**Comparative Reduction of Egg Yolk Cholesterol Using Anionic Chelating Agents**

**Minerva Bautista Villarreal 1 , Claudia T. Gallardo Rivera 1 , Eristeo García Márquez 2 , José Rodríguez Rodríguez 3 , María Adriana Núñez González 1 , Abelardo Chávez Montes 4 and Juan G. Báez González 1, \***


Academic Editors: Lillian Barros and Isabel C.F.R. Ferreira Received: 1 November 2018; Accepted: 2 December 2018; Published: 5 December 2018

**Abstract:** Egg yolk is used as an emulsifying agent. Nevertheless, its high concentration of cholesterol is linked to chronic degenerative diseases that cause cardiovascular disease. In this study, three methods for reducing the level of cholesterol in egg yolks were studied. The first method consisted of physical separation of the granules contained in the yolk (NaG). The second method applied was the use of anionic chelating biopolymers, such as arabic gum solution (AG) and mesquite gum solution (MG), and the third method was extraction with a solvent (SA). For this purpose, the cholesterol present in egg yolks, the microstructure, particle size, zeta potential, and its emulsifying capacity were determined. The amount of cholesterol removed was 97.24% using 1% mesquite gum (MG1%), and 93.26% using 1% Arabic gum (AG1%). The zeta potential was determined, and the isoelectric point (*ζ* = 0) of egg yolk was identified as pH 4.6. While, at this pH, the zeta potential of mesquite gum was −14.8 mV, the zeta potential for the arabic gum was −16 mV. The emulsifying capacity of MG1% was 62.95%, while the emulsifying capacity of AG1% was 63.57%. The complex obtained can be used in the development of functional foods reduced in cholesterol.

**Keywords:** egg yolk; cholesterol extraction; granules extraction; anionic chelating biopolymers

#### **1. Introduction**

Egg yolk is a good source of lutein, zeaxanthin, proteins, lipids, and vitamins in human nutrition and is made up of practically 50% solids. The major constituents of the solid matter are lipids (65–70% on dry basis) and proteins (30% on dry basis). The proteins present are livetins, lipoproteins [1], and some particles including high-density lipoproteins (HDLs), low-density lipoproteins (LDLs), and phosvitin [2,3].

Egg yolk is an efficient ingredient in many food products, and its functional properties include emulsifying, coagulating, foaming, and gelling properties [4]. Moreover, it contains proteins, vitamins, minerals, essential fatty acids, phospholipids, and other compounds. However, it has high cholesterol

content; one simple egg contains between 200 mg and 300 mg of cholesterol/100 g; therefore, it almost meets the dietary intake limit set by the American Heart Association of <300 mg/day [3,5]. Consuming products with a high amount of cholesterol can result in cardiovascular disease. Clinical studies demonstrate that dietary cholesterol may increase serum LDL in certain individuals (hyper-responders). This is generally accompanied by an increase in HDLs [6]. Sichittiano et al. [7] undertook a review of nutraceuticals and functional food ingredients that are beneficial to vascular health. Grape seeds can reduce blood lipid levels, since it includes proanthocyanidins (polyphenols), which seem to play the main role in this process. Proanthocyanidins reduce the levels of triacylglycerol in chylomicrons and in very-low-density lipoproteins (VLDLs) [8]. Other functional ingredients are anthocyanins, which act on LDLs and HDLs. The influence on the lipid profile of anthocyanin supplements obtained from berries was evaluated in dyslipidemic patients. A decrease in LDLs was observed in patients after 12 weeks of treatment [9]. Mesquite gum has some phenolic compounds that are trapped in the gum matrix, and these substances are involved in the defense of the plant; in addition, the polymerization of these compounds produces polyphenols, resulting in brown or yellow gum [10].

Several researchers worked on different ways of decreasing the amount of cholesterol present in egg yolks. Warren et al. [11] used solvents such as hexane. Hexane forms a blend with egg yolk solids, but it requires a process of filtering to remove the solvent and a prolonged drying period. The yield reported was 62.2% cholesterol. Borges et al. [12] used a ratio of 1:12 *w*/*w* (yolk/acetone), and the emulsifying properties were maintained.

Paraskevopoulou et al. [13] extracted cholesterol from egg yolk with ethanol/water 20:80 (*v*/*v*), and 1.5% (*w*/*v*) polysorbate 80; after that, the dispersion was then centrifuged, and the yolk precipitate had 7.1 ± 0.3 mg of cholesterol. Laca et al. [14] worked on the extraction of egg yolk granules. In this process, egg yolk was mixed with water (1:15 *v*/*v*), the pH was adjusted to 7, and it was kept overnight at 4 ◦C; the granules had a concentration of 291 mg of cholesterol/100 g of egg yolk, equivalent to a reduction of 77% cholesterol. Another method involved the reduction of cholesterol using β-cyclodextrin (β-CD) due to its affinity for non-polar molecules such as cholesterol [15]. Jeong et al. [16] reported that cholesterol removal was 92.76% when using 25% crosslinked β-CD at 40 ◦C. Chiu et al. [17] used the immobilization of β-cyclodextrin in chitosan beads (Ch-BCD) by cross-linking with 1,6 hexamethylene diisocyanate (HMDI) reagent for cholesterol absorption from egg yolk, removing 92% of the cholesterol. That cholesterol was removed using 1% *w*/*v* Ch-BCD for 2 h at a proportion 1:30 yolk to water, and a mixture of β-cyclodextrins with chitosan. Garcia et al. [18] used high-methoxyl pectin at a concentration of 3% *w*/*w*, ionic strength 0.39 M, and pH equal to 9.2, and subsequently obtained a reduction of 88.6% cholesterol. Hsieh et al. [19] developed a complex of egg yolk and acacia gum applying the following concentrations: 1%, 3%, 5%, and 10% (*w*/*w*) with cholesterol extraction rates of 70%, 86%, 79%, and 59%, respectively.

Meanwhile, anionic biopolymers, such as gum ghatti, gum tragacanth, gum karaya, xanthan gum, and especially, Arabic gum (AG), are chelating agents, forming complexes with lipoproteins in the yolk. Lipoprotein molecules are positively charged, and anionic polysaccharides are used as chelating agents, controlling pH and temperature [20].

The methods mentioned above for removing cholesterol are complicated because solvents are used, thus prolonging the extraction time. It was shown that the use of biopolymers, including commercial arabic gum, decreases the concentration of cholesterol present in the egg yolk. However, the shortage of arabic gum, due to drought in the regions where it is produced, stimulated the search for other botanical sources that offer greater security of supply and costs [21]. Therefore, the use of mesquite gum as an alternative to arabic gum is proposed in this study. Mesquite gum (MG) has an ability to stabilize colloidal particles (1–100 µm), which it disperses or emulsifies. This property is manifested due to protein arabinogalactans, which allow adsorption in liquid–liquid interfaces [22,23]. The aim of this work was to reduce the amount of cholesterol present in egg yolk by preparing a complex of biopolymer mesquite gum–yolk and arabic gum–yolk, before comparing the results with cholesterol extraction using a solvent, and with physical separation of the granule using sodium chloride.

#### **2. Materials and Methods**

#### *2.1. Materials*

Eggs were purchased from the local market (commercial brand name "El Dorado"; a box containing 30 white eggs, expiration date 9 April 2018, batch 043501). Salt, sugar, and vinegar were purchased at the supermarket. The separation of the yolk from albumen was undertaken manually. The vitelline membrane was then cut with a scalpel blade, and the content of the yolk was collected in a glass vessel. The yolk was then mixed gently with a glass rod. The solution was maintained at 4 ◦C in refrigeration (pH 7). The arabic gum and mesquite gum were purchased from Natural Products of Mexico (Yautepec, Morelos, México), and the gum was purified according to the method of Vernon et al. [23]. Sodium chloride, acetone, ethanol, and hexane reagent-grade chemicals were purchased from Desarrollo de Especialidades Químicas (Monterrey, Nuevo León, México); the cholesterol standard was obtained from Sigma-Aldrich Chemical (Toluca, México). Deionized water was used in all the experiments.

#### *2.2. Cholesterol Reduction in Egg Yolk*

The egg yolk was treated with physical separation, polysaccharides (arabic gum and mesquite gum), and a solvent for reducing the cholesterol content. The first process was the physical separation of the granules contained in the yolk (NaG) using aqueous salt solution and separation through centrifugation. Another process was the separation of cholesterol using complexation with biopolymers, arabic gum (AG) and mesquite gum (MG). In the third method, acetone extraction (SA) was used.

#### 2.2.1. Egg Yolk Granule Extraction (NaG)

The cholesterol was removed using the method of Laca et al. [14], with a few modifications; 8.5 g of egg yolk and 11.4 g of 0.15 M NaCl solution (1:1.34) was mixed with a vortex for 1 min at 25 ◦C. The content was centrifuged (Hermle Labnet Z326, Labnet International, Inc., Wehingen, Germany) at 10,000× *g* for 45 min at 25 ◦C. Finally, the compounds were separated carefully from the aqueous fraction through decantation. The product obtained was lyophilized, and was stored at −20 ◦C until analysis with GC.

#### 2.2.2. Anionic Polysaccharide/Egg Yolk Complexes

Complex formation was obtained based on the methodology reported by Hsieh et al. [19]. Stock solutions of arabic gum were prepared at 1% (AG1%), 3% (AG3%), and 10% (AG10%), in addition to 1% (MG1%), 3% (MG3%), and 10% (MG10%) mesquite gum. All solutions were maintained in constant agitation all night long. All concentrations are given in ratios of weight/weight (*w*/*w*).

Firstly, 3 g of egg yolk was mixed with 1 g of gum solution (at the concentrations mentioned above) and 4 g of water. The solution was mixed for 1 min in a vortex (Mixer Labnet International, Edison, NJ, USA); it was then centrifuged (Hermle Labnet Z326, Labnet International, Inc., Wehingen, Germany) at 6000× *g* for 15 min at 25 ◦C. After that, the supernatant was decanted, and then aggregated with 0.5 g of solution (0.9 M NaCl); this was mixed for 1 min in the vortex. Then, 6 g of ethanol was poured and mixed in the vortex for 1 min at 25 ◦C, and centrifuged at 6000× *g* for 15 min. After that, the solution was decanted; 6 g of ethanol was added to the lipoprotein/anionic biopolymer complexes, and the solution was mixed, before being carefully decanted. The precipitate complex was quantified [19]. Samples were lyophilized and stored at −20 ◦C until analysis with GC.

#### 2.2.3. Solvent Extraction

Extraction of cholesterol with a solvent (SA) was performed using the method described by Borges et al. [12] using a ratio (*w*/*w*) of 1:12 (yolk/acetone), and mixing at 100 rpm for 2 min in the

stirrer (EURO-ST 60 D S001, IKA, Wilmington, NC, USA). This permitted the separation of the sample after 10 min, and the solvent of the precipitate was carefully decanted. Finally, the precipitate was washed with water. The samples were lyophilized and stored at −20 ◦C until analysis with GC.

#### *2.3. Quantification of Cholesterol Using Gas Chromatography*

Method 26.052 of the Association of Official Agricultural Chemists (AOAC) [24] was used. For the acid hydrolysis, 0.2 g of the sample was mixed with 2 mL of methanol, and 7% H2SO<sup>4</sup> (*v*/*v*). Next, the sample was heated for 90 min at 80 ◦C; after that, the sample was cooled at 25 ◦C. Then, 3 mL of hexane was added and mixed for 1 min in the vortex (Mixer Labnet International, Inc., Edison, NJ, USA). The solution was kept for 15 min until the formation of two phases was completed; the process was performed twice. The supernatant recovered was mixed and diluted into a 10-mL flask with hexane. Subsequently, the solution was analyzed with GC (7890B, Agilent Technologies, Santa Clara, CA, USA), coupled to a mass spectrometer (5977A, Agilent Technologies, Santa Clara, CA, USA), and equipped with an HP-5MS capillary column (length: 30 m; inner diameter (ID): 0.25 mm; film thickness: 0.25 µm). The injected sample was 1 µL on split mode. The chromatographic conditions were as follows: column temperature 70 ◦C, kept for 1 min; increased to 200 ◦C at 10 ◦C/min, maintained for 2 min; increased to 300 ◦C at 10 ◦C/min, maintained for 7 min. The temperature of the injector was 250 ◦C, and the temperatures of the ion source and quadruple were 230 ◦C and 150 ◦C, respectively. The carrier gas helium flow rate was 1 mL/min. The ionization with electron impact was 70 eV and the scan acquisition mode had a range of 30 to 400 UMA. The calibration curve was done with a cholesterol standard of 20 to 120 ppm.

#### *2.4. Particle Size Measurement*

The particle size distribution of the samples (AG1%, AG3%, AG10%, MG1%, MG3%, MG10%, SA, and NaG), and the yolk were monitored using a Malvern Mastersizer 3000 (Malvern Instruments, Ltd, Worcestershire, UK) particle size analyzer with a unit of Hydro LV with water as a dispersant. The angular scattering intensity data were analyzed to calculate the size of the particles, creating a scattering pattern using the Mie theory of light scattering. The software calculated the particle size distribution (D(3,2)). Optical properties of the sample were defined as a refractive index 1.460 and an absorption of 0.1.

#### *2.5. Zeta Potential (ζ)*

The zeta potential was determined using dynamic light scattering equipment Zetasizer Nano ZS90 (Malvern Instruments, Worcestershire, England, UK). The measurements were carried out using a universal dip cell (ZEN 1002, Malvern Instrument, Worcestershire, UK) at 25 ◦C, using the diluted solutions. The zeta potential is related to the velocity of the solutions in an electric field. The equipment software converts the electrophoretic mobility measurements into zeta potential values using the Smoluchowski model [25]. The zeta potential of egg yolk/polysaccharide solutions at different pH levels was measured with the method of Navidghasemizad et al. [26]. About 0.1 g of sample was diluted to a final volume of 20 mL using distilled water, and the pH was adjusted to values of 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, and 10.0 using 0.1 M HCl or 0.1 M NaOH solution while stirring the samples. Both complexes, yolk/AG and yolk/MG, were monitored for the zeta potential. The zeta potential was calculated from the average of three measurements of the diluted solutions.

#### *2.6. Emulsifying Capacity*

Samples were prepared individually, and 10 g of sample was mixed (NaG, SA, AG1%, and MG1%, samples with high yield) with salt (1.59%), sugar (1.06%), and water (11%), separately. Then, the vinegar (3.17%) was added and kept under constant stirring at 300 rpm in the stirrer (IKA Eurostar 60 digital) for 10 min. The test ended when it was not possible to integrate more oil contained in the burette (*Voil*), and a layer was observed on the surface of the emulsion (*VEmulsion*) [27]. The values were estimated as percentage of emulsified oil (%*EC*) in total emulsion using Equation (1).

$$\% \text{EC} = \frac{V\_{oil}}{V\_{Emulus}} \times 100\tag{1}$$

#### *2.7. Microstructure Analysis*

The topology was analyzed employing scanning electron microscopy (SEM) obtained with the methodology reported by Valverde et al. [28]. Briefly, the samples were fixed overnight in 3% glutaraldehyde in 25 mM phosphate buffer (pH 3.25). After that, the samples were consecutively dehydrated with 20%, 40%, 60%, 80%, and 100% ethanol. Then, the ethanol was consecutively removed with 20%, 40%, 60%, 80%, and 100% acetone, and the samples were analyzed. Finally, they were dried at a pressure of 1 × 10−<sup>2</sup> Torr (soft vacuum) in a vacuum desiccator. The dried sample was placed on aluminum SEM stubs and coated with gold/palladium. The microscope used was a JSM-6490LV (JEOL, Tokyo, Japan).

#### *2.8. Color Analysis*

The granules were measured for color in the lightness *L\*,* redness (*a\**)*,* and yellowness (*b\**) system. Measurements were carried out using a ColorFlex EZ (Hunter Lab, Reston, VA, USA). A fixed amount of sample was poured into the measurement cell, and analyses were conducted in specular exclusion mode.

The color changes are expressed as **∆***E* with the color of the egg yolk as a reference sample [29]; hence, **∆***E* is the total color change due to the different contributions calculated using Equation (2).

$$
\Delta \mathbf{E} = \sqrt{(\Delta L^\*)^2 + (\Delta a^\*)^2 + (\Delta b^\*)^2} \tag{2}
$$

#### *2.9. Statistical Analysis*

All tmeasurements were performed in triplicate, and ANOVA was performed with a confidence level of 95% (*p* < 0.05) using SPSS 20 software (IBM, SPSS Inc, Chicago, IL, USA). To determine the statistically significant difference between values, a one-way variance analysis and a Tukey test were performed.

#### **3. Results and Discussion**

#### *3.1. Cholesterol Removal in Egg Yolk*

The process of extracting egg yolk granules, and chelating with arabic gum and mesquite gum was used as an alternative to preparing yolk with a high concentration of protein and low cholesterol as a functional ingredient in the food industry, especially due to its functional attributes, such as foaming capacity, high level of phosvitin and high-density lipoprotein, and emulsifying and binding properties [28,30]. Notwithstanding the process, other properties, such as the emulsifying of oil, are reduced. In order of importance, the removal of cholesterol from lower to highest was Na<sup>G</sup> < S<sup>A</sup> < AG < MG. The cholesterol removed from the egg yolk with Na<sup>G</sup> (Table 1) was the least effective method (51.43%). It was previously reported that the Na<sup>G</sup> method with ionic solvent removed cholesterol and increased the protein content. The method is relatively fast and inexpensive. The results obtained in the laboratory were compared with those obtained by Strixner and Kulozik [31]. Based on our data, the removal of cholesterol with solvent was the second best method for cholesterol removal (64.15%). However, Martucci and Borges [32] reported a six-stage extraction system of 92% cholesterol removal in a computer simulation study.



Note: Differing letters within a column are significantly different (*p* < 0.05) ± standard deviation (*n* = 3). NaG—physical separation of granules; AG—arabic gum; MG—mesquite gum; SA—solvent extraction.

The removal of cholesterol using acetone in the laboratory was lower than the 81%, as reported by Borges et al. [12]. The removal of cholesterol was lower due to the concentration of water in the fresh yolk, the extraction time, and the acetone ratio. It was reported that the use of non-polar organic solvents prevents protein denaturation, especially in emulsifying activity, which is related to the solubility of the protein. Meanwhile, polar organic solvents reduce the emulsifying activity because they break the hydrophobic interactions between lipids and proteins. Even so, the use of solvents is not completely accepted due to the residues that may remain in the product [33].

The use of arabic gum allowed cholesterol removal between 83.85% and 93.26% (Table 1). The concentrations AG3% and AG10% show significant differences (*p* < 0.05). The removal of cholesterol was greater when 1% arabic gum was used. Meanwhile, when 3% and 10% arabic gum was used, the cholesterol removal was almost 83.85% and 89.93%, respectively, despite the fact that pH values were similar to neutral. The structure of arabic gum is a branched heteropolysaccharide with anionic properties. The quantity of arabic gum–yolk depends on the polyanionic properties of the gum, especially of residues of D-glucuronic acid (~2%) and D-glucuronic acid (~21%) [2,34]. The arabic gum has properties of anionic polysaccharides, which may be used as chelating agents, forming insoluble electrostatic complexes (chelating agent/lipoprotein) [18]. Navidghsemizad et al. [35] used a ratio of 50 g of fresh yolk per gram of arabic gum to observe the separation of phases at different pH levels. They concluded that the nature of the polysaccharide and pH had important effects, resulting in the phase separation behavior.

Other anionic polysaccharides were used as chelating agents [19]; however, their low solubility and high viscosity reduce their practical applications. Mesquite gum is a polysaccharide which contains acidic residues of β-D-glucuronic and 4-*O*-methyl-β-D-glucuronic, bound to mono-sugars or oligosaccharide chains [36,37]. The removal of high cholesterol levels when the mesquite gum solution was used was 1% *w*/*w* (*p* < 0.05). From mesquite gum MG1%, MG3%, and MG10%, extractions of cholesterol of 97.24%, 96.68%, and 96.60%, respectively, were obtained. No significant differences were observed in the different concentrations where mesquite gum was used.

In light of our results, it can be stated that sodium chloride or acetone have a lower capacity to remove cholesterol compared to both anionic polysaccharides. The greater efficiency in cholesterol removal was obtained with mesquite gum at 25 ◦C. Nonetheless, obtaining complexes from mesquite gum–yolk allowed a high removal of bound cholesterol. We think that the complex obtained can be used as functional supplement, necessary for reducing the unwanted effects of cholesterol. Scicchitano et al. [7] mentioned the importance of reducing lipid levels, especially for coronary artery disease.

#### *3.2. Particle Size Measurement*

In natural conditions, the yolk is a supramolecular assembly of lipids and proteins, and a highly organized system with approximate size between 0.8 µm and 10 µm [38]. The insoluble structure of yolk has a size range between 0.3 µm and 2 µm [14]. Molecular assembly can be disorganized

into individual structures depending on the affinity of the solvent or polysaccharide used to remove cholesterol. The morphology and topology of the particle size determined using the light scattering method can be seen in Figure 1. Figure 1a,b show the granules and the complexes formed with arabic gum and mesquite gum (1%, 3%, and 10%), respectively.

−− −− −− −− −− −− − − −− −− −− − − −− −− −− −− −− −− −− **Figure 1.** The average particle size: (**a**) −− MG1%, − − MG3%, and −N− MG10%; (**b**) −− AG1%, − − AG3%, and −N− AG10%; (**c**) −− MG1%, − − AG1%, −− NaG, and −H− SA; (**d**) −− MG3%, − − AG3%, −− NaG, and −H− S<sup>A</sup> and (**e**) −− MG10%, − − AG10%,−N− NaG, −H− SA. MG—mesquite gum; AG—arabic gum; NaG—physical separation of granules; SA—solvent extraction.

Regarding the distribution of proteins, the structure of the supramolecular system contains many particles of different sizes, three of which are of special interest: the HDLs, which range from 7 to 20 nm; the micelles formed by LDLs in the egg yolk plasma, which range from 17 to 60 nm; and the LDL sources present in the yolk, which range from 80 to 350 nm [31,38]. Anton [39] and Hsieh et al. [19] mentioned that granules of yolk are composed of 70% HDLs, 16% phosvitin, and 12% LDLs. We believe that the granules obtained from solvents and polysaccharides follow the same distribution pattern. The granules and lipoprotein/anionic polysaccharide complexes had different particle size population profiles; these size profiles may be associated with the process of cholesterol removal, the anionic polysaccharides (arabic gum and mesquite gum), and concentration (Figure 1a,b; 1%, 3%, and 10%), and the ratio. The change in distribution of particle size may be due to the viscosity and the charge density of proteins diverse in egg yolk, between MG–yolk and AG–yolk complexes, including the concentration of reactive groups contained in both biopolymers, which form an electrostatic complex [40]. In macroscopic terms, the distribution of granules was separated into three different sizes, based on the chelate concentration. The concentration of chelating polysaccharides at 10% (arabic gum and mesquite gum) had a range of particle size distribution between 0.3 µm and 600 µm. Similarly, the same profile was obtained with low-molecular-weight chelates. The yolk–chelate ratio of 3% for both polysaccharides produced a range of granule size distribution from 0.3 µm to 300 µm (Figure 1d). Finally, when a high molecular weight at 1% concentration was used, the yolk–chelate reduced the size of granules to between 0.3 µm and 250 µm (Figure 1c). More specifically, the granules and lipoprotein complexes obtained had multimodal distributions, as described below. Figure 1e shows three different population distributions when NaG, SA, and AG and MG at 10% were used. The first range was from 0.3 µm to 1 µm, the second was from 1 µm to 4 µm, and the third range of distribution was from 10 µm to 300 µm, when using the Na<sup>G</sup> method. The removal of cholesterol using acetone had three different populations of granule size; the range of least distribution was between 0.3 µm and 0.9 µm,

and the greatest range showed a variation 0.9 µm to 10 µm. Finally, particle size distribution could be observed in the range of 10 µm to 200 µm. A similar range of distribution was observed when using concentrations of 3% and 1% (both polysaccharides and solvents).

#### *3.3. Zeta Potential (ζ)*

*ζ*

The interaction between biopolymers may be segregative, due to steric repulsion or associative interactions such as hydrophobic interactions and hydrogen bonding [41]. Electrostatic interactions are the most common force for the complex formation [42]. The pH affected the charge of biopolymers and proteins of the egg yolk, which influenced the zeta potential *(ζ)* as a function of pH; AG3% and MG3% were studied in the pH range of 2–10. *ζ*

The zeta potential of egg yolk was positive at pH values of 2 and 4 (+15.5 mV and +8.5 mV, respectively), while it was negative at pH values of 5 and 10 (Figure 2a); this was similar to the report by Navidghasemizad et al. [26], who obtained positive values in the range of pH 3–5, and, at pH 6, it was negative above the zeta potential. The isoelectric point (*ζ* = 0) of egg yolk was found to be pH 4.6, determined from the zeta potential. While, at this pH, the zeta potential of mesquite gum was −14.8 mV, the zeta potential for the arabic gum was −16 mV. The formation of insoluble complexes appears to occur at pH 3. At this point, the density of the opposite charge between egg yolk and polysaccharides (arabic gum and mesquite gum) has practically the same magnitude. However, the percentage of cholesterol removal was lower than at pH 7. The values are shown in Table 1. The mesquite gum had values of −2.25 mV to −24.61 mV, in the acidic to basic (2 to 10) pH range. We believe that the free and exposed glucuronic acid and protein residues present in mesquite gum reacted in the different media [43] with acidic residues present in the polymer, similar to arabic gum [1]. In Figure 2b, the zeta potential profile of the yolk–3% polysaccharide complex essentially modified the isoelectric point to pH 4 for both biopolymers. For values lower than pH 4, the load profile was positive; however, for values higher than pH 4, the profile was negative. *ζ* − − − −

**Figure 2.** Average zeta potential *(ζ)* of (**a**) −− egg yolk, − − mesquite gum, and −N− arabic gum, and (**b**) −− egg yolk, − − AG3%, and −N− MG3%.

#### *3.4. Emulsifying Capacity*

−− −− −−

The emulsifying properties only work in specific cases; these properties cannot be generalized. Therefore, we determined the emulsifying properties of the granules obtained, depending on the different treatments. As shown in Table 1, when comparing the emulsifying capacities of the methods discussed, the egg yolk method supports the highest percentage of oil, followed by NaG, which is similar to the findings obtained by Laca et al [14]. In AG3%, a lower emulsifying capacity than Na<sup>G</sup> was obtained. It was reported that this is because using arabic gum to remove cholesterol allows the loss of 66% of proteins, including the yolk's emulsifying proteins, along with lipids [18].

*ζ* −− −− −−

For the egg yolk–mesquite gum complex, the emulsifying capacity of MG3% was 62.95%. It is possible that the ethanol used in the washing of biopolymer–yolk complexes dissolves phospholipid cholesterol and diminishes the emulsifying capacity. A monolayer of water prevents the denaturation of the protein [44]; when using SA, the result obtained was 72.33%. Both products of Na<sup>G</sup> were proposed as additives with low cholesterol in products like muffins [4] and salad dressings [14]. The complex of egg yolk–mesquite gum with polyphenols can be used for the development of foods reduced in cholesterol, thereby helping avoid health problems like cardiovascular disease.

#### *3.5. Microstructure Analysis*

The microphotograph of Na<sup>G</sup> (Figure 3g) shows irregular structures with small aggregates; it is probable these are HDL–phosvitin complexes linked by phosphocalcic bridges between the phosphate groups [45]. The microstructures of treatments with arabic gum (AG1% AG3%, and AG10%; Figure 3a,b,c) show small aggregates, which are probably due to the interaction of the biopolymer carboxyl groups and the lipoproteins of egg yolk. In the microphotographs of different treatments with mesquite gum (MG1%, MG3%, and MG10%; Figure 3d,e,f), spherical structures within 2 to 5 µm of size were observed; it is probable that they correspond to the interaction between arabinogalactan proteins in mesquite gum and the lipoproteins of egg yolk [20].

The microphotograph of egg yolk shows an irregular structure because it is a complex system with several particles in suspension in a fluid that contains proteins. The microphotograph of egg yolk with acetone (SA) shows irregular structures with small aggregates.

**Figure 3.** SEM photographs using different cholesterol extraction methods: (**a**) AG1%, (**b**) AG3%, (**c**) AG10%, (**d**) MG1%, (**e**) MG3%, (**f**) MG10%, (**g**) NaG, and (**h**) SA.

#### *3.6. Color Analysis*

*Δ* The values regarding the lightness (*L\**), redness (*a\**), and yellowness (*b\**) values are shown in Table 2. There are no significant differences for lightness *L\**, while *a\** and *b\** values show significant differences (*p* < 0.05). The Na<sup>G</sup> shows low (∆*E*) total color change, which we can state to be similar to egg yolk; this method removes less pigment than SA. The redness (*a\**) and yellowness (*b\**) values for Na<sup>G</sup> were 1.7 and 22.73, respectively, greater than those reported for S<sup>A</sup> (8.91 and 49.65, respectively). The color of egg yolk is attributed to carotenoids (xanthophylls, including lutein, zeaxanthin, β-cryptoxanthin, and β-carotene) [46]. We suggest that, if a large amount of cholesterol is removed, more carotenoids are also removed.


**Table 2.** Color measurements of different methods for cholesterol reduction.

Note: Letters varying within a column are significantly different (*p* < 0.05) ± standard deviation (*n* = 3). *L\**—lightness; *a\**—redness; *b\**—yellowness; ∆*E*—total color change.

#### **4. Conclusions**

Separation efficiency of complex lipoproteins (HDLs)–mesquite gum shows a strong dependence on pH. The greatest cholesterol reduction was seen at pH 7.0. The amount of cholesterol removed was 97.24% using 1% mesquite gum (MG1%), and 93.26% using 1% arabic gum (AG1%). This is a consequence of the chemical composition in the chelate (mesquite gum or arabic gum) and yolk. The use of mesquite gum shows structural changes in the form of definite spheres with a low size in comparison with arabic gum, observed using SEM. The high removal of cholesterol contained in the egg yolk using mesquite gum or arabic gum reduced the primary emulsifying capacity of the egg yolk. The use of mesquite gum to remove cholesterol is an alternative method that does not require organic solvents. The use of 3% mesquite gum removed 12.83% more cholesterol than the same concentration of arabic gum. The complex obtained can be used in the development of functional foods reduced in cholesterol.

**Author Contributions:** M.B.V., J.G.B.G., and E.G.M. designed and led the research, and wrote the paper. M.A.N.G. collaborated in the experimental phase of research. J.R.R. collaborated in the experimental phase of research. A.C.M. collaborated in the experimental phase of research. C.T.G.R. collaborated in the experimental phase of research.

**Funding:** The authors would like to thank the "Consejo Nacional de Ciencia y Tecnología" (CONACyT) of Mexico for partial financing of this project through grant #277814, as well as through the grant Problemas nacionales-2015-01-1470 and the grant CONACyT CB-157511.

**Conflicts of Interest:** The authors declare no conflicts of interest.

#### **References**


#### **Sample Availability:** Not available.

© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Effect of Beta Cyclodextrin on the Reduction of Cholesterol in Ewe's Milk Manchego Cheese**

#### **Leocadio Alonso 1, \*, Patrick F. Fox 2 , María V. Calvo <sup>3</sup> ID and Javier Fontecha 3**


Academic Editors: Lillian Barros and Isabel C. F. R. Ferreira

Received: 21 June 2018; Accepted: 18 July 2018; Published: 20 July 2018

**Abstract:** Beta-cyclodextrin (β-CD) is a cyclic oligosaccharide consisting of seven glucose units and is produced from starch using cyclodextrin glycotransferase enzymes to break the polysaccharide chain and forming a cyclic polysaccharide molecule. The use of β-CD in food research for reduction of cholesterol is increasing due to its affinity for non-polar molecules such as cholesterol. The aim of this study was to evaluate the feasibility of using β-CD in cholesterol removal from pasteurized ewe's milk Manchego cheese and evaluate the effect on the main components of the milk, lipids, and flavor characteristics. Approximately 97.6% cholesterol reduction was observed in the cheese that was treated using β-CD. Physicochemical properties (fat, moisture and protein) were not changed by the β-CD treatment, except the soluble nitrogen and non-protein nitrogen that showed slight differences after the treatment. The amount of the different components of the lipid fraction (fatty acids, triglycerides and phospholipids) were similar in cheeses treated and not treated with β-CD. Flavor compound and short chain free fatty acids were not mostly significantly influenced by the effect of the β-CD. β-CD molecules are edible and nontoxic and as a result they can be used safely for cholesterol removal processing in cheese manufacturing. Therefore, the present study suggests that β-CD treatment is an effective process for cholesterol removal from Manchego cheese while preserving its properties.

**Keywords:** beta cyclodextrin; ewe's milk; cheese; Manchego; lipids; cholesterol

#### **1. Introduction**

Although dairy products in general have the image of being healthy foods, this is often not the case for products with a high fat content such as butter, cream and cheeses. The World Health Organization and the American Heart Association have recommended that consumers reduce their consumption of saturated fatty acids and cholesterol to lower the risk of coronary heart disease. This advice, coupled with radical opinions, have created a demand for low-cholesterol products [1]. Nowadays, there is a growing interest in the manufacture of cholesterol-reduced dairy products. Food companies have developed many methods to reduce cholesterol, however, most of these methods are relatively nonselective and remove flavor and nutritional components when cholesterol is removed. Moreover, some methods require high investment and operation costs. Methods for reducing cholesterol in foods have been developed including blending with vegetable oils [2,3], extraction by distillation and crystallization [4,5], adsorption with saponin and digitonin [6,7], assimilation of cholesterol by enzymes from microorganisms [8,9] and removal by supercritical carbon dioxide extraction [10,11]. In the last years, several studies have been published describing the use of β-CD in food applications [12–14]. It has been proved that the β-CD molecule can be used as non-toxic and non-digestible molecule to remove cholesterol effectively from milk and dairy products, egg yolk, and lard [15–20] with much less investment and operation costs. β-CD is a cyclic oligosaccharide consisting of seven glucose units and is produced from starch using cyclodextrin glycotransferase enzymes, to break the polysaccharide chains and form cyclic polysaccharide molecules. The molecule of β-CD is doughnut shaped and its central portion is a circular hydrophobic space similar in diameter to a cholesterol molecule, giving the molecule its affinity for non-polar molecules such as cholesterol [21,22].

Manchego cheese is one of the most representative of the Spanish hard cheeses. It is manufactured in the region of Castilla-La Mancha (Spain) using pure ewe's milk from local herds under conditions regulated by an origin appellation. Manchego cheese is a rich in fat (the fat content in the dry cheese is higher than 50%) [23,24], and possesses a characteristic sharp flavor, which increases with the ripening time. Its texture is smooth but consistent, and a few irregular holes randomly distributed in ivory-colored paste. Although the most of investigations for removing cholesterol in milk using β-CD were performed in cow's milk, no investigations have been reported on the effect of β-CD on reduction of cholesterol in ewe's milk. Therefore, the aims of this study was to evaluate the feasibility of the β-CD in cholesterol removal from pasteurized ewe's milk Manchego cheese and its effect on the main components of milk, focusing especially on the lipidic fractions, and flavor characteristics.

#### **2. Results and Discussion**

#### *2.1. Gross Composition*

Due to the structural characteristics of β-CD and processing conditions used during cholesterol removal with β-CD, it is possible that some of the milk constituents are also entrapped and removed along with cholesterol. Thus, it is important to investigate the compositional changes occurring during the cholesterol removal process in Manchego cheese.

The chemical composition and cholesterol removal rate of control cheese (CC) without β-CD in milk and the experimental cheese (EC) with 1% of β-CD in milk are presented in Table 1. We used 1% β-CD because in previous studies we studied different concentrations of β-CD in the range (0.1 to 1%) for the elimination of cholesterol in cow's milk fat. We found that in that study the optimal concentration to obtain cholesterol reduction higher than 90% was with a β-CD concentration approx. 0.8% [15]. Fat, moisture and protein content showed similar ratio between the CC and the EC (34.50 ± 1.12% vs. 32.51 ± 1.18%; 36.79 ± 1.65% vs. 38.15 ± 1.93%; 25.68 ± 1.04% vs. 25.10 ± 1.16%) respectively. Fat/dry matter and protein/dry matter (%) were slightly lower in EC with β-CD that the CC as a result of the higher moisture content, as suggested in the study by Seon et al. [20]. The lower fat content of the cholesterol reduced cheese than the control might be attained to the less incorporation with casein via a fat protein network, probably due to modification of the casein matrix by β-CD [25]. Soluble nitrogen (SN) and non-protein nitrogen (NPN) showed differences (*p* ≤ 0.05) between CC and EC cheese (4.76 ± 0.23% vs. 5.79 ± 0.32%; 2.41 ± 0.19% vs. 3.95 ± 0.24%), this could be due to the slight increase in the proteolysis in EC cheese that may reflect a higher peptidase activity in the EC by the influence of the β-CD [26]. During the ripening period proteolysis occurs which is an important biochemical event governing the sensory profile. The insoluble caseins are partially converted into polypeptides and amino acids. Treatment of the milk with β-CD from which cheese is manufactured results in modification of caseins matrix and thus altering the SN and NPN and consequently could be accelerate a little the ripening period of the cheese. The cholesterol removal rate of CC related to EC (195.67 ± 6.03 mg/100 g fat vs. 1.37 ± 0.19 mg/100 g fat) reached a reduction of 97.29% (Figure 1). Similar cholesterol removal were also found by Kwak et al. [27] in a study of removal of cholesterol from Cheddar cheese and Kin et al. [28] in blue cheese using β-CD. The remain β-CD showed also differences (*p* ≤ 0.05) between CC and EC with value of 0.31%. It confirms that cholesterol removal by β-CD does not affect the proximate chemical composition of Manchego ewe's milk cheese.


**Table 1.** Gross composition of the control and the experimental Manchego cheese by the effect of the β-CD.

CC, control cheese without β-CD in milk; EC, experimental cheese with 1% β-CD in milk; SN, soluble nitrogen (% as protein); NNP, non-protein nitrogen (% as protein); REE (%), relative experimental error; Mean standard deviation (*n* = 12); a,b Different letters in the same row mean significant differences (*p* ≤ 0.05).

**Figure 1.** Cholesterol profile by gas chromatography with flame ionization detector in control cheese and experimental cheese with 1% of beta cyclodextrin. Peaks: 1 = 5α-cholestane; 2 = cholesterol. Blue line: control cheese (CC); Red line: experimental cheese (EC).

#### *2.2. Lipid Characteristics*

Table 2 shows mean values of fatty acids (%) of CC and EC cheeses. Concentrations of individual fatty acids did not exhibit significant differences (*p* ≤ 0.05) between fat from the CC and EC cheese with β-CD. There are few reports regarding studies in manufacturing low cholesterol cheeses by β-CD and the effect on the lipidic fraction. Chen et al. [29], using supercritical fluid extraction with carbon dioxide for fractionating milk fat to remove cholesterol, observed that the fractionated milk fat showed considerable differences in fatty acids composition compared with the control cheeses. The amounts for short and medium chain fatty acids reported by these authors were 40% and 10%

less, respectively, in the extracted milk fat compared with the control milk fat. Similar results were found by Gonzalez et al. [10] in a study on solubility of fatty acids in cream from ewe's milk using supercritical fluid carbon dioxide


**Table 2.** Fatty acids composition (g/100 g fat) from the control and the experimental Manchego cheese by the effect of β-CD.

CC, control cheese without β-CD in milk; EC, experimental cheese with 1% β-CD in milk; REE (%), relative experimental error; Mean standard deviation (*n* = 12); <sup>a</sup> Different letters in the same row mean significant differences (*p* ≤ 0.05).

In our study using β-CD for removing cholesterol and the effect on the composition for short-(C4 to C8) (2.24 ± 0.19% vs. 2.14 ± 0.26%; 1.74 ± 0.06% vs. 1.68 ± 0.05%; 1.70 ± 0.05% vs. 1.66 ± 0.08%), medium-(C10 to C12) (5.02 ± 0.15% vs. 4.95 ± 0.13%; 3.19 ± 0.11% vs. 3.14 ± 0.18%), and long chain-(C14 to C18) (9.22 ± 0.84% vs. 9.21 ± 0.51%; 27.16 ± 1.52% vs. 27.41 ± 1.18%; 13.39 ± 0.55% vs. 13.59 ± 0.52%) fatty acids were no significantly different (*p* ≤ 0.05) between groups respectively. Similar results were found by Alonso et al. [15], in their study of using β-CD to decrease the level of cholesterol in milk fat.

Table 3 shows the mean values of the individual groups of triglyceride composition of fats of the CC and EC cheeses. The triglycerides of the fat cheese were resolved into 16 groups from C26 to C54. Each group is the sum of the different molecular species of triglycerides that contain the same number of carbon atoms. None of differences between control and experimental cheese with β-CD were observed (*p* ≤ 0.05), in the ∑ short-(C24–C32) (1.76 ± 0.20% vs. 1.84 ± 0.56%), ∑ medium-(C34–C48) (77.45 ± 0.85% vs. 77 ± 0.91%), and ∑ long-(C50–C54) (4.53 ± 0.39% vs. 4.43 ± 0.51%). No prior research studies have been reported on the triglycerides in cheeses treated with β-CD for removing cholesterol. Chen et al. [29], Bhaskar et al. [30], and Gonzalez et al. [10], using different techniques, found variations in triglycerides composition between control and experimental milks. The supercritical fluid extraction methods used by these investigators may have caused some variation in triglycerides composition because the triglycerides were removed by solvent extraction, that could have selectively extracted some triglycerides better than other.


**Table 3.** Triglycerides composition (g/100 g fat) from the control and the experimental Manchego cheese by the effect of the β-CD.

CC, control cheese without β-CD in milk; EC, experimental cheese with 1% β-CD in milk; REE (%), relative experimental error. Mean standard deviation (*n* = 12); a,b Different letters in the same row mean significant differences (*p* ≤ 0.05).

In relation to the phospholipid fraction, Table 4 shows the composition in phospholipids (%) of CC and EC Manchego cheeses. Analysis of variance did not reveal any significant difference (*p* ≤ 0.05) in relative composition of the different phospholipid classes among between groups of cheeses related to the total phospholipids. Phosphatidylethanolamine (42.42 ± 4.05% vs. 38.25 ± 1.40%) was the most predominant phospholipid followed by phosphatidylcoline (27.23 ± 0.74% vs. 1.04 ± 2.21%) and sphyngomyelin (26.70 ± 5.32% vs. 25.20 ± 1.53%). Similar results were obtained by Alonso et al. [31], in a study of the effect of the β-CD on phospholipids of the milk fat in pasteurized milk. These three species of phospholipids represented more than 80% of the total phospholipids in dairy products. One of the reasons why the β-CD did not affect to these components of the milk fat could be based on the fact that β-CD specifically forms an inclusion complex with cholesterol. The central cavity of β-CD is hydrophobic, giving the molecule its affinity for non-polar molecules such as cholesterol. The radius of the cavity is such as to accommodate a cholesterol molecule almost exactly, conferring the highly specific nature of the β-CD ability to form an inclusion complex with cholesterol. They are therefore accessible to β-CD in the aqueous phase forming the insoluble inclusion complex which can be removed by centrifugation [15].

**Table 4.** Phospholipids composition of the control and the experimental Manchego cheese by the effect of the β-CD.


CC, control cheese without β-CD in milk; EC, experimental cheese with 1% β-CD in milk; PLs, Phospholipids; PE, phosphatidylethanolamine; PI, phosphatidylinositol; PS, phosphatidylserin; PC, phosphatidylcoline; SM, sphyngomyelin; REE (%), relative experimental error; Mean standard deviation (*n* = 12); <sup>a</sup> Different letters in the same row mean significant differences (*p* ≤ 0.05).

#### *2.3. Flavor Characteristics*

Flavor compounds isolated from CC and EC cheeses with three months of ripening are shown in Table 5. A total of 13 flavor compounds were isolated in both cheeses and some differences were observed between samples. In all cheeses, 13 flavor compound were detected, including five ketones, three aldehydes and five alcohols. Analysis of the variance did not reveal any significant difference in the total amount of ketones (2505.61 ± 36.40 ppm vs. 2314.95 ± 26.07 ppm) aldehydes (1139.63 ± 18.68 ppm vs. 1377.45 ± 24.94 ppm) and alcohols (4235.77 ± 17.13 ppm vs. 4808.87 ± 23.79 ppm) between CC and EC cheeses. 3-methylbutanal (1121.42 ± 48.32 vs. 1358.96 ± 70.32) and ethanol (4107.60 ± 62.30 ppm vs. 4685.30 ± 95.79 ppm) were the only compounds significantly different (*p* ≤ 0.05) found in CC and EC cheeses, ethanol production was the highest among flavor compounds measured, similar to those found by Kwak et al. [27] in Cheddar cheese treated with β-CD. In the study by Jeon et al. [32] of the removal of cholesterol of cream cheese by β-CD, no differences were found in the overall flavour compounds in the treated cheese compared to the regular cream cheese.

Ketones with odd carbon number have typical odor characteristics and low perception thresholds. These compounds are formed by β-oxidation and decarboxylation of fatty acids. It is known that aldehydes are not the major compounds in cheeses, as they are rapidly converted to alcohols or their corresponding acids. Branched chain aldehydes like 3-methylbutanal are formed by the catabolism of branched chain amino acids by an aminotransferase [33], and this compound was the only statistically different (*p* ≤ 0.05) in EC comparing with the CC cheese together with the ethanol. 3-Methylbutanal is an intermediate in the catabolism of leucine. Lactic acid bacteria present in the cheese together with some yeast are involved in the formation of 3-methylbutanal and alcohols (ethanol) during ripening of the cheese [33], and in our study there is a high proteolysis in the EC with a high content in non-protein nitrogen (include aminoacids as leucine), and this is the main reason why the content of 3-methylbutanal is higher in EC than in the CC. Ethanol was also higher in EC than in CC, due that this compound is also an intermediate in the catabolism of aminoacids and in the fermentation of the residual lactose by the yeast and lactic acid bacteria [32].


**Table 5.** Volatile compounds (ppm) of the control and the experimental Manchego cheese by the effect of the β-CD.

CC, control cheese without β-CD in milk; EC, experimental cheese with 1% β-CD in milk; REE (%), relative experimental error; Mean standard deviation (*n* = 12); a,b Different letters in the same row mean significant differences (*p* ≤ 0.05).

The amounts of short chain free fatty acids (SCFFAs), acetic, propionic, butyric and caproic acids in the control and cholesterol reduced cheeses are shown in Table 6. There was no significant difference (*P* ≤ 0.05) in total and individual amounts of FFAs (149.14 ± 5.86 ppm vs. 154.70 ± 6.12 ppm) at the end of the three month ripened, between the CC and EC cheeses. These results indicate that there was no differences in the amounts of short chain FFAs between the control and the cholesterol reduced cheese made by β-CD. Similar results in the amount of short chain SCFFAs in the control and cholesterol reduced process and cheddar cheese made by β-CD were found by [27,34]. The release of butyric and caproic acid at the three months ripening contribute to the backbone characteristics of Manchego cheese [35,36].

**Table 6.** Short chain free fatty acids (SCFFA) (ppm) of the control and the experimental Manchego cheese by the effect of the β-CD.


CC, control cheese without β-CD in milk; EC, experimental cheese with 1% β-CD in milk; REE (%), relative experimental error; Mean standard deviation (*n* = 12); <sup>a</sup> Different letters in the same row mean significant differences (*p* ≤ 0.05).

The sensory attributes of CC and EC cheese for a maximum of 5 score are shown in Table 7. No significant differences (*p* ≤ 0.05) were observed in flavor (3.32 ± 0.44 vs. 3.07 ± 0.89), arome (3.59 ± 0.49 vs. 3.28 ± 0.83), color (3.69 ± 0.68 vs. 3.49 ± 0.73) and acceptability (3.45 ± 0.60 vs. 3.22 ± 0.76) between CC and EC cheese. These attributes are correlated with the production of SCFFAs acids and methyl ketones during ripening (3 months) in the CC and EC cheese, which were not affected the treatment with β-CD. Texture was significantly different (*p* ≤ 0.05) in the EC with respect to the CC (3.70 ± 0.57 vs. 3.29 ± 0.72). This could be due than in the experimental cheese resulted in a higher proteolysis due to a greater peptidase activity in the cholesterol reduced cheese, that is higher in the EC by the treatment with β-CD and an slight high moisture in the cheese treated, increased by β-CD, which resulted in a slow drainage, as suggested Metzge et al. [37]. The overall preference was maintained over the ripening period of three months and no differences were found between CC and EC for flavor, aroma, color and acceptability. This study indicates that even though some differences were observed, most of the sensory characteristics and overall preferences were comparable to those of the control and three months cheese ripened treated with β-CD. Therefore, we may suggest the possibility of cholesterol reduced Manchego cheese manufactured by β-CD.

**Table 7.** Sensory analysis of the control and the experimental Manchego cheese by the effect of the β-CD. Flavor, arome, color, texture and acceptability were evaluated on a five point scale (1 = poor to 5 = excellent).


CC, control cheese without β-CD in milk; EC, experimental cheese with 1% β-CD in milk; REE (%), relative experimental error; Mean standard deviation (*n* = 12); a,b Different letters in the same row mean significant differences (*p* ≤ 0.05).

#### **3. Materials and Methods**

#### *3.1. Chemicals*

α-Cyclodextrin (α-CD), β-cyclodextrin (β-CD) and all reagents grade were supplied by Sigma (St. Louis MO, USA). Deionized water was prepared by a water purification system (Millipore Co., Burlington, MA, USA).

#### *3.2. Manchego Manufacture*

Ewe's milk was previously treated with 1% β-CD by the method described by Alonso et al. [15]. One hundred L volumes of whole pasteurized milk (74 ◦C for 15 s) milk containing 1.0% *wt*/*vol* of β-CD were placed in a cold room at 4 ◦C and mixed by a stirrer (430 rcf) during 30 min. After mixing, the treated milk was left standing overnight at 4 ◦C (to allow time for binding the cholesterol) and precipitate the cholesterol-β-CD complex at the bottom of the tank. The upper layer without the complex was separated for making the cheeses. Manchego cheese was made by the procedure described by Fernández-García et al. [35]. Cheeses were ripened at a temperature of 12–14 ◦C with relative humidity of 85–90% during 3 months. The cheese-making experiment was carried out in triplicate for control and cheeses treated with 1% of β-CD.

#### *3.3. Gross Composition*

Fat, moisture and protein contents and nitrogen fractions were determined using the method by Alonso et al. [38].

#### *3.4. Beta Cyclodextrin Analysis*

β-CD was analysed by the method proposed by Alonso et al. [39]. Ten g of cheese was mixed with 5 mg of α-CD dissolved in one mL of water (internal standard for quantitative analysis). After shaking for 2 min at 40 ◦C it was centrifuged at room temperature for 40,000 rpm for 30 min, the upper layer was separated and filtered through a 0.45 µm membrane (Millipore Co.). A 30 µL aliquot of the supernatant spiked with the internal standard were transferred to the autosampler. A 10 µL aliquot of the supernatant was injected onto column for HPLC analysis.

The apparatus used for HPLC analysis was a Waters Alliance 2695 separation module coupled to a 410 refractive index (RI) detector, data acquisition and analysis were performed using the Empower 2 chromatography data software (Waters, Milford, MA, USA). Separation was carried out on YMC ODS-AQ column (Teknochroma, Miami, FL, USA). The mobile phase composition was a mixture of methanol and water (7:93) in isocratic condition at a flow rate 1 mL/min. The standard solutions were prepared in water to establish elution time and the quantification of β-CD was conducted by comparing sample peak area of β-CD with α-CD as the internal standard.

#### *3.5. Lipid Extraction*

Lipids were extracted from samples following a procedure described by an International Standard Method for milk and milk products [40]. Briefly, it consisted of an addition of an ammonia-ethanol solution to a test portion followed by lipid extraction using diethyl ether and hexane. Then, the upper layer was removed, and the solvent completely evaporated. The lipid extracts obtained were placed into amber glass vials, flushed with a stream of nitrogen and stored at −20 ◦C until analyzed.

#### *3.6. Determination of Cholesterol*

The technique chosen for cholesterol determination was as described by Alonso et al. [41] using direct injection of milk fat by capillary gas chromatography (GC). Approximately 30 mg anhydrous milk fat and 0.1 mL 5-α-cholestane as internal standard (3.5 mg/mL in hexane) was dissolved in 1 mL of hexane; 0.5 µL of the resulting solution was injected for GC analysis. For GC analysis for free cholesterol by this direct method we used an Agilent Technology 6890 chromatograph (Palo Alto, CA, USA) equipped with flame ionization detector. Analyses were performed using a HP-5 fused silica capillary column (30 m × 0.32 mm i.d. 0.25 µm thickness). Experimental chromatographic conditions were: He carrier gas at 17 psi head pressure; initial column temperature 280 ◦C, held for 1 min, increased to 355 ◦C at 3 ◦C/min. Injector temperature 350 ◦C and detector temperature was 360 ◦C. Peak identification was done by comparison of relative retention times with retention times of standards. Quantification of cholesterol was conducted by comparing sample peak area with of the 5 α-cholestane internal standard. The percentage of cholesterol reduction in milk fat was calculated by the formula [(100 − amount of cholesterol in milk fat) × 100]/amount of cholesterol in untreated milk).

#### *3.7. Fatty Acids and Triglycerides Analysis*

Fatty acids methyl esters (FAMES) were prepared by alkaline catalyzed methanolysis of the extracted lipids using 2 N KOH in methanol. The FAMES were analyzed on an Agilent Technology 6890 chromatograph (Palo Alto, CA, USA) with FID detector. Fatty acids were separated using CP-Sil 88 fused-silica capillary column (50 m × 0.25 mm i.d. × 0.2 µm film thickness, Chrompack, CA, USA) using the method described by Alonso et al. [42]. GC analysis of triglycerides by direct injection was performed on an Agilent gas chromatograph 6890 (Palo Alto, CA, USA) equipped with flame ionization detector. Analyses were performed using a WCOT fused silica capillary column (25 m × 0.25 mm × 0.1 µm film thickness) coated with OV 17 TRI (J.W. Scientific, Polson, CA, USA) using the method described by Alonso [43].

#### *3.8. Phospholipids Analysis*

Extractions of cheese fat were carried out with an Accelerated Solid Extraction ASE-200 extractor (Dionex Corp., Sunnyvale, CA, USA) using 2 g of freeze-dried cheese sample that was well mixed with 2 g of sea sand and loaded into a stainless steel extraction cell covered with filters on both sides. For the maximum cheese fat yield, the extraction included the use of dichloromethane-methanol solution (2:1, *vol*/*vol*) as solvent mixture and 10.3 MPa of pressure as fixed conditions described by Castro-Gómez et al. [44].

Separation of lipid classes was accomplished in an HPLC system (model 1260; Agilent Technologies Inc., Santa Clara, CA, USA) coupled with an evaporative light scattering detector (SEDEX 85 model; Sedere SAS, Alfortville CEDEX, France) using prefiltered compressed air as the nebulizing gas at a pressure of 350 kPa at 60 ◦C; the gain was set at 3. Two columns in series (250 × 4.5 mm Zorbax Rx-SIL column with 5-µm particle diameter; Agilent Technologies Inc.) and a precolumn with the same packing were used [44].

#### *3.9. Analysis of Volatile Compounds*

Analysis of volatile fraction was performed by headspace gas chromatographic mass spectrometric (GC-MS) method described by Alonso et al. [45]. To 10 g of previously homogenized cheese, 80 µL of aqueous solution of propionic acid ethyl methyl ester (1.14 mg/mL) as internal standard and anhydrous sodium sulphate (10 g) to retain water 176 were added. Individual standard dilutions in aqueous solution were prepared and were stored hermetically in sealed vials at 20 ◦C until their use. Prior to be analyzed in a static headspace apparatus (Model HSS 19395; Hewlett Packard), the samples were maintained at 80 ◦C for 60 min until the sample and gaseous phase reached the thermodynamic equilibrium. Apparatus was programmed as follows: 5 s pressurization, 18 equilibrium and filling and 2 min for injection. Helium was employed as carrier gas at a flow 18 rate of 17.5 mL/min. A Hewlett Packard GC Model 5890 coupled to selective MS Model 5972 was employed for volatile compounds analysis. Samples were injected in the split mode (split 18 rate of 7:1) on a capillary silica column with polyethylene glycol (HP Innovas, 60 m, 0.25 mm 18 ID, 0.25 µm film thickness, Hewlett Packard). Helium was used as carrier gas, at a flow rate of 18 36.5 cm/s. The column temperature program was: 33 ◦C for 5 min, increase at 1 ◦C/min up to 38 ◦C and then at 7 ◦C/min up to 210 ◦C, and held for

10 min. Injection was carried out at 200 ◦C and the interface line of MS at 280 ◦C. Electronic ionisation energy and photomultiplier voltage 18 were 70 eV and 1647 V, respectively.

### *3.10. Short Chain Free Fatty Acids*

For the analysis of SCFFAs, cheese sample (1 g) was homogenized in 20 mL of distilled water, centrifuge at 10,000 rpm for 10 min and filtered by 0.40 µm filter. A Hewlett-Packard model 5890 A equipped with a flame ionization detector on a capillary silica column (HP FFAP, 30 m, 0.25 mm ID, 0.25 µm film thickness, Agilent J & W) was used for analysis. Quantitative analysis were done using 2-ethylbutanoic acid as internal standard.

#### *3.11. Sensory Analysis*

Samples of Manchego cheese were cut in slices of approximately 8 × 8 cm of a thickness of approx. 1 cm and placed on white plates. Samples were tempered at ambient temperature (20 ± 2 ◦C) and then presented to the panelists. Twenty two trained sensory panelists from the members of the research Institute which, trained in sensory analysis of cheese, evaluated randomly coded cheeses. The testing conditions of the room for the sensory analysis were in conformity with the ISO requirements [46]. Flavor, aroma, color, texture and acceptability were evaluated on a five point scale (1 = poor to 5 = excellent).

#### *3.12. Statistical Analysis*

Experimental data were treated by analysis of variance (ANOVA) using the statistical software SAS (version 8.02, SAS Institute Inc., Cary, NC, USA). Differences among treatments were determined by statistical analysis using a Student t-test where *p* ≤ 0.05 was considered statiscally significant.

#### **4. Conclusions**

Approximately 97.6% cholesterol reduction was observed in the cheese that was treated using β-CD. Physicochemical properties (fat, moisture and protein) were not changed by the β-CD treatment, except the NS and NNP that showed slight differences attributed to the treatment. The amount of the different components of the lipid fraction (fatty acids, triglycerides and phospholipids) were similar in both, treated and untreated cheese with β-CD. Flavor compounds and short chain free fatty acids were mostly not significantly influenced by the β-CD. Although, the β-CD molecules are edible and nontoxic and a results they can be used safely for cholesterol removal processing. Therefore, the present study suggested that the treatment with the β-CD was an effective process for cholesterol removal from Manchego cheese, while preserving its nutritional properties. Further studies to evaluate the effect of the intake of the control and low cholesterol Manchego cheeses on the concentration of serum cholesterol would be of interest.

**Author Contributions:** L.A. conceived, designed the experimental and performed the experiments. P.F.F. revised the manuscript. M.V.C. designed the experimental and performed the experiments. J.F. designed the experimental and performed the experiments.

**Funding:** Authors thank to the Ministry of Economy and Competitiveness from Spain (grant number AGL-2014-56464; AGL-2017-84878).

**Acknowledgments:** The authors thanks to the Monte Toledo dairy cheese factory (Toledo, Castilla-La Mancha. Spain).

**Conflicts of Interest:** The authors declare that there are no conflicts of interest.

#### **References**

1. Hansel, B.; Nicolle, C.; Lalanne, F.F.; Brucket, E. Effect of low-fat, fermented milk enriched with plant sterols on serum lipid profile and oxidative stress in moderate hypercholesterolemia. *Am. J. Clin. Nutr.* **2007**, *86*, 790–796. [CrossRef] [PubMed]


**Sample Availability:** Samples of the compounds are not available from the authors.

© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Effect of Enzyme Modified Soymilk on Rennet Induced Gelation of Skim Milk**

#### **Kaixin Li 1,2 , Jianjun Yang 1,2 , Qigen Tong 1,2 , Wei Zhang 1,2 and Fang Wang 1,2, \***


Academic Editors: Lillian Barros and Isabel C.F.R. Ferreira Received: 29 October 2018; Accepted: 23 November 2018; Published: 26 November 2018

**Abstract:** In this study, soymilk was hydrolyzed to different degrees with flavourzyme, and then soymilk and enzyme modified soymilk at various levels were added to skim milk respectively, to generate a mixed gel using rennet. Rheological properties, scanning electron microscopy imaging, and physical and chemical indexes were examined to reveal the effect of enzyme modified soymilk on rennet induced gelation of skim milk. Results showed that soymilk inhibited the aggregation of skim milk, led to a decrease in storage modulus (G'), significantly increased moisture content and curd yield, and the resulting network was coarse. Enzyme modified soymilk with a molecular weight below 20 kDa led to a more uniform curd distribution, which counteracted the reduction of G' and allowed for the formation of a stronger gel. Both the moisture content and the curd yield increased with the addition of soymilk and enzyme modified soymilk, and overall the effect of adding a high degree of hydrolysis of enzyme modified soymilk was superior. Compared to untreated soymilk, the addition of a certain amount of enzyme modified soymilk resulted in a new protein structure, which would improve the texture of blend cheese.

**Keywords:** soymilk; enzyme modified soymilk; skim milk; rennet induced gelation; cheese; rheological properties

### **1. Introduction**

In recent years, with rising food price levels, increased dietary awareness and health concerns, people are becoming more interested in soy products and novel value-added food products. Consuming foods that contain casein and soy protein has many beneficial health effects, and gels made from these two proteins can achieve dual health effects of two products [1].

The gelation of milk often occurs in two ways. In the first mechanism, during acidification, colloidal calcium phosphate is dissolved, and caseins formed a self-supporting network near their isoelectric point [2]. In the other mechanism, rennet specifically hydrolyzes k-casein at the Phe105–Met106 bond, decreasing steric and electrostatic stabilization and resulting in casein aggregation [3,4]. Soymilk gels are typically prepared by heating soymilk, and gelation can also be induced by adding magnesium chloride or glucono-δ-lactone [5].

Many studies have focused on mixed skim milk and soymilk gels. Previous studies have shown that under acidification conditions, the viscoelastic properties and microstructure of generated protein gels of skim milk and soymilk mixtures are dependent on the concentration of skim milk powder and soy protein concentrate [6]. In addition, adding rennet to skim milk and soymilk systems, rennet will only hydrolyze the cow milk components [7]. It has been suggested that combined acid and rennet induced gelation of cow milk and soymilk mixtures would lead to the simultaneous aggregation of proteins [8].

Soy protein hydrolysate has been used in specialized adult nutrition formulations [9]. Many studies have shown when soy protein isolate was enzymed, its functional properties can be improved [10–14]. Rinaldoni et al. [15] added soy protein to skimmed cow milk to develop a spreadable cheese-like product. Their results showed that soy protein concentrate improved cheese yield, but the elastic properties were significantly reduced. Gao et al. [16] studied the enzymatic hydrolysis process of soymilk, and mixed enzyme modified soymilk with cow milk to make mozzarella cheese. It was found that the stretchability, elasticity and hardness of mozzarella cheese had been greatly improved. For most cheeses, rennet induced gelation is the key process affecting cheese yield and quality [17–19]. Therefore, the objective of this study was to explore the effect of different amounts and hydrolysis degrees of enzyme modified soymilk on renneting of skim milk, to provide a theoretical basis for the production of a new blend cheese. –

#### **2. Results**

#### *2.1. Protein Structural Characteristics*

The effect of enzymatic hydrolysis on protein profiles of skim milk, soymilk and enzyme modified soymilk is shown in Figure 1, and lanes 1–5 are the result of the marker, skim milk, soymilk, and high and low degree of hydrolysis of enzyme modified soymilk, respectively. Compared to soymilk, the distribution of bands of enzyme modified soymilk changed dramatically, indicating the decomposition of soy protein. After enzymatic hydrolysis of soy protein, the bands were mostly concentrated below 20 kDa, and its β-conglycinin, and acidic and basic subunits of glycinin almost disappeared entirely. Compared to the low degree of hydrolysis of enzyme modified soymilk, the molecular weight of soy protein further reduced after the high degree of hydrolysis. –

**Figure 1.** Sodium dodecyl sulfate polyacrylamide gel electropheresis for the control, soymilk and enzyme modified soymilk. M, marker; C, skim milk; S, soymilk; H, high degree of hydrolysis of soymilk; L, low degree of hydrolysis of soymilk. 7S and 11S indicate the relative proteins in soymilk.

#### *2.2. Rheological Properties*

– Rheological properties were used to monitor the effect of soymilk and enzyme modified soymilk on the coagulation process of skim milk. After adding rennet, the storage modulus of the control increased rapidly, and then changed slowly with gelation time (Figure 2a–c). The storage modulus of

β

experimental samples (except S2 and S3) showed the same trend. However, the storage modulus of experimental samples was lower than that of the control sample at the same renneting time (*p* < 0.05). The trend of the changes of loss modulus was consistent with that of storage modulus (data not shown). The loss tangent (the ration of loss modulus to storage modulus) decreased with renneting (Figure 2d–f). Compared to the control, gelation time of experimental samples lagged significantly (*p* < 0.05) (Table 1), and the delay effect in gelation time was more significant (*p* < 0.05) with the increase in soymilk and enzyme modified soymilk. The order of gelation time of experimental samples at the same concentration was L group ≥ H group > S group. At the end of gel formation, compared with control sample, the final strength of the curd determined by the storage modulus value at 60 min (G'60min) in experimental samples decreased significantly (*p* < 0.05) (Table 1), and the decrease was more significant with the increase in soymilk and enzyme modified soymilk (*p* < 0.05). The order of G'60min of the experimental samples at the same concentration was L group > H group > S group. – was L group ≥ H group > S group. At the end of G' G'

–

G') – – **Figure 2.** The changes in storage modulus (G') (**a**–**c**) and loss tangent (**d**–**f**) during rennet induced gelation. Curves are representative runs.

**G'**


**Table 1.** The rheological parameters of mixtures during renneting.

Means with different letters within the same column are significantly different (*p* < 0.05). <sup>1</sup> G'60min means storage modulus value of the gel at the end of renneting at 60 min. <sup>2</sup> Gelation time means the time point when storage modulus value of the gel was ≥1 Pa.

#### *2.3. Physical and Chemical Indicators*

The addition of soymilk and enzyme modified soymilk significantly increased the moisture content and curd yield (Table 2). With the increase in soymilk, curd yield showed a significant increase (*p* < 0.05), but the moisture content did not change significantly (*p* > 0.05). With the increase in enzyme modified soymilk, the moisture content and curd yield increased significantly (*p* < 0.05). At 5% and 10% addition, the moisture content and curd yield of the S group were significantly higher than those of L group and H group (*p* < 0.05). However, no significant differences were found in moisture content and curd yield among the three groups at 15% addition (*p* > 0.05), and moisture content and curd yield showed no significant differences between L group and H group at 20% addition (*p* > 0.05). Under 25% addition, the curd yield was higher in L group than in H group, and moisture content between the two samples showed no significant difference (*p* > 0.05).



Means with different letters within the same column are significantly different (*p* < 0.05).

### *2.4. Microstructure*

The above results showed that rennet induced gelation was affected by soymilk and enzyme modified soymilk. Therefore, the ratio of skim milk and soymilk/enzyme modified soymilk selected here was 85:15, to ensure their comparability. The caseins in the control were relatively smooth (Figure 3a). The soy protein in the sample with soymilk attached to the surface of caseins, formed rough and small clusters, and therefore resulted in a coarse structure (Figure 3b). Compared with the control and S3, the addition of enzyme modified soymilk caused a more uniform protein network (Figure 3c,d). Compared to the control, S3 and L3, the casein aggregates in H3 were smaller, and the small voids became denser. The soy protein in S3 and enzyme modified soy protein in L3 mainly existed in large caseins, but the distribution of enzyme modified soy protein in H3 was more uniform.

**Figure 3.** Scanning electron microscopy micrographs of rennet induced curds. **a**, control; **b**, S3; **c**, L3; **d**, H3 (magnification: ×15,000) (The arrow denotes soy protein and its hydrolysate).

#### **3. Discussion**

ĸ In our study, the pH values during renneting among mixtures showed no significant differences (data no shown), and the gel formation was the action of rennet. Under the conditions of our study, the addition of rennet to soymilk did not result in gelation, as monitored using a rheometer with no change in storage modulus (data not shown), which was agreed with previous studies [9,20]. The increase in the storage modulus and the decrease in loss tangent indicated that there was an interaction among caseins, and the increased contact area of caseins therefrom led to the formation of a gel network [21]. Previous studies have shown that rheological properties were mainly affected by casein concentration [22]. In experimental groups, the decrease in skim milk, which was partly replaced with soymilk or enzyme modified soymilk, resulted in the decrease in k-casein and therefore the number of action sites for rennet. Moreover, soymilk and enzyme modified soymilk set a barrier to the accumulation of caseins and further affected their aggregation (Figure 4). Thus, we speculated that these two effects would lead to a lag in gelation time and a decrease in final strength, which was also reported in previous literature [23,24]. Meanwhile, the delay in gelation time was more significant and final strength further decreased with the increase of soymilk and enzyme modified soymilk. Compared to the control with the final strength of 96.26 Pa, when soymilk content increased to 15%, the final strength was reduced to 3.55 Pa (Table 2), indicating that aggregation of caseins was

strongly hindered [8]. However, this effect can be overcome to a certain extent by adding enzyme modified soymilk, showing that the final strength of L3 and H3 was 75.32 Pa and 65.84 Pa, respectively. Compared with the control and S1, the decrease of k-casein in L1 and H1 had little effect on casein aggregation. The main contributor may be the reduction of protein size in enzyme modified soymilk as monitored by sodium dodecyl sulfate polyacrylamide gel electropheresis (SDS-PAGE), which was similar to the report of Luo et al. [25]. They studied the effect of native fat globule size on gel formation, and found storage modulus of the curd with small fat globules was higher than that of the curd with large fat globules. The high level of enzymatic hydrolysis increased the number of peptides, which may shield enzymatic sites and inhibit the aggregation of caseins (Figure 4), and therefore final strength in H group was significantly lower than that of L group. ĸ

**Figure 4.** Schematic diagram of rennet induced gelation of different mixtures. **a**, skim milk; **b**, skim milk and soymilk; **c**, skim milk and enzyme modified soymilk.

– Previous studies have shown that the addition of soy protein to milk can increase the moisture content of the resulting cheese [26], which was consistent with our result. Soy protein had water-holding properties [9], and therefore we speculated that soy protein trapped in the gel structure can retain more moisture through hydrogen bonds and other forces. Furthermore, our rheological results have shown that soymilk significantly inhibited rennet induced aggregation of caseins, affected dehydration of curds and increased moisture content and curd yield, which was in agreement with other studies [27–29]. Utsumi [30] indicated that the water-holding property of soy protein after enzymatic hydrolysis can be improved. However, the moisture content in L group and H group was lower than that in S group in our study, which may be ascribed to the decreased inhibition of gelation and the loss of small substances after enzymatic hydrolysis (Figure 4). As the amount of soymilk increased, the hydrophilicity of soy protein and the inhibition effect on casein aggregation were more significant, and therefore moisture content and curd yield further increased. There was no significant difference in moisture content between L group and H group under the same addition, but the curd yield of H group was lower than that of L group, and the losing of smaller peptide segments or amino acids with whey after a high degree of hydrolysis may be the major contributor.

The control curd showed a compact protein matrix, which was composed of thick chains and large clusters of caseins (Figure 3a), and it was consistent with previous report [31]. The soy protein adhered or was bound to the surface of caseins, formed a clustered structure and therefore inhibited the aggregation of caseins, causing the increased moisture content and curd yield. The research of Ingrassia et al. showed a similar phenomenon [32]. Both glycinin and β-conglycinin in soymilk can be effectively degraded by flavourzyme [33], which may decrease the inhibition of gelation and therefore contribute to a more compact and orderly gel structure. The increase in the number of small substances after a high degree of hydrolysis further inhibited the aggregation of caseins, and thus decreased the size of casein aggregates.

#### **4. Materials and Methods**

#### *4.1. Skim Milk Preparation*

Fresh milk was supplied by a local farm (Fuchun Farm, Beijing Sanyuan Food Co., Ltd., Beijing, China) and sodium azide (0.02%) was added immediately. The milk was centrifuged at 4000× *g* for 20 min at 4 ◦C (LYNX 2000, Thermo Scientific, Waltham, MA, USA) and then filtered three times through filters (Fisher Scientific, Whitby, ON, Canada), as far as possible to remove fat. The resulting skim milk was pasteurized at 63 ◦C for 30 min, cooled to room temperature, and stored in a refrigerator until use. The fat content of skim milk was 0.04 ± 0.01% by the AOAC method [34], and protein content was 3.21 ± 0.05% measured using the Kjeldahl method [19].

#### *4.2. Enzyme Modified Soymilk Preparation*

Soybeans (Helen soybean, Heilongjiang Black Soil Town Modern Agriculture Development Co., Ltd., Haerbin, Heilongjiang, China) containing 36% protein and 16% fat were obtained from a local market. Soymilk was prepared as previously reported [35], with slight modifications. Soybeans were soaked overnight in deionized water for hydration and mixed with a certain amount of deionized water to obtain the desired protein content. Subsequently, samples were passed through a household soymilk maker (RM-125, Ruimei, Wuxi, Jiangsu, China). After soymilk maker cycle was completed, soymilk was passed through a filter (Fisher Scientific, Whitby, ON, Canada) and passed through cheesecloth to remove okara. Then soymilk was boiled at 100 ◦C for 10 min and rapidly cooled to room temperature. The protein content of the obtained soymilk was 3.32 ± 0.07% measured using the Kjeldahl method [19].

Flavourzyme (0.28%; HP202474, Novozymes, Tianjin, China) was added to soymilk for 2 h and 4 h at 45 ◦C, respectively. The subsequent degree of hydrolysis of soymilk was 5.92% and 9.88% according to the Ninhydrin method [36]. L and H were used to represent the low (L) and high (H) degree of hydrolysis. The enzyme modified soymilk was then incubated at 85 ◦C for 15 min to inactivate the enzyme. When samples were cooled to room temperature, sodium azide (0.02%) was added. The obtained samples were stored at 4 ◦C until use.

#### *4.3. Electrophoresis*

The effect of enzyme treatments on protein profiles of soymilk and enzyme modified soymilk was determined using SDS-PAGE with separating and stacking gels containing 15 and 4% acrylamide, respectively. A molecular weight marker ranging from 11 to 180 kDa (PR1910, Solarbio, Beijing, China) was used as a standard. Skim milk, soymilk and enzyme modified soymilk were dissolved with sample buffer (10 mM DTT, pH 6.8, 1 mM EDTA, 1% SDS, 10% glycerol, and 0.01% bromphenol blue) and then boiled for 5 min. After centrifugation, the electrophoresis was carried out following the method of Lamsal et al. [37].

#### *4.4. Mixture Preparation*

The mixtures were prepared according to Table 3, and skim milk was selected as the control. For convenient reading, samples with different ratios of skim milk and soymilk/enzyme modified soymilk are indicated by **S1**–**S3**, **L1**–**L5** and **H1**–**H5**. The mass ratio selected were to ensure that rennet induced gelation would occur. After the mixtures of skim milk and soymilk/enzyme modified soymilk had been prepared, the samples were stirred at room temperature for 30 min, and the pH was adjusted to 6.7 with 0.1 M NaOH. All the samples were stored in a refrigerator until use.


**Table 3.** The mixtures with different mass ratios of skim milk and soymilk/enzyme modified soymilk.

#### *4.5. Curd Making*

All samples (weighed and recorded as W1) were first warmed at 32 ◦C for 30 min and then incubated with rennet (0.01%). Rennet (CHY-MAX powder Extra) was got from Chr. Hansen (Beijing, China) and the coagulation strength was ~2235 IMCU/g. After 1 h of incubation, the resulting gels were cut manually into small pieces (1 × 1 × 1 cm), and then set for 5 min to promote syneresis. Curds were subsequently centrifuged at 4000× *g* for 15 min at room temperature. The upper whey was removed and the curd was collected carefully and weighed (recorded as W2). The curd yield was calculated using following Equation:

$$\text{Curd yield} = \text{W}\_2/\text{W}\_1 \times 100\%$$

The moisture content of the curds was analyzed using an oven method [34]. All measurements were obtained in triplicate.

#### *4.6. Rheological Properties*

The coagulation process of the mixtures was monitored as described previously [38]. All samples were first maintained at 32 ◦C for 30 min. The gelation time was defined as the time point when storage modulus value of the gel was ≥1 Pa [6]. All measurements were determined in triplicate.

#### *4.7. Microstructure Determination*

The samples for microstructure determination were prepared as described earlier [39]. Then samples were critical point dried in a Technics Critical Point Dryer (CPD030, Leica Mikrosysteme GmbH, Wien, Austria), fixed on the sample table (MC1000, Hitachi, Ibarakiken, Japan) for spraying operation and then examined using a scanning electron microscope (SU8010, Hitachi, Ibarakiken, Japan) operated at 25 kV. Representative micrographs were selected for presentation.

#### *4.8. Statistical Analysis*

Statistical analysis of the data was performed using SPSS 18.0 (SPSS Inc., Armonk, NY, USA). The differences were compared at a significance level of *p* < 0.05.

#### **5. Conclusions**

Soymilk and enzyme modified soymilk have a significant effect on renneting of skim milk. Compared to skim milk, a certain proportion of soymilk and enzyme modified soymilk would

decrease storage modulus of the mixed gel, and significantly increase moisture content and curd yield. In addition, with the increase in soymilk and enzyme modified soymilk, the inhibition of gel formation was stronger, and moisture content and curd yield further increased. Compared to soymilk, enzyme modified soymilk with more small molecular substances decreased the impediment of rennet induced gelation, improved storage modulus of the gel, and the resulting curd had a more uniform structure.

**Author Contributions:** Data curation, K.L.; Funding acquisition, F.W.; Investigation, K.L., J.Y. and W.Z.; Project administration, Q.T. and F.W.; Writing—original draft, K.L.; Writing—review and editing, F.W.

**Funding:** This project was financially supported by the Beijing Municipal Commission of Education Research Program (KM201710020013).

**Acknowledgments:** We would like to thank Donghua Yu for assisting with the scanning electron micrographs at the Institute of Atomic Energy Utilization, Chinese Academy of Agricultural Sciences.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


**Sample Availability:** Samples of skim milk and soymilk/enzyme modified soymilk are available from the authors.

© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Article*
