Inclusion of starch in imitation cheese: Its influence on water mobility and cheese functionality

ARTICLE IN PRESS

FOOD

HYDROCOLLOIDS Food Hydrocolloids 22 (2008) 1612–1621 www.elsevier.com/locate/foodhyd

Inclusion of starch in imitation cheese: Its influence on water mobility and cheese functionality N. Noronhaa, E. Duggana, G.R. Zieglerb, E.D. O’Riordana,, M. O’Sullivana a

School of Agriculture, Food Science and Veterinary Medicine, College of Life Sciences, University College Dublin, Belfield, Dublin 4, Ireland b Department of Food Science, The Pennsylvania State University, 341 Food Science Building, University Park, PA 16802, USA Received 9 August 2007; accepted 7 November 2007

Abstract The impact of starch type and concentration on the nature of water in and the rheology of imitation cheese were investigated. Imitation cheese (55% moisture) containing four starches (native, pre-gelatinised, resistant or waxy corn) at inclusion levels of 1.9%, 3.9%, 5.8%, 7.8%, or 9.9% w/w were manufactured using a Brabender Farinograph-Es. The textural properties were assessed by torsion gelometry and dynamic rheology and the mobility of water by nuclear magnetic resonance (NMR) relaxation techniques. Cheese microstructure was assessed using light microscopy. Increasing the starch content changed the texture of cheeses from ‘soft’ to ‘brittle/ tough’ and significantly (po0.05) decreased the mobility of water. Cheese melt and hardness were influenced by the mobility of water. Matrices in which the water was more mobile produced good melting soft cheeses, while cheeses in which water was less mobile were tough and non-melting. Light micrographs showed that starch type influenced cheese microstructure. The native and pre-gelatinised starches became swollen and disrupted the continuity of the protein matrix, separating the matrix into a protein and starch phase. Resistant and waxy corn starches were present in the protein matrix as small discrete particles, appearing relatively intact, unswollen and relatively unchanged by the cheese manufacturing process. The study indicates that varying the level/type of starch alters the water mobility and thus the functionality of imitation cheeses. r 2007 Elsevier Ltd. All rights reserved. Keywords: Imitation cheese; Starch; NMR relaxometry; Light microscopy; Torsion

1. Introduction Consumer preference for tailor-made convenience products using molten cheese, i.e. pizza pie topping, sauces have necessitated the production of cheese analogues; imitation Mozzarella cheese being the product of choice (Jana & Upadhyay, 2003). The stability of imitation cheeses with respect to shredability, apparent viscosity and free oil during refrigerated storage make them appealing to food processors and food-service industries. Native and modified starches have been used to modify the functional applications of imitation cheese (Mounsey & O’Riordan, 2001, 2008, respectively). Native starch granules are insoluble in cold water; however with continued heat these granules swell and imbibe water (Davies, 1995). Corresponding author. Tel.: +353 1 7167016; fax: +353 1 7161147.

E-mail address: [email protected] (E.D. O’Riordan). 0268-005X/$ - see front matter r 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.foodhyd.2007.11.007

Pre-gelatinised starches are modified starches that are cold water soluble, which have immediate viscosity effects (Light, 1990). Both pre-gelatinised and native starches have a tendency to retrograde on cooling, with the linear amylose molecules re-associating, forming hydrogen bonds, which results in the formation of a gel. Waxy starch develops viscosity without the gelling characteristics generally associated with native starches (Thomas & Atwell, 1999). Resistant starch is a starch that has been designed to withstand gelatinisation (granule swelling) under most heating regimes (Sajilata, Singhal, & Kulkarni, 2006). Work conducted previously in this laboratory used native, pre-gelatinised and waxy maize starch to replace rennet casein in imitation cheese (Mounsey & O’Riordan, 2001). These authors suggested that starch type had an impact on the functionality of starch-containing imitation cheeses, with native starch reducing casein hydration to a

ARTICLE IN PRESS N. Noronha et al. / Food Hydrocolloids 22 (2008) 1612–1621

lesser extent than the pre-gelatinised starch. Recently our laboratories have produced a variety of imitation cheeses in which resistant starch has been used as a fat replacer (Montesinos, Cottell, O’Riordan, & O’Sullivan, 2006; Noronha, O’Riordan, & O’Sullivan, 2007). Incorporation of relatively high levels (17.3% w/w) of resistant starch had minimum effects on cheese functionality, e.g. meltability, which was attributed to the low water binding capacity of the resistant starch used in that study (Duggan, Noronha, O’Riordan, & O’Sullivan, 2008). Although work to date has hypothesised that starch addition influences the properties of imitation cheese by affecting casein hydration, there has been no direct evidence to support this theory. Many researchers have found that the mobility of water, as measured by nuclear magnetic resonance (NMR) relaxometry relates to the availability of water in complex systems (Chen, Long, Ruan, & Labuza, 1997; Ruan et al., 1997; Schmidt, 1990). In relaxometry, the spin-lattice and transverse relaxation times (T1, T2) are related to the mobility of water molecules with very mobile water molecules taking longer to reach their equilibrium state, thus having long relaxation times. The objective of this study was to replace different levels of fat with starches (native, pre-gelatinised, resistant, or waxy corn starch)—chosen on their ability to bind water differently—on certain functional properties of imitation cheese. In addition the effect of this replacement on the hydration characteristics of cheese was examined using NMR.

1613

Batches (150 g) were manufactured by agitating the vegetable oil with water in the Brabender Farinograph-Es (mixer bowl type: sigma mixer S50, mixer speed: 63 rev min1) (Brabender Instruments, Inc., S. Hackensack, NJ, USA) at 50 1C for 2 min. The temperature of the mixer jacket was regulated by an external water bath (55 1C). Trisodium citrate, disodium phosphate, sodium chloride and sorbic acid were then added and mixed for a further 2 min. The rennet casein was subsequently added, and following 2 min of mixing the temperature was increased to 80 1C by closing the valve to the water bath at 55 1C and opening the valve to another water bath set at 95 1C. The mixture was maintained at 80 1C until a homogeneous cheese mass was produced and no pockets of water were observed on visual inspection. Citric acid was added and mixed in for 1 final minute. The product was discharged into a plastic box, covered and then placed into a refrigerator at 4 1C for 24 h, after which the cheese was vacuum packed (Model C10H, Webomatics, Bochum, Germany) and stored at 4 1C until analysed. All samples were analysed within 1 week of manufacture. Using a similar process, imitation cheeses, containing 1.9%, 3.9%, 5.8%, 7.8%, or 9.9% w/w native, pregelatinised, resistant and waxy corn starch were manufactured. The starches were added in direct replacement (on a weight basis) of hydrogenated palm oil in the control formulation. The starch was incorporated into its respective batch once the temperature of 80 1C had been reached. The rest of the procedure was the same as for the control cheese.

2. Material and methods 2.1. Materials Kerry Ingredients Ltd. (Listowel, Co. Kerry, Ireland) provided rennet casein (80% protein). Hydrogenated palm oil and rapeseed oil were obtained from Trilby Trading (Ireland) Ltd. (Drogheda, Co. Louth, Ireland). Native (Numould), pre-gelatinised (Instant Pure-cote), resistant (Novelose 240) and waxy (Amioca) corn starches were all obtained from National Starch Food Innovations (National Starch and Chemicals, Bridgewater, NJ, USA). All chemicals, including anhydrous disodium phosphate (Albright and Wilson Ltd., Cheshire, England), trisodium citrate and anhydrous citric acid (Jungbunzlauer GmbH, Pernhofen, Austria), sodium chloride (Salt Union, Cheshire, England) and sorbic acid (Hoechst Ireland Ltd., Dublin, Ireland) were of food grade. 2.2. Manufacture of imitation cheese A control imitation cheese was manufactured according to the following composition (expressed as weight percentage (% w/w)): 54% water, 21.7% protein, 21.0% vegetable oil, 0.9% trisodium citrate, 0.4% disodium phosphate, 1.4% sodium chloride, 0.5% citric acid and 0.1% sorbic acid.

2.3. Compositional analysis of imitation cheese The moisture content of imitation cheese was determined by the oven drying method (IDF, 1958) and the pH was measured by inserting a calibrated Unicam glass/Ag/AgCl combination pH electrode attached to a pH meter (model 9450, Unicam, Cambridge, UK) directly into the cheese at three randomly chosen locations after equilibration of imitation cheese to room temperature for at least an hour. The nitrogen content of imitation cheese samples was determined by combustion in a nitrogen analyzer (Leco, St. Joseph, MI) according to manufacturer’s instructions. A Kjeldahl factor of 6.38 was used to calculate crude protein. 2.4. Differential scanning calorimetry (DSC) The gelatinisation temperatures of native, resistant, waxy and pre-gelatinised corn starch were assessed. Duplicate aqueous samples (10% (w/w) of 35 mg wet weight) were prepared in sealed stainless steel differential scanning calorimetry (DSC) pans (60 ml; Perkin-Elmer Instruments, Norwalk, CT). Each sample was then heated in a DSC 7 (Perkin-Elmer Instruments, Norwalk, CT) from 10 to 180 1C at 10 1C/min.

ARTICLE IN PRESS 1614

N. Noronha et al. / Food Hydrocolloids 22 (2008) 1612–1621

2.5. Water activity of imitation cheese Water activity was measured using a Decagon CX 1 (Decagon, Pullman, WA, USA). Cheese samples were grated and sealed in an airtight container and were allowed to equilibrate to 25 1C (approx. 3 h). Samples were measured at 25 1C on the day of preparation. 2.6. Nuclear magnetic resonance (NMR) All T1 and T2 experiments were performed using a Minispec mq 20 spectrometer (Bruker Spectroscopy, Milton, Ontario, Canada). The resonance frequency of 1 H was 19.475 MHz. The NMR spectrometer was heated or cooled by a constant gas flow (air or liquid nitrogen) delivered by a variable temperature unit. Grated samples (0.5 g) were placed in sealed NMR tubes (10 mm diameter). All measurements were conducted at 2570.5 1C. The classical inversion-recovery method using 1801-t-901 pulse sequence was performed to determine T1 (Ernst, Bosedhausen, & Wokaun, 1991). The repetition delay was chosen to be at least five times T1. The 901 pulse width was 7.7 ms. Eight scans were acquired for each measurement. T2 relaxation times were obtained using the Carr, Purcell, Meiboom and Gill (CPMG) pulse sequence with a relaxation delay of 2000 ms and a 90–1801 pulse gap of 1.0 ms. The 901 pulse was 2.20 ms and 1801 pulse was 4.52 ms in length. A total of 2094 points were acquired for each of the eight scans. T2 distributions were calculated using the CONTIN Laplace transform algorithm (Provencher, 1982) and a Gaussian model was used to fit the resulting components. 2.7. Light microscopy Samples were analysed by light microscopy (BX41, Olympus, Melville, NY) using objective lenses of 10  , 20  and 40  , connected to a digital camera (SPOT II, Bioscan, Warrendale, PA). Micrographs (bright field and polarised) were taken using the auto-exposure function. Cryo-sections 8 mm in thickness were taken from cubes of cheese (10 mm3) using a cryo-microtome (Bright Instrument Company Ltd., Huntingdon, England). The sections were placed on glass slides, and then dried in an incubator at 30 1C for 24 h. Following drying, three components of the imitation cheese specimens, protein, fat and starch were sequentially stained greenish blue, red and black using Fast Green (Clark, 1980), Sudan III (Drury & Wallington, 1967) and iodine, respectively. The specimens were first fixed with 2.5% aqueous glutaraldehyde solution for 3 min and subsequently well rinsed with distilled water. The fixed samples were then flooded with aqueous Fast Green (0.5 g 100 mL1) for 1 min, to stain protein and then rinsed with distilled water. The samples were subsequently well rinsed in aqueous triethyl phosphate (60 g 100 mL1) (TEP), then immersed

in a solution of Sudan III (1 g 100 mL1 TEP) for 18 min to stain the fat. Finally, to allow the examination of starch, sections were stained with iodine (0.3 g I2, 1.0 g KI 100 mL1) for 30 s and then rinsed with distilled water. 2.8. Dynamic rheology test Dynamic oscillatory measurements were performed on a controlled strain rheometer (model ARES, TA Instruments Inc., Schaumburg, IL, USA) fitted with a 25 mm parallel plate with a 1.8 mm gap according to the method of Mounsey and O’Riordan (1999). Disc-shaped samples of cheese (25 mm diameter, 1.9 mm thick) were prepared using a slicer and punch. The samples were placed on the lower plate and compressed 0.1 mm to prevent slippage. A very thin film of mineral oil surrounded cheese samples to avoid dehydration during analysis. All measurements were made at a frequency of 1 Hz and a strain of 0.1%, previously established to be within the linear viscoelastic range (not shown). The temperature of the samples was increased from 25 to 90 1C at 5 1C/min using a Peltier heating element, during which time the parameters storage modulus (G0 ), loss modulus (G00 ) and the loss angle (tan d (G00 /G0 )), and crossover temperature (Tc) where G0 ¼ G00 were measured. 2.9. Torsion gelometry of imitation cheese Shear stress and shear strain values were determined using a torsion gelometer (Gel Consultants, Raleigh, NC, USA) that was operated at 2.5 rpm. Three plugs were bored from the sample and milled to the appropriate capstan shape as described by Foegeding (1992). A specific milling configuration was used to ensure that the specimens were uniform, so that their geometry had minimal influence on the calculations. A texture map (fracture stress vs. fracture strain) was constructed using the torsional test data. 2.10. Statistical analysis A block randomised design was used for all treatments. Analysis of variance (ANOVA) was conducted using SPSS version 12 (SPSS Inc., Illinois, USA). Treatment means were considered significantly different at pp0.05 unless stated differently. When significant differences were indicated by ANOVA, Tukey pair-wise comparisons were performed to indicate where the differences between properties existed. 3. Results and discussion 3.1. Manufacture and compositional analysis The gelatinisation temperatures of all starches (native, pre-gelatinised, resistant or waxy corn starch) used in this study are shown in Table 1. Imitation cheeses were successfully manufactured on a small-scale in a Brabender

ARTICLE IN PRESS N. Noronha et al. / Food Hydrocolloids 22 (2008) 1612–1621

Farinograph-Es. A control (no starch) and cheeses containing starches at various concentrations (1.9–9.9% w/w) were manufactured. The use of starch during manufacture did not result in any difficulties in processing the cheese. The cook times of cheeses—i.e. time taken to form a homogeneous mass without the presence of visual water—were similar (4 min), irrespective of the presence, type or concentration of starch used. The cheese cook times

Table 1 Gelatinisation characteristics of native, waxy, resistant and pre-gelatinised corn starch Starch type

Gelatinisation onset temp (1C)

Native Pregelatinised Resistant Waxy

65.84 44.22 78.97 62.88

Gelatinisation peak temp (1C)

1615

were similar to those previously reported by Noronha et al. (2007), when resistant starch was used to replace fat in imitation cheese samples. The mean moisture, protein, aw and the pH of the control imitation cheese was 54.7370.21%, 21.8270.1%, 0.95270.01 and 5.9870.01, respectively. Fat contents decreased concomitantly from 19.1% to 10.7% w/w with increasing starch levels (1.9% to 9.9% w/w). In accordance with the experimental design, the mean moisture content of cheeses containing starch was 54.7470.24%, while protein levels were 21.8270.3%, and the mean cheese pH was 5.9870.01. The average cheese aw was 0.93670.01.

Energy required (J/g)

3.2. Light microscopy

70.54 79.44

10.39 7.92

100.97 71.88

20.77 16.41

Light micrographs of cheese containing no starch (Fig. 1A) are in agreement to those presented by Montesinos et al. (2006) and Noronha et al. (2007) in showing a continuous protein matrix (green/blue) interspersed with fat globules (red). Fat globule size ranged

Fig. 1. Light micrographs of control (starch-free) imitation cheese containing 55% w/w moisture (A), imitation cheeses containing 9.9% w/w native (B), pre-gelatinised (C), resistant (polarised image) (D) and waxy (E) corn starch, F: fat globule; P: protein matrix; S: starch particles.

ARTICLE IN PRESS N. Noronha et al. / Food Hydrocolloids 22 (2008) 1612–1621

from 0.5 to 20 mm. Starch inclusion influenced the cheese microstructure, in a way determined by starch type and level of concentration. Starch particles are both more visible and numerous in the high starch cheese (9.9% w/w) compared to the low starch cheeses (1.9% w/w) (micrographs not shown). Cheese containing 9.9% native corn starch are shown in Fig. 1B in which starch particles appeared to cluster. Starch particles were irregular in shape (5–40 mm) and were scattered unevenly throughout the protein matrix. Fat globules (1–15 mm) were largely spherical in shape and were very finely scattered throughout the matrix, which is in agreement to Mounsey and O’Riordan (2001) who also observed that imitation cheese products containing native corn starch appeared to have smaller fat globules to those present in cheeses containing no starch. The fat globules and proteins appeared to form a tight continuous matrix. A light micrograph of cheese containing 9.9% w/w pregelatinised starch is shown in Fig. 1C. There was evidence of fat (F) globules (5–30 mm) throughout the cheese. Starch particles (S) were irregularly shaped and were present in large groups (30–60 mm) and disrupted the continuity of the protein matrix (P). Mounsey and O’Riordan (2008) also observed irregular shaped starch particles which disrupted the protein structure, when pre-gelatinised starch was used to replace rennet casein in imitation cheese. A polarised image (Fig. 1D) of cheeses containing resistant corn starch shows there was clear evidence of bifringence/maltese crosses (an indicator of intact ungelatinised granules) throughout the cheese matrix. Fig. 1D also shows that starch was uniformly distributed throughout this cheese, with the average starch particle size being 10 mm. The fat globule size was o10 mm. For cheeses containing 9.9% waxy corn starch, the starch was present in the protein matrix as discrete particles (5–30 mm), appearing slightly swollen, having a tendency to cluster (Fig. 1E). Fat was scattered as large spherical globules, although some were irregular in shape, throughout the protein matrix of these cheeses. Fat globule size was noticeably bigger than seen in Fig. 1B, ranging in size from 0.5 to 25 mm. 3.3. NMR 3.3.1. Spin-lattice relaxation measurements (T1) The spin-lattice relaxation time (T1) for the control imitation cheese was 198.5 ms. This value was considerably shorter than those reported by Kuo, Gunasekaran, Johnson, and Chen (2001) when natural pasta filata Mozzarella (o720 ms) and natural non-pasta filata Mozzarella (o565 ms) were measured using a similar NMR technique, suggesting that the water in imitation Mozzarella cheese was less rotationally mobile than that in natural Mozzarella cheeses. On inclusion of starch, T1 significantly decreased with increasing starch concentration (Fig. 2, for clarity, only cheeses containing no starch, pre-gelatinised and native corn starch are shown). All starch-containing

300 Pre-gelatinised Native Control 250

T1 (ms)

1616

200

150

100 0

2

4

6

8

10

12

Starch Concentration (%) Fig. 2. The spin-lattice relaxation times (T1) for a starch-free imitation cheese and imitation cheeses containing 1.9–9.9% w/w pre-gelatinised and native corn starch.

cheese at a given starch concentration had similar T1 times, with the exception of the pre-gelatinised starch cheeses, which had lower T1 times. T1 times were in the range of 160–205 ms, which were considerably shorter than T1 for bulk water (2700 ms). Lower T1 times at higher starch concentrations suggest that the water present in these cheeses was less rotationally mobile than in cheeses with lower starch levels. The lower T1 of the pre-gelatinised starch cheeses compared to the other starch-containing cheeses (Fig. 2) indicates that this starch reduced the mobility of water to a greater extent than other starches. 3.3.2. Spin–spin relaxation measurements (T2) The spin–spin relaxation behaviour of starch-containing imitation cheeses did not follow a simple exponential decay. Two relaxation states in imitation cheeses were found, and accordingly a multi-component model was used. Budiman, Stroshine, and Campanella (2000) also reported that the mobility of water in cheese analogues exhibited multi-component behaviour, in which the individual components can be interpreted as representing different relaxation fractions in the system. Fig. 3 shows a typical T2 profile for the starch-free cheese (control). The component with the shorter relaxation time, T2,tb, corresponds to protons in a less mobile fraction of water within the cheese sample, corresponding to water that is tightly bound. The effect of starch type and increasing concentration on the mobility of the tightly bound water fraction in imitation cheese samples is shown in Table 2. The second component, T2,f, was ascribed to protons from the fat phase of imitation cheese. To verify this, a cheese containing 0% fat was analysed and was not found to have this component. Liquid oil was also assessed and its relaxation time (80–90 ms) correlated with T2,f. The association of these relaxation fractions/components to water held by the protein and the fat phase are in agreement to those reported by Budiman et al. (2000) when the relaxation behaviour of cheese analogues was examined.

ARTICLE IN PRESS N. Noronha et al. / Food Hydrocolloids 22 (2008) 1612–1621

1617

22

30 T2,tb

20

25

16 T2,tb (ms)

Intensity (a.u)

18 20 15

R2 = 0.9620

14 12

10 10 T2,f

5

8 6

0 0

50

100

150

200

250

0

2

T2 (ms)

4

6

8

10

12

Resistant Starch Concentration (%)

Fig. 3. Distribution of spin–spin relaxation times in a starch-free imitation cheese.

Fig. 4. The spin–spin relaxation times (T2,tb) for the cheeses containing 1.9–9.9% w/w resistant corn starch.

Table 2 Spin–spin relaxation (T2,tb) times (ms) of imitation cheese (55% w/w moisture) containing different levels of starch (1.9–9.9% w/w)

16

Starch type

0%

1.9%

3.9%

5.8%

7.8%

9.9%

12

Native Pre-gelatinised Resistant Waxy

19.64 19.64 19.64 19.64

18.41 17.09 19.03 17.88

15.85 15.39 17.83 15.87

15.80 13.30 16.71 14.15

12.87 10.16 13.63 14.78

13.24 11.81 12.73 16.12

10

T2,tb (ms)

14

8 6 4

T2,tb decreased with increasing starch concentration for all cheeses containing native, pre-gelatinised and resistant starches, e.g. T2,tb decreased linearly (R2 ¼ 0.9620) from 19 to 12.5 ms with increasing resistant starch concentration (1.9–9.9%) (Fig. 4). This indicated that the lowmobility water molecules became more tightly bound with increasing starch concentration. For cheeses containing waxy starch, T2,tb initially decreased with increasing starch concentration, however was found to increase when levels X7.8% of starch were added. Cheeses containing pregelatinised starch had the lowest T2,tb times (Fig. 5, for clarity, only cheeses containing starch at 7.8% are shown). Unlike T2,tb, T2,f—representing protons from the fat phase—did not follow any obvious trend with either starch type or concentration. All T2,f times were in the range of 90–111 ms, irrespective of cheese composition. Although starch concentration had no apparent effect on T2,f values, T2,f peak profiles—i.e. peak width—varied with cheese composition (Figs. 3 and 6, correspond to a starchfree and a cheese containing pre-gelatinised starch, respectively). In general, peak widths relate to the uniformity of the sample, with broader peaks corresponding to less uniform samples. For all cheeses, T2,f peak widths ranged from 20 to 50 ms. Increasing the starch concentration in cheeses tended to narrow peak width, especially for cheeses containing native corn starch. Therefore, inclusion of this starch increased the uniformity

2 0 Native

Pregelatinised

Resistant

Waxy

Starch Concentration (7.8 %) Fig. 5. The spin–spin relaxation times (T2,tb) for cheeses containing 7.8% w/w of various starches.

of the fat phase in starch-containing imitation cheeses. Light microscopy correlates with this NMR data, showing that well distributed fat globules which were generally smaller in size were evident in cheeses containing native corn starch. Fat within such finely distributed small fat globules would be expected to be more uniform in its mobility than fat in larger fat globules. A third component, T2,mb, was observed in cheeses with higher concentrations of pre-gelatinised (Fig. 6) and native corn starch. T2,mb times were approximately six times longer (540–660 ms) than T2,f, though still noticeably shorter than free water (T2 ¼ 2700 ms at 25 1C). This component corresponds to moderately bound water within the cheese matrix. Overall the inclusion of starch in imitation cheese decreased both T1 and T2. However differences were observed in the responses of the relaxation times to starch type and concentration. The decreases observed in T1

ARTICLE IN PRESS N. Noronha et al. / Food Hydrocolloids 22 (2008) 1612–1621

25

T2,tb

Intensity (a.u)

20

15

10 T2,f

5

T2,mb 0 0

200

400

600

800

1000

T2 (ms) Fig. 6. Distribution of spin–spin relaxation times in imitation cheese containing 9.9% w/w pre-gelatinised corn starch.

indicated that increasing starch concentration decreased the water mobility of the cheese; however the effect of starch type on this parameter could not be greatly distinguished. T2 differentiated between different states of water binding (tightly and moderately bound water) and the uniformity of the fat proton component within the cheese matrix. These components were dependent on starch type. Tightly bound water can be largely attributed to water held by the protein matrix as this component (T2,tb) was also evident in the starch-free imitation cheese. Kuo et al. (2001) also attributed the shortest relaxation fraction to water that is associated with the cheese protein matrix. However this water fraction was somewhat affected by starch type. The resistant starch used in this study (Novelose 240) has less capacity to absorb water compared to other common corn starches, because of its limited ability to gelatinise in the cheese manufacturing process (Fig. 1C). However previous work from this laboratory has shown that addition of this starch (10%) to a 52% moisture imitation cheese bound an additional 1.7 g water/ 100 g dry matter in the monolayer compared to a starchfree cheese (Duggan et al., 2008). This study also found that the monolayer value increased with increasing starch content. In the present study increasing resistant starch concentration decreased the mobility of the tightly bound water fraction (T2,tb). This may suggest that water in the T2,tb component is related to monolayer water. The fat component, T2,f, was quite similar to that of the control cheese (no starch), indicating that resistant starch (irrespective of concentration) did not alter how the fat was held by the cheese matrix. As the fat phase was little affected by resistant starch concentration (i.e. fat globule size remained similar) it was not unexpected that T2,f times were similar to those of the starch-free cheese. In contrast, the native corn and pre-gelatinised starches influenced both the water and fat fractions of imitation cheese, making these fractions both more tightly bound

and uniform. This may be due to gelatinisation of these starches during cheese manufacture (Table 1). Gelatinisation is a process that breaks down the inter-molecular bonds of starch molecules in the presence of water and heat, allowing an increase in water binding. Thus, these starches would imbibe increasingly more water than starches that do not gelatinise during cooking (e.g. resistant corn starch). The appearance of a new component (T2,mb) for the native corn and pre-gelatinised starch cheeses suggests that new physical and chemical environments were formed within the cheese system. Post-gelatinisation these starches have a tendency to retrograde (the change in gelatinised starch from an amorphous to a crystalline state) resulting in a loss of water holding capacity. This may have caused a shift in water molecules from being tightly bound (T2,tb) to more moderately bound (T2,mb). The mobility of water in cheese containing waxy corn starch decreased at lower concentrations, in line with other starches. However at higher concentrations (X7.8%) the mobility increased. The reason for this behaviour is not clear. 3.4. Rheology The values of the moduli G0 and G00 for the control imitation cheese decreased when the temperature was increased from 25 to 90 1C, showing that the cheese matrix began to soften with increasing temperature. Tan d values increased with increasing temperature, peaking at 85 1C at a value of 2.1. The melting temperature (Tc) (tan d ¼ 1) of the control cheese was 50.170.3 1C. Fig. 7 shows the effect of starch type on G0 at 25 1C and shows that the cheese containing the pre-gelatinised and waxy starch had the highest and lowest G0 values at 25 1C, respectively. This is most likely due to the effect these starches had on the water mobility, as the cheese containing pre-gelatinised starch had the most tightly bound water

1000000

G' (Pa)

1618

1000 Native

Pre-gelatinised Resistant

Waxy

Starch Type Fig. 7. Effect of starch type on the G0 value of all 1.9% w/w starchcontaining imitation cheeses (55% w/w moisture) at 25 1C.

ARTICLE IN PRESS N. Noronha et al. / Food Hydrocolloids 22 (2008) 1612–1621

while water in cheese containing waxy starch was more mobile. As temperature was increased from 25 to 90 1C all cheeses containing starch softened in a similar manner to the control cheese, however the effect of the added starch was to increase the values of both G0 (Fig. 8A) and G00 and to decrease tan d values over the temperature range studied, irrespective of starch type. Increasing starch concentration in cheeses led to increases in both G0 and G00 values (Fig. 8B); the cheeses with higher starch levels (7.8 and 9.9% w/w) had noticeably different G0 and G00 values, which did not decrease on heating to the same degree as the cheeses containing lower levels of starch. The moduli for these cheeses in fact levelled off or increased at temperatures 470 1C, with the exception of the cheeses containing waxy starch. This increase in G0 and G00 may be due to an increase in the level of starch hydration or gelatinisation at these temperatures; such an increase in G0 has previously been reported by Montesinos et al. (2006) and Noronha et al. (2007) for imitation cheeses containing resistant starch. The temperature at which tan d ¼ 1, i.e. G0 ¼ G00 , has been used as an indicator of the melting temperature of imitation cheeses (Mounsey & O’Riordan, 1999, 2001;

100000 no starch pre-gelatinised resistant

G' (Pa)

10000

1000

100

Table 3 Crossover temperatures (Tc) of imitation cheese (55% w/w moisture) containing different levels of starch (1.9–9.9% w/w) Starch type

0%

1.9%

3.9%

5.8%

7.8%

9.9%

No starch Native Pre-gelatinised Resistant Waxy

50.10 NA NA NA NA

NA 52.28aw 54.03cw 53.03bw 52.05aw

NA 62.39cx 64.49dx 55.84ax 59.16bx

NA 75.83cy No Tc 57.34ay 66.84by

NA No Tc No Tc No Tc No Tc

NA No Tc No Tc No Tc No Tc

For each column, means having the same superscript letter a, b, c, d, or e did not differ significantly. For each row, means having the same superscript letter w, x, y, or z did not differ significantly.

Montesinos et al., 2006; Noronha et al., 2007). Increasing the concentration of starch in imitation cheeses resulted in an increase in the melting temperature (Table 3). The cheese containing the native starch had significantly (po0.001) higher melting temperatures compared to all other cheeses at the same concentration. All cheeses containing X7.8% starch did not exhibit a melting temperature; while cheeses containing 5.8% pre-gelatinised starch also failed to exhibit a crossover temperature. The melting temperature is influenced by the availability of free water to plasticise the cheese matrix (McMahon, Payne, Fife, & Oberg, 1996). Our NMR data showed that the pre-gelatinised corn starch reduced the mobility of water in the cheese to the greatest extent, which would in turn reduce the availability of water for plasticisation. The resistant starch cheese exhibited the smallest incremental increase in Tc with increasing concentration of starch compared to all other cheeses. 3.5. Torsion

10 20

30

40

50

60

70

80

90

Temperature (°C) 1000000

1.9% 5.8% 9.9%

100000 G' (Pa)

1619

10000

1000

100 20

30

40

50

60

70

80

90

Temperature (°C) Fig. 8. Effect of temperature on the G0 value of cheeses (55% w/w moisture) containing no starch or 1.9% w/w pre-gelatinised and resistant corn starch (A) and increasing levels of 1.9%, 5.8% and 9.9% native starch (B).

Torsion gelometry is a fundamental rheological test in which specimens are twisted until they fracture. Torsion produces what is termed a pure stress, a condition that maintains sample shape and volume during the test. Shear stress and strain in torsion have been used to construct ‘texture maps’ of food materials (Hamann & MacDonald, 1992). Foods exhibiting low shear stress at failure, (smax) are termed soft if the shear strain, (gmax) is low, or rubbery if gmax is high. Foods with high smax values can be considered brittle if gmax is low and tough if gmax is high. A texture map, which is a plot of shear stress vs. shear strain, provides a graphical representation of product texture (Truong & Daubert, 2001). The shear stress and strain values at failure for the starch-free cheese were 109.46 kPa and 1.77, respectively, which were similar to the values reported by Bowland and Foegeding (1999) for model processed cheese. Addition of starch, with a concomitant reduction in fat, to the cheese increased the shear stress while decreasing the strain values at failure. Gwartney, Foegeding, and Larick (2002) reported similar observations when the relationship between fracture stress for natural and processed cheese was

ARTICLE IN PRESS N. Noronha et al. / Food Hydrocolloids 22 (2008) 1612–1621

1620

Fracture Stress (KPa)

400

Brittle

4. Conclusion

Tough 5.8% Native 5.8% Waxy 5.8% Pre-gelatinised

300

5.8% Resistant

1.9% Pre-gelatinised

200

1.9% Waxy 1.9% Native 1.9% Resistant No starch

100

Rubbery

Soft

0 0

0.5 1 1.5 Fracture Strain (arbitrary units)

2

Fig. 9. Texture map of a starch-free and cheeses containing 1.9% and 5.8% w/w of various starches.

examined, in which reduced-fat samples had higher fracture stress values than their equivalent full-fat products. At starch concentrations 47.8% only cheeses containing resistant starch fractured on torsion. Thus the texture map could only be constructed using the values of shear and strain for 1.9% and 5.8% starch cheeses (Fig. 9). According to the texture map, the starch-free cheese, comprising mainly protein and fat, was more rubbery than soft. All cheeses containing 1.9% corn starch were plotted quite closely together suggesting that they had similar texture profiles, with the exception of the cheese containing the pre-gelatinised corn starch. This cheese was more brittle than all other cheeses containing 1.9% starch. The 1.9% resistant starch cheese was the closest in texture to the cheese without added starch (control), reaffirming that this starch had little interaction with the protein matrix, hence minimally affecting the texture of the cheese. Increasing the starch content to 5.8% changed the texture profile of cheeses from being slightly rubbery to more brittle. The textural properties of the cheeses containing 5.8% resistant corn starch did not change to the same extent as the other cheeses (Fig. 9). As previously discussed, the pre-gelatinised and native corn starch reduced the mobility of water to the greatest extent and this reduction in water mobility is most likely what caused these cheeses to become more brittle as the lower the mobility of water the less effective a plasticiser it is. Cheeses containing 5.8% waxy starch also became brittle, having values similar to the native and pre-gelatinised starch cheeses. This result is perhaps surprising given that this cheese had similar water mobilities to that of cheese containing resistant starch. Waxy starch consists for the most part of highly branched amylopectin chains and their presence may have influenced the cheese texture. Light microscopy images showed this starch tended to cluster. These clusters could provide areas most susceptible to ‘failure’, which would contribute to the brittleness of the cheese.

Nuclear magnetic resonance relaxometry proved to be a useful technique in investigating the behaviour of water in starch-containing imitation cheese. Imitation cheese (starch-free) was found to have two distinct components, corresponding to tightly bound water and protons associated with the fat phase. Addition of starch was found to decrease the mobility of water and this decrease was dependent on starch type. A second water binding component, corresponding to moderately bound water, was found for cheeses containing starches which had a greater tendency to retrograde. Cheese functionality, i.e. melting temperature and hardness, was associated with the mobility of water within the cheese matrix. Matrices in which the water was more mobile produced good melting, softer cheeses, while cheeses in which the water was more tightly bound were brittle and non-melting. Therefore the use of various starches with different gelling abilities can be used to tailor imitation cheese products to meet specific needs, i.e. less melting cheeses suitable for applications in which flow resistance is required. Acknowledgments This work was funded through the Food Institutional Research Measure (FIRM) administered by the Irish Department of Agriculture, Food and Rural Development. References Bowland, E., & Foegeding, E. (1999). Factors determining large-strain (fracture) rheological properties of model processed cheese. Journal of Dairy Science, 82, 1851–1859. Budiman, M., Stroshine, R. L., & Campanella, O. H. (2000). Stress relaxation and low filed proton magnetic resonance studies of cheese analog. Journal of Texture Studies, 31, 477–498. Chen, P. L., Long, Z., Ruan, R., & Labuza, T. P. (1997). Nuclear magnetic resonance studies of water mobility in bread during storage. Food Science Technology/Lebensmittel-Wissenchatt Technology (LWT), 30, 178–183. Clark, G. (1980). Staining procedures (p. 415). Davies, L. (1995). Starch-composition, modifications, applications and nutritional value in foodstuffs. Food Technology Europe, June/July, 44–52. Drury, R. A. B., & Wallington, E. A. (1967). Carleton’s histological technique. London, UK: Oxford University Press. Duggan, E., Noronha, N., O’Riordan, E. D., & O’Sullivan, M. (2008). Effect of resistant starch on the water binding properties of imitation cheese. Journal of Food Engineering, 84, 108–115. Ernst, R. R., Bosedhausen, B., & Wokaun, A. (1991). Principles of nuclear magnetic resonance in one and two dimensions (p. 203). Foegeding, E. A. (1992). Rheological properties of whey protein isolate gels determined by torsional fracture and stress relaxation. Journal of Textural Studies, 23, 337–358. Gwartney, E. A., Foegeding, E. A., & Larick, D. K. (2002). The texture of commercial full-fat and reduced-fat cheese. Journal of Food Science, 67, 812–816. Hamann, D. D., & MacDonald, G. A. (1992). Rheology and texture properties of surimi and surimi-based foods. In T. C. Lanier, & C. M. Lee (Eds.), Surimi technology (pp. 429–500). New York: Marcel Dekker, Inc.

ARTICLE IN PRESS N. Noronha et al. / Food Hydrocolloids 22 (2008) 1612–1621 IDF. (1958). Determination of dry matter in cheese and processed cheese. Standard FIL-IDF 4. Brussels, Belgium: International Dairy Federation. Jana, A. H., & Upadhyay, K. G. (2003). Mozzarella cheese analogue—A review. Journal of Food Science Technology (Mysore), 40, 1–10. Kuo, M. I., Gunasekaran, S., Johnson, M., & Chen, C. (2001). Nuclear magnetic resonance study of water mobility in pasta filata and non-pasta filata Mozzarella. Journal of Dairy Science, 84, 1950–1958. Light, J. M. (1990). Modified food starches: Why, what, where and how. Cereal Foods World, 35, 1081–1092. McMahon, D. J., Payne, M. C., Fife, R. L., & Oberg, C. J. (1996). Use of fat replacers in low fat Mozzarella cheese. Journal of Dairy Science, 70, 1911–1921. Montesinos, C., Cottell, D. C., O’Riordan, E. D., & O’Sullivan, M. (2006). Partial replacement of fat by functional fibre in imitation cheese. Effects of rheology and microstructure. International Dairy Journal, 16, 910–919. Mounsey, J. S., & O’Riordan, E. D. (1999). Empirical and dynamic rheological data correlation to characterize melt characteristics of imitation cheese. Journal of Food Science, 64, 701–703. Mounsey, J. S., & O’Riordan, E. D. (2001). Characteristics of imitation cheese containing native starches. Journal of Food Science, 66, 586–591.

1621

Mounsey, J. S., & O’Riordan, E. D. (2008). Influence of pre-gelatinised maize starch on the rheology, microstructure and processing of imitation cheese. Journal of Food Engineering, 84, 57–64. Noronha, N., O’Riordan, E. D., & O’Sullivan, M. (2007). Replacement of fat with functional fibre in imitation cheese. International Dairy Journal, 17, 1073–1082. Provencher, W. S. (1982). CONTIN: A general purpose constrained regularization program for inverting noisy linear algebraic and integral equations. Computer Physics Communications, 27, 229. Ruan, R. R., Zou, C., Wadhawan, C., Martines, B., Chen, P. L., & Addis, P. (1997). Studies of hardness and water mobility of cooked wild rice using nuclear magnetic resonance. Journal of Food Process and Preservation, 21, 91–104. Sajilata, M. G., Singhal, R. S., & Kulkarni, P. R. (2006). Resistant starch—A review. Comprehensive Reviews in Food Science and Food Safety, 5, 1–17. Schmidt, S. J. (1990). Characterisation of water in foods by NMR. In J. W. Finley, S. J. Schmidt, & A. S. Serianui (Eds.), NMR application in biopolymers (p. 415). New York, NY: Plenum Press. Thomas, D. J., & Atwell, W. A. (1999). In D.J. Thomas, &W.A. Atwell (Eds.), Starch. Minnesota: The American Association of Cereal Chemists, Inc.; St. Paul, MN: Eagan Press. Truong, V. D., & Daubert, C. R. (2001). Textural characterization of cheeses using vane rheometry and torsion analysis. Food Engineering and Physical Properties, 66, 716–721.