325 (Warner Jenkinson Company, St Louis, MO), and NAT Cherry Flavor WONF WS USL-32895 (Systems BioIndustries, Inc., Trevose, PA). Moisture content of gelatin and pectin powders was determined to be 7.03 and 11.95% by weight, respectively, by vacuum oven drying using A.O.A.C. method 14.003 (A.O.A.C., 1980). Pectin concentration in the Unipectin Yellow Ribbon sample was quantied to be 77.4 ^ 0.2%, by precipitation with three volumes of 80.0% ethanol, ltration, drying at 1008C for 5 h, and weighing (Kertesz, 1951; Lin & Humbert, 1978). The degree of esterication for the Unipectin Yellow Ribbon pectin was 5965% (Systems Bio-Industries, Inc., Waukesha, WI). Percent solids of the
corn syrup was determined to be 81.9% using an ABBE Mark II Digital refractometer (Cambridge Instruments, Inc., Buffalo, NY). 2.2. Preparation of gummy gels 2.2.1. Experimental design A completely randomized 3 4 factorial design was proposed with two factors: concentration of gelatin (3.0 wt%, 4.5 wt%, and 6.0 wt%) and concentration of high-methoxyl pectin (0.0 wt%, 0.5 wt%, 1.0 wt%, and 1.5 wt%). However, the treatment containing 6.0 wt% gelatin and 1.5 wt% pectin was omitted because entrapped air bubbles could not be easily removed and interfered with rheological measurements. Two replicate batches of each gel were prepared. 2.2.2. Formulation and production Table 1 shows the gummy gel recipe. Fig. 1 depicts a ow diagram for the preparation of the gels. Each gel was made using the same amount of gelling slurry (72.6 g). Gelling slurries were formulated such that nal gels contained the desired amount of gelatin and pectin adjusted for the actual polymer content in the commercial sample. Citric acid
645
Fig. 2. Aluminum molds for cast oval-shaped gel rings (outer length of oval 5.90 cm, center-to-center length 2.48 cm, radius of outer arc 1.71 cm, width of gel leg 1.03 cm, depth of gel 1.0 cm).
solution (50% w/v) was added to obtain a nal pH of 3.35 ^ 0.05. Gelatin and pectin powders were mixed together and added to 120.0 g boiling water in a Waring Commercial Blender (Dynamics Corp. of America, New Hartford, CT). The blender was set at low speed and connected to a 120 V rheostat (Staco Energy Products Co., Dayton, OH), initially set at 48 V and turned down to 28 V after 5 s. The slurry was deaerated in loosely covered glass jars at 908C in a water bath (Model W19, Haake, Karlsruhe, Germany). During slurry deaeration, corn syrup, sucrose, and water were
Table 1 Recipe for gelatin:pectin gummy gels (250.0 g batches) Ingredient Water 42 DE corn syrup Sucrose Gelling slurry a Citric acid solution (50% w/v) b Flavor Mass (g) 26.0 91.0 83.5 72.6 variable 0.5
brought to a total mass of 178 g at t1258C. Evaporation after the addition of gelling slurry, citric acid, and avor brought the solution to the nal desired mass of 250.0 g. For molding, the nal gel solution was poured into a foam-insulated Pyrex funnel (inner stem diameter 1/ 8 in.) and deposited into oval-shaped aluminum molds (Fig. 2). A thin layer of PAM w (American Home Foods, Milton, PA) was sprayed on the surface of the gel to prevent drying. Gels were stored at 198C and instrumental tests performed after 24 h. 2.3. Instrumental tests Twenty-four hours after preparation, the oval-shaped gels were stretched in tension at 5 mm/s on two 1.40 cm aluminum dowel pins attached to a TA.XT2 Texture Analyzer (Stable Micro Systems, Halsemere, Surrey, UK). Nine measurements of stress and strain at fracture were made for each batch. Dowel pins and inside surfaces of the samples were lightly coated with mineral oil to prevent friction. Force and deformation at fracture were used to calculate stress at fracture (s ) and strain at fracture (e ) using Eqs. (1) and (2), respectively. F 2DL s 11 1 2A 0 Ca
a Composition of gelling slurry depends on nal gelatin and pectin concentrations. The quantities of commercial gelatin and pectin preparations were adjusted to obtain the desired concentrations of gelling polymer. b Citric acid solution (50.0% w/v) was added so nal pH of gel was 3.35 and varied from 2.4 to 6.3 g depending on the polymer concentration.
646
(Bagley, 1983; Tschoegel, Rinde, & Smith, 1970) 2D L e ln 1 1 Ca
(van Vliet, Luyten, & Walstra, 1991).where: s is the true stress at fracture; F is the force at fracture; A0 the initial cross sectional area (m 2); DL the change in leg length (m); Ca the average circumference; and e the Hencky or true strain at fracture. 2.3.1. Statistical analysis of instrumental data Stress and strain at fracture were analyzed by a two-way ANOVA using gelatin concentration and pectin concentration as factors in the GLM procedure (SAS Institute, Inc., 1985). The two replicates and nine observations for each treatment were used in the analysis and the experimental error was used in the F tests. The interaction term from analysis of variance for both stress and strain at fracture was used to see if the main effects of factors (i.e. gelatin concentration and pectin concentration) could be isolated. The interaction term was statistically signicant (P , 0.05) for both. Therefore, main effects of gelatin and pectin concentration could not be isolated, and multiple comparisons of individual means by pairwise t-tests using a Bonferroni adjustment on the P value were made for stress at fracture and strain at fracture to the separate means at a 0.05. 2.4. Sensory analysis by free choice proling Seven gelatin:pectin gels (3.0:0.0, 3.0:1.0, 3.0:1.5, 4.5:0.0, 4.5:1.0, 6.0:0.0, and 6.0:1.0) that showed different stress and strain at fracture from instrumental measurements were chosen for evaluation by sensory analysis. A reference gel, 4.5:0.5, was chosen for sensory evaluation to maintain consistency in panelists' evaluations (Oreskovich, Klein, & Sutherland, 1990). Each gel treatment was replicated. Two sensory measurements and four stress and strain measurements were made on each replicate. Identical ingredients were used for sensory analysis except Staley 1300 42 DE corn syrup (A.E. Staley Manufacturing Company, Decatur, IL) was substituted. Batch size was increased to 1500 g. To accommodate this increase, some procedural alterations were made. Heating of the water, corn syrup, and sucrose was started at the same time as the preparation of the gelling slurry because more time was needed to bring the sugar mixture to 1258C. Gelatin and pectin powders were mixed in 720.0 g water for 10.0 min instead of 2.5 min to ensure complete hydration. A Lightninw Lab master mixer, model MS 1500, (Mixing Equipment Co., Inc., Rochester, NY) was used to mix the gelling slurry because the blender could not mix this larger volume. Mixing times after slurry addition and after avor and citric acid addition were similar to those for the 250.0 g batches. For molding, 1000 mL were poured into a 17 11 3/4 in. non-stick baking tray, covered with plastic
wrap, and stored at 198C, and four ovals cast for measurements of stress and strain at fracture. The panel was composed of 10 graduate students or staff employees from the Department of Food Science at the Pennsylvania State University. All panelists had prior experience in quantitative descriptive analysis. Training took place during three sessions over 3 days. On the rst day, panelists sampled a range of gels and developed an individual list of texture and avor attributes. During the next session, the panelists rened their attribute lists, developed individual denitions for each attribute, and practiced scaling products on 15.0 cm line scales anchored at the extremes. On the third day, panelists continued to practice scaling and scaled the reference gel. Following the completion of training, each panelist had developed a list containing personal denitions and intensity scores for the reference product. Testing was conducted at the Sensory Evaluation Laboratory in the Department of Food Science, Penn State University. Evaluations took place in partitioned booths illuminated with red lights to mask any color intensity differences. Samples were stored and evaluated at 198C. One batch of each treatment was prepared and evaluated one week and another batch was prepared and evaluated the following week. Each batch was evaluated twice by the sensory panel. Gels were cut into 1.5 3.0 0.8 cm. rectangles and at least four samples were placed in medium size plastic cups with lids coded with 3-digit random numbers. Fourteen samples from each replicate of seven gels were randomized and evaluated in ve sessions over 3 days. Three samples were evaluated during the rst four sessions and two at the remaining session. At each session, panelists were given their denition sheet which included intensity scores for the reference, individual score sheets, reference samples, experimental gel samples, unsalted crackers, and water. 2.4.1. Statistical analysis of sensory data Sensory data were separated by texture and avor and analyzed separately by Generalized Procrustes Analysis (GPA) (Gower, 1975) using the Procrustes-PC v 2.2 computer package (Oliemans Punter & Partners, 1991). Statistical analysis of the rst principal component values was similar to the analysis of the instrumental data. The rst principal component values for the two replicates and two observations of each gel were analyzed by two-way ANOVA using the GLM procedure (SAS Institute, Inc., 1985). F tests were done using the experimental error term as the denominator. Multiple comparisons using pairwise t-tests with a Bonferroni adjustment on the p value were applied to the separate means at a 0.05. 2.5. Microscopy The following four gelatin:pectin mixed gels, chosen because they had different rheological and sensory properties,
647
Fig. 3. Stress at fracture as a function of gelatin and pectin concentration.
were examined by light and electron microscopy: 3.0:1.0; 3.0:1.5, 4.5:0.5, and 4.5:1.0. Each gel was stored at 198C for 4 days and then xed for microscopy as described below. One-millimeter cubes were xed in 3.0% glutaraldehyde in 0.10 M sodium cacodylate buffer (pH 7.4) for 2 h, and then post-xed in 1.0% osmium tetroxide in 0.10 M sodium cacodylate buffer (pH 7.4) for 1 h at 198C. Gels were stored in 0.10 M sodium cacodylate buffer overnight. Gel pieces were stained, en bloc, in 1.0% uranyl acetate in 0.10 M sodium acetate buffer overnight, dehydrated in a gradient series of acetone, and embedded in Spurr's low viscosity medium. Thick sections (500 nm) were cut with a microtome (Ultratome 3, LKB Pharmacia Biotechnology, Gaithersburg, MD), stained with a 1.0% toluidine blue in 1.0% sodium borate, and examined using a light microscope (Carl Zeiss, Inc., Thornwood, NY) at 400 and 1000 magnication.
Thin sections (6080 nm) were cut with a microtome (LKB Ultratome 3), stained with 3.0% uranyl acetate in 7.5% methanol, post-stained in 2.0% lead citrate, and observed with a transmission electron microscope (Model 11E, Hitachi, Japan) at 80 kV.
3. Results and discussion 3.1. Instrumental properties Figs. 3 and 4 show the stress and strain at fracture, respectively, as a function of gelatin and pectin concentrations. The effect of pectin on the fracture properties of gelatin gels was not simple. This complexity in rheological properties is typical for two-phase mixed gels, and can be explained by a modied Takayanagi approach (Clark,
Fig. 4. Strain at fracture as a function of gelatin and pectin concentration.
648
Fig. 5. Light micrographs of gelatin:pectin gels: A. 3.0:1.0, B. 4.5:1.0 (bar 10 mm).
Richardson, Ross-Murphy, & Stubbs, 1983; Kasapis, Morris, Norton, & Clark, 1993; McEvoy, Ross-Murphy, & Clark, 1985b; Walkenstrom and Hermansson, 1994; Ziegler & Rizvi, 1989). Rheological properties of mixed gels follow upper bound behavior when the stronger gel forms the continuous phase, follow lower bound behavior when the weaker gel forms the continuous phase, and fall in between bounds at or near the phase inversion point. What can be inferred from Fig. 3, is that at 3 and 4.5% gelatin, the addition of even 0.5% pectin caused phase inversion from a gelatin-supported network to a pectinsupported network and a transition to the lower bounds. Between 1 and 1.5% pectin, this transition was complete, and the stress at fracture began to increase again. At 6% gelatin, the stress at fracture increased between 0 and 0.5% pectin seeming to indicate upper bound behavior. The stress then decreased between 0.5 and 1.0% pectin as phase inversion began. This is sensible since one would expect a greater amount of pectin to be required to cause phase inversion at higher gelatin concentrations. However, `rheology without morphology is theology' (Stein, personal communication). Fig. 4 reveals that the strain at fracture decreased as pectin was added to a gelatin gel. This would seem to
conrm the proposition that pectin `shortens' a gelatin network (Poppe, 1995). The reduction in strain at fracture from 6.0:0.0 to 6:0.5 is similar to the behavior of agar/gelatin gels (McEvoy et al., 1985b). They found that as the percentage of gelatin in the gelatin-rich phase decreased toward phase inversion, the strain at fracture of the mixed gel also decreased. The decreasing strain at fracture corresponds with the statements of Poppe (1995) that gelatin:HM pectin gels have properties intermediate between elastic gelatin gels and brittle pectin gels. 3.2. Microscopy Figs. 5 and 6 show light and electron micrographs, respectively, of two gelatin:pectin mixed gels (3.0:1.0 and 4.5:1.0). These gelatin:pectin mixed gels have phase separated into a gelatin-rich phase and a pectin-rich phase. Identication of the gelatin-rich phase as the dark area and pectin-rich phase as the light area was supported by staining tests. Osmium tetroxide liquid and vapor staining tests showed that gelatin confectionery gels stain dark and pectin confectionery gels do not stain at all. Behram (1985) stated that polysaccharides do not react to a signicant extent with
649
Fig. 6. Transmission electron micrographs of gelatin:pectin gels: A. 3.0:1.0 (bar 1.0 mm), B. 4.5:1.0 (bar 0.5 mm).
osmium tetroxide. Uranyl acetate staining en bloc is used to improve contrast of the oxmium tetroxide staining (Hayat, 1989). Toluidine blue does not stain pectin after it has been acidied with oxalic acid (Flint, 1994). For the mixed gels in this study, the dark areas stained with toluidine blue after acidication with oxalic acid indicating they were not pectin or at least that they contained substantial gelatin. Furthermore, a hydroxylamine-ferric chloride reaction (Reeve, 1959), stains pectin red. The dark areas in the mixed gelatin:HM pectin gels did not stain red for this test. In all the cases examined, the pectin-rich phase was continuous with a dispersed gelatin-rich phase as predicted by rheological measurements. Furthermore, the morphology in Fig. 6(A) is nearly identical to that observed for high-impact polystyrene produced by the phase-inversion method (Sperling, 1992). An extensively-coalesced dispersed phase was observed in the 3.0:1.0 and 4.5:0.5 gels, while a more homogenous structure revealing less dispersed phase coalescence was observed at 3.0:1.5 and 4.5:1.0. This can be explained by the competing effects of the rate of coalescence, inuenced by the phase viscosity, and the rate of gelation, affected by the gel concentration. At lower total polymer concentration, the rate of coalescence increases and the rate of gelation decreases favoring larger dispersed phase morphology, the inverse being true at higher total polymer concentration. Dickinson (1995) simulated microstructures of mixed
particulate gels and predicted that mixed gel microstructure is inuenced by rate of gelation. He also predicted that microstructure is inuenced by attractive or repulsive forces between molecules because they inuence the rate of separation. Dickinson's simulations were consistent with the structures observed in Fig. 6. Pectin has a higher afnity for water than does gelatin (Tolstoguzov, 1991), suggesting that pectin would tend to form the continuous phase. Other work on protein/polysaccharide mixtures supports the idea that pectin will concentrate gelatin in a gelatin-rich phase. Zhuravskaya, Kiknadze, Antonov, and Tolstoguzov (1986) showed that pectin `concentrated' milk protein ve to 12 times its initial concentration. Tolstoguzov (1988) showed that aqueous polysaccharide/protein mixtures separate into a dilute polysaccharide phase and a concentrated protein phase. However, phase separation is rarely complete. Fig. 6 shows that gelatin (dark strands) is present in the pectin-rich phase (lighter area). The presence of gelatin in the pectin-rich continuous phase is important. Although it is present at lower concentrations than in the dispersed phase, it may still be concentrated enough to gel. In this case, the pectin-rich continuous phase could be considered an interpenetrating polymer network (IPN), because both gelling agents form continuous networks and are load bearing (Stading, 1993). When pure pectin gels were placed in osmium tetroxide they dissolved; however, mixed gels
650
L.L. DeMars, G.R. Ziegler / Food Hydrocolloids 15 (2001) 643653 Table 2 Adjectives scored in the assessment of gels by the 10 panelists Panelist 1 Flavor Sweet Tart Fruity Off avor Sweet Cherry Sour Aftertaste Acid Fruity Sweet Stale/Clean Cherry Staleness Sweet Tart Acidity/Sourness Fruity Sweetness Texture Soft/Firm Chewy Solubility Sticky Hardness Pasty Size of particles Sticky Time to break into particles Soft/Tough Brittle/Chewy Smooth/Rough Smooth Bite/Tough bite Creamy/Clumpy
remained intact, inferring that gelatin forms a continuous network within the pectin-rich phase. The interface between the dispersed gelatin-rich phase and the continuous pectinrich phase looks almost continuous as shown in Fig. 6. Stronger binding between the ller and matrix has been shown to increase the stress and strain at fracture (Brownsey, Ellis, Ridout, & Ring, 1987). The pectin-rich phase forms the continuous phase even at low pectin concentrations probably because it attracts more solvent and gels at a higher temperature. Because the continuous phase bears most of the load during deformation, it contributes most to the mixed gel properties (Clark, Richardson, Robinson, Ross-Murphy, & Weaver, 1982; Walkenstrom & Hermasson, 1994; Ziegler & Rizvi, 1989). Therefore, the interpenetrating polymer network within the pectin-rich phase can help explain the complicated rheological fracture properties. Both the brittle pectin network and the more elastic gelatin network contribute to the mixed gel properties. 3.3. Sensory analysis Knowledge of structure and rheology is most useful when it can be related to human sensory perception. As per the method of Free Choice Proling (FCP), the 10 panelists developed their own list of descriptors (Table 2). One assumption in FCP is that panelists experience common underlying physical phenomena regardless of the individualistic words used to describe such. Generalized procrustes analysis is then used to statistically derive the commonalties in the form of principal components. For texture, the rst two principal axes accounted for 82.5% of the variance in the transformed data. Table 3 shows the loading values for attributes of individual panelists and Fig. 7 displays the consensus plot and general interpretation of the rst principal axis. The addition of 1.0% pectin to all three concentrations of gelatin resulted in gels described variously as softer, more brittle, easier to break apart, having smaller particles, and faster solubility than pure gelatin gels, as shown by the lower rst principal axis scores. The texture of the 3.0:1.0 and 3.0:1.5 gels was not signicantly different (P . 0.05). One interpretation of this failure to observe a signicant difference is that the rmness and brittleness of these two gels were indistinguishable, but that the pulpy, pasty and grainy descriptors, that some of the panelists used, referred to the large gelatin agglomerates in the 3.0:1.0 gel. The addition of 1.0% pectin softened gelatin gels at 4.5 and 6.0% so that they were indistinguishable from 3.0 and 4.5% gelatin alone on the rst principal axes. However, the second principal axis was effective in distinguishing 3.0:0.0 from 4.5:1.0 gels and 4.5:0.0 from 6.0:1.0 gels. For avor, the rst two principal axes accounted for 58.2% of the variance in the transformed data. Table 4 and Fig. 8 display the loading values and consensus plot for avor. The rst principal axis effectively differentiated
Hardness (Firmness) Chewiness (Rubbery) Smoothness Stickiness Clean Soft/Hard Sticky Elastic Soft/Firm Gummy Crumbly Smooth/Grainy Force for rst fracture Looseness Soft Consistency Resistance Rubbery Pulpy Solubility Toughness Rubbery Mouth coating
Sour Sweet Fruity Sweet Sour Fruity Sweet Cherry Tart
Sweetness Fruitiness Tartness Chemical Sour Cherry Sweetness
10
gels based on the gelatin:pectin ratio, with those having relatively more pectin described as more fruity, more sweet and more tart, i.e. generally more intense avor overall. Plain gelatin gels were differentiated along the second principal axis primarily by the difference in sweetness intensity. Sweetness intensity was inversely proportional to gelatin concentration. The rst principal component values for sensory texture correlated with strain at fracture (r 0.90), log (stress at
L.L. DeMars, G.R. Ziegler / Food Hydrocolloids 15 (2001) 643653 Table 3 Denition of rst two principal axes for texture in terms of panelists' original descriptions Panelist 1 2 3 4 5 6 7 8 9 10
a
651
Principal axis 1 Firmness (0.66) 1 Chewy (0.69) Hardness (0.47) 2 Pasty (0.56) 1 Size of Particles (0.50) 1 Time to Break (0.46) Soft/Tough (0.67) 1 Brittle/Chewy (0.58) 1 Smooth/Rough (0.46) Smooth (0.72) 1 Creamy (0.70) Hardness (0.63) 1 Rubbery (0.67) 1 Smooth (0.32) Soft/Hard (0.61) 2 Sticky (0.38) 1 Elastic (0.69) Soft/Firm (0.54) 2 Gummy (0.59) 1 Crumbly(0.59) 2 First Fracture (0.50) 2 Looseness (0.50) 1 Soft (0.48) 1 Resistance (0.48) Rubbery (0.62) 2 Pulpy (0.54) 2 Solubility (0.57) Tough (0.64) 1 Rubbery (0.57) 2 Mouth Coating (0.52)
a
Principal axis 2 2 Solubility (0.89) 2 Chewy (0.37) Hardness (0.63) 1 Pasty (0.45 1 Sticky (0.57) Soft/Tough (0.72) 2 Brittle/Chewy (0.35) 2 Rough/Smooth (0.60) 2 Smooth (0.70) 1 Creamy (0.72) 2 Smooth (0.91) 1 Hardness (0.37) Sticky (0.89) 1 Soft/Hard (0.44) Smooth (0.85) 1 Soft/Firm (0.47) Soft (0.46) 2 Consistency (0.60) 2 Resistance (0.59) Rubbery (0.52) 1 Pulpy (0.83) 2 Solubility (0.21) Mouth Coating (0.82) 1 Tough (0.55)
Figures in parentheses refer to vector loading for individual assessors.
Table 4 Denition of rst two principal axes for avor in terms of panelists' original descriptions Panelist 1 2 3 4 5 6 7 8 9 10
a
Principal axis 1 Sweet (0.51) a 1 Fruity(0.84) Sweet (0.45) 1 Cherry (0.56) 1 Sour (0.44) 1Aftertaste (0.54) Stale/Clean (0.82) 1 Acid (0.48) Cherry (0.76) 1 Tart (0.63) Acidity (0.56) 1 Fruity (0.59) 1 Sweet (0.58) Sour (0.69) 1 Sweet (0.66) Sour (0.68) 1 Fruity (0.74) Sweet (0.67) 1 Tart (0.64) Sweet (0.50) 1 Fruitiness (0.83) Sour (0.91) 1 Cherry (0.39)
Principal axis 2 2 Tart (0.91) 2 Off Flavor (0.32) 2 Aftertaste (0.84) 2 Acid (0.44) 2 Stale (0.70) 2 Sweet (0.61) 2 Sweet (0.81) 2 Fruity (0.86) 2 Sweet (0.5) 2 Sweet (0.86) 2 Sweet (0.74) 1 Cherry (0.46) 1 Tart (0.5) 2 Sweet (0.82) 1 Tartness (0.44) Cherry (0.77) 2 Articial (0.58)
Figures in parentheses refer to vector loading for individual assessors.
Fig. 7. Consensus plot for gel texture attributes.
652
Fig. 8. Consensus plot for gel avor attributes.
fracture) (r 0.87), and the rst principal component values for avor (r 0.83). As gels were described as more `rm', `hard', `rubbery', or `tough', they were also described as less `fruity', `sweet', and `tart'. The increased avor perception at higher pectin concentrations could potentially be accounted for by the difference in texture or by the inherent fruity avor of pectin. An inverse relationship between gel hardness and overall gel avor has been previously reported (Clark, 1990; Jaime et al., 1993). Morris (1989) suggested that the faster the avor/taste particles can be replaced on the surface of the tongue, the greater the avor perception. 4. Conclusions The addition of pectin to a gelatin gummy gel reduces the strain at fracture and inuences the stress at fracture depending on the pectin and gelatin concentration. Sensory evaluation revealed that gels with pectin were more brittle, less chewy, more smooth, and easier to break up into smaller particles. Pectin also increases the fruity, sweet, and tart avor of the gels. The instrumental technique of fracture through tension was successful for measuring the broad range of gel textures prepared. The stress at fracture and strain at fracture measurements had good correlation with sensory rmness and sensory elasticity or brittleness, respectively. Structure evaluation by microscopy helped in understanding the rheological and sensory properties. The pectin-rich phase formed the continuous phase for all of the mixed gels except perhaps 6.0:0.5. At low pectin concentrations, both the weak pectin network and dilute gelatin network in the pectin-rich phase contributed to the fracture properties of the gels. At high total polymer and pectin concentrations, phase separation and coalescence was incomplete probably because of the increased phase viscosity and gelation rate. Free choice proling was an effective sensory method that required less time for training and testing than conventional descriptive analysis. Although the results were
general and subject to the interpretation of analyst, they provided a range of scaled descriptors consumers might use to describe and differentiate between gummy products. Furthermore, FCP identied descriptors such as `pulpy' that might not otherwise have been scaled. This suggests the possibility of using mixed gel systems to simulate fruitlike textures. References
Association of Ofcial Analytical Chemists (1980). Ofcial methods of analysis, (13th ed). Arlington, VA: A.O.A.C. Bagley, E. B. (1983). Large deformation in testing and processing of food materials. In M. Peleg & E. B. Bagley, Physical properties of foods (pp. 325342). Westport, CT: AVI Publishing Co. Behram, E. J. (1985). The chemistry of osmium tetroxide xation. In: J. P. Revel, T. Barnard, G. H. Haggis, S. A. Bhatt (Eds.), Science of Biological Specimen Preparation for Microscopy and Microanalysis: Proceedings of the 2nd Pfefferkorn Conference, Sugar Loaf Mountain Resort, Traverse City, MI (pp. 15). AMF O'Hare, IL: Scanning Electron Microscopy, Inc. BeMiller, J. N. (1996). Gums/hydrocolloids: analytical aspects. In A. -C. Eliasson, Carbohydrates in food (pp. 265281). New York: Marcel Dekker. Brownsey, G. J., Ellis, H. S., Ridout, M. J., & Ring, S. G. (1987). Elasticity and failure in composite gels. J. Rheology, 31 (8), 635649. Clark, A. H., Richardson, R. K., Robinson, G., Ross-Murphy, S. B., & Weaver, A. C. (1982). Structure and mechanical properties of agar/ BSA co-gels. Prog. Fd. Nutr. Sci., 6, 149160. Clark, A. H., Richardson, R. K., Ross-Murphy, S. B., & Stubbs, J. M. (1983). Structural and mechanical properties of agar/gelatin co-gels. Small deformation studies. Macromolecules, 16, 13671374. Clark, R. C. (1990). Flavor and texture factors in model gels systems. In A. Turner, Food technology international Europe 1990 (pp. 271277). London: Sterling Publications International, Ltd. Clewlow, A. C., Rowe, A. J., & Tombs, M. P. (1995a). Pectin-gelatin phase separation: the inuence of polydispersity. In S. E. Harding, S. E. Hill & J. R. Mitchell, Biopolymer mixtures (pp. 173191). Nottingham, UK: Nottingham University Press. Clewlow, A. C., Clark, A. H., Rowe, A. J., & Tombs, M. P. (1995b). Predicting the phase diagram for a gelatin-pectin system. Biochemical Society Transactions, 23, 498S. Dickinson, E. (1995). Simulated structure of mixed particle gels. J. Chem. Soc. Faraday Trans., 91, 5157.
L.L. DeMars, G.R. Ziegler / Food Hydrocolloids 15 (2001) 643653 Dickinson, E., & McClements, D. J. (1995). Advances in food colloids, New York: Blackie Academic & Professional. Flint, O. (1994). Food microscopy: A manual of practical methods using optical microscopy, Oxford, UK: BIOS Scientic Publishers. Gower, J. C. (1975). Generalized procrustes analysis. Pscyometrika, 40 (1), 3351. Hamann, D. D., & Foegeding, E. A. (1994). Analysis of torsion, compression, and tension for testing food gel fracture properties. In: Rheology of Food Biopolymer Gels: a Practical Approach. Research Park Triangle, NC: Department of Food Science North Carolina State University. Hayat, M. A. (1989). Principles and techniques of electron microscopy biological applications, Boca Raton, FL: CRC Press, Inc. Ipsen, R. (1995). Mixed gels made from protein and kappa-carrageenan. Carbohydrate Polym., 28 (4), 337339. Jaime, I., Mela, D. J., & Bratchell, N. (1993). A study of texture-avor interactions using free-choice proling. J. Sensory Stud., 8, 177188. Kalab, M., Allan-Wojtas, P., & Miller, S. S. (1995). Microscopy and other imaging techniques in food structure analysis. Trends in Food Science and Technology, 6, 177186. Kasapis, S., Morris, E. R., Norton, I. T., & Clark, A. H. (1993). Phases equilibria and gelation in gelatin/maltodextrin systemsPart IV: composition-dependence of mixed-gel moduli. Carbohydrate Polym., 21, 269276. Kertesz, Z. I. (1951). The pectin substances, New York: Interscience Publishers, Inc. Kinsella, J. E., Rector, D. J., & Phillips, L. G. (1994). Physicochemical properties of proteins: texturization via gelation, glass and lm formation. In R. Y. Yada, R. L. Jackman & J. L. Smith, Protein structurefunction relationships in foods (pp. 121). New York: Blackie Academic & Professional. Langley, K. R., Green, M. L., & Brooker, B. E. (1991). Mechanical properties and structure of model composite foods. In E. Dickinson, Food polymers, gels, and colloids (pp. 383403). Norwich: The Royal Society of Chemistry. n, E. Paradao Lillford, P. J. (2000). Food composites. In J. E. Lozano, C. An novas, Trends in food engineering (pp. 65 Arias & G. V. Barbosa-Ca 75). Lancaster, PA: Technomic Publishing Company. Lin, M. J., & Humbert, E. S. (1978). Gelation characteristics of acidprecipitated pectin from sunower heads. Can. Inst. Food Sci. Technol. J., 11 (3), 113116. McEvoy, H., Ross-Murphy, S. B., & Clark, A. H. (1985a). Large deformation and ultimate properties of biopolymer gels: 1. Single biopolymer component systems. Polymer, 26, 14831492. McEvoy, H., Ross-Murphy, S. B., & Clark, A. H. (1985b). Large deformation and ultimate properties of biopolymer gels: 2. Mixed gel systems. Polymer, 26, 14931500. Meyers, F. S., & Wenrick, J. D. (1974). A comparison of tensile stress strain data from dumbbell, ring and oval specimens. Rubber Chem. Technol., 47, 12131233. Morris, E. R. (1989). Polysaccharide solution properties: origin, rheological characterization, and implications for food systems. In J. BeMiller & R. C. Karan, Frontiers in carbohydrate research. I. Food applications (pp. 132163). New York: Elsevier Applied Science. Oliemans Punter & Partners (1991). Procrustes-PC v.2.2, Utrecht, Netherlands: OP&P Software Development. Oreskovich, D. C., Klein, B. P., & Sutherland, J. W. (1990). Procrustes
653
analysis and its applications to free-choice and other sensory proling. In H. T. Lawless & B. P. Klein, Sensory science theory & applications in foods (pp. 353393). New York: Marcel Decker, Inc. Owen, A. J., & Jones, R. A. L. (1998). Rheology of a simultaneously phase-separating and gelling biopolymer mixture. Macromolecules, 31, 73367339. Poppe, J. (1995). New approaches to gelling agents in confectionery. In: Editorial Staff of The Manufacturing Confectioner (Eds.), Proceedings of the 49th Annual Production Conference, Hershey, PA (pp. 6878). Center Valley, PA: Pennsylvania Manufacturing Confectioners' Association. Reeve, R. M. (1959). A specic hydroxylamine-ferric chloride reaction for histochemical localization of pectin. Stain Technology, 31, 209211. Rodgers, W. A. (1973). Tension tests with cut ring specimens. Rubber Chem. Technol., 46, 586. Ross-Murphy, B. S. (Ed.). (1994). Physical techniques for the study of food biopolymers. New York: Blackie Academic & Professional. SAS Institute, Inc. (1985). User's guide: Statistics, Cary, NC: SAS Institute, Inc. Sperling, L. H. (1992). Fig. 4.14, p.143. Introduction to physical polymer science, (2nd ed). New York: John Wiley & Sons. Stading, M. (1993). Rheological behavior of biopolymer gels in relation to structure, Goteborg: Chalmers Tekniska Hogskola. Stein, R. University of Massachusetts, personal communication to Ralph Colby. Tolstoguzov, V. B. (1988). Some physico-chemical spects of protein processing into foodstuffs. Food Hydrocolloids, 2 (5), 339370. Tolstoguzov, V. B. (1990). Interactions of gelatin with polysaccharides. In G. O. Phillips, Gums and stabilizers for the food industry (pp. 157175), Vol. 5. London: Elsevier Applied Science Publishers. Tolstoguzov, V. B. (1991). Functional properties of food proteins and role of protein-polysaccharide interaction. Food Hydrocolloids, 4 (6), 429468. Tschoegel, N. W., Rinde, J. A., & Smith, T. L. (1970). Rheological properties of wheat our doughs IMethod for determining the large deformation and rupture properties in simple tension. J. Sci. Fd. Agric., 21 (2), 6570. Tschumak, J. A., Wajnermann, E. S., & Tolstoguzov, V. B. (1976). The structure and properties of complex gels of gelatin and pectin. Die Nahrung, 20, 321328. van Vliet, T., Luyten, H., & Walstra, P. (1991). Fracture and yielding of gels. In E. Dickinson, Food polymers, gels, and colloids (pp. 392403). Norwich: The Royal Society of Chemistry. Walkenstrom, P., & Hermansson, A. M. (1994). Mixed gels of ne-stranded and particulate networks of gelatin and whey proteins. Food Hydrocolloids, 8 (6), 589607. Williams, A. A., & Langron, S. P. (1984). The use of free-choice proling for the evaluation of commercial ports. J. Sci. Food Agric., 35, 558568. Zhuravskaya, N. A., Kiknadze, E. V., Antonov, Yu. A., & Tolstoguzov, V. B. (1986). Concentration of proteins as a result of the phase separation of water-protein-polysaccharide systems. Part 2. concentration of milk proteins. Die Nahrung, 30, 601613. Ziegler, G. R., & Rizvi, S. S. H. (1989). Predicting the dynamic elastic modulus of mixed gelatin-egg white gels. J. Food Sci., 54 (2), 430436.
Food Hydrocolloids 14 (2000) 579590 www.elsevier.com/locate/foodhyd
Quantitative assessment of phase composition and morphology of twophase gelatinpectin gels using uorescence microscopy
T.S. Nordmark, G.R. Ziegler*
Department of Food Science, The Pennsylvania State University, 116 Borland Laboratory, University Park, PA 16802 USA Received 14 January 2000; revised 2 June 2000; accepted 2 June 2000
Abstract A technique for quantitative determination of the concentrations of polysaccharide and protein in two-phase mixtures by uorometry has been developed and compared with chemical analysis. In the rst case, a general method for uorescent labeling of carbohydrate polymers was developed. For the latter purpose, two micro-assays were developed on the basis of recent polymer macro-assays. A blend of lowmethoxyl pectin and gelatin B was used as a model system. The commercial components were subjected to multi-step purication procedures, and phase separation was initiated by the addition of NaCl to aqueous solutions containing the two polymers. Samples were withdrawn for microscopy after various holding times at 60C. Tie-lines were determined using both the uorescent and chemical methods. The results from these methods were in fair agreement with each other and with literature data. A three-phase region was discovered in the pseudo-ternary phase diagram. The morphology of double labeled gels was also studied in two and three dimensions using confocal scanning laser microscopy. The results show promise for the quantitative assessment of phases that contain carbohydrate polymers and in the study of morphological changes that occur during thermo-mechanical processing. 2000 Elsevier Science Ltd. All rights reserved.
Keywords: Gum; Low-methoxyl pectin; Gelatin; Fluorescence; Morphology; Phase diagram
1. Introduction Structure-forming polysaccharides and proteins provide desired functional properties to a wide range of foods (Kinsella, Rector, & Phillips, 1994). The characteristics of blends of such hydrocolloids both in the liquid and gel states have lately been the subject of an increasing number of investigations because of the prospect of discovering useful synergistic effects (Ipsen, 1995). This research has begun to demonstrate how physical properties of blends can be related to phase morphology and quantitative relationships can be established (Owen & Jones, 1998). A number of different analytical approaches have been employed to elucidate structurefunction relationships (BeMiller, 1996; Dickinson & McClements, 1995; Kalab, Allan-Wojtas, & Miller, 1995; Ross-Murphy, 1994). Direct determination of phase-composition is an obvious major target for future research on biopolymer co-gels (Kasapis, Morris, Norton, & Clark, 1993b). The present work was undertaken to expand the use of uorescence-based analytical methods to research on phaseseparated food materials and the relation of morphological
* Corresponding author. Tel.: 1-814-863-2960; fax: 1-814-863-6132. E-mail address: grz1@psu.edu (G.R. Ziegler).
and compositional features to rheological properties. Fluorescence and uorescence microscopy in biology are expounded upon by Ichinose, Schwedt, Schnepel, and Adachi (1991), Ploem (1993), and Rost (1991). Applications of uorescence in food research have been recently reviewed (Blonk & van Aalst, 1993; Strasburg & Ludescher, 1995; Vodovotz, Vittadini, Coupland, McClements, & Chinachoti, 1996) and the properties of gelatin pectin gels have been the focus of research (Al-Ruqaie, Kasapis, & Abeysekera, 1997; DeMars, 1995; Gubenkova, Somov, & Shenson, 1988). Gelatin is derived from denatured collagen that has been further processed, and the dominating amino acids are glycine, proline, and hydroxyproline. Thermoplastic gels are formed upon cooling, and a blend of ne and coarse networks can be found (Ziegler & Foegeding, 1990). Gelatins of type B have their isoelectric points close to pH 4.9 and below this may form complex coacervates with negatively charged polysaccharides. Pectic substances are linear, partly methylesteried polygalacturonic acid chains, where neutral sugars like rhamnose may be present as side chains or inserted in the main chains (da Silva & Goncalves, 1994). Both smooth and hairy chains, the latter with side chains of arabinogalactan or other oligosaccharides, may exist (Aman & Westerlund, 1996). The galacturonic acid residues
0268-005X/00/$ - see front matter 2000 Elsevier Science Ltd. All rights reserved. PII: S0268-005 X( 00)00 037-0
580
T.S. Nordmark, G.R. Ziegler / Food Hydrocolloids 14 (2000) 579590
contain vicinal diols, which we employ in the protocol for uorescent labeling. There are fewer alternatives for the uorescent labeling of polysaccharides than for the labeling of proteins, in particular for general and quantitative purposes. Food proteins are frequently labeled with FITC (uorescein isothiocyanate), Rhodamine, or Texas Red (Blonk & van Aalst, 1993). Traditional dyes such as Calcouor White and Anilin Blue often attach to specic carbohydrate residues or linkages (Fulcher, Faubion, Ruan, & Miller, 1994). The lectins from pea tree (Caragana arborescens) have been shown to bind agarose beads and can be marked with FITC (EY Laboratories, 1990). Polysaccharide side chains can be made more reactive by using transferases (Brossmer & Gross, 1994; Gahmberg & Tolvanen, 1994) or by using galactose oxidase (EC 1.1.3.9), which predominantly, but not always exclusively, acts upon terminal, non-reducing dgalactose residues (Goudsmit, Matsuura, & Blake, 1984; Mazur, 1991; Wilchek & Bayer, 1987). In this paper we present a protocol for covalent labeling of pectin and many other carbohydrate polymers with the uorescent probe BODIPY FL hydrazide. This probe has recently been used for the quantication of progesterone and other 3-keto steroids by HPLC (Katayama et al., 1998). We labeled gelatin covalently with the succinimidylester of carboxytetramethylrhodamine by slightly modifying an existing protocol for labeling of globular proteins that contain aliphatic amines. This probe has lately been conjugated with peptides for inclusion and detection in degradable poly(lactic acid) (PLA) microspheres (Brunner, Minamitake, & Gopferich, 1998). The phase behavior of mixed polysaccharides and proteins has been investigated by employing centrifugation of the phases, chemical analyses, osmotic pressure measurements, light scattering, FTIR, and turbidimetry (Antonov, Lashko, Glotova, Malovikova, & Markovich, 1996; Clewlow, Rowe, & Tombs, 1995b; Durrani, Prystupa, Donald, & Clark, 1993; Vinches, Parker, & Reed, 1997). Improvements in most of these techniques cannot compensate for the difculties that arise due to increased viscosity in more concentrated mixes. Accordingly, there is an interest in the development of methods that could be used in situ, e.g. Blonk, van Eendenburg, Koning, Weisenborn, and Winkel (1995) discussed the use of confocal scanning laser microscopy in combination with uorescent double-labeling of alginate and caseinate (max. 2 and 10%, respectively) in the liquid state. The objective of this work was the development of a method for the in situ measurement of phase composition and morphology in highly viscous biopolymer mixtures.
remove electrolytes and particles. Commercial low-methoxyl citrus pectin (LM290 NA95 from SKW Biosystems, Inc., Waukesha, WI) with 31.9% degree of esterication and gelatin type B (Sigma Chemical Co., St. Louis, MO) made from bovine skin tissue were used as polymeric raw materials. Polymers were puried from cations and sugars using the following procedure. The pectin and gelatin were dispersed in cold 25 and 40% aqueous ethanol, respectively, dialtered 4 times, and treated batch-wise with stirred AG50W-X8 (20/50 mesh) cation exchange resin (Bio-Rad Laboratories, Hercules, CA). The desalted pectin and gelatin dispersions were decanted, the resin was washed with aqueous ethanol until clean of polymer, and the used wash liquid was added to the polymer dispersions. The dispersions were then ltered through a 50 mm fritted glass lter funnel, dialtered with 100 ml aq. ethanol (40 and 25% concentration for gelatin and pectin, respectively), titrated to pH 5.5 (using 100 mM hydrochloric acid or sodium hydroxide), slowly precipitated with ethanol, freeze-dried for 48 h, ground in an analytical mill, and stored in a dessicator at 20C (Berth, 1988; Doner & Douds, 1995; Walter & Sherman, 1983). The uorescent probes D-2371 (BODIPY FL hydrazide), C-1171 (TAMRA succinimidylester), and T6105 were purchased from Molecular Probes, Inc. (Eugene, OR). Supor-450 polysulfone membrane lters (0.45 mm) from Gelman Sciences (Ann Arbor, MI) were used when ltering the polymer solutions. Warm, gas tight syringes (Hamilton Co., Reno, NV) were utilized for the quantitative transfer of high viscosity polymer solutions at 50C. Disposable 10 DG polyacrylamide size exclusion chromatography (SEC) columns (Bio-Rad Laboratories, Hercules, CA) and Centricon-10 centrifugal concentrators (Amicon, Inc., Beverly, MA) were used in the protocols for uorescent labeling. An LSM 410 inverted Laser Scan Microscope (Carl Zeiss, Inc., Thornwood, NY) with argon- and helium neon lasers and photomultiplier tube detectors was employed for confocal uorescence microscopy. A Nikon Diaphot 300 inverted uorescence microscope (Nikon Inc., Melville, NY), a 75 W xenon lamp, and a liquid cooled CCD-camera of type CH 250 (Photometrics Ltd., Tucson, AZ) were used for wide-angle uorescence microscopy. Cytoseal 60 (Stephens Scientic, Riverdale, NJ) was generally employed as sealant for gels mounted for microscopy. The image processing software was IPLab Spectrum HSU2, v. 2.5.7 (Signal Analytics Corporation, Vienna, VA). An IEC Model CL centrifuge (International Equipment Company, Needham Heights, MA) was employed in the reference experiments of the quantitative study.
3. Methods 2. Materials The water was treated in a NANOpure water purication system (Barnstead/Thermolyne, Inc., Dubuque, IO) to Puried pectin and gelatin were covalently labeled with uorescent probes as described below. The concentration of the uorescent solution was estimated using

Disperse 30 mg purified LM pectin in 2 ml cold deionized water while using a stir bar at high speed and continue stirring at moderate speed for 1 hour Boil for 30 seconds Cool to 60 C and filter 1 ml Cool to 20 C Add 0.5 ml periodate solution while using a stir bar at moderate speed and continue stirring at low speed for 30 minutes in darkness Add 50 L of 2 M aq. glycerol while stirring and wait for 5 minutes Filter through a 6 kD SEC column and elute with 10 mM PBS buffer pH 7.4 Concentrate to 1.2 ml by centrifugation at 5,000 g Add 50 L solvent, stir, and slowly add 33 L of fluorescent solution Stir at low speed for 3 hours in darkness and overnight in a refrigerator Add 1-2 mg sodium cyanoborohydride and stir at low speed for 40 minutes Filter twice through 6 kD SEC columns and elute with deionized water Concentrate and deaerate to 1.1 ml by centrifugation at 5,000 g Heat to 60 C; evaporate to 1.0 ml; solution may be stored in a refrigerator
581
3.1. Pectin Pectin was uorescently labeled with the non-ionic dye BODIPY FL (4,4-diuoro-5,7-dimethyl-4-bora-3a,4adiaza-s-indacene-3-propionylhydrazide), which is more photostable than uorescein and has high extinction coefcient and quantum yield (Haugland, 1996). Its emission spectrum is reasonably distant from the excitation spectrum of the TAMRA dye, which was used for labeling of gelatin. The hydrazide group forms conjugates with ketogroups (Hermanson, 1996). A uorochrome concentration of 250 mg/ml in the reaction mixture was chosen (in similarity with the work by Katayma et al., 1998).Reagent solutions. Periodate solution: Take 0.5 ml of a 10 mM stock solution of sodium metaperiodate, and add 1.75 ml deionized water and 65 mg PBS (phosphate buffered saline)-powder as buffer salt (Sigma Chemical Co., St. Louis, MO). Fluorescent solution: Dissolve 5 mg uorescent probe (D-2371) in 500 ml methanol (Omnisolv from EM Industries, Inc., Gibbstown, NJ) by vortexing. Dimethylformamide (DMF) (ACS Reagent from Sigma Chemical Co., St. Louis, MO) may be employed as solvent if the probe solution will be used within one day. This solvent was used in the qualitative study. Fluorescent labeling: See Fig. 1.
Fig. 1. Fluorescent labeling of LM pectin.
3.2. Gelatin Gelatin was uorescently labeled with mixed isomers of TAMRA SE (5- and 6-carboxytetramethylrhodamine succinimidylester), which are among the most photostable uorescent dyes available and emit uorescent light of high intensity. A protein conjugate made from a succinimidyl ester of TAMRA is considered to be more chemically stable than a conjugate made from the isothiocyanate derivatives commonly used (Haugland, 1996). The overlap between its excitation spectrum and the emission spectrum of the BODIPY FL dye is reasonably small. The CSLM heliumneon laser delivers light at 543 nm, which closely matches the 546 nm excitation wavelength of the TAMRA SE. It may form non-uorescent dimers when attached to proteins and is more prone to degrade in moist environments (liquid or solid state) than the BODIPY probe is (Molecular Probes, Inc. Eugene, OR). The protocol below was followed.Reagent solutions. Fluorescent solution: Make a solution of 10 mg probe C-1171/ml DMF in a conical microcentrifuge tube and vortex it until the probe is dissolved. Hydroxylamine solution: Dissolve 420 mg dry fresh hydroxylamine hydrochloride in 2 ml deionized water and carefully adjust the pH to 8.5 while stirring. Dilute the solution with water so that the nal concentration will be 1.5 M. Fluorescent labeling: See Fig. 2. All treatments of the gelatin solution were carried out in a 40C environment in order to avoid gelation.
spectrophotometry before the start of each labeling procedure. Puried and labeled polymer was characterized using the micro-assays described below.
Disperse 40 mg purified gelatin in 2 ml cold deionized water while using a stir bar at moderate speed and continue stirring at low speed for 1 hour Boil for 30 seconds Cool to 60 - 70 C and filter 1.3 ml Cool to 40 C Slowly add 130 L of fresh dissolved fluorescent probe while using a stir bar and continue stirring the covered solution at low speed for 1.5 hours Add 130 L fresh hydroxylamine solution and stir the covered mix at low speed for 1 hour Filter through a 6 kD SEC column and elute with PBS buffer pH 7.4 Filter through a 6 kD SEC column and elute with deionized water Concentrate and deaerate to 1.1 ml by centrifugation at 5,000 g Heat to 60 C and evaporate to 1.0 ml; the covered solution may be stored in a refrigerator
Fig. 2. Fluorescent labeling of gelatin.
582
Disperse purified LM pectin and purified gelatin separately in deionized water using a stir bar at high and moderate speed, respectively. Continue to stir at moderate speed for 1 hour Boil each dispersion for 30 seconds and filter through a 0.45 m membrane Cool each dispersion to 40 - 50 C Mix the dispersions at 40 C Concentrate to approximately final polymer concentrations or slightly less by evaporation at 70 C Adjust to pH 5.5 if required Stir and add a warm 20 % (w/w) solution of sodium chloride Boil for 30 second Evaporate at 60 C until final concentrations (including 1 M NaCl) are obtained Pour into a pre-weighed conical PP centrifuge tube and weigh again Keep the capped tube in a 60 C water bath for 3 hours minus the evaporation time in the previous step Spin at 1,100 g and 60 C for 10 minutes in a preheated centrifuge Immediately cool the tubes briefly in ice-water and refrigerate the tubes for 30 minutes Cleave the tube transversely with a hot knife; weigh the gel containing part Cleave this part longitudinally; place each phase in a separate, pre-weighed graduated tube and weigh the covered tubes and the original tube halves Dilute each phase with 60 C water as required for the polymer micro-assays and remove duplicate samples Analyze the remnants for moisture content by oven drying at 100 C for 2 hours
glasses (thickness #1, rectangular bottom 35 50 mm2 ; circular top 25 mm). Standards and mixed gels were prepared under red light to prevent photobleaching. The sample size was 25 ml. The mix compositions (unlabeled pectin and gelatin, respectively, per 2.5 ml of 1 M aq. sodium chloride) employed for the construction of quasi-ternary phase diagram from chemical assay were (mix a:) 0.75%/5.3%, (mix b:) 1.25%/5.7%, (mix c:) 1.75%/6%, and (mix d:) 4.12%/4.6%. 3.4. Centrifuged gel mixes Procedure when making gels for chemical assay: See Fig. 3. 3.5. Micro-assays The ISO hydroxyproline assay (ISO, 1994) based on the approach by Stegemann and Stalder (1967) was modied for the analysis of microgram quantities of gelatin as trans4-hydroxy-l-proline. Modications included the withdrawal of only 10 ml sample solution, hydrolysis for 24 h in a bath, elimination of the ltration step, and neutralization of the added acid. A potential problem was that the absorption peaks of the assay chromogen and the uorochrome are very close (555 and 562 nm, respectively). Separate experiments showed that the uorochrome is chemically destroyed under assay conditions before the absorption is recorded and, thus, no interference takes place. While paying attention to the stability of the reagents, we found our assay well reproducible and relied on one calibration graph during all experiments. The coefcient of variation for any determination of gelatin samples was 1.7%. We have also successfully used this assay for the assessment of collagen in egg shell membranes. The micro-assay, which was linear up to 600 mg gelatin, should, with a changed dilution factor, permit a determination of less than 1 mg gelatin. The method by Scott (1979) for determination of polygalacturonic acid was adopted by AOAC in 1995 as a part of the Ofcial First Action analysis of total dietary ber (Theander, Aman, Westerlund, Andersson, & Pettersson, 1995). We modied this assay to allow a quantitation limit of 30 mg LM pectin as d-galacturonic acid residues. Modications included the withdrawal of only 20 ml sample solution, use of cold and relatively less acid, pre-heating of the acidied sample before hydrolysis for 70 min, and elimination of the ltration step. We found this micro-assay to be linear up to at least 900 mg pectin and well reproducible. The coefcient of variation for any determination of pectin samples was 1.3%. 4. Results and discussion The procedure for uorescent labeling of pectin described in Section 3 should function for all carbohydrate polymers that contain vicinal diols (cis or trans isomers) or when such
Fig. 3. The procedure for separation of phases by centrifugation.
3.3. Phase diagrams: making standards and mixed gels Duplicate standards for microscopy were made from aqueous dispersions of labeled polymer. Concentrations were determined using the micro-assays described below. Desired concentrations were obtained by evaporation in a convection oven at 5060C. Calibration points in the concentration ranges between 04% (pectin) and 010% (gelatin) were employed. The linear correlation between uorescence intensity and concentration was higher for pectin (average R2 0:982 than for gelatin (average R2 0:848: Phase-separated mixes of labeled materials for microscopy were made by adding an aqueous solution of sodium chloride to the stirred polymer mix above the gelation temperature of 30C. The desired nal concentrations (including 1 M sodium chloride) were obtained by evaporating the mix as described above. The mix was held covered at 60C for additional 0.53 h to allow the separation to proceed further. During the quantitative study, samples were placed between horizontal, parallel, pre-heated cover
583
Fig. 4. Phase separation in a gelatinLM pectin gel as shown by uorescent labeling. The sample is illuminated in an ordinary uorescence microscope with excitation light for the uorochrome attached to gelatin (left) and pectin (right). The size of each image is 67 70 mm2 :
groups can be introduced into the molecule. The rst step is the oxidation with periodate (Guthrie, 1962; Jackson, 1944), which has been employed for the oxidation of, for instance, corn- and potato-starch (Jackson, 1944). The periodic acid reaction has been used in combination with the Schiffs reagent, but histochemical and molecular modeling studies have shown that this combined reaction (called the PAS reaction) is not a quantitative test for polysaccharides (Puchtler, Meloan, & Brewton, 1974). We hypothesized that the use of a bright, less bulky reagent with a spacer and carefully controlled reaction conditions would allow quantitative results. The protocol discussed in this paper results in a 6-atom spacer being located between the uorochrome and the polysaccharide chain, whereby both steric constraint and transfer of light energy to the chain should be considerably reduced. Katayama et al. (1998) mention that hydrochloric and triuoroacetic acid have been used as catalysts for the conjugation of hydrazide and steroids in methanol or ethanol solution. However, the use of DMF followed by stabilization of the bonds by treatment with cyanoborohydride resulted in very bright labeling of the pectin without the employment of catalysts. The use of methanol, without catalyst, in the quantitative protocol resulted in a lower but acceptable extent of labeling for quantitative purposes. Commercial DMF was found to have an amine-containing contaminant that slowly reacted with the hydrazide-containing probe to form a dark brown compound. Thus, solutions of BODIPY in DMF should be made fresh for each experiment. Heat treatment in excess of 94C for 12 min. was avoided during the making of uorescent gels, since the BODIPY probe may otherwise not remain stable. The labeling of gelatin and the preparation of samples for microscopy was performed at elevated temperature to avoid gelation and lower the viscosity to facilitate phase separation and handling of the gelatin solutions. A raised temperature was also necessary for phase separation to occur since the phase diagram is inverted with a lower critical solution temperature above the gelation temperature of gelatin (Antonov et al., 1996; Tolstoguzov, 1990). Environmental temperatures in the range 5060C satised the liquefaction, phase separation, and viscosity requirements. Although
the solvent DMF tends to precipitate gelatin out of aqueous solutions, no such problems were observed. The use of only two SEC columns in the labeling protocol appeared adequate, since only a very small amount of uorescent probe remained after the rst column. Most of the TAMRA probe could not be eluted from the columns, implying that the probe reacted with the column packing. In addition, gelatin yields lower than 100% indicated that some gelatin was trapped in the column. The procedure for uorescent labeling of gelatin was, as an alternative, also carried out using a derivative (T-6105) of TAMRA SE with a 7-atom aminohexanoyl spacer between the uorophore and the reactive group. However, the spacer-equipped uorophore offered no advantages compared to the standard uorophore in terms of brightness, gelation, or stability. In a qualitative study, sodium chloride was added to an aqueous mix composed of 1.7% labeled pectin and 9.6% labeled gelatin so that its concentration was 2 M. The mix was held at 55C for 30 min. The fact that sodium chloride actually induced phase separation in a mix of homogeneously labeled gelatin and pectin was demonstrated by withdrawing samples for uorescence microscopy before and after the addition of sodium chloride. In the rst case, only diffuse light emerged from the sample and no contrasts were observed upon excitation of the BODIPY uorophore. In the second case, complementary light and dark regions were observed when the sample was illuminated with excitation light for each of the two uorophores. Thus, the light regions represented assemblies of labeled pectin and gelatin rich material, respectively. Both ordinary uorescence microscopy and CSLM revealed the presence of complementary regions of uorescence upon excitation of each of the two uorochromes with suitable light. Digital images of a double-labeled gel are shown in Fig. 4. Each image is the negative of the other, since the same area was illuminated with either blue or green light aimed for BODIPY and TAMRA, respectively. The displayed morphology should be the result of both an initial demixing and a subsequent mixing process, which was halted by gelation. Confocal microscopy was used to obtain a color representation of the distribution of uorescent labels (Fig. 5). The colors were separated and the intensity values of each color quantied. In the continuous phase the concentration of uorophore attached to pectin varied more than the concentration of uorophore attached to gelatin (Fig. 6). This variation, which was on a scale similar to the resolution limit of the instrument, suggests that individual uorophores were detected, since the labeling of the pectin was sparse. It may not be excluded that clusters of uorophores were present and detected as peaks. The relatively low apparent uorescence signal from the gelatin label in the pectin-rich phase may be due to the sigmoidal response curve of the photo-multiplier tube (PMT)-detector of the CSLM equipment. Differences in uorescence intensity might also be related to uneven illumination within the sample or to
584
Fig. 5. Phase separation in a gelatinLM pectin gel as viewed in the confocal scanning laser microscope. The thickness of the optical slice is about 0.5 mm. The image size is 70 70 mm2 : The colors represent: green: pectin label; red: gelatin label; yellow: intermediate composition.
uneven detector response. The CSLM technique permits an estimate of the maximum size of the dispersed particles when the slice thickness is small relative to the diameter of the particle. In the current case, dispersed phase diameters were 20 mm or (usually) less. Fig. 6 shows that the thickness of the interphase region can be estimated at 3 mm. The curvature of the phase boundary will introduce little error (1%) when the thickness to diameter ratio is 0.5:20 as in this case. A portion of the phase-separated mix from which the previous samples had been taken was centrifuged at
20,000g and 40C for 10 min and then immediately cooled to 5C. A small, buff colored phase at the bottom and a sizable, reddish phase of lower density were found in the centrifuge tube. The latter phase was the gelatin rich phase, since only the uorochrome attached to gelatin is red. According to the phase diagram (Fig. 7) and the results from uorescence microscopy, the gelatin-rich phase should occupy the largest volume as indeed it did. Previously, while using centrifugation, an unlabeled mix of the same composition had been shown to phase separate with a similar proportion between the phase volumes. Thus, no
585
A
240 220 Green Va lue 200 180 160 140 120 100 17 1 9 25 33 41 49 57 65 73 81 89 97 105
Distance
250 200 Red Va lue 150 100 50 0 1 9 17 25 33 41 49 57 65 73 81 89 97
Distance
Fig. 6. Fluorescence intensity of (A) BODIPY FL and (B) TAMRA versus distance across the interphase region in Fig. 5 (1 mm 7:5 distance units).
inuence of the labeling itself on the phase separation behavior was observed. Preceding a quantitative study, we determined that a 10% solution of gelatin B at pH 5.5 was not precipitated by the addition of 1 M sodium chloride. This was done because sodium chloride at a concentration of 2 M was found to precipitate gelatin in a 0.56% gelatin solution (Finch & Jobling, 1977). The quantitative study included centrifugation and uorometry, and, in both cases, micro-assays. The chemical analysis of the different phases in the four centrifuged mixes revealed a typical segregative (i.e. representing polymers of low thermodynamic compatibility) phase diagram (Fig. 7). This is expected, since at pH 5.5 both polymers are similarly charged polyelectrolytes, and the phase behavior resembles that of an aqueous mixture of non-ionic polymers. The analysis also disclosed the occurrence of a three-phase region at high polymer concentrations. The phase-separated region occupies most space in the diagram especially at low solvent concentrations, and this reects the high molecular weights of pectin and gelatin compared with the molecular sizes associated with the solvent. The apex of the binodal curve and the whole twophase area are located closer to the gelatin axis. This is consistent with a lower solvent compatibility (x prs) for gelatin than for LM pectin (x pss) and can be regarded as a Dx -effect (Grinberg & Tolstoguzov, 1997). The value of x prs for gelatin and water has been estimated to 0.46 (Ziegler, 1988; Ross-Murphy, 1995), and the x pss for pectin and water should be slightly lower. The asymmetry of the diagram could also be related to a higher molecular weight
AQUEOUS SOLVENT
105
Threshold point slightly above critical point
2 4
2 4 6 8
Bulk composition of mix Composition of phase
6 8 % GELATIN 10 12 14 16 18
a b
c
% PECTIN 10 12
C
14 16 18
Fig. 7. Ternary phase diagram of an aqueous gelatinLM pectin system with 1 M sodium chloride and at 60C obtained by using centrifugation and chemical assays.
586
of pectin than of gelatin. The tie-lines are slightly skewed towards the gelatin axis, and the critical point (the locus of the tie-lines) is separated from the phase separation threshold point. This is generally expected when xprs xpss ; implying a higher water-binding capacity of pectin than of gelatin. The location of the threshold point in a gelatin pectin mix was determined by Clewlow, Clark, Rowe, and Tombs. (1995a) to be 3.85% gelatin and 0.4% pectin. Our threshold point is located at approximately 4.15% gelatin and 0.9% pectin. The phase diagram is in other respects consistent with results presented by Clewlow et al. (1995a), who determined the phase behavior of a gelatin pectin mix at 80C, pH 5.5, and 0.5 M sodium chloride concentration. Addition of salt in excess of the amount required for phase separation is not expected to have inuenced the phase diagram signicantly, but the rate of separation is likely to increase when the concentration of salt is raised. This behavior with respect to salt has been observed on aqueous gelatinoligosaccharide mixtures (Vinches, Parker, & Reed, 1997). Some depolymerization may take place during the heat treatments of pectin and gelatin, but the effects on the diagram are likely to be small (Clewlow et al., 1995a). In this work, the pectin-rich phase was generally found to be the dispersed phase. This was likely due to use of LM instead of HM pectin and to higher gelatin concentration when compared with the work of DeMars (1995). LM pectin is reported to be less hydrophilic than HM pectin (Walter, 1991) and, thus, could have a higher x pss-value, vis-a-vis HM pectin. This would increase the symmetry in the binodal, reduce the width but raise the threshold point of the phase separated area, and likely move the point of phase inversion (Fig. 7). When Dx becomes smaller, the slope of the tie-lines are expected to decrease. It is reasonable to assume that a decreased hydrophilicity, vis-a-vis HM pectin, would make the LM pectin-rich phase less likely to form a continuous phase in aqueous medium. This would be consistent with a shorter distance between the point of phase inversion and the right branch of the binodal. Thus, substitution of LM pectin for HM pectin would make the gelatin-rich phase more likely to be continuous. If the bulk composition point is close to the point of phase inversion, a change in the relative positions of these points will easily take place after such a substitution. The location of the phase inversion point is known to vary considerably with the type of system. Gelatinagarose and gelatinmaltodextrin gels have phase inversion points at 0.6% agarose/4% gelatin (Tolstoguzov, 1995) and 15% maltodextrin/5% gelatin (Kasapis, Morris, Norton, & Brown, 1993a), respectively. It is reasonable to assume that this variation is related to the higher molecular weight of agarose compared to the molecular weight of maltodextrin. In practice, the location of the point and the solvent partitioning are kinetically controlled and therefore vary with the thermal treatment. For instance, both cooling rate and nal temperature inuence the location of the phase-inversion point in a
gelatinmaltodextrin system (Alevisopoulos, Kasapis, & Abeysekera, 1996). At approximately 60% sugar, DeMars (1995) found gelatin (4.5%) to be dispersed in HM-pectin (0.5%). However, at 6.0% gelatin and 0.5% pectin concentrations rheological analysis indicated that phase inversion had occurred and gelatin was the continuous phase. In the current case, and in some regions of the sample, none of the two phases seemed to have an apparent dominance. This is consistent with DeMars nding. Thus, both the current research and the work by DeMars support that a relatively high gelatin concentration (69%) is essential for gelatin to be the continuous phase when the pectin concentration is 0.51.5%. The potential for microuorometry to replace the centrifugal technique and determine the phase diagram in situ was explored by using confocal and ordinary wide-eld equipment. There were several reasons for this approach. The latter equipment is less expensive but may still allow attainment of equally good results. The presence of a pinhole in a confocal system limits the light gathering capacity. The contrast and, thus, the resolution that can be obtained in practice may therefore be affected by the chosen pinhole size. In fact, confocal microscopy, which can provide superior image quality in many instances, may produce results that are inferior to ordinary microscopy when sensitivity is a major concern (Stelzer, 1998). The presence of detectors with different sensitivities and ranges in the two systems further motivated a comparative study. In the rst case, a confocal scanning microscope with argon- and helium neon-lasers, and state-of-the-art PMT was used. Such equipment was also used by Kumar, Laird, Srikant, Escher and Patel (1997) in quantitative double-labeling experiments using FITC and rhodamine as uorescent dyes. In the second case, an inverted uorescence microscope and a liquid-cooled CCD-camera were employed. Methanol was used to solubilize the uorescent dye aimed for pectin because of poor stability over time of the dye dissolved in DMF. This change resulted in a more sparse derivatization of the pectin and lower uorescence intensity. However, this level of derivatization was sufcient because of the higher sensitivity of the CCD detector. Furthermore, a sparse derivatization of the polymers helped avoid self-quenching and possible effects on the phase separation behavior at high concentrations. The observed uorescence intensity was within the dynamic range of the CCD-camera. In order to avoid uorescence resonance energy transfer (FRET) in quantitative double-labeling experiments, only one of the polymers in a phase-separated sample was uorescently labeled. The effects of photobleaching and sample age on the uorescence intensity were considered. New concentration calibration data for the gelled, labeled polymer were recorded in each experiment. Previous attempts to obtain homogeneously distributed concentrations by evaporation of the polymer solutions on the microscopy slides had not been successful enough. The methodology outlined here allowed two tie-lines to be
587
Fig. 8. Phases of labeled pectin 50 min after initialized phase separation in a mix of 1.7% LM pectin, 4.25% gelatin B, and 1 M sodium chloride. The image size is 67 45 mm2 :
determined while employing phase separation conditions that closely resembled those previously used during the centrifugation of polymer mixes. The results from microuorometry are compared with the results based on centrifugation and chemical assays and with literature data. The mix compositions 1.7% LM pectin/4.25% gelatin B and 2.0% LM pectin/5.0% gelatin B were chosen for microscopy. Based on previous results, it was estimated that both these compositions would allow the gelatin rich phase to be the continuous phase. Therefore, all samples would be in the gel state during microscopy at ambient temperature. Pectin only was labeled in two experiments, in which the calibration gels contained between 0.4 and 4% pectin. Gelatin only was labeled in two other experiments, in which the calibration gels contained between 2 and 10% gelatin. In both cases, several samples with holding times varying from 0.5 to 3 h at the temperature of phase separation were studied. The dynamic range of the CCD camera response covered all uorescence data, although a few low intensities were located close to the non-linear part of the response curve. Photobleaching of the uorochromes occurred but
Fig. 9. Coalescence of unlabeled pectin phases 50 min after initialized phase separation in a mix of 1.7% LM pectin, 4.25% gelatin B, and 1 M sodium chloride. The image size is 67 85 mm2 :
at low rates especially in the case of TAMRA. Accordingly, no correction of intensity data was deemed necessary. Apart from photobleaching, there was an initial decline in uorescence followed by a signicant decrease over a period of several weeks despite that the gel state was maintained. Thus, all measurements of uorescence intensity were performed within 13 days after the initial decline. For any focal depth, the range of light collection was of a magnitude similar to or larger than the thickness of the gel, which was 610 mm. Thus, focal setting and minor moisture evaporation from the sample were less critical. Fluorescence intensities were preferentially recorded in the central region of the circular sample and from phases with dimensions at least as large as the gel thickness. Examples of gel morphology are shown in Figs. 810. An early stage of separation (i.e. a holding time of 50 min) of pectin-rich phases is shown in Fig. 8. Beginning coalescence of pectin-rich phases is shown as a negative image in Fig. 9, since only gelatin is labeled. Fig. 10 shows a later stage of separation and clustering of pectin rich material when only gelatin is labeled. It also displays that the composition of the gel is likely to be close to the point of phase inversion. Table 1 shows the results from the uorometric quantitation of pectin and gelatin concentrations in the phases appearing in the thin gels. Comparative data from the centrifugation and chemical assays of bulk mix (Fig. 7) are also displayed. Since it was not possible to expose (all parts of) the samples for centrifugation and uorometry to identical thermal histories, the results of the two methods are not necessarily identical. Fluorometry and centrifugation gave similar results in the case of labeled pectin. The agreement between the two methods in the case of labeled gelatin is also fair with the exception of the gelatin content of the dispersed phase in the rst mix. We believe that a slight non-linearity in the calibration curve could explain this deviation. Generally, the determination of pectin concentrations were associated with smaller statistical error than the determination of gelatin concentrations. Consistent
588
when the polymer concentration rises. Thus, there is likely a limit for how high concentrations one can determine, and this limit is expected to be different for each system of mixed food polymers. 5. Conclusions Both the centrifugation and uorometric methods are capable of yielding results of good accuracy when applied to gelled systems. The former method does not include chemical treatment of the polymers, but it includes a centrifugation step and a manual removal of the phases. The centrifugal separation of the phases may not have gone to completion especially if the difference in density between the phases is small or the viscosity of the mix is high. Manual removal of phases is associated with material losses and separation error. These problems are absent when uorometry is used. In the latter case, the thermal history of the small sample is easily controlled and the integrity of the sample can be preserved during storage. Morphology and kinetic effects can therefore be studied in contrast to when centrifugation is used. Fluorescently labeled polymer may be stored in dry form. The prospect of employing this technique in higher concentration ranges remains and the accuracy of the results will depend on the properties of the particular polymer system. Some further development would be required regarding the preparation of representative slides for uorometric calibration. Viable analytical approaches based on either ordinary uorescence microscopy or CSLM can most likely be designed for the determination of the biopolymer concentrations. In the rst case, the use of thin samples and a CCDcamera offers the advantages of low-impact uorescent labeling, minor interfering optical effects, and low cost. In the second case, a confocal scanning microscope with a detector in photon-counting mode could be employed for whole-scanning of bulk samples. An advantage in this case would be a less laborious collection and interpretation of data, if interfering effects can be kept small. Acknowledgements We are grateful to Dr Simon Gilroy and the Department of Biology for advice and providing opportunities for the use of their equipment for uorescence microscopy.
Fig. 10. Coalesced phases of unlabeled pectin 100 min after initialized phase separation in a mix of 2.0% LM pectin, 5.0% gelatin B, and 1 M sodium chloride. The image size is 280 350 mm2:
with the lower correlation among the calibration standards between uorescence intensity and concentration of labeled gelatin compared to labeled pectin, there could be a more nonlinear behavior at higher concentrations. We believe that these facts are related to a more pronounced associative behavior among molecules of labeled gelatin than among molecules of labeled pectin in aqueous solution. Such phenomena would particularly inuence those regions where aggregation of polymer molecules takes place before the start of gelation. Association of chromophore-containing material is likely to occur since water-soluble polymers with hydrophobic groups are known to form microdomains above a critical aggregation concentration (CAC) (Fischer et al., 1998), and clusters or micelles in aqueous solution (Winnik & Winnik, 1993). For instance, proteins conjugated with tetramethylrhodamine are prone to aggregation (Bioprobes 27, Molecular Probes, Inc., OR). This fact suggests that local gelation may take place at a higher temperature than the ordinary gelation temperature of gelatin and freeze in early morphological stages. Thus, variability in observed morphology within samples could be explained on such premises. Although the sparse labeling of the polymers is likely to reduce the inuence of the uorochromes on chain conformation and intramolecular selfquenching, there are increased opportunities for competitive mechanisms by which loss of excitation light energy can occur
Table 1 Phase composition data obtained either from uorometry or from chemical assay. The mixes contained low-methoxyl pectin (P), gelatin B (G), 1.0 M sodium chloride, and water Mix 1.7% P/4.25% G 2.0% P/5.0% G Phase Dispersed Continuous Dispersed Continuous % Pectin (Fig. 7) 2.4 0.5 3.5 0.5 % Pectin (uorometry) 2.3 0.7 3.6 0.8 % Gelatin (Fig. 7) 3.5 5.5 3.4 7.0 % Gelatin (uorometry) 2.2 5.1 3.6 7.5
589
References
Alevisopoulos, S., Kasapis, S., & Abeysekera, R. (1996). Formation of kinetically trapped gels in the maltodextringelatin system. Carbohydrate Research, 293, 7999. Al-Ruqaie, I. M., Kasapis, S., & Abeysekera, R. (1997). Structural properties of gelatinpectin gels. Part II: effect of sucrose/glucose syrup. Carbohydrate Polymers, 34, 309321. Aman, P., & Westerlund, E. (1996). Cell wall polysaccharides: structural, chemical, and analytical aspects. In A. -C. Eliasson, Carbohydrates in food (pp. 207208). New York: Marcel Dekker. Antonov, Y. A., Lashko, N. P., Glotova, Y. K., Malovikova, A., & Markovich, O. (1996). Effect of the structural features of pectins and alginates on their thermodynamic compatibility with gelatin in aqueous media. Food Hydrocolloids, 10 (1), 19. BeMiller, J. N. (1996). Gums/Hydrocolloids: analytical aspects. In A. -C. Eliasson, Carbohydrates in food (pp. 265281). New York: Marcel Dekker. Berth, G. (1988). Studies on the heterogeneity of citrus pectin by gel permeation chromatography on Sepharose 2 B/Sepharose 4 B. Carbohydrate Polymers, 8, 105117. Blonk, J. C. G., & van Aalst, H. (1993). Confocal scanning light microscopy in food research. Food Research International, 26 (4), 297311. Blonk, J. C. G., van Eendenburg, J., Koning, M. M. G., Weisenborn, P. C. M., & Winkel, C. (1995). A new CSLM-based method for determination of the phase behaviour of aqueous mixtures of biopolymers. Carbohydrate Polymers, 28, 287295. Brossmer, R., & Gross, H. J. (1994). Fluorescent and photoactivatable sialic acids. Methods in enzymology, Vol. 247. New York: Academic Press (pp. 177193). Brunner, A., Minamitake, Y., & Gopferich, A. (1998). Labeling peptides with uorescent probes for incorporation into degradable polymers. European Journal of Pharmaceutics and Biopharmaceutics, 45 (3), 265273. Clewlow, A. C., Clark, A. H., Rowe, A. J., & Tombs, M. P. (1995a). Predicting the phase diagram for a gelatinpectin system. Biochemical Society Transactions, 23, 498S. Clewlow, A. C., Rowe, A. J., & Tombs, M. P. (1995b). Pectingelatin phase separation: the inuence of polydispersity. In S. E. Harding, S. E. Hill & J. R. Mitchell, Biopolymer mixtures (pp. 173191). Nottingham: Nottingham University Press. da Silva, J. A. L., & Goncalves, M. P. (1994). Rheological study into the ageing process of high methoxyl pectin/sucrose aqueous gels. Carbohydrate Polymers, 24, 235245. DeMars, L. L. (1995). Texture and structure of gelatin/pectin-based gummy confections. MSc thesis, The Pennsylvania State University. Dickinson, E., & McClements, D. J. (1995). Advances in food colloids, New York: Blackie Academic and Professional (pp. 1112). Doner, L. W., & Douds, D. D. (1995). Purication of commercial gellan to monovalent cation salts results in acute modication of solution and gel-forming properties. Carbohydrate Research, 273, 225233. Durrani, M., Prystupa, D., Donald, A., & Clark, A. H. (1993). Phase diagram of mixtures of polymers in aqueous solution using Fourier transform infrared spectroscopy. Macromolecules, 26, 981987. EY-Laboratories. (1990). Biotechs. San Mateo, CA: EY Laboratories. Finch, C. A., & Jobling, A. (1977). The physical properties of gelatin. In A. G. Ward & A. Courts, The science and technology of gelatin (pp. 249 294). New York: Academic Press. Fischer, A., Houzelle, M. C., Hubert, P., Axelos, M. A. V., Geoffroy , M. C., Viriot, M. L., & Dellacherie, E. (1998). Chapotot, C., Carre Detection of intramolecular associations in hydrophobically modied pectin derivatives using uorescent probes. Langmuir, 14 (16), 4482 4488. Fulcher, R. G., Faubion, J. M., Ruan, R., & Miller, S. S. (1994). Quantitative microscopy in carbohydrate analysis. Carbohydrate Polymers, 25, 285293. Gahmberg, C. G., & Tolvanen, M. (1994). Nonmetabolic radiolabeling and
tagging of glycoconjugates. Methods in enzymology, Vol. 230. New York: Academic Press (pp. 3244). Goudsmit, E. M., Matsuura, F., & Blake, D. A. (1984). Substrate specicity of d-galactose oxidase. The Journal of Biological Chemistry, 259 (5), 28752878. Grinberg, V. Y., & Tolstoguzov, V. B. (1997). Thermodynamic incompatibility of proteins and polysaccharides in solutions. Food Hydrocolloids, 11 (2), 145158. Gubenkova, A. A., Somov, V. K., & Shenson, V. A. (1988). Physicochemical properties of pectin and pectin-based solutions and jellies. Pishchevaya Promyshlennost, USSR, 5, 1316 (abstract in English). Guthrie, R. D. (1962). Periodate oxidation experimental conditions. In R. L. Whistler & M. L. Wolfrom, Methods in carbohydrate chemistry (pp. 432435). New York: Academic Press. Haugland, R. P. (1996). Handbook of uorescent probes and research chemicals, . (6th ed.)Eugene, OR: Molecular Probes (pp. 14, 30). Hermanson, G. T. (1996). Bioconjugate techniques, San Diego: Academic Press. Ichinose, N., Schwedt, G., Schnepel, F. M., & Adachi, K. (1991). Fluorometric analysis in biomedical chemistry, trends and techniques including HPLC applications. Chemical analysis series, Vol. 109. New York: Wiley. Ipsen, R. (1995). Mixed gels made from protein and kappa-carrageenan. Carbohydrate Polymers, 28 (4), 337339. ISO. (1994). Meat and meat products determination of hydroxyproline content. ISO 3496:1994(E). Geneve, Switzerland: International Organization for Standardization. Jackson, E. L. (1944). Periodic acid oxidation. In R. Adams, Organic reactions (pp. 341375). , Vol. 2. New York: Wiley. Kalab, M., Allan-Wojtas, P., & Miller, S. S. (1995). Microscopy and other imaging techniques in food structure analysis. Trends in Food Science and Technology, 6, 177186. Kasapis, S., Morris, E. R., Norton, I. T., & Brown, R. T. (1993a). Phase equilibria and gelation in gelatin/maltodextrin systems part III: phase separation in mixed gels. Carbohydrate Polymers, 21, 261268. Kasapis, S., Morris, E. R., Norton, I. T., & Clark, A. H. (1993b). Phase equilibria and gelation in gelatin/maltodextrin systems part IV: composition-dependence of mixed-gel moduli. Carbohydrate Polymers, 21, 269276. Katayama, M., Nakane, R., Matsuda, Y., Kaneko, S., Hara, I., & Sato, H. (1998). Determination of progesterone and 17-hydroxyprogesterone by high performance liquid chromatography after pre-column derivatization with 4,4-diuoro-5,7-dimethyl-4-bora-3a,4a-diaza-s-indacene-3propiono-hydrazide. Analyst, 123, 23392342. Kinsella, J. E., Rector, D. J., & Phillips, L. G. (1994). Physicochemical properties of proteins: texturization via gelation, glass and lm formation. In R. Y. Yada, R. L. Jackman & J. L. Smith, Protein structure function relationships in foods (pp. 121). New York: Blackie Academic and Professional. Kumar, U., Laird, D., Srikant, C. B., Escher, E., & Patel, Y. C. (1997). Expression of the ve somatostatin receptor (SSTR1-5) subtypes in rat pituitary somatotrophes: quantitative analysis by double-label immunouorescence confocal microscopy. Endocrinology, 138 (10), 4473 4474. Mazur, A. W. (1991). Galactose oxidase selected properties and synthetic applications. ACS symposium series, Vol. 466 (pp. 99110). Owen, A. J., & Jones, R. A. L. (1998). Rheology of a simultaneously phaseseparating and gelling biopolymer mixture. Macromolecules, 31, 7336 7339. Ploem, J. S. (1993). Fluorescence microscopy. In W. T. Mason, Fluorescent and luminescent probes for biological activity. Biological techniques (pp. 111). New York: Academic Press. Puchtler, H., Meloan, S. N., & Brewton, B. R. (1974). On the binding of Schiffs reagent in the PAS reaction: effects of molecular congurations in models. Histochemistry, 40, 291299. Ross-Murphy, B. S. (1994). Physical techniques for the study of food biopolymers New York: Blackie Academic and Professional.
590
T.S. Nordmark, G.R. Ziegler / Food Hydrocolloids 14 (2000) 579590 Vinches, C., Parker, A., & Reed, W. F. (1997). Phase behavior of aqueous gelatin/oligosaccharide mixtures. Biopolymers, 41 (6), 607622. Vodovotz, Y., Vittadini, E., Coupland, J., McClements, D. J., & Chinachoti, P. (1996). Bridging the gap: use of confocal microscopy in food research. Food Technology, 50 (6), 7482. Walter, R. H. (1991). Analytical and graphical methods for pectin. In R. H. Walter, The chemistry and technology of pectin (pp. 207208). San Diego: Academic Press (pp. 215216). Walter, R. H., & Sherman, R. M. (1983). The induced stabilization of aqueous pectin dispersions by ethanol. Journal of Food Science, 48, 12351237 (see also p. 1241). Wilchek, M., & Bayer, E. A. (1987). Labeling glycoconjugates with hydrazide reagents. Methods in enzymology, Vol. 138. New York: Academic Press (pp. 429442). Winnik, M. A., & Winnik, F. M. (1993). Fluorescence studies of polymer association in water. In M. W. Urban & C. D. Craver, Structure-property relations in polymers. Spectroscopy and performanceAdvances in chemistry series (pp. 485505). , Vol. 236. Washington, DC: American Chemical Society. Ziegler, G. R. (1988). Relationship between the dynamic shear modulus of mixed gelatinegg white gels and their thermodynamic compatibility in an aqueous solvent. PhD thesis, Cornell University. Ziegler, G. R., & Foegeding, E. A. (1990). The gelation of proteins. In J. E. Kinsella, Advances in food and nutrition research (pp. 203298). Vol. 34. New York: Academic Press.
Ross-Murphy, S. B. (1995). Small deformation rheological behavior of biopolymer mixtures. In S. E. Harding, S. E. Hill & J. R. Mitchell, Biopolymer mixtures (p. 95). Nottingham: Nottingham University Press. Rost, F. W. D. (1991). Quantitative uorescence microscopy, Cambridge: Cambridge University Press. Scott, R. W. (1979). Colorimetric determination of hexuronic acids in plant materials. Analytical Chemistry, 51 (7), 936941. Stegemann, H., & Stalder, K. (1967). Determination of hydroxyproline. Clinica Chimica Acta, 18, 267273. Stelzer, E. H. K. (1998). Contrast, resolution, pixelation, dynamic range and signal-to-noise ratio: fundamental limits to resolution in uorescence light microscopy. Journal of Microscopy, 189 (1), 1524. Strasburg, G. M., & Ludescher, R. D. (1995). Theory and applications of uorescence spectroscopy in food research. Trends in Food Science and Technology, 6, 6975. Theander, O., Aman, P., Westerlund, E., Andersson, R., & Pettersson, D. (1995). Total dietary ber determined as neutral sugar residues, uronic acid residues, and Klason lignin (The Uppsala Method): collaborative study. Journal of AOAC International, 78 (4), 10301041. Tolstoguzov, V. B. (1990). Interactions of gelatin with polysaccharides. In G. Phillips, P. A. Williams & D. Wedlock, Gums and stabilisers for the food industry (pp. 157175). , Vol. 5. Wrexham, UK: IRL Press. Tolstoguzov, V. B. (1995). Some physico-chemical aspects of protein processing in foods. Multicomponent gels. Food Hydrocolloids, 9 (4), 317332.
Carbohydrate Polymers 34 (1997) 309-321
0 1998 Published by Elsevier Science Ltd All rights reserved. Printed in Great Britain
ELSEVIER PII: SO144-8617(97)00107-O Oh&l-8617/97/%17.00 + .OO
Structural properties of pectin-gelatin gels. Part II: effect of sucrose/glucose syrup

Ibrahim M. Al-Ruqaie, Stefan Kasapis* and Rukmal AbeysekeraB
Department of Food Research and Technology, Cranfield University, Silsoe College, Silsoe, Bedfordshire h4K45 4DT, UK bInstitute for Applied Biology, University of York, Yorkshire YOI 5DD, UK
(Received 3 December 1996; revised version received 17 February 1997; accepted 17 February
1997)
Small deformation dynamic oscillation and bright field microscopy were used to examine the structural properties of single and mixed high methoxy pectin and gelatin systems in the presence of sucrose/glucose syrup blends. Co-solute concentrated (278%) systems of the polysaccharide form rubbery structures which are readily transformed into glassy consistencies according to the timetemperature superposition principle. Increasing amounts of co-solute in the gelatin samples induce changes in viscoelasticity from that of conventional hydrogels to mechanical traces that cover much of the plateau region and the beginning of the glass transition area. Furthermore, manipulation of the protein/ sugar ratio can result in strong crystalline matrices, or viscoelastic solutions where the co-solute forms the continuous phase and the gelatin inclusions can undertake a conformational transition. The properties of the single components were used to rationalise the phase behaviour of their mixtures. Upon triggering the gelation of pectin, mixtures can be made where either gelatin or both components form a continuous phase. Results are discussed in the light of evidence obtained from the ethylene glvcol work in Part I. 0 1998 Published by Elsevier Science Ltd. All rightsreserv&l INTRODUCTION
The aim of our work on gelatin/high methoxy pectin/ co-solute samples is to develop model systems which are close to real foods, with the polymeric component being studied in the presence of cosolutes, and not merely of biopolymers themselves. Upcoming uses for mixed preparations of gelatin and pectin in the presence of sucrose/glucose syrup include wine gums and fruit pastilles. To develop a general understanding and to contrast the behaviour of different co-solutes, Part I of the series deals with the effect of ethylene glycol (EG) on the structure and mechanical properties of these biopolymers (Chronakis et al., 1997). Preparation of gelatin samples with O-70% EG results in an immediate rise in gel strength followed by a subsequent network weakening with a maximum
point at approx. 30% co-solute. The increase in network strength might be rationalised on the basis of
unfavourable interactions between EG and protein segments which can be minimised by enhancing the
*To whom correspondence should be addressed. 309
gelatin self-associations (Gekko and Timasheff, 198 1). However, a thermodynamically stable gelatin helix requires a surrounding layer of hydration whose diminishing presence leads to the ultimate drop in rigidity at the top range of ethylene glycol concentrations (Privalov and Tiktopulo, 1970). Compared with gelatin, high methoxy pectins require a subtle balance of hydrophilic and hydrophobic interactions to sustain a stable gel structure. These involve formation of aggregated helices supported by hydrogen bonds and grouping of methyl ester groups within a cage of water molecules (Walkinshaw and Arnott, 1981; Rolin, 1993). At 30% EG in the system, pectin is capable of forming a soft network which is further reinforced with addition of co-solute up to 60%, presumably due to increasing hydrogen bonding between polymeric segments. Higher levels of ethylene glycol, however, cause a reduction in gel strength since the co-solute can disrupt the water calyx and solvate the methyl clusterings. Mixtures of gelatin and high methoxy pectin with an EG content of less than 30% form protein continuous gels, as judged from their gelling and melting profiles, which are congruent with the corresponding traces of
310
I.M. Al-Ruqaie et al. product of Systems Bio-industries, an acid pigskin extract with an isoelectric point of pHx8. The citrus peel pectin sample is a high methoxy variety (70% degree of esterification), and came from Hercules (GENU B). It is standardised to specific gel properties by blending with sucrose. The pectin content was found to be approx. 68.1%, and an allowance was made for the polymer and co-solute concentration in the final preparations. Sucrose was of food grade. Cerestar provided glucose syrup with a dextrose equivalent of 42 and a water content of 19%. Gelatin samples were prepared by soaking the granules in distilled water overnight and then heating to 60C. Pectin samples were dissolved at 90C with gentle agitation for I5 min. Then sucrose or sucrose plus corn syrup was added and the pH was adjusted to 3 with 2~ HCl. Binary systems were prepared by mixing appropriate amounts of stock solutions at 75C. Dynamic oscillation measurements were made on a parallel plate geometry (40mm diameter; 1 mm gap) of a commercial high-torque Carri-Med CSL 500 rheometer or a cone and plate arrangement (0.02 rad; 50mm diameter) of an in-house sensitive prototype (Richardson, 1991). Samples were loaded onto the preheated platen of the rheometer and the edges were covered with silicone fluid to prevent evaporation. Cooling and heating runs to 5C and back were carried out at a scan rate of lC/min, a frequency of 1.6 Hz (corresponding to an angular frequency of M10 rad/s), and 1% strain (well within the linear viscoelastic region). Mechanical spectra over three or four decades of frequency (in the range from 0.001 to 10Hz) were also recorded at selected temperatures to extend the window of observation of viscoelastic parameters.
single gelatin networks (Chronakis et al., 1997). Nevertheless, ordered segments of pectin exclude the gelatin chains from their domain, thus inducing a phase separated arrangement of increased mechanical strength. At higher concentrations of EG, the ordered assemblies of pectin are capable of forming a network at higher temperatures than gelatin, with the cooling profiles now showing a bimodal building of structure. Microscopy evidence argues that gelatin manages to create a continuous network alongside the pectin matrix, the mixture thus being a phase separated bicontinuous system. Regardless of the variation in rigidity of gelatin and high methoxy pectin structures as a function of ethylene glycol composition, the viscoelastic ratio of loss modulus (G) to storage modulus (G), tan 6, is only slightly affected. For example, the tan 6 values of gelatin and pectin gels containing 70% EG are respectively 0.040 and 0.074 (YC), and fall well within the range expected (tan S<O.l) for aqueous biopolymer networks (Almdal et al., 1993). This result might be non-specific thermodynamically due to the unfavourable interaction between EG and these biopolymers that reduces the area of solvent-gelatin/ pectin contact, and depending on EG concentration, it allows an extensive or limited formation of relatively similar intermolecular associations. By contrast, addition of sugars to the gellan polysaccharide generates a maximum in the gel modulus vs co-solute graph and, in addition, transforms the viscoelastic ratio (Papageorgiou et al., 1994). Thus, gellan samples in the presence of 50% sucrose plus 20% glucose syrup not only form weaker networks than those at peak strength (30% sucrose), but also exhibit substantial frequency dependence of shear moduli and a tan 6 value of 0.906 at 5C (marker frequency of lOrad/s). In contrast to the sigmoidal solution (G > G) + gel (G > G) transition of gelatin/pectin-EG samples, the gellan-sugar system exhibits rubbery viscoelasticity with G > G at the highest experimentally accessible temperature of 90C which develops gradually to an extremely viscous solution (glass) during cooling to subzero temperatures (G > G). It was argued that sugars and gellan chains interact closely to reduce crystallinity/aggregation, thus transforming the network to an assembly of flexible polysaccharide segments where the entropic contribution to elasticity is dominant (Whittaker et al., 1997). In the present work we observe a metamorphosis in the behaviour of single gelatin and high methoxy pectin systems by using sucrose/glucose syrup as the co-solute, and attempt to rationalise its effect on both single and binary mixtures.
RESULTS AND DISCUSSION Formation of gels and the transition from rubber- to glass-like consistency in high methoxy pectin induced by increasing amounts of sucrose/&Iucose syrup blends As outlined in the Introduction to this work, gel formation in pectin/ethylene glycol systems is seen at co-solute concentrations as low as 30%. However, replacement of ethylene glycol with blends of sucrose/ glucose syrup raises the co-solute requirement for structure development. Table 1 shows the variation in viscoelastic parameters as a function of sugar content induced during cooling from 90 to 5C at lC/min, holding there for lOOmin, and heating to 95C at the same scan rate. At 58% co-solute, the onset of gel formation is seen as a steep rise in storage modulus and occurs at 34C. This gel is thermoreversible and melts at approx. 86C. Addition of co-solute up to 74% induces an earlier network formation (~72C) and reinforces the elastic properties of the systems
MATERIALS
AND METHODS
The polymeric ingredients have been described in some detail in Part I of the series. Briefly, gelatin was a
Pectin/Gelatin/Sugar
Systems
311
which now remain thermally irreversible at the highest accessible temperature (tan 6 values at 95C are less than 0.5). The transformation from a liquid-like (tan 6 > 1 at 95C) to a solid-like (tan 6 < 1 at 5C) behaviour, reported so far, has also been observed during cooling of the water/EG mixtures of gelatin and pectin, and is the standard process occurring in aqueous polysaccharide preparations, e.g. of agar 1993) and rc-carrageenan (Selby and Whistler, (Therkelsen, 1993). In contrast to the pectin-water/EG mixtures where the values of G peak at 60% ethylene glycol and decline continuously thereafter, the upward trend in network strength is not affected at 66 and 74% sucrose/glucose syrup blends. In Part I of the series, we argued that the drop in G at this range of co-solute is due to disruption of the hydrogen bond cages around the methyl groups which are now dissolved in ethylene glycol. This statement is endorsed by the current results, where the polyhydric compounds should preserve the hydrogen bond calyxmethyl clustering arrangement, thus reinforcing the networks rigidity. The general form of modulus development is entirely different when the top two concentrations of co-solute in Table 1 are employed. During sample preparation, it appeared that coherent structures can be formed at any temperature below the boiling point, an observation which is verified by a solid-like response on the rheometer at 90C (e.g., the tan 6 of 1% pectin plus 78% co-solute is equal to 0.168). The antithesis with the results in the preceding paragraph persists at the other end of the cooling run (5C) where a predominantly liquid-like response is now obtained; e.g. the tan 6 at 86% co-solute amounts to 1.941. This unusual behaviour is illustrated in Fig. 1, where the experimental constraints (scan rate of lC/min; frequency of 1.6 Hz) reveal a gradual transformation in viscoelasticity, the viscous component eventually
5.5 -
5.0 -
g E
ii
4.0 -
45
3.0 2.5 +
0
EJ 3.5 -
20
40
60
80
100
Tempexature (T) Fig. 1. Development of storage and loss moduli during cooling of 1% pectin sample with 78 and 86% co-solute (bottom and top spectra, respectively). The scan rate is lC/ min and the frequency is 1.6Hz.
312
I.M. Al-Ruqaie et al.

(tan (5= 1) and finally levels off at the upper end of the frequency range, indicating that the superposition extends well into the glass transition area (Ferry, 1980). This rather spectacular alteration in the viscoelasticity of high sugar pectin samples will be contrasted with the mechanical profiles of high sugar gelatin samples, and together used as a baseline of behaviour in the discussion of binary mixtures. Small deformation properties of solutions, gels, and crystalline systems comprising gelatin and sucrose/ glucose syrup blends Gelatin samples were prepared at concentrations of 1.5, 3. and 5% with a view to covering a range of product applications from soft table jellies to firm confectioneries. For each gelatin series, samples were cooled from 70 to 5C at lC/min, left there to equilibrate for 100 min, monitored under increasing angular frequency, and finally heated back. The usual frequency-dependent markers of network formation sharp rise in the G trace) and collapse (tmelt; G (fii,& falling below G during heating) were used. Let us reiterate here that our values of tform should not acquire the sense of the gelling point according to the Te Nijenhuis-Winter criterion (i.e., identification of the gelling point with the frequency-independent loss tangent; Te Nijenhuis and Winter, 1989) since, in combination with the tmelt values, their use is that of an operational convenience for rationalisation of the phase continuity in biopolymer mixtures. Table 2 reproduces the changes in viscoelastic functions with increasing levels of co-solute for the 5% gelatin series. The expected stabilisation of a gelatin network is noted as rising storage modulus with increasing co-solute concentrations up to 58%, and it has been associated with the generation of extra collagen-like triple helix structures (Oakenfull and Scott, 1986). There is, however, a significant drop in the rigidity of the gelatin network in the presence of 66% co-solute, which should be attributable to the destabilisation of the triple helix conformation at this level of water shortage (Privalov and Tiktopulo, 1970). Fig. 3(a) shows the asymmetric cooling and heating traces (see also tform and r,,it values in Table 2) separated by the isothermal run, at the highest, 66%, sucrose/glucose syrup content. Both the graph and the tabulated values indicate that the tan 6 at 5C (0.156) is unusually high for an aqueous gelatin gel (0.011 at 0% sucrose). The high sol-fraction is due to the large quantity of co-solute and its slowing-down effect on the helix fraction of gelatin chains. As a result, a rubbery spectrum of gelatin at 66% co-solute is obtained with the G trace showing a pronounced frequency dependence and the G values start taking off at frequencies higher than 1 Hz (Fig. 3b). The changes in mechanical characteristics observed
becoming predominant. Upon heating, both moduli trace back their cooling spectra with no signs of thermal hysteresis, which is evident from the differences in the trorm and t,,it temperatures at levels of co-solute up to 74% (Table 1). As mentioned in the Introduction, high sugar gellan with preparations show cooling a similar transformation from solid- to liquid-like behaviour. This, of course, has been reported before for the transition from rubberto glass-like consistency in amorphous synthetic polymers, and rationalised with the combined WLF/free volume theoretical framework (Williams et al., 1955). The essence of this approach is a change in state, but not in phase of the material, which is manifest in the glass transition prolile. The development in viscoelasticity can be used to infer the changing free volume which collapses to approx. 3% of the total volume at the glass transition temperature (Ferry, 1980). It follows that in the absence of a disorder-to-order transition, the vitrification process at a low temperature can be reproduced at a higher temperature, as long as measurements can be carried out at extremely short timescales (usually in the order of tens of kHz). Since conventional rheological measurements cannot be performed at such high frequencies, viscoelastic parameters obtained at regular intervals during a cooling run are shifted along a logarithmic frequency axis, thus depicting a composite curve over a wide frequency window (at least 6-7 decades) for an arbitrary chosen reference temperature (time-temperature superposition; TTS). Application of the TTS principle to aqueous preparations of biopolymers fails because changes in free volume are usually swamped by other temperature-dependent effects, such as enthalpic interactions between the chains (Lopes da Silva et al., 1994). Following the synthetic polymer approach, therefore, frequency sweeps were recorded at 90, 70, 50, 30, and 10C for the 1% pectin sample in the presence of 86% co-solute. Figure 2(a, b) illustrates the variation in storage and loss moduli with decreasing temperature. At the top temperature range (90-50C) and frequency of oscillation, up to 0.1 Hz the spectra of G remain relatively flat. Beyond this range, however, there is a clear build up of structure culminating at the sharp frequency dependence of G at 10C. The G values also rise steeply covering three orders of magnitude and eventually overtake those of G at 10C. The outcome of superimposing our data at 90C (reference temperature) is shown in Fig. 2(c). These were fitted in the form of a composite curve and its viscoelastic ratio, tan 6 = G/G, covering a frequency range from 10-s to 104Hz (Fig. 2d). There is a clear progression from a rubbery plateau to a glass transition region with increasing frequency in the way induced by cooling in Fig. 1. As a result, the ratio of loss to storage modulus crosses the threshold from solid- to liquid-like response
Pectin/Gelatin/Sugar 7.0
i 1. 6.5
l l
Systems
313
6.5
l
6.0
l
6.0
l
$ V 3 IJ
5.5
0 l
$ V 8 -1
5.5
l
l l
A A A
5.0 4.5 4.0 3.5 0.001 0.01 0.1 1 10
5.0 4.5 4.0 3.5 0.001 0.01 0.1 1 10
Frequency (Hz)
Frequency (Hz)
1.08 6.5 3 $ 2 2 5.0 6.0 5.5
C
6.5 6.0
1.06
0.96 3.5 I -3 3.5 I -2 -1 0 1 2 3 4 -3 1 -2 1 -1 I 0 I I I 2 I 3 : 0.94 4
Log (Frequency/Hz)
Log (Frequency/Hz)
Fig. 2. Frequency sweeps of 1% pectin plus 86% co-solute at 90 (m), 70 (Cl), 50 (A), 30 (A), and 10C (0) for the storage (a) and loss (b) moduli. The time-temperature superposition for the experimental points and a fit with the tan 6 trace are shown in (c) and (d), respectively. Table 2. Sucrose Small deformation viscoelastic parameters for 5% gelatin preparations made with increasing levels of co-solute Glucose syrup Total solute Pseudoequilibrium modulus (G; Pa) 5C
0 -
Tan 6 at1 Hz
tfcm
((3
GneJt ((3
(%)
(%)
(%) 6713 7352 8637 9084 9239 9553 6909
(5OC) 0.011 0.012 0.014 0.016 0.017 0.058 0.156
lC/min 17 20 22 24 25 28 37
lC/min 27 30 31 33 34 38 41
20 30 40 50 50 50 The concentration
8 16
20 30 40 50 58 66
of glucose syrup refers to dry solids.
314
I.M. Al-Ruqaie et al. (tmeit=2S0C) characteristics of a 3% gelatin network (Part I), the present values see a rise in tmelt of more than ten degrees centigrade with increasing sucrose/ glucose syrup solids from 0% to 62%. Similarly, the interactive nature of gelatin-sucrose/glucose syrup molecules raises steadily the sol-fraction of these gels, whereas in the presence of, for example, 70% ethylene glycol, a rather conventional tan 6 value of 0.040 is recorded. The high tan 6 value (0.643) and the rapidly drooping G value (1869 Pa) at 70% sucrose/glucose syrup emphasises the destabilisation of gelatins native structure at low levels of water. Figure 4(a) illustrates the asymmetric development and demise of viscoelasticity with the G trace struggling and eventually overtaking that of G only after 35min within the isothermal run (5C). Nevertheless, a true structure has taken shape at the end of the isothermal run (lOOmin; frequency of 1.6 Hz) which produces a characteristic shoulder during subsequent heating, and clearly demonstrates the development of thermal hysteresis between the belated network formation and melting (tmeit= 34C) of this preparation. The outcome of dropping the polymer to co-solute ratio from approx. 0.08 (Fig. 3b) to 0.04 in the sample of 3% gelatin with 70% sugar is shown in Fig. 4(b). This time the values of storage modulus depart from the plateau of the rubbery zone at much lower frequencies (in the order of 0.01 Hz) and converge with those of loss modulus. Eventually, the G overtakes the G at the high frequency end, an indication of the advent of the glass transition region. The general form of modulus development during our experimental routine changes dramatically when samples of 3% gelatin are made with 50% sucrose and 28% glucose syrup. Initially, a viscous solution response is obtained during a cooling run and frequency sweeps at elevated temperatures, e.g. at 70C (not shown here). As depicted in Fig. S(a), however, at about 10C (scan rate of lC/min) the
0.5
-I
0 50
100 150
200
Time (min)
0.01
01
IO
Frequency(Hz) Variation in storage and loss moduli for 5% gelatin in the presence of 66% co-solute during a coolingisothermal-heating run (a) and a frequency sweep (b). The scan rate is lC/min, the isothermal run lasted lOOmin, and the frequency of (a) is 1.6 Hz.
Fig. 3.
for the 3% gelatin series during the coolingisothermal-heating routine are given in Table 3. The
use of lower levels of protein allowed formulation of systems with up to 78% co-solute. Compared with the inert ethylene glycol molecules which did not 14C) and melting alter the setting (trorm z
Table 3.
Small deformation viscoelastic parameters for 3% gelatin preparations made with increasing levels of co-solute
Sucrose
Glucose syrup
Total solute
Pseudoequilibrium modulus (G; Pa) 5C
Tan 6 at1 Hz
tform
(Cl
Gnelt
(C)
W)
0
W)
W) 1628 2177 2776 2966 2868 2904 2467 1869 1159000
(5C) 0.034 0.046 0.042 0.048 0.054 0.084 0.117 0.643 0.334
1C/min 12 14 20 21 22 24 26 35min at 5C 10
1C/min 25 28 31 32 33 35 36 34 10
20 30 40 50 50 50 50 50 The concentration
8 12 20 28
20 30 40 50 58 62 70 78
80
Systems
315
,la
I\
I
70 60
- 40 -30 - 20
B
E g
zt
50
100
150
200
50
100
150
200
Time.(min)
Time (mm)
4.0
3 ?i
;;i % 3 E B a
3.5
is -1
3.0
2.5
q 0
-I
0.01
0. I
IO
G"
10
Frew=y 0-W Fig. 5. Variation in storage and loss moduli for 3% gelatin in the presence of 78% co-solute during a cooling-isothermalheating run (a) and a frequency sweep (b). Conditions as in Fig. 3.
2.0 0.01
Fig. 4. Variation in storage and loss moduli for 3% gelatin in the presence of 70% co-solute during a cooling-isothermal-
heating run (a) and a frequency sweep (b). Conditions as in Fig. 3.
modulus traces cross-over and the solid-like response achieves values in excess of a million Pascals whereas the liquid-like response takes a sudden dive. The picture remains unchanged during the isothermal run and reverses itself on subsequent heating. As a result a perfectly symmetrical time-temperature profile is obtained with no signs of thermal hysteresis pertaining to formation and melting of polymeric networks. This, and the formation of a rigid structure at 5C, contrasts strongly with the asymmetric profile of the weak and rubbery gelatin gel in Fig. 4; compare also the viscoelastic parameters for the top two co-solute samples in Table 3. Furthermore, the frequency dependence of the G and G at 5C (Fig. 5b) reproduces metamorphosis from a viscous solution to a rigid solid seen at the final stage of the cooling run. Very
similar mechanical. spectra have been recorded for concentrated sucrose/glucose syrup blends which were accompanied by endothermic peaks during heating of the samples in a calorimeter (Al-Ruqaie et al., 1997; Ong et al., in press). Work by Dea et al. (1984) has also reported this pattern of behaviour for ice cream at subzero temperatures, where it is expected that lactose crystallisation has occurred. Therefore, phase inversion has occurred from a rubbery gelatin consistency at 70% co-solute to a continuous, 78%, co-solute phase whose saturation point is reached during cooling, thus resulting in partial crystallisation and structural reinforcement of the system. In the last gelatin series we reduced further the protein to co-solute ratio and looked for additional patterns of viscoelastic behaviour. Table 4 summarises the effect of sucrose/glucose syrup blends on the properties of 1.5 % gelatin samples. Increasing the sugar content from 0 to 66% results in rubbery gelatin gels with a swollen sol-fraction, as seen for the 3% gelatin series (Fig. 4b) before the occurrence
316
Table 4.
I.M. Al-Ruqaie et al.

Small deformation viscoelastic parameters for 1.5% gelatin preparations made with increasing levels of co-solute
Sucrose
Glucose syrup
Total solute
Pseudoequilibrium modulus (G; Pa) 5C
Tan 6 at1 Hz
tfcm P-7
Gnelt (C)
W) 0 20 30 40 50 50 50 50 50 The concentration
(%)
(%) 102 85 139 181 130 139 87
(5C) 0.036 0.090 0.083 0.098 0.190 0.282 0.803 1.336 2.352
I C/min
5 7 12 15 22 28 80min at 5C
1C/min
21 29 30 33 33 34 30
8 16 24 28
20 30 40 50 58 66 74 78
of crystallisation. There, crystallinity was evidenced by a decrease in the tan 6 and a mounting of the G values. This time, topping up the amount of cosolute from 66 to 74 and 78% maintained the upward trend in tan 6, and values now are higher than 1. In addition, systems remain viscoelastic liquids throughout the experimental routine, as illustrated for the 1.5% gelatin plus 78% co-solute in Fig. 6(a). Over the allocated time at 5C there is a slow but steady growth in G which in combination with the relatively flat G creates a skewed profile. Mechanical spectra at the end of the isothermal run (Fig. 6b) show a rather unconventional frequency response with a shear thinning dynamic viscosity, although the moduli traces are well separated from each other. Subsequent heating of our system creates a sigmoidal trace, especially from the side of G whose values collapse well below those of G. Combined evidence from Fig. 6(a, b) argues that in the 1.5% gelatin series there is a transformation from a rubbery structure to a quite viscous solution. Clearly, the low levels of gelatin keep the sugars in solution which now form the continuous phase. It seems, though, that the gelatin inclusions undergo conformational ordering and gradual ageing/strengthening during the isothermal run at 5C. Since blends of sucrose and glucose syrup remain Newtonian during a frequency ramp (Papageorgiou et al., 1994), disruption of the pattern of flow (rubbing overlapping fields of surfaces, countercurrents) between adjacent gelatin particles with increasing frequency might account for the shear thinning behaviour of these composites. Deconvolution of the gelatin helices during subsequent heating of the sample will eliminate the solid-like component along the sigmoidal path monitored in Fig. 6(a). As in the case of a continuous gelatin network (e.g. Table 2), the high melting temperature of gelatin inclusions (mid-transition temperature is between 40 and 45C) argues that the continuous co-solute phase also imbeds in and crosscontaminates heavily the suspended gelatin particles. By
50
100
150
200
Time (min)
b
G
q
.
-1.8 -0.8 0.2 1.2
Log (Frequency/Hz) Fig. 6. Variation in G, G and n* for I .5% gelatin in the presence of78%co-soluteduringacooling-isothermal-heatingrun(a)anda frequency sweep (b). Conditions as in Fig. 3.
way of concluding this section, we hope that the multitude of viscoelastic profiles exposed by the reduction in gelatin to co-solute ratio and its proposed rationalisation will facilitate a stimulating debate in the area of high solids gelatin systems.
Systems
4.0 80 70 60 50 fi e 2.5 2.0 1.5 1.0 0.5 0 50 100 150 200 40 30 20 10 0 $ 2 8.
317
I I I I I I I IZ
3 2 a g
3.5 3.0
z il
Time (min)
_l
II
II
-2
-1
Log (Frequency/Hz)
Fig. 7. Variation in storage and loss moduli for the mixture

of 1% pectin and 3% gelatin in the presence of 58% co-solute during a cooling-isothermal-heating run (a) and a frequency sweep (b). Conditions as in Fig. 3. The effect of sucrose/glucose syrup on the phase behaviour of high methoxy pectin/gelatin mixtures
I I I I I=JZzi%
The understanding achieved at the above work of single biopolymers was put to practice in mixtures containing 1% pectin and 3% gelatin. As shown in Table 5, the first point at issue is the significant change in the values of tfom and tmert with increasing co-solute content. Thus, addition of sugar up to 58% generates a range of structure formation and melting indices which is covered by the temperature behaviour of gelatin gels. Under our experimental settings, the small-deformation results at 58% co-solute are illustrated in Fig. 7. Gratifyingly, the onset of structuring and the completion of melting, the network reinforcement during the isothermal run, the viscoelastic ratio at the end of the curing process at YC, and the ensuing frequency dependence at this temperature are congruent with the properties of a single, 5%, gelatin sample with 66% co-solute (Fig. 3 and Table 2). On the other hand, pectin structures melt at elevated temperatures (L86C in Table 1). Furthermore, the
318
2.0 -r 0 50 100 I50 200
r0 250
Time (mm)
Fig. 8.
The development and demise of G and G for the mixture of 1% pectin and 3% gelatin with 78% co-solute during a cooling-isothermal-heating run. Conditions as in Fig. 3.
mixed gel network is reinforced (~7 kPa) beyond the level expected from the corresponding sum of rigidities of single pectin and gelatin gels (~0.5 and 2.9 kPa, respectively). It appears, therefore, that mixtures have phase separated into a continuous gelatin phase suspending the pectin inclusions. At a sucrose/glucose syrup concentration of 58% the pectin domains are capable of gelation, thus excluding efficiently and concentrating up the gelatin matrix, with the system achieving the enhanced elasticity of a single 5% gelatin gel at 66% co-solute. The step from the gelatin-related tf,, and t,,lt values to those in the range of a pectin gel is achieved when the co-solute content is increased to 66 and 74% (Table 5). In the blend with 74% sugar, for example, the onset of gel formation is recorded at 65C and its structure remains thermally stable at the end of a heating run; tan 6 at 90C is 0.557. Therefore, despite the early melting profile of gelatin gels, mixtures retain overall cohesion at elevated temperatures, a result which argues that the polysaccharide has now formed a continuous network. The mixture with the highest possible amount of sucrose/glucose syrup (78%) forms, of course, a pectin continuous phase but its behaviour is distinct from that in the preceding paragraph and akin to the transformation observed for single pectin samples at 78 and 86% co-solute (Table 1). Its overall behaviour in our experimental routine is reproduced in Fig. 8. This time a dominant elastic response is recorded at the start of the cooling run (tan 6 at 85C is 0.851) but, as in Fig. 1, the reduction in temperature triggers a metamorphosis in viscoelasticity with the tan 6 value at
5C being equal to 1.831. In the absence of further development during the isothermal run, the gradual reversal to dominant G values on subsequent heating concludes a symmetric time-temperature profile. Mirror image temperature runs (i.e., no thermal hysteresis) were also reported for the 3% gelatin plus 78% co-solute in Fig. 5(a). There, however, we observed a transformation from liquid- to solid-like behaviour, owing to co-solute crystallisation, whereas in Fig. 8 the opposite occurs. Following the methodology of Fig. 2, mechanical spectra of the most concentrated mixture were taken at 70, 50, 30, and 5C. As in the case of single pectin preparations with 278% sugar, the patterns of G and G advance with decreasing temperature/increasing frequency from relatively flat to quite sharply sloping spectra (Fig. 9a and b). The time-temperature superposition principle was implemented in Fig. 9(c), starting from the reference temperature of 70C. Clearly, the mixture exhibits a mechanical transition from rubbery to glass-like consistency, as was noted for the high solids pectin systems in Fig. 2(c). Indeed, it is quite remarkable to see that whereas sugar in the presence of gelatin comes out of solution (crystallises) and gradually imparts solid-like properties to the sample of Fig. 5, pectin uses the co-solute as an innate part of its network, thus preventing it from separating out in the form of crystals and allowing the rubber to In Table 1 and Table 5, the values of pseudoequilibrium modulus in parentheses indicate the dominance of the viscous
component at 5C and the strong frequency dependence of the solid component, as opposed to the relatively flat response of G in the plateau region discussed at other places in the paper.
Pectin/Gelatin/Sugar 6
Systems
319
A
A
A
A A A A q n A
0 0
AL
c
I n
A
0
A
0
n
A 0 n
0 n
q n
10
2
0.01 0.1 1
Frequency 6 ,
(Hz)
a
2
&
lJ
n
2
i
q n
I 0.1
I 1
Frequency
(Hz)
C
06
00
08
0,
()(
glass transformation to occur in the mixture of Fig. 9. Although the rheological work has shown comprehensively the formation of a continuous pectin phase at the top range of co-solute (>66%), it is difficult to say if this is a true phase inversion point, i.e., a transformation from a gelatin to a pectin supporting matrix. To determine the topology of the gelatin phase we used bright field microscopy and histochemical dyes to provide contrast to the specific chemical groups of the two polymeric components. These were: Sirius red, which contains sulphonic acid groups and reacts with the basic groups of collagen and Toluidine blue, a metachromatic dye that stains pectin reddish pink (Abeysekera and Robards, 1995). Samples were cooled to 5C and matured there for 2 h. The dyes reached their target by diffusion for 10min and gel pieces of 10x10x0.5mm were taken for examination. Figure 10 reproduces the images of single and mixed gelatin/pectin systems in the presence of co-solute. Single polymer preparations appear as a featureless background, which undoubtedly constitutes an isotropic phase (top and middle micrographs in Fig. 10). Although pectin does not gel at concentrations of co-solute below 58% (Table l), the reinforced rigidity of the supporting gelatin matrix in mixtures with sugar levels up to 50% (compare values in Table 3 and Table 5) argues that the ordered polysaccharide sequences effectively exclude the protein from their domain. That must correspond to a scale of phase separation below the resolution of our light microscope (maximum magnification of 400x), since mixtures with a low cosolute concentration also appeared as a single entity. Pectin gelation at high levels of sucrose/glucose syrup should further enhance steric exclusion phenomena and more clearly differentiate one phase from another. Thus, in the bottom micrograph of Fig. 10, a phase separated system is obtained with discontinuous inclusions in the range of 1-2~. Staining of the samples strongly suggests that gelatin forms a continuous featureless background, with the darker pectin spots being the excluded phase. Therefore, rheology and microscopy have emphasised the phase continuity of the polysaccharide or the protein component at the top range of co-solute (266%) and together build up a case for a phase separated bicontinuous structure.
-2
-1
CONCLUSIONS Addition of co-solute, especially in the high solids regime (60-85%), to single and mixed systems of high methoxy pectin and gelatin results in dramatic changes of viscoelasticity. It is encouraging that the methodology of dynamic oscillation and microscopy employed extensively in the description of phase behaviour of
Log (Frequency/Hz)
Fig. 9. Mechanical spectra of 1% pectin plus 3% gelatin with 78% co-solute taken at 70 (D), 50 (El), 30 (A), and 5C (A)for the storage (a) and loss (b) moduli (the sample was scanned down between frequency sweeps at lC/min). The timetemperature superposition of both moduli at the reference temperature of 70C is shown in (c).
320
I.M. Al-Ruqaie et al. integral part of a pectin structure and at highly concentrated levels of co-solute, a rubbery polymersugar-water entity is stabilised which can undergo a glass transition. Within our experimental settings, rubbery gelatin-sucrose-glucose syrup systems show signs of vitrification. It appears, however, that the protein-sugar interaction is less stable than with pectin and upon increasing the sugar-to-protein ratio, cosolute precipitates out and creates strong crystalline structures. At conditions below the saturation point, blends of sucrose/glucose syrup can establish a continuous liquid phase with the gelatin inclusions being able to undergo a conformational transition. By comparison, increasing concentrations of ethylene glycol hardly change the viscoelastic ratio (tan S) of aqueous pectin and gelatin gels, hence suggesting a nonconformative type of hydroxyl groups (mainly cis gauch rotamers about the C-C bond with internal hydrogen bonds) for the development of consequential interactions with the two biopolymers.
ACKNOWLEDGEMENTS The authors are grateful to their colleague, Dr M. W. N. Hember for his critical evaluation of this manuscript.
REFERENCES Abeysekera. R. M. and Robards, A. W. (1995) Microscopy as an analytical tool in the study of phase separation of starch-gelatin binary mixtures. In Biopolymer Mixtures, eds S. E. Harding, S. E. Hill and J. R. Mitchell, pp. 143-160. Nottingham University Press, Nottingham. Al-Ruqaie, I. M., Kasapis, S., Richardson, R. K. and Mitchell, G. (1997). The glass transition zone in high solids pectin and gellan preparations. Polymer, 38, 5685-5694. Almdal, K., Dyre, J., Hvidt, S. and Kramer, 0. (1993) Towards a phenomenological definition of the term gel. Polymer Gels and Networks 1, 5517. Chronakis, I. S., Kasapis, S. and Abeysekera, R. (1997). Structural properties of gelatin-pectin gels. Part I: Effect of ethylene glycol. Food Hydrocolloids, 11, 271-279.
Dea, I. C. M., Richardson, R. K. and Ross-Murphy, S. B. (1984) Characterisation of rheological changes during the processing of food materials. In Gums and Stabilisers for the Food Industry 2, eds G. 0. Phillips, D. J. Wedlock and P. A. Williams, pp. 3577366. Pergamon Press, Oxford. Ferry, J. D. (1980) Dependence of viscoelastic behavior on temperature and pressure. In Viscoelastic Properties of Pol,vmers, pp. 264320. John Wiley & Sons, New York. Gekko, K. and Timasheff, S. N. (1981) Mechanism of protein stabilisation by glycerol: preferential hydration in glyceroll water mixtures. Biochemistry 20, 46674676. Lopes da Silva, J. A., Goncalves, M. P. and Rao, M. A. (1994) Influence of temperature on the dynamic and steady-shear rheology of pectin dispersions. Carbohydrate Polymers 23, 77-87. Oakenfull, D. and Scott, A. (1986) Stabilization of gelatin by sugars and polyols. Food HydrocoIloids 1, 163-175.
50pm
Fig. 10. Bright field micrographs for 1% pectin plus 70% co-solute (top), 3% gelatin plus 70% co-solute (middle), and 1% pectin-3% gelatin plus 70% co-solute (bottom). Images of 400x magnification were acquired with a Nikon inverted microscope Diaphot-TMD.
aqueous systems can be utilised in the area of high solids, where there is renewed scientific interest. In addition, the synthetic polymer approach adapted for the particular needs of biopolymers can guide us in the time/ temperature-induced transformation of samples from rubbery to glassy consistencies. Sugar constitutes an
Ong, M. H., Whitehouse, S., Abeysekera, R., Al-Ruqaie, I. M. and Kasapis, S. (1997) Glass transition-related or crystalline forms in the structural properties of gelatin/oxidised starch/glucose syrup mixtures. Food Hydrocolloidr;, in press. Papageorgiou, M., Kasapis, S. and Richardson, R. K. (1994) Glassy-state phenomena in gellan-sucrose-corn syrup mixtures. Carbohydrate Polymers 25, 101-109. Privalov, P. L. and Tiktopulo, E. I. (1970) Thermal conformational transformation of tropocollagen. I. Calorimetric study. Biopolymers 9, 127-139. Richardson, R. K. (1991) Rheological characterisation of biopolymer systems. PhD Thesis, Cranfield Institute of Technology. Rolin, C. (1993). Pectin. In Industrial Gums, eds R. L. Whistler and J. N. BeMiller, pp. 257-293. Academic Press, London. Selby, H. H. and Whistler, R. L. (1993) Agar. In Industrial Gums, eds R. L. Whistler and J. N. BeMiller, pp. 87-103. Academic Press, London.
Systems
321
Te Nijenhuis, K. and Winter, H. H. (1989) Mechanical properties at the gel point of a crystallizing poly(viny1 chloride) solution. Macromolecules 22, 411414. Therkelsen, G. H. (1993) Carrageenan. In Industrial Gums, eds R. L. Whistler and J. N. BeMiller, pp. 145-180. Academic Press, London. Walkinshaw, M. D. and Arnott, S. (198 1) Conformations and interactions of pectins II. Models for junction zones in pectinic acid and calcium pectate gels. Journal of Molecular Biology 153, 1075-1085. Whittaker, L. E., Al-Ruqaie, I. M., Kasapis, S. and Richardson, R. K. (1997) Development of composite structures in the gellan polysaccharide/sucrose system.
Carbohydrate Polymers, 33, 39-46.
Williams, M. L., Landel, R. F. and Ferry, J. D. (1955) Temperature dependence of relaxation mechanism in amorphous polymers and other glass-forming liquids.
Journal 3706. of the American Chemical Society 77, 3701-
Food Hydrocolloids
Vol. 11 no. 3 pp. 271-279, 1997
Structural properties of gelatin-pectin gels. Part I: Effect of ethylene glycol

Joannis S.Chronakis, Stefan Kasapis! and Rukmal Abeysekera>
Department of Food Research and Technology, Cranfield University, Silsoe College, Silsoe, Bedfordshire MK45 4DT and 2Institute for Applied Biology, University of York, Yorkshire YOISOD, UK ITowhom correspondence should be addressed
Abstract The role of ethylene glycol (EG) in the gelation mechanisms of acid pigskin gelatin and high methoxy pectin has been monitored and used as a baseline for the investigation of mixed gelatin-pectin gels in various mixed ethylene glycol-water solvents. The addition of EG did not alter the gelation (tgelr::! 14C) and melting (tmel r::! 28C) temperatures of an aqueous gelatin network, the strength of which, however, was first increased, and then reduced at concentrations of co-solute higher than 30%. The reduction in values of storage modulus (G) was attributed to a decrease in the proportion of polypeptide chains involved in the formation ofjunction zones. By contrast, increasing levels of ethylene glycol encouraged formation ofpectin gels at high temperatures (e.g. tgel was 63C at 800/0 EG) which largely retained their structure upon subsequent heating. The network strength increased rapidly and peaked at 60% co-solute followed by a subsequent reduction in storage moduli at conditions of low water activity (60-800/0 EG). On the basis of a model for gel formation involving a two-step process, it was proposed that ethylene glycol promotes the ordered structure of contiguous pectin chains (first stage) but 'dissolves' the hydrophobic clusterings ofmethyl groups (second stage). Differential scanning calorimetry demonstrated that thermodynamic incompatibility between the two polymers is the driving force behind the phase behaviour of mixed preparations. Based on the mechanical properties of single components, it was argued that increasing amounts of EO, within the 0-25% range, promote pectin's conformational ordering, which becomes more and more effective in excluding, concentrating up and strengthening the continuous gelatin phase. At higher levels of co-solute (from 30 to 70%), pectin can form a thermally stable network and during cooling it does so before gelatin's gelation at lower temperatures. Light microscopy work strongly suggests that gelatin also forms a continuous network throughout the body of the sample. Therefore, the latter mixtures can be described as phase-separated, bicontinuous arrangements.
Introduction
Mixtures of protein and polysaccharide are used increasingly in the manufacturing of food products with a low calorific value and novel textural properties (e.g. low-fat spreads and confectioneries). Mixing of two different biopolymers in solution commonly results in heterotypic complexation or steric exclusion, depending on the balance between the entropy of mixing plus mutual energetic associations and the enthalpy of self-interaction (1). Favourable interactions occur, for example, between anionic polysaccharides and proteins below their isoelectric point. Thus the sulphate groups of carrageenan can form electrostatic bonds with the positively charged amino groups of the basic sweet protein, thaumatin, leading to complex formation and a subsequent reduction in the sweetness intensity of thaumatin (2). Alternatively, polyanion-polycation interactions can produce an insoluble precipitate, like the acidic complex of gelatin and gellan in deionized water (3), behaviour widely used in encapsulation technology. In the absence of favourable interactions, concentrated solutions of two biopolymers will often immediately become cloudy, then resolve gradually into two clear layers, each containing most of one polymer and little of the other. Upon
Oxford University Press
272
IS. Chronakis. S. Kasapis and R. Abeysekera
cooling the mixture before bulk phase separation, composite gels are formed, with the network of the faster gelling biopolymer forming the continuous, supporting phase and the second species confined to discontinuous inclusions. Work on a number of biopolymer mixtures of relevance to the food industry has shown that the thermal regime dictates the extent of phase separation, the development of mechanical properties and the polymer composition at which phase inversion occurs in the blend (4). Slower cooling rates, e.g. a scan rate of l/min as opposed to quenching, reduce the overall levels of polymer needed for phase separation in a mixture, which can be rationalized on the basis of physical properties of the individual components (5). The gelatin-polysaccharide-water system has attracted the attention of food scientists and it has been demonstrated for >80 mixtures that steric exclusion or heterotypic interactions could be induced by changes in bulk concentration, pH, temperature and ionic strength (6). At relatively low polymer concentrations, thermodynamic incompatibility is the overriding influence in mixtures of gelatin with neutral polysaccharides. Thus for composite gels of 1% agar and 1% gelatin, light microscopy shows a continuous polysaccharide phase surrounding the protein inclusions, with the system phase inverting to a protein continuous network at -3% gelatin (7). On the other hand, steric exclusion between the positively charged gelatin chains (pH below its isoelectric point) and an anionic polysaccharide occurs only when the important electrostatic interactions between the two polymers are suppressed. Work on mixed systems of gelatin and gellan, a polysaccharide with a carboxyl group per repeat sequence, has documented the striking transformation from a precipitating coacervate in the absence of salt (3) to a clear composite gel in the presence of sodium cations (added to stoichiometric levels and beyond), which screen the negative charge of the gellan chain (8). Recently, the above reasoning was used to provide a baseline for the development of high solids confectionery products using gelatin-gellan formulations at levels of co-solute (sucrose plus corn syrup) of 60-85% in the mixture (9,10). Steric incompatibility between the two polymers resulted in a stable, two-phase system and a rubbery gellan network was obtained at temperatures close to the boiling point, which on subsequent cooling started transforming into a body with a glass-like consistency. In the present study we replaced gellan with pectin, a polysaccharide traditionally used for structuring high solids products (l I). To extend our knowledge of the effect of polyhydric compounds on the mechanical properties of mixed gels, ethylene glycol (EG) was used as the co-solute in part I of this investigation, with part II being a full account of the work on gelatinpectin-sugar-corn syrup mixtures.
Materialsand methods
The gelatin sample used was kindly supplied by Systems Bio-Industries (Baupte, France). It was the acid pretreatment
product of pigskin with an isoelectric point of pH 8. Stock solutions (25% polymer) were prepared by soaking the granules at room temperature overnight and then heating to 60C. Before each experiment, the pH of the single gelatin preparations (3% protein plus 0-70% EG) was adjusted to 3 with 2 mol/dm! HCl. The citrus peel pectin used was a commercial product kindly supplied by Hercules, LilIe-Skensved, Denmark (GENU B). It is a rapid set variety with a galacturonate content (esterified or nonesterified) of 83% and a high degree of esterification (70%). Since these commercial materials are 'standardized' to specific gel properties by blending with sucrose, they were dialysed extensively against distilled water (four changes) and freeze-dried to produce pure polysaccharide samples. The pH value in 1% pectin with 0-80% EG solutions was adjusted to 3 (2 mol/dm! HCl). For preparation of binary systems, appropriate amounts of gelatin and pectin solutions were mixed at 75C followed by addition of EG and the remaining distilled water to bring the sample to the required composition (stilI at pH 3). Differential scanning calorimetry (DSC) experiments were carried out using a Setaram batch and flow microcalorimeter. Single or mixed polymer preparations with ethylene glycol were weighed directly (-0.8 g) on a pan which was then sealed hermetically. A second pan containing the same amount of EG-water solution was used as a reference. Samples were equilibrated at 5C for 2 h and then heated to 90C at a scan rate of 0.2/min. Small deformation oscillatory measurements were made using a parallel plate geometry (40 mm diameter; I mm gap) on a controlled stress Carri-Med CSL 500 rheometer. The hot single or mixed solutions (75C) were loaded on the platen of the machine preset at the same temperature. The exposed edges of samples were covered with a light silicone oil (50 cs) to prevent evaporation. Temperature was controlled by a Pt 100 thermometer in a Peltier device. Network development was monitored on cooling to 5C at a scan rate of l/min and frequency of 1.6 Hz (-10 rad/s). The imposed strain was fixed at 1%. Strain sweeps on a few selected gels demonstrated that the working deformation was well within the linear viscoelastic region (extending up to 20% strain). Cooling was followed by an isothermal scan (5C) for 2 h and a frequency sweep between 0.01 and 10 Hz. The rheological routine was completed with a heating run from 5 to 95C, thus generating a complete picture of change in storage modulus (G), loss modulus (G') and dynamic viscosity (n"). For the microscopy work, samples were allowed to set up over 12 h at 5C. Pieces of dimension 10 x 10 x 0.5 mm were cut from the composite gels using a sharp scalpel. Unstained samples were placed on a microscope slide and a cover slip was lowered gently onto the gel surface, excluding any trapped air, and examined immediately. For the stained specimen, a drop of 0.05% w/w aqueous Sirius red (contains sulphonic acid groups that can react with basic groups of collagen) or Toluidine blue (a metachromatic dye that stains
Mixed gels of gelatin-pectin in the presence of ethylene glycol
273
- 0.2
- 1.4
- 3.6
-S
~
Q
- 0.3
- 1.5 - 3.7
G:
-0.4
- 1.6 - 3.8
Q,I
- 0.5
- 1.7
d
- 0.6
10
- 1.8
- 3.9 10
30
SO
70
30
50
10
- 3.4
- 2.6
- 2.4
s =
~ ~ ;:
2.5
- 3.6
2.8
- 3.8
- 2.6
- 2.7
e
- 4.0 10 -3.0 - 2.8 10
30
50
70
30
50
10
- 1.6
- 2.0
~
~
Q
-;:
E
- 1.7
- 2.05
....
~
- 1.8
- 2.1
- 1.9
- 2.15 - 2.2
10
- 2.0 30 50 70
f
10
30
50
70
Temperature (0C)
Temperature (OC)
Figure 1 Differential scanning calorimetry work on: 3% gelatin with 0 (a; left y axis) and 70% (a; right y axis) ethylene glycol; 1% pectin with
30 (b; left y axis) and 50% (b; righty axis) EG ; 1% pectin with 70% EG(c); and 3% gelatin plus 1% pectin with 30 (d), 50 (e) and 70%(f) EO. Heating scan rate of O.2/min.
274
I.S. Chronakis. SiKasapis and RAbeysekera
pectin) solution was placed on the upper surface of the composite gel. Ten minutes were allowed for staining to take place and then the samples were processed like their unstained counterparts. Images were acquired on TMax 100 photographic film using a Nikon inverted microscope Diaphot-TMD with bright field optics.
Results and discussion

Steric incompatibility versus heterotypic interactions
Mixing of clear gelatin and pectin solutions (pH 3) results in a turbid blend which indicates microscopic phase separation and formation of discontinuous inclusions capable of scattering visible light. Mixtures remain cloudy upon addition of ethylene glycol. Centrifugation of 3% gelatin-l % pectin with 40 or 50% EG at 45C (4600 g; 30 min) results in bulk phase separation with two discrete layers. The bottom layer is rich in pectin and has a characteristic creamy colour, whereas the top one is relatively colourless, as is typical for gelatin-rich preparations. DSC (Fig . la) provides thermodynamic information about the melting profiles of gelatin gels prepared in the absence and maximum level of ethylene glycol (70%). The presence of co-solute stabilizes the gelatin helices, with the temperature at maximum heat flow (Tmax) being shifted to 28C from -23C for an aqueous gelatin gel. The heating thermograms of pectin gels denote a less cooperative order-to-disorder transition which proceeds over the majority of the experimental temperature range (Fig. lb and c). The broad endotherms make it difficult to define accurately the mid-point of a transition but there is an apparent shift of Tmax values from -38 and 44 to 49C at ethylene glycol concentrations of 30, 50 and 70% respectively. For the same
amounts of co-solute (F ig. ld-f) the endothermic spectra of mixed gels reproduce in terms of Tmax position and general band form the sharp deconvolution of gelatin chains and the broader loss of order seen for individual pectin assemblies at higher temperatures. To recap, centrifugation shows evidence of phase separation in solution, and in the gel state calorimetry indicates that the two polymeric components form separate molecular associations in the manner observed for the single preparations. By contrast, heterotypic interactions between two biopolymers (e.g. in x-carrageenan-konjac mannan blends) have been shown to either generate a new endothermic peak or to distort the characteristic position and form of the single component transitions (12). A 70% ester pectin has a pKa of 3.6 (13), and on this basis Morris and co-workers (14) estimated that only 5.6% of the total uronate residues are ionized at pH 3. Furthermore, the reduction in gel strength at lower or higher degrees of esterification suggested that 70% is the ideal ester content for maximum stabilization of pectin's interchain junctions. The above findings support our experimental evidence of polymer exclusion and the ensuing homotypic intermolecular associations at the expense of potential electrostatic interactions between gelatin and pectin.
Effect of ethylene glycol on single-component gelatin gels
Recording the mechanical properties of gelatin and pectin networks as a function of increasing ethylene glycol concentration constitutes the most informative baseline for understanding the macromolecular organization of the mixed system. Although the first priority of this work was to examine the phase behaviour of blends, the same exercise gives some indications of the nature of the intermolecular interactions responsible for network formation.
3.5
80 70
3.5
c:l.o ;;:;
~ 2.5
'-'
o '-' so
::I
ll.I L.
60
40
- ~ .. 1
E 8'-'
2S . 2 1.5
~
~~
...:l
2 1.5
TEMP.
30 20 10 0 12000
e
ll.I
...:l
Eo<
~c
30
1Ie.!Vi
1"1
3000
6000
9000
10
20
40
Time (sec)
Temperature ('C)
Figure 2 Cooling-isothermal (a) and heating (b) profiles of G(.) and G'(O) for 3% gelatin with 70% ethylene glycol. Frequency of 1.6 Hz, scan rate of l/min. Between temperature runs samples were left at SoC for 2 h.
275
Figure 2a shows combined cooling and isothermal runs for 3% gelatin in the presence of 70% ethylene glycol. Network formation is indicated by a sharp increase in G, the onset of which has been taken as the gelation temperature (tgel). Samples were left at 5C for 2 h, thus allowing G to approach asymptotically a constant value (i.e. attainment of a 'pseudoequilibrium' modulus). The mechanical changes are reversible on heating, albeit shifted to higher temperatures, thus producing thermal hysteresis (Fig. 2b). As an index of network liquefaction, we have considered the temperature at which G' >G (tmel)' Figure 3 shows that the onset of a steep rise in G is not affected by increasing amounts of co-solute; tgel remains at 14 I"C. Similarly, the temperature on completion of structure melting remains constant throughout the experimental range of EG (tmel is 28 I"C). With increasing EG concentration, however, there is initially an increase in the G values at the end of the 2 h time sweep, with a substantial fall when the co-solute level rises to >30% w/w. As shown in Figure 4, the storage modulus of gelatin gels increases from 1.7 to 2.2 kPa at 0 and 30% EG in the blend, and then it drops to just below I kPa at 70% of the co-solute. To describe qualitatively the effect of ethylene glycol molecules on gelatin's supramolecular structure, one may start with the conception of an infinite network of triple helicesstabilized by neighbouring water molecules (15). Then one can consider the densimetric investigation of Gekko and
100
90
"LU
~
8-
80
->:
4 4
Timasheff, which demonstrated that addition of several different proteins to an aqueous glycerol solvent makes the polymer-solvent interactions thermodynamically unfavourable and increases the chemical potential of glycerol (16). They proposed that glycerol enhances the self-association of protein molecules by being preferentially excluded from their immediate domain. On the basis of the non-specific nature of the aforementioned process, it was suggested that the same mechanism determines the interactions of a number of polyols with gelatin (17). Therefore, a possible explanation of the effect of EG on gelatin is that the co-solute is excluded from the domain of the protein, which results in an increase in the chemical potential of ethylene glycol. Consequently, the surface of contact between gelatin and solvent is minimized, the polymer-EG interaction becomes thermodynamically unfavourable and the polypeptide chain folds increasingly in the helical form which creates extra junction zones and results in the initial rise of the storage modulus seen in Figure 4. Cramer and Truhlar (18), using quantum chemical conformational analysis, have looked at the prevailing type of conformer and hydrogen bond in ethylene-glycol-water solutions. Calculated relative solvation free energies showed that 92% of the rotamers were in the gauche form and 77% of the hydrogen bonds were internal. They concluded that the intramolecular hydrogen bonds found between vicinal hydroxyl groups in the gas phase are only slightly disrupted in aqueous solutions in order to permit additional intermolecular hydrogen bonding with foreign hydroxyl groups. Obviously, a limited extent of intermolecular
4.5
U 70
5 60 ...
... 50
!4O
30 20 10
0
~_':~-~~-3
0 10 20 30 40 50 60 70 80 90 2.5
--o
10
Ethylene Glycol (% w/w)

(tgel) and melting (tmel) temperatures as a function of ethylene glycol content for: 3% gelatin with tgel and tmel values at 14 and 28C respectively (--); 1% pectin with the tgel depicted as a solid-line curve and the tmel as a long arrow; 3% gelatin plus 1% pectin showing the tgel as (~) and the tmel as (A), with the short arrow indicating that mixed gels become thermally irreversible like their pectin counterparts.
20
30
40
50
60
70
80
Figure 3 Gelation
Ethylene Glycol (% w/w)

Figure 4 Storage modulus (G) development with increasing levels of ethylene glycol for 3% gelatin (--), 1% pectin (-), and 3% gelatin plus 1% pectin (.). Values of G were taken at the end of the 2 h isothermal sweep (SoC).
276
IS. Chronakis, S.Kasapis and RAbeysekera
hydrogen bonds between EG and gelatin is the chemical basis for thermodynamically unfavourable interactions between the two components argued in the preceding paragraph. The junction zones of gelatin chains cannot be stabilized only by hydrogen bonds between the polypeptide chains. Calorimetric work on the collagen's helix-to-coil transition has shown that the enthalpy of elastically active network chains increases linearly with imino acid content (15). Steric considerations dictated that the stabilization was due to hydrogen bonds between the imino acid groups and water molecules adjacent to the collagen structure. Replacing water bridges with rather inert ethylene glycol molecules would increasingly destabilize the helices and weaken the gelatin network in the way observed in Figure 4 for co-solute levels beyond 30%. In other words, we suggest that the reduction in gel rigidity is not the result of a secondary effect at higher structural level (i.e. excessive aggregation, effective dehydration and formation of precipitated gels) but rather the outcome of reduction in the extent of intermolecular bonding in gelatin chain. This idea is in agreement with the data of Oakenfull and Scott (17), who reported a reduction in the proportion of ordered polypeptide chains involved in the formation of junction zones from 60 to 55% as a result of increasing the EG content in the mixture (from 35 to 45%). Finally, the contrasting behaviour of T max (Fig. la) and tmel (Fig. 3) as a function of ethylene glycol concentration might reflect respectively the molecular conformation or energetic content of the helical structure of gelatin segments, and the absence of aggregated junction zones (of the kind seen in polysaccharide gelation) responsible for the additional thermal stability of a three dimensional network. Thus increasing amounts of ethylene glycol might encourage polymer-polymer interactions and impart thermal stability to the triple helix (rise of Tmax in DSC experiments) but the lack of extensive inter-helical association in gelatin systems would prevent the formation of junction zones in the form of crystalline aggregates with increasing thermal hysteresis and higher tmel values during heating on the rheometer.
Effect of ethylene glycol on pectin gels
4,-------------...,
o
~
~aa a
80
20
40
60
tOO
Temperature (DC)
Figure 5 Storage (G) and loss (G') moduli variation for 1% pectin with 70% ethylene glycol on cooling [G (.); G' (D)] and heating (G (L); G' (A)]. Frequency of 1.6 Hz, scan rate of IOlmin.
Temperature sweeps of the storage and loss moduli and their variation with frequency for 1% pectin plus 25% EG showed that the 'phase lag', 8, (tan 8 = G'IG) is >45, a result which indicates the absence of a continuous, three-dimensional structure. At higher amounts of co-solute (~30%), however, networks are formed with a temperature course of gelation and thermal stability of resulting gels that are entirely different from those of the gelatin counterparts. Figure 5 demonstrates the cooling and heating profiles of a pectin sample at 70% co-solute. Gelation is characterized by a sharp increase in solid-like character at a tgel value of -58C. Samples are left at 5C without significant enhancement of storage modulus, and during subsequent heating they maintain substantial structure even at 95C, thus showing that there is a large degree of thermal hysteresis between gel formation and dissociation in high methoxy
pectin networks. Figure 3 shows the thermal stabilization of pectin's associations, in terms of tgel and t me!> with increasing concentration of ethylene glycol. In Figure 4 the improved gelation observed from 30 to 60% EG (the G at 5C is 0.6 and 7.2 kPa respectively) is followed by a weakening of the gel structure at higher levels of co-solute (eg. G drops to 4 kPa at 80% EG). The above mechanical evidence reveals a reduction in gel rigidity under high ethylene glycol conditions (>60% in Fig. 4) occurring in parallel with an increasingly stable ordered structure (progressiverise in tgel in Fig. 3). This paradox may be resolved if we consider that network development in high methoxy pectins involves formation of 3-fold right-handed helices (19) which aggregate to form supramolecular ordered structures (20). Computer modelling indicates two types of bonding holding together these structural elements in triangular packing arrangements: hydrogen bonds between hydroxyl groups of neighbouring ordered chains and clustering of methyl ester groups within a calyx of water molecules. According to this model, the enhancement of viscosity with increasing degree of esterification is due to more effective structuring of water molecules around the methyl groups. The acidic environment of this investigation (pH 3) protonates most of the unesterified carboxylate groups, minimizes electrostatic repulsions between chains and allows additional intermolecular hydrogen bonding. Obviously, hydrogen bonds between pectin segments and adjacent water molecules also occur with the effect of counterbalancing the crystalline polymer-polymer aggregation. As argued for the gelatin case, ethylene glycol is a poor solvent for hydrophilic
277
polymer parts. Thus, the co-solute reduces the water activity and enhances the polymer-polymer hydrogen bonding , with the concomitant increase in network strength seen for pectin gels with an EG content of up to 60% (Fig. 4). However, the existence of methyl substituents introduces a second dimension in the effect of ethylene glycol on pectin gelation since the co-solute is a good solvent for hydrocarbons and related hydrophobic assemblies (21). High levels of EG , therefore, will facilitate disruption of the water cages and dissolution of the methyl clusterings which are a requisite of gel strength, and this is therefore reduced beyond 60% co-solute (Fig. 4). Overall ethylene glycol promotes helix
formation and antagonizes hydrophobic aggregation . Its direct involvement in the gelation process of pectin might explain the continuous stabilization of the ordered structure, judged by the progressive increase in temperature at which it forms, as compared with the non-specific exclusion with gelatin that generates a constant t gel value.
Phase phenomena in gelatin-pectin-ethyIene glycol gels
The experimental approach of the investigation in single gelatin and pectin systems was also employed in the phase characterization of their mixed gels. Figure 6 reproduces the
3.5 3
3.5 3
~.....
..,mCCa:J
80 70
,-...
,-...
g.,
~
ell
~ 2.5
'-'
s
Q
OIl
..J
2 1.5
.~ lID
I
ca::JCP"
60 U 0
,-...
50 ~
40
'-'
E 8-
~ i
ell
2.5 2 1.5
'-'
0 3000
ItI
<U 30 Eo-
~ ..J
~.
c
20
TEMP.
10 0 12000 0 10 20
~ ItI
lIgi c
30
6000
9000
40
Time (sec)
Temperature ~C)
Figure 6 Cooling-isothermal (a) and heating (b) profiles of G (_) and 0" (D) for 3% gelatin plus 1% pectin with 20% ethylene glycol. Frequency of 1.6 Hz, scan rate of la/min . Between temperature runs samples were left at SoC for 2 h.
4.5
80
4.5
::. "3
'C
Q
g.,
CIl
3.5 3 2.5
2
---
70 60 50
40
'-'
U 0

Cc
~ ..J
S '-'
= Cl.
<U
~ ;;;.
3.5 3 2.5 2 1.5

DC
C
30 S 20
TEMP.
"3 'g S
ccc
..J
Cc cCaa C a
CC~ \
1.5
10 0 12000
IlJ
100
3000
6000
9000
20
40
60
80
Time (sec)
Temperature ~C)
Figur e 7 Cooling-isothermal (a) and heating (b) profiles of G (_) and G' (D) for 3% gelatin plus 1% pectin with 70% ethylene glycol. Frequency of 1.6 Hz, scan rate of la/min . Between temperature runs samples were left at 5C for 2 h.
278
IS. Chronakis, S. Kasapis and R Abeysekera
progression of gelation and melting with decreasing, constant and increasing temperature at 20% ethylene glycol. The temperatures for the onset of the sol-gel process and completion of the gel-sol transition (17 and 28C respectively) are similar to tgel and tmel values of single gelatin preparations. Increasing concentrations of co-solute, however, alter dramatically the time-temperature profile of the gelatin-pectin blend. As shown in Figure 7a, the sol-gel transition moves to higher temperatures (tgel is 58C at 70% EG) and a second 'wave' of structure appears at -15C. During subsequent heating (Fig. 7b) there is a partial loss of structure at low temperatures but then a broad endothermic process takes over which sees a predominant elastic behaviour even at the highest temperature available to the rheometer (i.e. tmel is >95C). The immediate conclusion from this evidence is that a high ethylene glycol content in the blend promotes the formation of a pectin network and gelation of gelatin at the temperature ranges associated with the structuring of the individual components. Similarly, the bimodal heating profile comprises a melting step of the thermally metastable gelatin structure and a gradual weakening of the pectin network. Figures 3 and 4 map out the effect of ethylene glycol on the gelation temperature, melting process and elastic modulus at 5C for our system. At low levelsof co-solute (0-25%), where pectin shows no evidence of network formation, the tgel and tmel values of the mixtures remain close to those of single gelatin gels. The rigidity of the blends, however, diverges from the mechanical strength of gelatin networks with increasing ethylene glycol concentration. Taking into account that thermodynamic incompatibility determines the phase behaviour of this mixture, increasing amounts of EG should promote pectin's conformational ordering which becomes more and more effective in excluding gelatin, hence creating a more concentrated and stronger protein continuous phase. The mechanical analogue of this composite system is that of an isostrain arrangement (5), where a strong continuous phase (gelatin) is penetrated by much weaker discontinuous inclusions (pectin). The structuring and melting characteristics of samples with a higher EG content (from 30 to 70%) follow the onset of gelation and the extent of thermal hysteresis observed for single pectin gels. The change in the pattern of tgel and tmel values in Figure 3 is coincident with a sharp increase in gel strength from 3.4 to 12.3 kPa occurring at 25 and 30% co-solute respectively (Fig. 4). Obviously, gelation of pectin at 30% EG transforms the filler inclusions of pectin, argued for lower levels of co-solute, into a continuous network. Of course, the topology of gelatin in these mixtures comes into question. A continuous gelatin structure that interpenetrates the pectin network, with both polymers not 'seeing' each other, will create a type of mixed gel known as interpenetrating networks (22). In this case, both networks span the entire system, with each one effectively having a phase volume equal to one. The modulus of the blend should be estimated from the sum of the moduli of the individual
components. Scanning through the data of Figure 4, however, reveals that for our system this is not the case; for example, the mechanical strength of gelatin, pectin and the mixed preparation at 50% EG is 1.5, 6.3 and 15.5 kPa respectively. Therefore, thermodynamic incompatibility between the two polymers prevents the formation of two separate interpenetrating networks, a result which is in agreement with the steric exclusion phenomena in mixed solutions of pectin and gelatin, observed in the form of micro- and bulk phase separation. Steric incompatibility in the second range of EG samples (30-70%) may result in phase separation in the form of a gelatin filler surrounded by a continuous pectin phase (composite gel) or of a bicontinuous arrangement. In either case, each component tends to exclude the other from its domain, the solvent is partitioned between the two phases and the effective concentrations of both polymers are raised. As a result the sum of moduli of the effective concentrations, scaled down by the phase volume of each network, can accommodate the high values of the storage modulus recorded for the binary gels of Figure 4. Ouantitative analysis of mechanical properties for phase-separated biopolymer gels has been attempted in the past (7,23), based on the isostress and isostrain equations of the blending laws (5) and the concentration-dependence of the modulus of the single gelling components (24). In doing so, a computerized procedure was devised to assess every possible distribution of water between the two phases. However, the introduction of a fourth component (ethylene glycol) to the analysis will require a rather complicated computerization to address the problem of solvent partition, and a combined polymer concentration-ethylene glycol-storage modulus relationship Which, in our opinion, may prove difficult to achieve and to interpret. Qualitatively, our understanding of the macromolecular organization at high levels of co-solute was refined by using histochemical stains to identify the gelatin and pectin phases in conventional transmission microscopy. Gross differences in images were detected either as a reduction in the intensity of monochromatic light on passing through optically dense regions of the unstained specimen or by the colours associated with different structures when Sirius red (gelatin-specific) and Toluidine blue (which labels pectin) were used. The emerging picture of the microscopy work is as follows: individual polymer gels under conditions of low or high ethylene glycol show homogeneous networks across the whole sample, which undoubtedly constitute a single phase. Mixed polymer gels at low co-solute concentrations show structures which are similar to the molecular organizations of single gelatin-EG preparations. The absence of a structured pectin image is, of course, in accordance with the rheological evidence that pectin does not form a gel at levels of co-solute of <30%. Greater amounts of co-solute promote gelation of pectin, large assemblies of which are clearly detectable in mixed systems. Figure 8 shows the pattern of phase separation at
279
4. Kasapis,S., Alevisopoulos,S., Abeysekera,R., Manoj,P., Chronakis,I.S. and Papageorgiou,M. (1996) In Phillips,G.O ., Wedlock,D.l and Williams,P.A. (eds), Gums and Stabilisers for the Food Industry, 8. Oxford University Press, Oxford, pp. 195-206. 5. Takayanagi,M., Harima,H. and Iwata,Y. (1963) Mem.
Fac. Eng. Kyusha Univ. , 23, 1-13 . 6. Tolstoguzov.Vb. (1990) In Phillips,G.O., Wedlock,D.l and Williams,P.A. (eds), Gums and Stabilisers for the Food Industry, 5. Oxford University Press, Oxford, pp. 157-175.
7. Clark,A.H., Richardson,R.K., Ross-Murphy,S.B. and Stubbs,lM. (1983) Macromolecules, 16, 1367-1374. 8. Papageorgiou,M., Kasapis,S. and Richardson,R.K. (1994) Food Hydro coli., 8, 97-112. 9. Papageorgiou,M. and Kasapis,S. (1995) Food Hydrocoll.,
9,211-220. 10. Papageorgiou,M ., Kasapis,S. and Richardson,R.K. (1994) Carbohydr. Polym., 25, 101-109. 11. Rolin,C. (1994) The European Food and Drink Review, Autumn, 61-65. 12. Williams, P.A., Clegg,S.M., Langdon,M.l, Nishinari,K. and Piculell,L. (1993) Macromolecules, 26,5441-5446. 13. Plaschina,I.G. , Braudo,E.E. and Tolstoguzov.VB. (1978) Carbohydr. Res., 60, 1-8. 14. Morris,E.R., Gidley,M.l, Murray,E.l, Powell,D.A. and Rees,D.A. (1980) Int. 1 Bioi. Macromol., 2, 327-330. 15. Privalov,P.L. and Tiktopulo,E.I. (1970) Biopolymers, 9, 127-139. 16. Gekko,K. and Timasheff,S.N. (1981) Biochemistry, 20, 4667-4676. 17. Oakenfull,D. and Scott,A. (1986) Food Hydrocoll., 1, 163-175. 18. Cramer,C.l and Truhlar,D.G. (1994) 1 Am. Chem. Soc., 116, 3892-3900. 19. Walkinshaw,M.D. andArnott,S. (1981) 1 Mol. Bioi., 153, 1075-1085. 20. Rolin,C. (1993) In Whistler,R.L. and BeMiller,lN. (eds), Industrial Gums. Academic Press, London, pp. 257-293. 21. Tanford,C. (1980) The Hydrophobic Effect. Formation of Micelles and Biological Membranes. 1 Wiley and Sons
Figure 8 Light microscopy picture of a mixed gel at 3% gelatin plus

1% pectin and 60% ethylene glycol. The magnification is x200 .
60% ethylene glycol, with the Toluidine blue staining the
irregularly shaped particle-like structures of pectin (with an average size of Sum). The unstained continuous background is due to the gelatin structure. Treatment of the sample with only Sirius red stains this continuous phase exclusively. Therefore, the light microscopy observations strongly suggest that there is a continuous gelatin network in addition to the three-dimensional structure of pectin detected from the mechanical measurements. Furthermore, we have learnt by experience that composite systems of a well defined protein filler and a continuous polysaccharide assembly, that bind clearly to specific stains, should show an intense staining contrast. However, in the case of gelatin and pectin the colour contrast across the sample, from one region to another, was clearly quite faint, supporting the argument for two closely intertwined networks in the form of a bicontinuous system.
Acknowledgements
We are grateful to Dr Claus Rolin/Copenhagen Pectin Division of Hercules Inc. for providing analytical values for the citrus pectin samples, and to our colleagues Prof. E.R .Morris and Dr M .WN.Hember for stimulating discussions.
Inc., New York.

22. Morris, V.I (1986) In Phillips,G.O., Wedlock,D.l and Williams,P.A. (eds), Gums and Stabilisers for the Food Industry, 3. Elsevier, London, pp. 87-99. 23. Kasapis,S. (1995) In Harding,S.E., Hill,S.E. and Mitchell,lR. (eds), Biopolymer Mixtures. Nottingham University Press, Nottingham, pp. 193-224. 24. Ross-Murphy,S.B. and McEvoy,H. (1986) Br. Biopolym. 1 ,18,2-7.
References
1. Jones,R. (1995) Physics World, March, 47-51. 2. Ohashi,S., Ura ,F., Takeuchi,M., Iida,H., Sakaue,K., Ochi,T., Ukai,S. and Hiramatsu.K. (1990) Food
Hydro coll., 4, 323-333 . 3. Chilvers,G.R. and Morris,Y.l (1987) Carbohydr. Polym., 7, 111-119.
Accepted on December 4, 1995

Pektin & Gelatin

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Pektin & Gelatin

Uploaded by

Copyright:

Available Formats

Food Hydrocolloids 15 (2001) 643653

Texture and structure of gelatin/pectin-based gummy confections

L.L. DeMars, G.R. Ziegler / Food Hydrocolloids 15 (2001) 643653

L.L. DeMars, G.R. Ziegler / Food Hydrocolloids 15 (2001) 643653

L.L. DeMars, G.R. Ziegler / Food Hydrocolloids 15 (2001) 643653

(Bagley, 1983; Tschoegel, Rinde, & Smith, 1970)   2D L e ln 1 1 Ca

L.L. DeMars, G.R. Ziegler / Food Hydrocolloids 15 (2001) 643653

Fig. 3. Stress at fracture as a function of gelatin and pectin concentration.

Fig. 4. Strain at fracture as a function of gelatin and pectin concentration.

L.L. DeMars, G.R. Ziegler / Food Hydrocolloids 15 (2001) 643653

Fig. 5. Light micrographs of gelatin:pectin gels: A. 3.0:1.0, B. 4.5:1.0 (bar 10 mm).

L.L. DeMars, G.R. Ziegler / Food Hydrocolloids 15 (2001) 643653

Sour Sweet Fruity Sweet Sour Fruity Sweet Cherry Tart

Sweetness Fruitiness Tartness Chemical Sour Cherry Sweetness

Figures in parentheses refer to vector loading for individual assessors.

Figures in parentheses refer to vector loading for individual assessors.

Fig. 7. Consensus plot for gel texture attributes.

L.L. DeMars, G.R. Ziegler / Food Hydrocolloids 15 (2001) 643653

Fig. 8. Consensus plot for gel avor attributes.

Food Hydrocolloids 14 (2000) 579590 www.elsevier.com/locate/foodhyd

T.S. Nordmark, G.R. Ziegler / Food Hydrocolloids 14 (2000) 579590

T.S. Nordmark, G.R. Ziegler / Food Hydrocolloids 14 (2000) 579590

Fig. 1. Fluorescent labeling of LM pectin.

Fig. 2. Fluorescent labeling of gelatin.

T.S. Nordmark, G.R. Ziegler / Food Hydrocolloids 14 (2000) 579590

Fig. 3. The procedure for separation of phases by centrifugation.

T.S. Nordmark, G.R. Ziegler / Food Hydrocolloids 14 (2000) 579590

T.S. Nordmark, G.R. Ziegler / Food Hydrocolloids 14 (2000) 579590

T.S. Nordmark, G.R. Ziegler / Food Hydrocolloids 14 (2000) 579590

250 200 Red Va lue 150 100 50 0 1 9 17 25 33 41 49 57 65 73 81 89 97

Threshold point slightly above critical point

Bulk composition of mix Composition of phase

T.S. Nordmark, G.R. Ziegler / Food Hydrocolloids 14 (2000) 579590

T.S. Nordmark, G.R. Ziegler / Food Hydrocolloids 14 (2000) 579590

T.S. Nordmark, G.R. Ziegler / Food Hydrocolloids 14 (2000) 579590

T.S. Nordmark, G.R. Ziegler / Food Hydrocolloids 14 (2000) 579590

Carbohydrate Polymers 34 (1997) 309-321

Structural properties of pectin-gelatin gels. Part II: effect of sucrose/glucose syrup

I.M. Al-Ruqaie et al.

5.0 4.5 4.0 3.5 0.001 0.01 0.1 1 10

5.0 4.5 4.0 3.5 0.001 0.01 0.1 1 10

1.08 6.5 3 $ 2 2 5.0 6.0 5.5

0.96 3.5 I -3 3.5 I -2 -1 0 1 2 3 4 -3 1 -2 1 -1 I 0 I I I 2 I 3 : 0.94 4

(%) 6713 7352 8637 9084 9239 9553 6909

(5OC) 0.011 0.012 0.014 0.016 0.017 0.058 0.156

of glucose syrup refers to dry solids.

Pseudoequilibrium modulus (G; Pa) 5C

W) 1628 2177 2776 2966 2868 2904 2467 1869 1159000

of glucose syrup refers to dry solids.

heating run (a) and a frequency sweep (b). Conditions as in Fig. 3.

I.M. Al-Ruqaie et al.

Pseudoequilibrium modulus (G; Pa) 5C

(%) 102 85 139 181 130 139 87

of glucose syrup refers to dry solids.

Fig. 7. Variation in storage and loss moduli for the mixture

2.0 -r 0 50 100 I50 200

Vol. 11 no. 3 pp. 271-279, 1997

Structural properties of gelatin-pectin gels. Part I: Effect of ethylene glycol

Oxford University Press

IS. Chronakis. S. Kasapis and R. Abeysekera

Mixed gels of gelatin-pectin in the presence of ethylene glycol

I.S. Chronakis. SiKasapis and RAbeysekera

(Bagley, 1983; Tschoegel, Rinde, & Smith, 1970) 2D L e ln 1 1 Ca