High protein and low-fat chips (snack) made out of a legume mixture

ABSTRACT Snacks tend to be unhealthy products, so it is important to develop more nutritious alternatives. For this reason, the aim of this research was to develop legume chips for use as a snack food and evaluate their nutritional properties and sensory characteristics. Common bean, broad bean, and textured soy bean flour mixtures at different ratios were used as ingredients in the chips. The formulation most accepted by the consumers was made of 65% of broad bean and 35% of soy bean. This formulation had the highest level of protein 34.17 g/100g, with 11.53 g/100 g of fiber, and 3.52 g/100 g of lipids. 65/35 BB/S (broad bean/soy bean) was the product with greater protein digestibility 72 g/100 g. The results show that it is feasible to develop a legume snack without the use of cereals that is pleasant to the senses, offering nutritional alternatives and natural potential functional food. Graphical Abstract


Introduction
A snack is anything consumed between main meals and tends to be energy-dense but nutrient-poor (Hartmann, Siegrist, & Van der Horst, 2013;Piernas & Popkin, 2010). The main reasons for unhealthy snacking include enjoying a special occasion, opportunity-induced eating and gaining energy (Aukje, Marieke, de Vet, Fennis, & de Ridder, 2015). The consumption of unhealthy snacks (chips, cakes, cookies, sweets, soft drinks, and other high-calorie products) is on the rise. These unhealthy products contain high concentrations of sugar, fats, and refined grains, and are directly related to overweight and obesity (Piernas & Popkin, 2010).
The World Health Organization (WHO) catalogued obesity as a global health epidemic in 2003, and in 2014, the statistics for this organization indicated that 13% of adults 18 years old or older worldwide are obese with 39% being overweight, also 41 million children under the age of five are obese or overweight. Therefore, an epidemic of these proportions is a large-scale health problem (Madhusoodanan, 2017). Obesity is associated with health problems, including diabetes, heart disease, disorders of the locomotor system, and some cancers. However, not all snacks are unhealthy; snacking on whole fruit, vegetables, and crackers has been found to contribute to better overall diet quality (Hartmann et al., 2013).
Nutritious snacks can increase access to healthy, lowenergy food choices. In order to avoid obesity and all diseases related to this epidemic, it can help to establish healthy snacking habits and assess the nutritional quality of products (De Vlieger, Collins, & Bucher, 2017). A healthy snack can be obtained by supplementing with legumes, which have a high protein content compared with cereals and provide soluble and insoluble dietary fibers, vitamins, and minerals (Qayyum, Butt, Anjum, & Nawaz, 2012).
For example, beans contain 21 to 24 g/100 g of protein and also contain significant quantities of minerals and multiple vitamins, while being low in fat (Mitchell, Lawrence, & Hartman, 2009). Studies report that pinto beans, in a reasonable quantity, provide health benefits that include lowering serum lipoproteins and reducing risks of coronary heart disease (Winham, Hutchins, & Johnston, 2007).
Broad beans are widely available and used as a source of protein and add complex carbohydrates, soluble fiber, and essential vitamins and minerals to the diet, yet are low in fat, sodium and are cholesterol free (Akubor, 2003). They are a source of 3,4-dihydroxyphenylalanine, a major ingredient in medicines used to treat Parkinson disease patients (Apaydin, Ertan, & Ozekmekçi, 2000).
Soybeans are an economical source of protein (40 g/ 100 g) and essential amino acids. They have high polyunsaturated fat content and absence of cholesterol and lactose. Soybeans are also a good source of minerals and isoflavones, phytosterols, protease inhibitors, inositol hexaphosphate, and saponins (Hertrampf & Piedad-Pascual, 2000). Soy is associated with appetite and obesity control and prevention of cancer (Hawrylewicz, Zapata, & Blair, 1995).
Moreover, legumes are consumed in lower quantities than is recommended by nutrition institutions. For this reason, some researchers have developed snacks or other products that are supplemented with legumes. Ochoa-Martínez, Castillo-Vázquez, Figueroa-Cárdenas, Morales-Castro, and Gallegos-Infante (2016) added bean flour to corn flour at different concentrations when manufacturing corn chips to produce superior nutritional properties and similar texture attributes to those of corn snacks that are currently found on the market. Man and Păucean (2013) reported that supplementation of wheat flours with legume flours that have a high protein content provides an improvement in the nutritional quality of baked goods. Legume consumption in various forms that is different from traditional consumption may be a way to contribute to a healthier diet; however, it is important to ensure the stability of the nutrients throughout the production process and during the shelf life of the product (Gulia, Dhaka, & Khatkar, 2014). In addition to the nutritive value, sensory characteristics should be taken into account to develop products that are attractive to people. In this context, the aim of this investigation was to develop a chips snack based on legumes and evaluate its nutritional properties and sensory characteristics.

Preparation of legume chips
A common bean (Phaseolus vulgaris L.) of the pinto variety, broad bean (Vicia faba) and textured soy bean (Glycine max L.) were purchased at a store in S.L.P., México. Beans and broad beans were dried in a drying oven (Binder, Germany) at 55°C for 6 h. The dried seeds were ground by passing them through a mill (Corona, China) adapted with a 400 rpm motor at ambient temperature (23 ± 2°C). The openings of the sieves that the flour passed through were 0.5 mm in diameter. Isolated soy beans were ground directly in the mill. Once the flours were obtained, they were mixed in combinations (w/w) as shown in Table 1.
The formulation to prepare the dough for the chips included 47.5 g/100 g flour, 47 g/100 g water, 5 g/100 g vegetable oil and 0.5 g/100 g salt. The flour mixtures were manually combined with water, salt, and oil to produce the dough, which was in turn passed through a manual tortillarolling machine (TM-G, México) to obtain 1 mm-thick chips. These were then shaped using a 3 × 4 cm triangular cutter. Afterwards, they were baked in an oven (San-Son, Mexico) at 100°C for 100 min and allowed to cool for 2 h before final storage in cellophane bags. The chips were stored at room temperature for 24 h before being analyzed.

Sensory analysis of the chips
The chips were evaluated by an untrained panel of 100 judges between 18 and 80 years old who consume chips at least once per month. These consumers evaluated the degree of acceptance using a 9-point hedonic scale (Since 1 = "I dislike them extremely" from to 9 = "I like them extremely"). The evaluation was applied on 3 different days (5 samples per day). The mean ± std. deviation of the hedonic scale was obtained (from 1 to 9) for each of the analyzed samples. The top 5 formulations according to the sensory analysis underwent moisture content, protein, fat, total carbohydrate, dietary fiber, amino acid profile and protein digestibility analysis.

Proximate composition
The proximate composition of the raw materials and chips was determined by following the American Oil Chemists' Society recommended official methods (AOAC, 1990). The crude protein content was estimated following the method 954.01 to determine the nitrogen content (N). The factor used to convert N to the estimated protein content for soy bean was 5.65 and 5.4 for broad bean and bean (De Almeida Costa, Queiroz-Monici, Machado Reis, & Costa de Oliveira, 2006; Moss, 1990). Total carbohydrates were obtained according to the Official Mexican Standard (NMX-F-089-1978) technique. The total fat proportion was determined following the Official Mexican Standard (NMX-F-312-1978). The moisture content was determined in a drying oven (APSA, México) at 105°C using the AOAC 930.15 method, the samples were previously defatted. All samples were made in triplicate, and the means were reported with their standard deviations.
As a control, the proximate composition, total dietary fiber and protein digestibility of two commercial tortilla chips are provided. The first two were given by the supplier and the digestibility was determined under the same conditions as those of the legume chips. The shaded rows correspond to selected formulation for the experimental design * Bfrijol, Sfrijol de soya, BBhaba Las filas sombreadas corresponden a la formulación seleccionada para el diseño experimental.

Total dietary fiber (TDF)
TDF was determined using the gravimetric-enzymatic method (985.29 of the AOAC), consisting of the simulation of the digestion of the pre-treated sample with the use of enzymes (Megazyme International, Wicklow, Ireland). 1 g of ground, defatted sample was used. Samples were gelatinized with α-amylase (95°C, pH 6, 15 min) and then enzymatically digested with protease (60°C, pH 7.5, 30 min), after which they were incubated with amyloglucosidase (60°C, pH 4.5, 30 min). Then the samples were filtered and simultaneously washed with 95% ethanol and acetone. The recovered filtrate was dried, and protein and ash content were determined to calculate the TDF content. TDF was determined in quadruplicate.

Amino acid determination
The snack's nutritional value was calculated by amino acid profile using UPLC-UV. Between 2.6 to 3.1 mg, previously crushed, dried and defatted samples were hydrolyzated and derivatizated and after this a UPLC-UV analysis were determined on an Acquity UPLC system (Waters, Milford, CT, USA).
The system consisted of a binary solvent manager, a sample manager and a photodiode array detector. The output signal was monitored and processed using the Waters Empower 2 software. The column was a Waters Acquity UPLC HSS C18, which was held at 30°C at a flow rate of 0.18 mL/min; the injection volume was 10 µL. UV detection was set to 210 nm. The mobile phase consisted of eluents (A and B) purchased from Waters (composition proprietary) used according to the manufacturer's specifications. The gradient elution was as follows: 0 min, 60% B; 4 min, 80% B; 9 min, 100% B; 11 min, 100% B; 11.1 min, 60% B; and 15 min, 60% B.

Protein digestibility of the chips (in vitro)
Protein was digested using the method proposed by Yang et al. (2014) with some modifications. Briefly, 4 g of the ground (approx. 0.5 mm) oily sample was mixed with 50 mL of distilled water, and the pH was adjusted to 2 with 6 N HCl. Then, 2 mL of pepsin (10% w/v, SIGMA, USA) was added, and the mixture was incubated at 37°C over the course of 2 h. Then, the pH was adjusted to 5.3 with 0.9 NaHC0 3 and to 7.5 with 1 M NaOH. Then, 5 mL of pancreatine was added to 5% w/v (SIGMA, USA), mixed and incubated at 37°C for 2 h. The enzymes were inactivated by submerging the tubes in boiling water for 10 min. The samples were cooled to room temperature and centrifuged at 4500 rpm for 20 min at 15°C. A blank that only included the enzymes was simultaneously run. The g/100 g of in vitro protein digestibility was calculated as ((total protein-residue protein))/total protein) * 100.

Statistical analysis
Sensory evaluation was statistically analyzed using the Kruskal-Wallis non-parametric statistics test for independent samples. For the statistical analysis of the proximal data and dietary fiber, the five best formulations selected in the sensory study and the raw materials were used. The experiment design used was a ternary mixture of Taguchi (Table 1, shaded rows). ANOVA was performed with a confidence level of 95% (p < .05) using Design-Expert software package, version 7.1.5 (Stat-Ease, Inc., Minneapolis, Minn, USA). Multiple regression analysis was used to fit a full polynomial equation (Eq. 1) to each of the evaluated dependent variables: (1) where y is the studied variable, β0 to β6 are the regression coefficients of the model and B, S and BB are the bean, soy bean, and broad bean flours, respectively. The difference between the five chips selected in the different analyses was compared using a least square means test (orthogonal contrasts) that was performed to detect significant differences at p < .05 (for the proximate compositions of the legume chips that had different formulations). Statistical analysis was applied using STATISTICA software version 7.0 (Stat Soft, Tulsa, Oklahoma, USA), considering a significance level of α = 0.05.

Sensory evaluation of the chips
The fifteen chip formulations with different concentrations of legume flour (Table 1) were prepared and evaluated by the consumer judges. Nonparametric tests showed that with a p ≤ 0.001, the null hypothesis of the test should be rejected. Therefore, the difference in the results of the sensory analyses is due to the formulation of the samples. Consumers' product acceptance (hedonic test) is shown in Table 1. In general, the legume chips were accepted by the consumers in all formulations, with a mean above 5, indicating on the hedonic scale that the consumers liked the product. The chips that were evaluated best were made of 65 g/100 g broad bean: 35 g/100 g soy bean (65/35 BB/S), 35 g/100 g bean: 65 g/100 g soy bean (35/65 B/S), 50 g/ 100 g bean: 50 g/100 g soy bean (50/50 B/S), 75 g/100 g broad bean: 25 g/100 g bean (75/25 BB/B) and 35 g/100 g broad bean: 65 g/100 g bean (35/65 BB/B). The 65/35 BB/S formulation was the best evaluated and was between 7 and 8 on the hedonic scale, suggesting that this formulation was liked much by the consumers. The other 4 best formulations did not exhibit a significant difference in the sensorial analysis. The five best formulations were analyzed for moisture content, protein, fat, total carbohydrate, dietary fiber, amino acid profile, and digestibility.

Proximate composition
Proximate composition was determined for the raw materials and the samples that obtained the best results in the sensory evaluation; the controls were the values of 2 commercial brands of tortilla chips. According to the analysis of variance (Table 2), it was observed that soy bean was the most important factor that affected the protein content in the different formulations produced with mixtures of legumes flour, as well as the interactions when it was included (p < .0001). With respect to carbohydrate content, bean and soy bean were the most important factors that affected these biomolecules (p < .0001) as were soy bean and broad bean for the fiber content (p < .0001). However, the lipids and moisture content were not significantly affected by these factors (p = .393 and p = .55, respectively).
In Table 3 the results of proximate composition are shown; the protein values for the legume chips ranged from 21.39 to 34.17 g/100 g, and the difference was significant (p < .05) among all formulations and the controls. The legume chips with the highest protein content had 27.84 g/ 100 g more protein than the control. The results of the protein contents were expected because of the high protein content of legumes and were in agreement with previous reports in which legume flour was added to their formulations (Man & Păucean, 2013). This high protein content is an important result because in recent years people have begun to focus on increasing protein in the diet; additionally, many people are trying non-animal-based protein sources (Nosworthy et al., 2017).
The carbohydrate contents of the legume chips were between 35.03 and 55.98 g/100 g, which is low compared with the controls that had 72.22 and 78.56 g/100 g (p < .05). All legume formulations and the control had significant differences in carbohydrate content (p < .05). These legume chips had between 16 and 44 g/100 g, fewer carbohydrates than market products. This finding is highlighted because it is well known that diets high in carbohydrates are associated with chronic degenerative diseases. Additionally, increased consumption of carbohydrates is associated with hyperglycemia, which is a factor in the development of insulin resistance and causes obesity, high blood pressure, dyslipidemias and atherosclerosis. These alterations lead to type II diabetes or vascular damage (Souza et al., 2015). Fat content was low, at between 2.26 g/100 g and 3.5 g/100 g, and similar to high-carbohydrate diets, those high in fats are associated with chronic degenerative diseases, and a high consumption of saturated and trans fatty acids elevate the levels of lowdensity cholesterol or decrease the levels of high-density cholesterol, which is associated with predisposition to type-II diabetes (Souza et al., 2015).
An energy-restricted, low-fat and low-carbohydrate but high-protein diet provides nutritional and metabolic benefits that are equal to or sometimes greater than those observed with a high-carbohydrate diet (Noakes, Keogh, Foster, & Clifton, 2005).
The moisture content in all the legume chips and in the control was low (between 3.52 g/100 g and 5.31 g/100 g) following the baking treatment of this product. This low moisture favors the conservation of the product over a longer period and the prevention of microbial growth.
This research also evaluated the fiber content of the legume chips, and this was influenced by the formulation of the chips. All formulations had significant differences in the fiber content between them and the control, except for the chips made with 75/25 BB/B, 35/65BB/S and 65/35 BB/S in which there were no significant differences between them (p < .05). Depending on the formulation, the legume chips include between 3.23 and 19.21 g/100 g more fiber than the controls, and the controls are what people currently consume. It is appropriate to increase the fiber content in the diet since studies indicate that dietary fiber reduces the risk of cardiovascular disease, diabetes, obesity, colon cancer, and a variety of other diseases (Matos-Chamorro & Chambilla-Mamani, 2010). An adequate total fiber intake is 38 and 25 g per day for men (aged 14-50 years) and women (aged 19-50 years), respectively. Consuming 30 g (normal amount that people consume per day) of the legume chips with more fiber Table 2. Regression coefficients and variance analysis of the polynomial model to evaluate the variation of protein, carbohydrates, lipids, moisture, and fiber content based on the interaction of bean, soy bean and broad bean flours.
Tabla 2. Coeficientes de regresión y análisis de varianza del modelo polinomial para evaluar la variación del contenido de proteínas, carbohidratos, lípidos, humedad y fibra en función de la interacción de las harinas de frijol, soya y haba.  , 2007). Importantly, high-fiber and high-protein foods, such as legume chips, have the potential to promote satiety (Gupta & Premavalli, 2012). The chips' proximal content is attributed mostly to the compositions of the raw materials, as is shown in Table 3. Based on the percentages in the formulations, we calculated the theoretical protein content of the raw materials of each formulation and compared them with the protein percentages in the final legume chips, and we observed a 3 to 5 g/100 g protein loss in these products. This reduction is perhaps because of cross-link and Maillard reactions; the Maillard reaction is a nonenzymatic browning reaction that mainly occurs between reducing sugars and free amino groups with lysine being the most susceptible. This is similar to the results of Cruz-Solorio, Villanueva-Arce, Garín-Aguilar, Leal-Lara, and Valencia Del Toro (2018). The protein content was reduced after the drying and baking process. Some investigations report similar protein reductions in their products during these processes (Gutiérrez-Dorado et al., 2008).

Parameters
To establish the effect of baking on the protein quality in the legume chips prepared from blended flours from pinto beans, broad bean and soy bean, the amino acid profile was determined by UPLC-UV.

Amino acid contents
The amounts of almost all amino acids in the legume chips were lower compared with the amounts in the raw materials, indicating that processing reduced them in the final products (Table 4). The amino acid contents of the raw materials were similar to those reported by Nosworthy et al. (2017).
The amino acids that decreased most through processing were Thr, Ala, Pro, Lys and Tyr (between 7 and 20 g/100 g). Between 8 and 14 g/100 g of Thr and between 13 and 20 g/ 100 g of Lys were lost in the different formulations. Lys is known to be sensitive to heat and is lost during some processes, possibly due to Maillard or cross-linking reactions (Cruz-Solorio et al., 2018).
The remaining amino acids suffered minor losses of between 0 and 8 g/100 g. The amino acids that were less affected included Ile with a maximum loss of 3 g/100 g in one of the formulations and Met with a loss of 4 g/100 g. This finding is important because these are essential amino acids (EA); additionally, sulfur amino acid is limited in this kind of product. People must have an adequate balance of amino acids in their diets since they act as agents against chronic human diseases and are conducive to maintaining general well-being; for example, the consumption of Arg is beneficial for promoting skeletal muscle gain and reducing whole body fat accretion, may have a cholesterol-lowering effect, inhibits platelet aggregation, may reduce atherosclerosis, and improves peripheral endothelium dependent dilatation (Aji et al., 1997). Additionally, Arg, Gly, Ala, Ser, Cys and Thr provide protective effects against cardiovascular diseases and cancer (Krajcovicova-Kudlackova, Babinska, & Valachovicova, 2005). Residues of His have been associated with proliferation of hepatocytes during regeneration of the liver following mechanical damage, and it has been identified as crucial for rapid fibrillation to prevent some diseases, such as type 2 diabetes, Alzheimer's and Parkinson's (Besant & Paul, 2012). Sulfur amino acids are involved in DNA transcription and RNA translation and may play a role in reducing the risk of CVD, dementia, cirrhosis, and immunomodulation (Townsend, Tew, & Tapiero, 2004). Leu has many roles, such as in the synthesis of new proteins, as a modulator of insulin and as a donor to produce alanine and glutamine in skeletal muscle (Baum et al., 2005).
It has been documented that cereal proteins, which are a common raw material in chips (corn), are deficient in certain essential amino acids, particularly in Lys (Anjum, Ahmad, Butt, Sheikh, & Pasha, 2005) and that legumes contain sufficient amounts of this amino acid (Sai-Ut, Ketnawa, Chaiwut, & Rawdkuen, 2009). Therefore, it is important to include these legumes to improve the protein and amino acid contents. Millward, Layman, Tomé, and Schaafsma (2008) mentioned that in addition to identifying the optimal dietary amino acid pattern in terms of specific amino acids, for the determination of protein quality, it is important to evaluate the digestibility of a protein or the capacity to provide metabolically available nitrogen and amino acid to tissues and organs; for this reason, we evaluated the protein digestibility of these chips.

Protein digestibility
The protein digestibility values of the chips are shown in Table 5. The highest value of protein digestibility was observed for chips prepared with broad bean and soybeans Table 3. Proximate composition of the legume chips (g/100 g) wet base.
Tabla 3. Composición proximal de los totopos a base de leguminosas (g/100 g) base húmeda. **Mean ± Std. deviation; n = 3 ***n = 4. Means within a column followed by different letters are significantly different (p < 0.05). + The data was given by the provider ** Media ± desviación estándar; n = 3 *** n = 4. Los promedios dentro de una columna seguidos por letras diferentes son significativamente diferentes (p < 0.05). + Los datos fueron proporcionados por el proveedor.g/100 g base húmeda (65/35), and the lowest was for chips made with bean and soybeans (35/65). In general, when broad bean was included in the formulation, a higher digestibility was obtained. Our results are in agreement with those reported by Tazrart, Lamacchia, Zaidic, and Haros (2016), who described that the inclusion of broad bean in fresh pasta improved protein digestibility. This improvement is probably because the inclusion of broad-bean flour increased the content of more digestible proteins (globulins). It's important to notice that chips prepared with broad bean and soybeans (65/35), had the same values of digestibility as those of the control. The low digestibility observed in the sample with bean and soybeans (35/65) is possibly because of protein aggregation during heat processing (Carbonaro, Cappelloni, Nicoli, Lucarini, & Carnovale, 1997). Low digestibility of protein legumes is due to the presence of antinutritional factors (Alonso, Aguirre, & Marzo, 2000). In these chips these antinutritional compounds were probably not completely removed following the baking process. However, according to De Vlieger et al. (2017), these legume snacks could be categorized as a nutritious food because they provide a higher proportion of essential nutrients (approx. 50 g/ 100 g of the EA of the total amino acids) relative to the energy content or portion size, high fiber content (between 11 and 19 g/100 g) and low fat content (between 2.26 and 3.49 g/100 g). They will also have a long shelf life due to their low moisture content. The demand for nutritional food is high, and this is one alternative. Other advantages include the increased nutritional value of the snack without an increase in cost because the process only involves raw material sources that are already available. Thus, there is no need to invest in expensive processes, such as genetic modifications. This product also supports farmers because it uses their products. Additionally, legume chips have a potential use as a natural functional food due to their chemical composition.

Conclusion
The elaboration of a legume snack without the use of cereals was accepted by consumers, which offers a nutritious alternative to fried or baked corn flour goods high in calories and low in protein content. In this investigation, the formulation most accepted by the consumers was elaborated with 65% of broad bean and 35% of soy bean flour. This formulation had 34.17 g/100g of protein (approx. 50% are essential amino acids), 11.53 g/100 g of fiber and 3.52 g/100 g of Table 4. Amino acids content in the raw materials and in the legume chips (g/100 g protein).
Tabla 4. Contenido de Aminoácidos en la materia prima y en los totopos a base de leguminosas (g/100g proteína bean, S -soy bean, BB -broad bean. The shaded rows correspond to the essential amino acids.
lipids. The legume chips developed via baking resulted in a depletion of protein between 3 and 5 g/100 g, which does not represent a major loss due to the high protein content. 65/35 BB/S (broad bean/soy bean) was the product with greater digestibility 72 g/100 g, like commercial tortilla chips. In general, when broad bean was included in the formulation, a higher digestibility was obtained. In future investigations the legumes could be pre-cooked to improve the digestibility in this kind of products or add small quantities of cereals to increase the content of some essential amino acids such as Met.

Disclosure statement
No potential conflict of interest was reported by the authors.