Effects of Different Chemically Modified Starches on the Rheological Properties of Stirred Non-fat Yoghurt

Abstract

The use of chemically modified starches to improve the physical properties of non-fat stirred yoghurt was investigated. Yoghurts were prepared from non-fat milk powder, sucrose, water and tapioca starch acetates (TSA-1, TSA-2, TSA-3), tapioca distarch phosphates (TDP-1, TDP-2, TDP-3) or native tapioca starch at 1% (w/w). Syneresis, particle size distribution and viscoelastic properties of the yoghurts were determined, and flow behaviour was described using the Herschel-Bulkley model. Furthermore, interactions between milk proteins and modified starches attributed to protein surface hydrophobicity were characterized. Results showed that yoghurts with starch acetates exhibited higher yield stress, consistency coefficient (K) values, hysteresis loop area, storage modulus (G′) and loss modulus (G″). Protein surface hydrophobicity was significantly influenced by the addition of starch acetates, and TSA-3 yoghurt exhibited the lowest values. This study concluded that the addition of TSA-3 starch showed better results in terms of the rheological properties of non-fat yoghurt.

Introduction

Yoghurt is produced from fresh milk by a fermentation process using lactic acid starter cultures to give a pH value of 3.8–4.6 (Tamime, 2003). Low-fat or non-fat yoghurts are popular due to their nutritional characteristics. However, reducing the fat content of yoghurt alters its structural and mechanical characteristics, resulting in poor food texture characteristics and high syneresis (Sandoval-Castilla et al., 2004; Pereira et al., 2006). The addition of dairy-based ingredients, non-dairy ingredients or a combination thereof increases the total solids content of the milk to produce desirable low-fat or non-fat yoghurts (Sandoval-Castilla et al., 2004). In the past, starches have proven useful for their role as gelling, stabilising and thickening agents in different food applications. However, in the dairy industry, native starches are not preferred since they possess low shear and thermal resistance and have a strong tendency to retrograde (Corredig et al., 2011). These shortcomings of native starches could be overcome by introducing functional groups into the molecules (Singh et al., 2007). In yoghurt production, chemically modified starches can exert some positive effects (Cui et al., 2014; Bravo-Núñez et al., 2019; Sharma et al., 2018; Pang et al., 2019). During heat treatment of milk, whey proteins, particularly β-lactoglobulin, are denatured, leading to the formation of soluble and micelle-bound whey protein aggregates through hydrophobic interaction, electrostatic interaction and disulphide bonding (Krzeminski et al., 2011). These aggregates interact with casein micelles during milk acidification. A decrease in pH leads to an increase in the attachment of whey proteins to casein micelles. On the other hand, starches gelatinize when heated in excess water, resulting in disruption of the granular structure, swelling and then hydration and solubilization of starch molecules. The combined effect of swollen starch granules, starch-adsorbed milk proteins and changes in the structure of milk proteins during heating and acidification produces a firm yoghurt structure (Oh et al., 2007; Noisuwan et al., 2011). The use of modified starches may increase yoghurt viscosity and strengthen the rigidity of the casein network by binding water and interacting with other milk constituents, such as proteins, thereby inhibiting syneresis. Although the addition of modified starches has been investigated in the production of low-fat yoghurt, it is difficult to independently evaluate the effect of the starches on yoghurt properties, since the addition of starch prior to fermentation also changes the final starch and total solids contents. Therefore, the objective of this research was to analyse the syneresis, rheological properties and protein surface hydrophobicity of non-fat stirred yoghurts produced using starch acetates and distarch phosphates at different levels of modification. The starch was added after acidification and prior to heat treatment of yoghurt samples.

Materials and Methods

Materials    Three types of tapioca starch acetates with different degrees of substitution [TSA-1 (0.019), TSA-2 (0.026) and TSA-3 (0.068)], three types of tapioca distarch phosphates with different degrees of substitution [TDP-1 (0.0058), TDP-2 (0.0063) and TDP-3 (0.0081)] and native tapioca starch were obtained from J-Oil Mills Inc., Tokyo, Japan. Low-heat-treated non-fat dry milk was obtained from Megmilk Snow Brand Co., Ltd, Tokyo, Japan. Freeze-dried non-ropy producing yoghurt culture containing Streptococcus thermophilus and Lactobacillus bulgaricus (YC-X11 Yo-Flex^®) was obtained from Chr. Hansen A/S, Hørsholm, Denmark. 8-Anilinonaphthalene-1-sulfonic acid (ANS; MP Biomedicals LLC, Illkirch, France) was used as the fluorescent probe.

Yoghurt production    Reconstituted milk was prepared by dissolving low-heat skim milk powder (12% (w/w)) and sugar (6% (w/w)) in deionized water. Milk samples were heated to 90 °C for 5 min and then cooled to 43 °C before inoculation with 0.002% (w/w) yoghurt starter. Fermentation was carried out in a water bath (NCB-3100; Tokyo Rikakikai Co., Ltd., Tokyo, Japan) at 43 °C for 5 h until pH 4.6 and 0.95–0.98% titratable acidity was reached. Starch solutions (1% w/w) were added, after which the yoghurts were heated to 72 °C for 10 min to stop the fermentation process. The yoghurt was then cooled to 5 °C for further analysis. Non-fat yoghurt without starch was used as the control.

Syneresis    Yoghurt (20.0 g) was centrifuged at 100×g for 15 min at 4 °C. The clear supernatant was poured off, weighed and expressed as% weight relative to the original weight of yoghurt.

Particle size analysis    The particle size distribution of the samples was obtained using a laser diffraction particle size analyser (LS230; Beckman Coulter, Small-Volume Mode, Fullerton, CA, USA). Yoghurt samples were dispersed in deionized water and measurements of particle size were obtained at an obscuration of 14–15% and polarization intensity differential scattering (PIDS) of 45–55%.

Protein surface hydrophobicity    The protein surface hydrophobicity of the samples was determined according to Bonomi et al. (1988). The relative fluorescence of the samples was measured using a fluorescence spectrometer (FP-8600; Jasco Corporation, Tokyo, Japan) at room temperature (20–24 °C) with ANS as the fluorescent probe. An excitation wavelength (λ_ex) of 390 nm and an emission wavelength (λ_em) of 480 nm were used. Yoghurts were diluted with 50 mM phosphate buffer (pH 6.8) containing dipotassium hydrogen phosphate (K₂HPO₄) and potassium dihydrogen phosphate (KH₂PO₄) to a protein concentration of approximately 2 g/L. The diluted samples were then titrated with increasing concentrations of ANS solution (final concentration of ANS in diluted samples varied from 0–150 µM) until no further increase in fluorescence was observed. Samples without ANS were measured as the blank.

From ANS curves, protein surface hydrophobicity (PSH) was calculated as follows:   

where F_max is the maximum fluorescence, K_d is the ANS concentration required to obtain half the value of F_max and P is the protein content.

Rheological measurements    Rheological tests were carried out using an ARES-G2 rheometer (TA Instruments, New Castle, DE, USA) with the cone and plate geometry measuring system (ø 25 mm, cone angle 1°) at 20 °C.

1. Flow behaviour

Flow curves were obtained with a shear rate range from 0.1 to 200 s⁻¹ for 90 s (2.2 s⁻¹/s, rising curve) and 200 to 0.1 s⁻¹ for 90 s (descendent curve) at 20 °C and the shear stress values were then recorded. Flow behaviour was determined using the Herschel–Bulkley model as follows:   

where σ is the shear stress (Pa), σ₀ is the yield stress (Pa), K is the consistency index (Pa·sⁿ), γ is the shear rate (s⁻¹) and n is the flow behaviour index, which indicates the closeness to Newtonian flow (n<1 indicates shear-thinning liquid). The hysteresis loop area, which is the area between the upward and downward curves of shear stress against shear rate, was also calculated.

2. Dynamic viscoelasticity testing

Amplitude sweep tests were performed using a strain sweep of 0.001–2.5 (6.28 rad/s) at 20 °C to determine the linear viscoelastic range for the yoghurt samples. Frequency sweeps were carried out under an angular frequency (ω) range of 0.1 to 100 rad/s (0.005 strain). The storage modulus, G′, the loss modulus, G″ and the loss angle (tan δ = G″/G′) were obtained as a function of ω.

Statistical analysis    IBM SPSS Statistics 21 software (IBM Corporation, Armonk, NY, USA) was used to perform statistical analyses. Analysis of variance (ANOVA) and Tukey's HSD test were performed to determine significance at P < 0.05.

Results and Discussion

Syneresis    Syneresis is a common defect observed during the storage of yoghurt and primarily occurs due to the rearrangement of casein particle aggregates in the gel network. Modified starches have been used to reduce yoghurt syneresis by holding substantial quantities of water in weak gel structures (Luo and Gao, 2011). The effect of modified starch addition on whey loss of the non-fat yoghurt was therefore measured and the results are shown in Table 1. The results showed that the addition of TSA-3 starch significantly decreased the level of whey loss by 4.98% compared to the control. This could be attributed to its relatively high acetyl content, which had a greater effect on starch granule swelling compared to cross-linking. It has been reported that the introduction of acetyl groups could disrupt hydrogen bonds in starch granules and disorganize the intragranular structure, thereby increasing the water binding of starch chains (Sodhi and Singh, 2005). Mirmoghtadaie et al. (2009) and Sodhi and Singh (2005) observed that a high degree of cross-linking could lead to strong bonding between the starch chains to restrict the mobility and swelling of the granules, thereby increasing syneresis. TDP-3 exhibited the highest syneresis, although it did not significantly differ from yoghurts with native, TDP-1, TDP-2, TSA-1 starches and the control. Besides TSA-3, syneresis for the other yoghurts did not differ significantly from the control.

Table 1. Syneresis of yoghurt samples.

Yoghurt samples	Syneresis (%)
Control	22.81 ± 1.27ab
Native tapioca	23.97 ± 0.69ab
Tapioca starch acetates:
TSA-1	23.24 ± 0.68ab
TSA-2	21.03 ± 0.35bc
TSA-3	17.83 ± 1.286c
Tapioca distarch phosphates:
TDP-1	24.85 ± 1.69a
TDP-2	22.80 ± 1.42ab
TDP-3	25.48 ± 0.88a

Values followed by the same superscript letter in the same column, for each measured parameter, are not significantly different at P > 0.05.

n=3

Particle size distribution    Figure 1 shows the particle size distribution of the yoghurt samples. Analysis of particle size distribution of yoghurts with added starches showed two or three characteristic peaks; the first peak was observed at about 15 µm, the second one at about 30–40 µm and the third one at 100 µm. The control, on the other hand, had only one peak at 42 µm, attributed to casein-whey protein aggregates. In yoghurts with added starches, the peaks shifted to larger diameters (greater than 90 µm) due to milk proteins adsorbed onto the gelatinized starch granules. The peculiar two or three characteristic peaks of these yoghurts could be attributed to some swollen starch granules being incorporated into the gel network and some of the starch gel fragments being unevenly distributed in the yoghurt independent of the protein network (Sandoval-Castilla et al., 2004).

Fig. 1.

Particle size distribution of yoghurt samples.

Protein surface hydrophobicity (PSH)    The fluorescent probe binding method was used to characterize the surface hydrophobicity of the milk proteins. The fluorescent marker ANS, a hydrophobic dye with an affinity for hydrophobic regions of proteins, was used in the protein hydrophobicity study (Bonomi et al., 1988). When milk protein molecules unfold, the exposed inner hydrophobic groups react with the ANS probe to form an “ANS–milk protein” complex. Variations in fluorescence intensity and PSH values could be indirectly related to protein-starch interactions. The effect of the addition of modified starches to non-fat yoghurts are given in Table 2 and ANS titration curves are shown in Fig. 2.

Table 2. Fmax, K_d and PSH of yoghurt samples

Yoghurt sample	Fmax	Kd	PSH
Control	367.74 ± 21.93a	23.37 ± 3.04	3.5 ± 0.02a
Native tapioca	362.39 ± 17.13a	25.37 ± 2.09	3.2 ± 0.00a
Tapioca starch acetates:
TSA-1	291.17 ± 15.36c	22.43 ± 2.29	2.9 ± 0.00ab
TSA-2	285.68 ± 18.12c	24.77 ± 3.14	2.5 ± 0.00b
TSA-3	251.42 ± 17.52c	25.57 ± 3.41	2.2 ± 0.00c
Tapioca distarch phosphates:
TDP-1	320.33 ± 17.33b	24.58 ± 0.19	2.8 ± 0.01ab
TDP-2	333.75 ± 17.28b	25.12 ± 2.46	2.9 ± 0.00ab
TDP-3	349.12 ± 20.30b	24.33 ± 2.91	3.2 ± 0.00a

Values followed by the same superscript letter in the same column, for each measured parameter, are not significantly different at P > 0.05.

n=3

Fig. 2.

Changes in fluorescence intensity as a function of ANS concentration.

Regarding F_max, which is the maximum fluorescence that could be attained as well as the maximum allowable hydrophobic surface sites that ANS could bind, TSA-3 yoghurt had significantly lower values compared to all the other yoghurts (Fig. 2). These results suggest that decreases in F_max values due to blocking of hydrophobic surface sites by TSA starches could be associated with the existence of stronger attractive interactions between proteins and the starches. Chi et al. (2008) noted that acetylated starches exhibited increased hydrophobicity due to reduced hydrophilicity of esters, which was attributed to the replacement of hydrophilic hydroxyls by the relatively hydrophobic acetyl groups. Compared to the control, a decrease in PSH indicated greater hydrophobic bonding between milk proteins and starch. The PSH decreased significantly (P < 0.05) following the introduction of starch acetates (Table 2). The lowest values were obtained with TSA-2 and TSA-3. The results indicated that differences exist in the macromolecular interactions between starch granules and caseins.

Rheological properties of yoghurt    The thixotropic behaviour of yoghurt samples, as determined by the hysteresis loop, is shown in Fig. 3. The yoghurt samples, except for those with distarch phosphates, showed a yield point followed by shear-thinning behaviour. Yoghurts with distarch phosphates exhibited shear-thickening with a slight increase in shear rate after the yield point before the onset of shear-thinning.

Fig. 3.

Flow behaviour of yoghurt samples. Flow curves are comprised of upward curves and downward curves of shear stress against shear rate.

Values of Herschel–Bulkley flow model parameters used to describe the flow curves are summarized in Table 3. Values of determination coefficients (R²) show that the Herschel–Bulkley flow model showed a good fit with the flow curves. Yield stress, a rheological property, is defined as the minimum shear stress required to initiate flow, which characterizes the firmness of yoghurt. The yield stress and K values were observed to be higher in yoghurts with native and starch acetates. The cross-linking depressed the swelling of the starch granules and thus the viscosity was decreased, while the native and acetylated starch granules swelled to a larger size, resulting in higher viscosity (Kurakake et al., 2009). Values of the flow behaviour index (n) proved non-Newtonian flow and shear-thinning behaviour of the samples since the values were less than 1. Yoghurt made with starch acetates had the highest hysteresis loop area compared to the other samples. Shi et al. (2016) reported that a larger hysteresis loop area reflected the higher mechanical stability of yoghurt.

Table 3. Flow behaviour indices (n), consistency indices (K), regression coefficient (R²), yield stress (σ₀) and hysteresis loop area (A_hys) of yoghurt samples predicted by the Herschel–Bulkley flow model.

Yoghurt sample	σ0 (Pa)	K (Pa·sn)	n	R2	Ahys (kPa/s)
Control	416.28	446.68	0.30	0.98	29,676.50
Native tapioca starch	546.90	602.56	0.36	0.96	30,878.00
Tapioca starch acetates:
TSA-1	682.00	660.69	0.35	0.97	37,893.60
TSA-2	503.91	616.60	0.34	0.97	37,380.70
TSA-3	733.38	891.25	0.33	0.95	56,945.40
Tapioca distarch phosphates:
TDP-1	346.51	467.74	0.38	0.96	24,374.80
TDP-2	285.58	389.05	0.35	0.94	20,949.50
TDP-3	266.45	316.23	0.26	0.98	21,674.40

Figure 4 shows the storage modulus (G′) and loss modulus (G″) of the yoghurt samples. G′ values were higher than G″ and both values increased with increasing frequency for all samples, indicating a typical weak viscoelastic system. The viscoelastic modulus (G′, G″) versus frequency of a typical viscoelastic system is characterised by G′ dominating G″ with increasing frequency (Lobato-Calleros et al., 2014). In addition, weak gels exhibit elastic solid behaviour; however, the elastic contribution to the viscoelastic system is not as pronounced as in the case of strong gels (Zupančič-Valant and Žumer, 2001). Yu et al. (2016) observed a similar trend. Based on G′ and G″, the results indicated that the yoghurts can be classified into two groups. Yoghurts with starch acetates and native starches had higher G′ and G″. At a higher level of starch acetylation (TSA-3), the interaction between the starch molecules and casein micelles formed a stronger network, making the viscoelastic characteristics more significant.

Fig. 4.

Storage modulus, G′ (▲) and loss modulus, G″ (■) of control (a), native (b), TSA-1 (c), TSA-2 (d), TSA-3 (e), TDP-1 (f), TDP-2 (g) and TDP-3 (h) yoghurt samples.

Tan δ, calculated from G″/G′, is used to interpret the viscoelastic behaviour of a semisolid food. Tan δ values higher than 1 indicated viscous behaviour, while those lower than 1 indicated elastic behaviour (Loveday et al., 2013; Seth et al., 2018). In the present study, the tan δ values of all samples were less than 1, confirming that all samples were more elastic than viscous, which is in agreement with the values of G′ and G″. The results showed that the tan δ of yoghurts increased with the addition of native starch, but the other yoghurts did not differ significantly from the control. Kaur et al. (2006) and Singh et al. (2007) reported higher tan δ maximum values in native maize and potato starches.

Conclusion

The incorporation of native starch and starch acetates had a positive impact on the flow and viscoelastic properties of yoghurt, while the addition of distarch phosphates had an adverse effect. Therefore, TSA-3 starch, which had a high level of acetylation, was considered to be the most suitable stabilizer in non-fat stirred yoghurt. The addition of TSA-3 starch to yoghurt produced lower syneresis and better rheological properties, which was attributed to the interaction between milk proteins and starch chains.

Acknowledgements    The authors thank the Japanese Ministry of Education, Culture, Sports, Science and Technology (MEXT) for financial support, J-Oil Mills Inc. (Tokyo, Japan) for supplying the tapioca starches and Megmilk Snow Brand Co., Ltd. (Tokyo, Japan) for supplying the low-heat-treated non-fat dry milk powder.

References

•  Bonomi,  F.,  Iametti,  S.,  Pagliarini,  E., and  Peri,  C. (1988). A spectrofluorimetric approach to the estimation of the surface hydrophobicity modifications in milk proteins upon thermal treatment. Milchwissenschaft (Germany, FR).
•  Bravo-Núñez,  Á.,  Pando,  V., and  Gómez,  M. (2019). Physically and chemically modified starches as texturisers of low-fat milk gels. Int. Dairy J., 92, 21-27.
•  Chi,  H.,  Xu,  K.,  Wu,  X.,  Chen,  Q.,  Xue,  D.,  Song,  C., and  Wang,  P. (2008). Effect of acetylation on the properties of corn starch. Food Chem., 106, 923-928.
•  Corredig,  M.,  Sharafbafi,  N., and  Kristo,  E. (2011). Polysaccharide–protein interactions in dairy matrices, control and design of structures. Food Hydrocoll., 25, 1833-1841.
•  Cui,  B.,  Lu,  Y. M.,  Tan,  C. P.,  Wang,  G. Q., and  Li,  G. H. (2014). Effect of cross-linked acetylated starch content on the structure and stability of set yoghurt. Food Hydrocoll., 35, 576-582.
•  Lobato-Calleros,  C.,  Ramírez-Santiago,  C.,  Vernon-Carter,  E. J., and  Alvarez-Ramirez,  J. (2014). Impact of native and chemically modified starches addition as fat replacers in the viscoelasticity of reduced-fat stirred yogurt. J. Food Eng., 131, 110-115.
•  Loveday,  S. M.,  Sarkar,  A., and  Singh,  H. (2013). Innovative yoghurts: Novel processing technologies for improving acid milk gel texture. Trends Food Sci. Technol., 33, 5-20.
•  Luo,  Z. and  Gao,  Q. (2011). Effect of enzyme-modified carboxymethyl starch as a fat replacer on the functional properties of sausages. Starch–Stärke, 63, 661-667.
•  Kaur,  L.,  Singh,  J., and  Singh,  N. (2006). Effect of cross-linking on some properties of potato (Solanum tuberosum L.) starches. J. Sci. Food Agric., 86, 1945-1954.
•  Krzeminski,  A.,  Großhable,  K., and  Hinrichs,  J. (2011). Structural properties of stirred yoghurt as influenced by whey proteins. LWT - Food Sci. Technol., 44, 2134-2140.
•  Kurakake,  M.,  Akiyama,  Y.,  Hagiwara,  H., and  Komaki,  T. (2009). Effects of cross-linking and low molecular amylose on pasting characteristics of waxy corn starch. Food Chem., 116, 66-70.
•  Mirmoghtadaie,  L.,  Kadivar,  M., and  Shahedi,  M. (2009). Effects of cross-linking and acetylation on oat starch properties. Food Chem., 116, 709-713.
•  Noisuwan,  A.,  Hemar,  Y.,  Wilkinson,  B., and  Bronlund,  J. E. (2011). Adsorption of milk proteins onto rice starch granules. Carbohydr. Polym., 84, 247-254.
•  Oh,  H. E.,  Anema,  S. G.,  Wong,  M.,  Pinder,  D. N., and  Hemar,  Y. (2007). Effect of potato starch addition on the acid gelation of milk. Int. Dairy J., 17, 808-815.
•  Pang,  Z.,  Xu,  R.,  Luo,  T.,  Che,  X.,  Bansal,  N., and  Liu,  X. (2019). Physiochemical properties of modified starch under yogurt manufacturing conditions and its relation to the properties of yogurt. J. Food Eng., 245, 11-17.
•  Pereira,  R.,  Matia-Merino,  L.,  Jones,  V., and  Singh,  H. (2006). Influence of fat on the perceived texture of set acid milk gels: A sensory perspective. Food Hydrocoll., 20, 305-313.
•  Sandoval-Castilla,  O.,  Lobato-Calleros,  C.,  Aguirre-Mandujano,  E., and  Vernon-Carter,  E. J. (2004). Microstructure and texture of yogurt as influenced by fat replacers. Int. Dairy J., 14, 151-159.
•  Seth,  D.,  Mishra,  H. N., and  Deka,  S. C. (2018). Effect of hydrocolloids on the physico-chemical and rheological properties of reconstituted sweetened yoghurt powder. J. Sci. Food Agric., 98, 1696-1702.
•  Sharma,  M.,  Yadav,  D. N.,  Singh,  A. K.,  Vishwakarma,  R. K., and  Sabikhi,  L. (2018). Impact of octenyl succinylated pearl millet (Pennisetum typhoides) starch addition as fat replacer on the rheological, textural and sensory characteristics of reduced-fat yoghurt. Int. J. Dairy Technol., 71, 723-733.
•  Shi,  J.,  Li,  D., and  Zhao,  X. H. (2017). Quality attributes of the set-style skimmed yoghurt containing enzymatic cross-linked or thermal polymerized whey protein isolate. CyTA-J. Food, 15, 34-40.
•  Singh,  J.,  Kaur,  L., and  McCarthy,  O. J. (2007). Factors influencing the physico-chemical, morphological, thermal and rheological properties of some chemically modified starches for food applications-A review. Food Hydrocoll., 21, 1-22.
•  Sodhi,  N. S. and  Singh,  N. (2005). Characteristics of acetylated starches prepared using starches separated from different rice cultivars. J. Food Eng., 70, 117-127.
•  Tamime,  A. Y. (2003). Fermented milks: a historical food with modern applications–a review. Eur. J. Clin. Nutr., 56, S2.
•  Yu,  H. Y.,  Wang,  L., and  McCarthy,  K. L. (2016). Characterization of yogurts made with milk solids nonfat by rheological behavior and nuclear magnetic resonance spectroscopy. J. Food Drug Anal., 24, 804-812.
•  Zupančič-Valant,  A. and  Žumer,  M. (2001). Viscoelastic properties of hydrophilic polymers in aqueous dispersions. Acta Chim. Slov., 4, 469-486.