Resistant starch and soy protein isolate in instant noodles obtained by conventional and vacuum frying

https://doi.org/10.1590/1981-6723.23918 Abstract Health concerns with the consumption of high-fat products and with respect to the intake of fibre are important for consumers. Vacuum frying process is an alternative frying process that increases the quality of foods. The objective was to develop instant noodles aiming to reduce carbohydrate and fat content by adding soy protein isolate (SPI) and resistant starch (RS3), using conventional and vacuum frying processes. A 2 2 central composite rotational design was used for each type of process, and the formulations were optimized using the Response Surface Methodology. The noodles were characterized with respect to fat absorption, cooking time, final resistant starch content and firmness. The noodles produced at the optimized point were also characterized for their isoflavone content, amino acid profile and by scanning electronic microscopy (SEM). The response surfaces showed that the addition of SPI and RS3 to the formulation resulted in reduced fat absorption and texture. Comparing both frying processes it was found that noodles obtained by vacuum frying absorbed 3% less fat and were less firm than those produced by conventional frying process. The SEM analysis showed that the noodles obtained by conventional frying had a more porous structure, which allowed an easier fat penetration, whereas those produced by vacuum frying showed a more closed structure with fat impregnated on the surface. Regardless of the process used, it was possible to obtain noodles with improved nutritional value, since they showed high fibre (8%) and protein (approximately 11%) contents, considerable isoflavone content (8.20 mg/100 g) and a complete amino acid profile, due to a significant increase in nearly all the amino acids, especially


Introduction
Health concerns with the consumption of high-fat products and with respect to the intake of fibre are important for consumers. In an attempt to increase the consumption of fibre, several alternatives have been proposed, including the development of new products that employ new ingredients that can increase the nutritional and functional value of foods. Resistant starch (RS) is a starch fraction that does not produce glucose in the body since it resists enzymatic digestion in the small intestine, but can be fermented in the colon by bacterial microflora producing mainly gas and short chain fatty acids. Due to this characteristic, resistant starch is usually considered a dietary fibre (Fuentes-Zaragoza et al., 2010). Nowadays, five types of resistant starch (RS1-RS5) have been identified and can be classified according to the nature of the starch, such as the structure of starch granules, cooking and cooling operations and complexation with other components (Remya et al., 2018). RS1 are physically non-accessible starch grains encapsulated in the food matrix, basically because of the cell walls and proteins. Whole or partially ground cereal grains, legumes and other materials containing starch in which the size or composition impede or retard the action of the digestive enzymes, belong to this group. RS2 are native grains, which are highly resistant to digestion by the enzyme α-amylase, present especially in green banana, raw potato and high-amylose content starch. RS3is formed after retrograding, due to cooling and ageing of the gel, forming a partially crystalline, insoluble structure that is resistant to enzymatic digestion and different from the initial conformation. It is the most common form and the most important from a technological point of view. RS4is a chemically modified starch (Sajilata et al., 2006). RS5 is formed by complexation of the amylose fraction of starch with lipids, fats and surfactants (Raigond et al., 2015).
The soybean is known as an excellent protein source for human and animal consumption, richer in protein than any other cereal or legume. Soy proteins have a good amino acid balance when compared to other vegetable proteins. However, legumes, such as soy, show a reduced amount of the sulfur amino acids and an elevated lysine content. Cereal proteins show the inverse situation with respect to their amino acid composition, and hence the combination of legumes, such as soy, with cereal products made from corn, rice, wheat etc. is advisable, to allow for complementation of the essential amino acids (Delcour & Hoseney, 2010).
Deep-frying is one of the most common and popular processes used in the food industry and consists of immersing the food in hot oil to remove water. The frying process includes heat and mass transfer between the food and the frying medium causing some changes in the sensory characteristic of food (Saguy & Dana, 2003). In vacuum frying the food is heated under pressure, reducing the boiling point of the frying oil and the water present in the product, and lowering the presence of oxygen during the frying process which inhibits some reactions such as lipid oxidation and enzymatic browning. Moreover, vacuum frying may help to preserve color and nutrients present in the fried foods (Fan et al., 2005).
The objective of the present work was to develop instant noodles with functional and nutritional properties, aiming to reduce the calorie content by reducing both oil absorption and the digestible carbohydrate content by the addition of soy protein isolate (SPI) and resistant starch type 3 (RS3), using conventional and vacuum frying processes.

Characterization of raw material
The quantitative analysis of the isoflavones was carried out following the procedure used by Berhow (2002) using HPLC. The Amino acid composition was determined after 24h hydrolysis in 6M HCl at 100 °C, the amino acids were reacted with phenylisothiocyanate and the derivatives separated using a Luna C-18, 100Å, 5 µ, 250 m × 4.6 mm (00G-4252-E0, Torrance, CA, USA) column, at 50 °C. Quantification was carried out by comparison with a standard mixture (Thermo Scientific, Rockford IL, USA) and DL-2aminobutyric acid from Sigma-Aldrich Corp., St. Louis, MO was used as the internal standard. The running time was 24 min. The free amino acids were determined by extracting 1.25 g flour samples in an 80% ethanol solution containing 0.1 M HCl in a 5 mL volumetric flask and adding 500 µL of aminobutyric acid as the internal standard. The mixture was sonicated for 10 min and homogenized for a further 1 h, followed by centrifugation at 8500 g for 15 min. The supernatant was filtered through a 0.22 mm membrane and a 40 µL aliquot derivatized as described above, injecting 20 µL of the derivatized extract into the HPLC.

Experimental design
Two 2 2 central composite rotational designs (CCRD) were used to study the effects of adding RS3 (0-16%) and SPI (0 to 16%) in the formulations of instant noodles obtained by conventional and vacuum frying. A total of 12 trials were conducted in each design: 4 for the factorial design, 4 for the axial points and 4 replicates at the central point to estimate error. Finally, an α = ± 1.4142 value was selected to ensure the rotatability of the model (Ribotta et al., 2008). The maximum levels were chosen in preliminary tests, adding as much amounts of resistant starch and soy protein isolate as possible to obtain an adequate dough (sufficiently stretchable and elastic) to produce instant noodles. For noodle production, the methodology described by Vernaza & Chang (2017) was used.
The instant noodles were characterized for oil absorption (method nº 30-25.01, American Association of Cereal Chemists, 2010), moisture content (method nº 44-15.01, American Association of Cereal Chemists, 2010) and resistant starch content using the methodology proposed by Goñi et al. (1996). These analyses were determined in four replicates. Firmness was determined in seven replicates using a TA-XT2i texturometer (Stable Micro Systems, England) with the A/LKB probe (maximum load of 50 kg) and with the following conditions: measured in compression force, pre-test and test speed: 0.17 mm/s; post-test speed: 10 mm/s; cut distance: 4.7 mm; trigger type: button and trigger force 10 g. For data analysis, the Texture Expert software was used.

Determination and characterization of the optimal point
The optimal point chosen using the experimental designs was processed for each type of frying (conventional and vacuum). The noodles obtained were characterized for their proximate composition, optimal cooking time, resistant starch content in the fried product and after the optimal cooking time (rehydration), isoflavones content, total amino acid content and firmness, according to the methodologies described above. Scanning electronic microscopy (SEM) analyses of the surface and cross-section of the noodles were also carried out with x 500 magnification and x 200 magnification, respectively, in order to compare the internal structure of the noodles produced by the two frying processes. A Scanning Electron Microscope (LEO 440i -Cambridge, UK) was used, where the energy and beam current were 10 kV and 50 mA, respectively. The samples were placed on aluminum supports and covered with a 92 Å gold layer prior the analysis.

Data analysis
When necessary, the results obtained in the characterization of the raw materials and optimal point were statistically analyzed by the Analysis of Variance (ANOVA) and Tukey's test, using the Statistica version 5.0 program (Stat Soft, Inc., Tulsa, USA), with the objective of identifying significant differences (p ≤ 0.05) between the samples.
Response Surface Methodology was used to analyze the results obtained from the CCRDs. The statistical calculations were carried out by the standard error using the Statistica version 5.0 software (Stat Soft, Inc., Tulsa, USA).

Isoflavones content and total amino acid profile of raw material
Research on soybean isoflavones have shown their protective effect against certain health problems related to menopause, osteoporosis, cancer and cardiovascular diseases (Schmiele et al., 2014). During processing to obtain soybean sub-products, there are changes and losses in the isoflavones originally present in the soybean, nevertheless the total isoflavones content of the soy protein isolate used in this study was 108.10 mg/100 g. The concentrations of glycosylated, malonylglycosylated, acetylglycosylated and aglycone forms were 27.8%, 33.5%, 6.3% and 32.4%, respectively. Table 1 shows that soy protein isolate had a high percentage (32.4%) of isoflavones in the aglycone form, due to the fact that during the production of soy protein isolate the glycosylated isomers can be hydrolyzed (Wang & Murphy, 1996). The physiological properties are predominant when the isoflavones are present in the aglycone form (without glucose) instead of glycosylated (conjugated to glucose) (Ribeiro et al., 2006). Some researchers reported isoflavone contents around 690 μg/g in dried basis in soy protein isolates produced and industrialized in Brazil (Lui, et al., 2003). The concentration of the isoflavone profile of soybean is affected by genetic and environmental factors, as well as by the processing conditions used to obtain products derived from or based on soybean. Table 2 shows the results obtained for the total amino acid analyses of the wheat flour and SPI. It can be observed that the values obtained for the amino acids of SPI were much higher than those of the wheat flour, since the SPI contains approximately 88% protein and thus the amino acids are much more concentrated, whereas in the wheat flour they are more diluted. Wheat flour protein presented higher glutamic acid, proline and sulfur amino acid (methionine and cysteine) contents whereas the SPI showed higher values for lysine, arginine and aspartic acid.
Soy protein is poor in sulfur amino acids, being methionine its limiting amino acid. However, it is rich in lysine, an amino acid deficient in most cereals. Thus, a combination of the proteins from legumes, such as soybean, and from cereals, such as wheat, is advantageous, since they complement each other with respect to methionine and lysine.   Table 3 shows the results obtained for water absorption, cooking, moisture content, fat absorption, resistant starch content and firmness for each type of frying.

Water absorption
The minimum and maximum values obtained for water absorption in the 12 trials were 61.4% and 74%, respectively. It can be seen that the equation that explain the water absorption was quadratic, the interaction factor was not significant and eliminated after fitting. The analysis of variance showed that the coefficient of determination was 0.99 and the ratio between Fcalculated and Ftabulated was 107 at 5% significance level, allowing to obtain the model (Equation 1) and generate the respective response surface (Figure 1). The response surface showed that the water absorption increased significantly when increasing the SPI content. According to the mathematical model (Equation 1), it can be observed that the addition of SPI was the variable most affecting water absorption. The moisture level or added water depends on the protein content and quality, starch content, especially damage starch, flour granulometry and pentosans. If too little water is added, the pasta will be striped and sometimes the laminar surface of the pasta is modified, leaving the pasta fragmented (not smooth). These results obtained weak strips that easily broke during drying due to the presence of noncohesive zones. The addition of too much water causes problems in the lamination process due to excessive gluten development (Hou, 2010).

Cooking time
The same cooking time was used for the noodles obtained using the two frying processes, since there was no significant difference between the samples. The cooking time varied between 144 and 210 s as the minimum and maximum values, respectively for the 12 trials (Table 3). It can be seen that the Equation 2, which explains the cooking time, was quadratic, the linear factor of resistant starch and the interaction were not significant and hence eliminated after fitting. The analysis of variance showed that the coefficient of determination was 0.81 and the ratio between Fcalculated and Ftabulated was 3.78 at 5% significance level, which allow to obtain the mathematical model (Equation 2), and the generation of the respective response surface (Figure 1). The response surface showed that the shortest cooking times were obtained with intermediate levels of resistant starch and intermediate to low levels of soy protein isolate (below 8%). Also, cooking time increased significantly when increasing the RS3 and SPI contents. This increase can be explained by the fact that the particles of SPI and of RS3 took longer to hydrate, leaving the noodles with a raw appearance when not completely hydrated. Cooking time s =156+24.56×RS +11.41×SPI+8.06×SPI (2)

Moisture content and fat absorption
A response surface could not be generated for the response of moisture content, since the data obtained (Table 3) were extremely close to each other for both conventional and vacuum frying. It is important to maintain the moisture content below 10% in the samples to avoid microbial growth and maintain a longer product shelf life. Drying by frying is a very fast process. Water vaporizes quickly from the surface of the noodles upon dipping into the hot oil. Dehydration of the exterior surface drives water to migrate from the interior to the exterior of the noodle strands. Eventually, some of the water in the noodles is replaced by oil. Many tiny holes are created during the frying process due to the mass transfer, and they serve as channels for water to get in upon rehydration in hot water (Hou, 2010). Therefore, the shelf life of instant noodles is related also with the fat content and fat degradation. Because of the dehydration process (frying) the water activity of instant noodles could be in the range of 0.3, which is a stable range that avoid microbial growth.
Fat absorption during frying is important from both the economic and nutritional points of view. Fat absorption increases production costs and decreases the product shelf life due to the rancidity of the oil absorbed by noodles (Wu et al., 2006). The values for fat absorption obtained using conventional frying (FAC) varied from 18.08% to 25.09% and for vacuum frying (FAV) varied from 18.28% to 22.2% (minimum and maximum values, respectively) for the 12 trials. It can be seen that the equations that explain the fat absorption of noodles obtained by conventional (Equation 3) and vacuum frying (Equation 4) was quadratic, and the quadratic factor of SPI and the interaction factor were not significant and were eliminated after fitting. The analysis of variance for the conventional process showed that the coefficient of determination was 0.95 and the ratio between Fcalculated and Ftabulated was 12.4 at 5% significance level; for vacuum frying the analysis of variance showed that the coefficient of determination was 0.84 and the ratio between Fcalculated and Ftabulated was 4 at the 5% level of significance, which allow to obtain the mathematical models (Equation 3 and Equation 4) and allow the generation of the respective response surfaces (Figure 1).
The analysis of the two models obtained showed the same tendency. On analyzing the regression coefficients of the mathematical models, a negative influence of the variables on fat absorption can be seen, such that on adding resistant starch or soy protein isolate, the fat content decreased. Thus, the lowest fat absorption values could be obtained with the highest values of these two variables. Table 3 shows the results obtained by conventional and vacuum frying for fat absorption. The difference between both processes was not as high as expected. Other authors reported a reduction of up to 45% in fat absorption using vacuum frying process (Garayo & Moreira, 2002;Mariscal & Bouchon, 2008;Dueik, et al., 2010). Several authors explained the mechanism of fat absorption in vacuum-fried technology (Garayo & Moreira, 2002;Fan et al., 2005;Troncoso et al., 2006). Troncoso et al. (2006) explained that it is necessary to have a de-oiling mechanism to reduce fat absorption of vacuum-fried products and they explained that the increase in fat content after vacuum frying can be characterized mainly by two steps: i) the frying period; and ii) the pressurization period. In vacuum frying, the heat and mass transfer rates are high due to a decrease in the boiling point of the water. In the first stage, the high rate of mass and heat transfer produces partial evaporation of the water in the food, which moves from the inside of the product to the oil medium, and is partially replaced by oil. In the second stage (pressurization), which proceeds after the product has been removed from the oil but is still inside the vessel at the frying pressure and temperature, there is a rapid increase in pressure in the pores, causing the oil adherence to the surface of the product to penetrate into the food, generating a sponge effect, until the pressure at the pores matches the atmospheric pressure.

Final resistant starch content
After conventional frying, the values for final resistant starch (RSC) varied between 4.44% and 8.47% as the minimum and maximum values, for the 12 trials. The analysis of variance showed that the coefficient of determination was 0.85 and the ratio between Fcalculated and Ftabulated was 6.2 at 5% significance level, which allow to obtain the mathematical model (Equation 5) and allow the generation of the respective response surface (Figure 1).
After vacuum frying, the final values for resistant starch (RSV) varied between 4.61 and 10.12% as the minimum and maximum values for the 12 trials. It can be seen that the equation that explains the final resistant starch content was quadratic and was represented solely by the effects of the variable of resistant starch. The analysis of variance showed that the coefficient of determination was 0.89 and the ratio between Fcalculated and Ftabulated was 10 at 5% significance level, which allow to obtain the mathematical model (Equation 6) and allow the generation of the respective response surface (Figure 1). Soy protein isolate had no effect on the residual resistant starch content after the two frying processes used. Vernaza & Chang (2017) reported around 30% loss of resistant starch type II in noodles with addition of green banana flour while the loss was around 18% in noodles with addition of resistant starch type III.

Firmness
The values for firmness of the noodles obtained by conventional frying varied between 5.98N and 7.65N as the minimum and maximum values, respectively, for the 12 trials. It can be seen that the equation that explain the firmness was quadratic, but the quadratic factor for resistant starch and the interaction factor were not significant and were eliminated after fitting. The analysis of variance showed that the coefficient of determination was 0.83 and the ratio between Fcalculated and Ftabulated was 3 at 10% significance level, which allow to obtain the mathematical model (Equation 7) and allow the generation of the respective response surface (Figure 1).
The values for texture of the noodles obtained by vacuum frying varied between 5.3N and 8.4N as the minimum and maximum values, respectively, for the 12 trials. However, response surfaces and a mathematical model could not be generated, since an analysis of the effects showed that the only parameters with a significant effect (p < 0.10) were the linear and the quadratic effects for soy protein isolate. Thus, the mathematical model and response surface were not considered, since, on ignoring the non-significant effects, the coefficient of determination (R 2 ) showed a low, non-significant value (0.68).
From the results obtained, it can be concluded that the addition of other ingredients to the formulation of instant noodles can affect the texture, independently of the process used. Wheat flour contains unique proteins, gliadins and glutenins with hydration and mechanical energy form gluten. The use of wheat flour in the production of various products is mainly influenced by the quantity and quality of the gluten, more than by the other ingredients. Thus, the addition of any other ingredient that is not a constituent of wheat or gluten, will significantly alter the rheological characteristics of the mass. It can be seen that the addition of RS3 and SPI diluted the gluten network forming proteins (gliadin and glutenin), favoring the formation of a weaker dough structure, resulting in noodles that were less firm.

Characterization of the optimal points
After analyzing the results, mainly fat absorption and resistant starch content, the formulation with the addition of 12% resistant starch (RS) and 8% soy protein isolate (SPI) was chosen and tested with the two frying processes, since it was found to guarantee a medium fat absorption with high resistant starch content.
The only component showing significant differences in the proximate composition was the fat content (Table 4). With vacuum frying, the total fat content of the noodles was reduced by 3%. The consumption of instant noodles has been increasing considerably in Brazil, mostly consumed by adolescents and children. The addition of 8% SPI resulted in a product of important source of protein. In addition, there was an increase in the isoflavone content (Table 1), differentiating this product from the traditional ones found on the market.
The histology of the samples could be analyzed with the aid of the images obtained by scanning electronic microscopy ( Figure 2). The surfaces and cross-sections of the noodles obtained by the two frying processes were analyzed by electronic microscopic. In the microscopic images, the fat is the whitish layer covering the noodles. In the surface views (Figure 2a), it can be observed that the noodles obtained by vacuum frying had a greater fat content on the surface than those obtained by conventional frying. On the other hand, the images of the cross-sections (Figure 2b) showed that the vacuum fried noodles had a more closed structure, whereas those obtained by conventional frying had a more porous structure. The instant noodles on sale nowadays are produced from wheat flour, salt and in some cases very low proportions of guar gum, showing no health benefit. However, in this study, in addition to these raw materials, soy protein isolate and resistant starch were added. Table 1 shows that the final product obtained in this study contained approximately 8 mg isoflavones per 100 g product, since in the vacuum frying process, the isoflavones were not degraded. If the instant noodles had been produced solely with wheat flour, the total amino acid contents would have been smaller. Table 2 shows the comparison of the total amino acid contents found for the wheat flour with those obtained for the instant noodles. It can be seen that the majority of the amino acids increased significantly with the exception of the sulfur amino acids glutamic acid and proline. This can be explained by the fact that soy protein isolate contains low contents of these amino acids, diluting those found in the wheat flour. It is important to mention that the lysine content doubled in the final product, and thus, it can be concluded that the addition of soy protein isolate increased the nutritive value of the instant noodles, since the combination of the proteins from legumes and cereals are complementary in relation to lysine. Other amino acids that increased considerably were aspartic acid and arginine.
The resistant starch content was also analyzed in the cooked, ready-to-eat samples, that is, after cooking for 170s in boiling water. The values obtained were 6.11 ± 0.47% and 7.04 ± 0.81% for the noodles fried under atmospheric pressure and under vacuum, respectively. The percent loss of resistant starch, during cooking (rehydration), of the samples was approximately 12%. Despite the loss during rehydration, the instant noodles obtained by both processes presented resistant starch contents above 6%.
Moreover, the experimental value of each of the response was compared with those predicted by the equations of the model (Table 5). When comparing the experimental values with those predicted by the obtained equations, we observe in Table 5 that the deviations were relatively low (< 10%), validating the mathematical models obtained in this study.

Conclusions
Fat absorption was reduced because of the addition of soy protein isolate (SPI) and resistant starch (RS3) during the atmospheric and vacuum frying process. However, higher fat content and firmer noodles were obtained with atmospheric frying. On the other hand, noodles obtained by vacuum frying showed less fat content. Thus, with the help of the mathematical models and the response surfaces, it can be concluded that the addition of SPI and RS3 affected the texture (firmness) of the noodles due to dilution of the glutenforming proteins, resulting in less firm noodles. In the characterization of the noodles obtained at the optimal point (12% RS3 and 8% SPI), the noodles obtained by vacuum frying presented approximately 3% less fat in the total mass. The noodles obtained by both processes presented considerable protein contents (about 15% on a dw and defatted basis) and high fibre contents. The addition of SPI allowed to obtain instant noodles with considerable isoflavone contents and to increase the contents of all the amino acids except the sulfur ones. Finally, independently of the frying process used, it was possible to develop a new formulation of noodles, with better nutritional properties because of the high fibre and protein contents, the presence of isoflavones and an improved amino acid profile.