Effect of wheat flour substitution and popped amaranth flour content on the rheological, physicochemical and textural properties of hot-press wheat-oat-quinoa-amaranth composite flour tortillas

ABSTRACT The aim of this research was to evaluate the effect of the popped amaranth flour on the rheology, chemical composition, protein digestibility and physical properties of hot-press tortillas elaborated with a wheat, oat and quinoa flour, compared with refined wheat flour. Mixolab C1 (dough stability) increased 33% with amaranth addition. Results showed that tortillas containing 5:5:20 oat:quinoa:amaranth increased 31.4% and 27.1% soluble dietary fiber and protein than control, respectively. Protein “in vitro” digestibility increased 1.02% and crude fat reduced 35.6%. Tortilla treatment 5:5:20 required same force to deform and to rupture at day 6, and showed a lower L* (16.1%) and higher a*(27.4) and b* (23.1%) than control. Texture acceptance was the same, and odor acceptance was lower for tortilla treatment 5:5:20, compared to control. The partial replacement of wheat flour with amaranth:quinoa:oat flour positively influences the protein content, total dietary fiber content and protein digestibility of tortillas.


Introduction
A popular staple food consumed in Northern Mexico and the United States is wheat flour tortilla (Serna-Saldivar et al., 2019). This unfermented flatbread, similar to other wheat-based foods, is rich in carbohydrates of high glycemic index and contains low proportions of minerals and fiber (Pérez-Carrillo et al., 2015). Due to the consumer concern about the implications that foods such as wheat flour tortillas have in their health, an increased demand for wheat tortillas with better nutritional values has been raised (Asiyanbi-Hammed & Simsek, 2020). Wheat flour substitution with sprouted wheat, yam and legumes flours has been used as strategy to improve tortilla's nutrimental quality (Asiyanbi-Hammed & Simsek, 2020;Heredia-Olea et al., 2015;Liu et al., 2017;Montemayor-Mora et al., 2018;Serna-Saldivar et al., 2019). However, these substitutions have been detrimental for their physical properties.
On the other hand, oat grain flour is characterized by a high concentration of β-glucans, a substantial content of fat, including unsaturated fatty acids and proteins with an equilibrated amino acid composition (Amundsen et al., 2003). Because of this, products containing this flour have hypocholesterolemic and hypotensive effects (Amundsen et al., 2003). Heredia-Olea et al. (2015) evaluated the use of oatmeal and inulin in wheat flour tortillas with good results on their physical properties. Other grains from the Amaranthaceae family, such as quinoa (Chenopodium quinoa), have generous amounts of starch, fiber and high-quality proteins (balanced and sulfur-rich, essential aminoacidic composition) (Contreras-Jiménez et al., 2019;Martínez-Villaluenga et al., 2020). These edible seeds are also a good source of other nutrients with potential health benefits, such as: minerals (calcium, iron and zinc), vitamins and phytochemicals (polyphenols, phytosterols, phytosteroids and betalains) (Alvarez-Jubete et al., 2010). Amaranth (Amarantus sp), is an important source of protein, which provides equilibrium of essential amino acids and with quinoa and oat could offer a nutraceutical food product. Moreover, amaranth is a cultural heritage of Mexico and has been one of the most important foods of Mesoamerican cultures . It is economically accessible and has a higher nutritional value than the main cereals in the human diet (corn, wheat and rice) with a higher protein content, 93% digestibility and is gluten-free (Aragón Gutiérrez et al., 2018).
Because of their excellent nutritional and nutraceutical value, the use of composite flours from these grains is a current trend for the development of functional food in human diets (Martínez-Villaluenga et al., 2020). Their inclusion has been found to improve the nutrient profile in bread and other baking products. However, to our knowledge, this has not been evaluated in wheat flour tortillas. The development of functional composite flour tortillas is a challenge, because the use of flours from other sources, other than refined wheat flour, could have negative effect in their physical properties and shelf life (Asiyanbi-Hammed & Simsek, 2020) .
The objective of this work is to evaluate the effects of the wheat flour substitution with amaranth, quinoa and oat flours on the rheological properties, chemical composition, starch and protein digestibility, microstructure, color and texture properties of wheat flour tortillas. In order to evaluate these effects, wheat flour was reduced 20%, 25% and 30%, to incorporate 10%, 15% and 20% of amaranth flour, respectively, and quinoa and oat flour in 5% (each one), as starting point. The flour blends, as raw material, and the different tortilla treatments were evaluated. Popped amaranth flour was employed as it has been reported to improve the rheological properties of composite flours (De La Barca et al., 2010).

Raw materials
Commercial refined all-purpose wheat (Triticum aestivum L.) flour (Harina Selecta by Molinera de México SA de CV), popped amaranth flour, quinoa flour and oat flour (Vitely S. A. de C.V.) were employed for the preparation of tortillas.

Flour evaluation
Four different tortilla treatments were tested: control (100% refined wheat flour) and three experimental tortillas systems containing in all case 5% oat flour and 5% quinoa flour, and 10% amaranth flour (T1), 15% amaranth flour (T2) and 20% amaranth flour (T3) and wheat flour. Levels of oat flour and quinoa were selected in a previous study with basis in dough subjective observation. Moisture content, crude protein, sedimentation and Mixolab tests were assayed using AACC (2010) Methods 44-15-A, 46-13.01, 56-60, and 54-60.01, respectively. Briefly, the sedimentation test measures the sedimented volume of a flour slurry after 5 min resting. High gluten flours yield higher sedimentation values because of its water absorption and swelling capacity. Dough rheological properties were determined with the Mixolab 2 (Chopin Technologies, Cedex, France). This method determined in the same assay the dough rheological properties and the amylograph behavior during heating and cooling cycles. The starch gelatinization was followed during the increase of temperature from 35° to 90°C at a rate of 2°C/min. The texture profile analyses (TPA) of optimally developed doughs were assessed with the TA.XT2i Texture Analyzer (Stable Micro Systems, Godalming, England) according to Montemayor-Mora et al. (2018). Tests were conducted with a cylinder probe acting with a force of 4.9 g with doughs previously placed on a cylinder mould (cylindrically formed doughs with a height of 4.5 cm and diameter of 4.6 cm). Resulting doughs were placed on the base and, for each test, the probe moved downward at a speed of 5 mm/s to determine hardness, adhesiveness, cohesiveness, and springiness.

Preparation of composite flour tortillas
The baker's formulation employed consisted of 2000 g wheat flour or composite flour (14% mb), 200 g shortening, 30 g iodized salt, 20 g low fat dry milk, 30 g double acting baking powder, 6 g sodium stearoyl lactylate, 4 g fumaric acid and 2 g potassium sorbate. Dry ingredients were first blended at low speed with hook attachment for 4 min in a Hobart Mixer H120 (Hobart, Ohio, United States). Then, a determined volume of distilled water was added, by the Mixolab equipment, at 40°C and mixed at slow speed for 1 min and then changed to medium speed until attaining dough development. Each of the treatments was made with 50%, 52% and 54% of humidity to evaluate the effect it has on their physical characteristics. Water content was selected with subjective evaluation of dough quality Processing was according to Heredia-Olea et al. (2015). Briefly, resulting doughs were divided into 30.00 ±0.05 g pieces, hand-rounded and immediately allowed to rest in a proof cabinet (National Manufacturing Co., Lincoln, NE) adjusted to 29°C and 85% RH for 30 minutes. Then, each dough ball was flattened using a commercial inclined hot press for 3.13 s. The temperature of the hot plates was adjusted at 187°C ± 5°C and the gap between the hot plates adjusted to 1.75 mm. The resulting flattened tortilla discs were baked on a four-pass circular moving griddle set at different temperatures (Manufacturas C&D Industriales, Monterrey, NL, Mexico). The average residence time of the tortillas on each pass was 11 s. The consecutive baking temperatures for the four stages of the oven were 200°C ± 5°C, 260°C ± 5°C, 265°C ± 5°C and finally 230°C ± 5°C. Tortillas were flipped over when changing from each of the baking plates, so each side of the tortillas baked for 22 s. The baked tortillas were first cooled on a cooling rack for about 20 s and then on a table at room temperature (25°C ± 3°C) for 30 minutes. The resulting tortillas were immediately placed inside sealed polyethylene bags and kept at room temperature for further evaluations (Serna-Saldivar, 2012).

Physical, chemical and textural characterization of tortilla's quality
Ten tortillas from each treatment were randomly selected from 100 tortillas and measured for diameter and thickness. The diameter of the tortillas was the average of two diagonal measurements with a plastic rule. Thickness was measured by measuring five piled tortillas and dividing the result by five. Five tortillas from each treatment were randomly selected, and their colour was measured using the Konica Minolta Spectrophotometer model CM-700d/600d (Japan).

Microscopic studies
Control treatment and T3 treatment stored at 0, 1 and 6 days were freeze-dried (Labconco FreeZone 6 plus, Kansas City, USA) at a pressure of 0.024 mbar and a temperature of −84° C for 24 hours, then, milled to pass the 40 US mesh using a cyclone sample mill (UDY Corporation, Fort Collins, CO, USA). An EVO MA 25 scanning electron microscope (Zeiss, Germany) was employed to observe the microstructural characteristics of the samples, as in Wu et al. (2010). Dried material was poured on double-sided tape mounted on aluminum stubs. The samples were coated with 10 nm of gold film in a Q150R ES rotary pumped coater (Quorum Technologies Ltd, East Sussex, UK) and examined under high vacuum conditions at an accelerating voltage of 5 kV.

Enzymatic measurements
Total dietary soluble and insoluble fibres were determined using the commercial kit Total Dietary Fibre provided by Megazyme (Ireland), following approved methods 32-45.01 and 32-50.01 of the AACC (2020). In vitro protein digestibility was measured following the multienzymatic method of Hsu et al. (1977) and starch digestibility according to Montemayor-Mora et al. (2018).

Organoleptic acceptability
Organoleptic evaluation was performed in a sensory evaluation laboratory furnished with individual booths. Thirty untrained panellists between 20 and 30 years evaluated the overall acceptability of the control and the T3 tortilla systems. Each panellist was given the hot tortilla coded samples along with a glass of water and a ballot. Panellists were asked to evaluate texture, flavour, and overall acceptance on a seven-point hedonic scale were 1 = like very much and 7 = dislike very much (Serna-Saldivar, 2012).

Statistical analysis
Flour characteristics and tortilla data were analyzed using a randomized experimental design using Analysis of Variance (ANOVA) to determine if there were statistical differences among the values measured for each property of the pizzas with a 95% confidence interval. All measurements were carried out with a minimum of 3 replicates per sample, except for the microscopic studies, where a random sample from the replicated control and T3 treatments was selected to observe under microscope. In the case of the hedonic scale result, the average of the 30 panelist evaluations of control and T3 treatments was compared. Measurements are presented as their average and standard deviation. Tortilla texture measurements were analyzed with non-parametric tests (Kruskal Wallis). Tukey method (p-value <0.05) was used to determine if there was a significant difference by comparing the means of the studied data. Obtained data of flour and tortilla characteristics was analyzed using Minitab 16 Statistical Software (Minitab LLC, State College, Pennsylvania, USA). Table 1 shows the results of composite flour characterization. In terms of moisture, there is no difference in content between flours. Crude protein increased 17%, 14.4% and 13.6% in T1, T2 and T3 composite flours, respectively. However, no significant difference among composed flour treatments was found. Sedimentation increased while amaranth content increased. Guardianelli et al. (2019) observed that, when adding amaranth flour to a wheat dough, the quantity of dry gluten (DG), as indicator of water-insoluble protein, increased. In this case, no gluten was added to the flour, since oat, quinoa and amaranth had no gluten. Thus, the observed results (crude protein content and sedimentation volume) may be associated with the increase in waterinsoluble proteins in the dough formulation. Furthermore, the popping process of amaranth could have increased the amount of water-insoluble proteins, which may stabilize the gluten matrix (Guardianelli et al., 2019) Water absorption was reduced from 63% to 59% and time of development was reduced to a half in all composite flours with respect to control. Dough stability time was reduced from 8.8 min to 0.5 min.

Dough characterization
The more amaranth flour is added, the less water and the less time are needed for the flour to be processed into tortillas, as well, the stability of the dough at its maximum consistency is decreased (Table 1). This may have been the result of the decrease in real gluten content. This behavior could be the result of the increase in the proportion of globular proteins (Avanza & Añón, 2007;Quiroga et al., 2009), due to the effect of dilution of gluten proteins of wheat flour that amaranth protein cannot compensate (Guardianelli et al., 2019). However, no significant difference was observed among treatments, which may be owed to the content of β-glucan and dietary fiber, present in the quinoa and oat flours (Habtamu, 2012).
Hardness and adhesiveness in T3 increased almost three times with respect to control. Cohesiveness increased directly with amaranth substitution. Springiness has no difference between flours. At this level of composition, the high fiber content and the water level may have an influence and make tortillas more cohesive and harder. This was in accordance with previous works using composed flours in wheat dough (Guardianelli et al., 2019;Habtamu, 2012).
The Mixolab C1 parameter (dough stability) increased according to the increased amaranth content, being the C1 value of the T3 treatment, 33% higher than the control flour. In case of C2 (protein weakening), C4 (hot gel stability), C5 (starch retrogradation in the cooling phase), C2-C1 (protein weak range), C5-C4 (gelling) and Alpha parameters, values were reduced when amaranth content increased. On the other hand, C3 (starch gelatinization), C3-C2 (starch gelatinization range), C4-C3 (cooking stability range), Beta and Gamma parameters, they increased or decreased in T1 with respect to control flour but with higher amaranth substitution these differences were reduced. This difference in dough rheology could be explained by the molecular mobility of water in dough. As mentioned by Guardianelli et al. (2019), water with high molecular mobility in dough links to other components of dough in a weak form (high-energy mobile state), and leads to a more labile gluten structure.

Tortilla's characterization
The size and dimensions of tortillas were not different between treatments (Table 2). However, in terms of color, L decreased with an increased amaranth content. In that way, T3 was 15.31% less luminous than control. The highest values of a and b were observed in T3 (4.94 and 23.8, respectively). The amount of pigments in the composed flour and the particle size of the flour may also have an effect in the color of tortillas. Commonly, a and b values increase when adding quinoa flour to wheat flours (El-Sohaimy et al., 2019). Although the addition of oat flour may have a darkening effect and a loss of brightness, the addition of popped amaranth flour may have an inverse effect. In a study of bread fortification with amaranth flour, Miranda-Ramos et al. (2019) and Sanz-Penella et al. (2013) observed the same effect due to amaranth flour addition to bread systems.
In terms of chemical composition, moisture and fat content were not different among treatments. But, as was expected, T3 had the highest ash and protein content. The effect on ash and protein content due to amaranth addition was previously reported in studies in bread (Banerji et al., 2018;Miranda-Ramos et al., 2019;Sanz-Penella et al., 2013). Rollability value had no significant difference among the treatments.
On day zero, the toughness value obtained was the highest for the control treatment and the lowest for T3, the latter being half of the former (Table 3). After 6 days of storage,  the toughness value for T3 did not present statistically significant changes; however, for control it was reduced to almost half of the value from day 0. In the case of extensibility, a similar behavior is observed. The extensibility for T3 at 6 days was reduced only by 20.00% and for Control, this value was reduced by 61.68%. In studies on wheat bread, Sanz-Penella et al. (2013) and Miranda-Ramos et al. (2019) observed a similar effect on texture. In their study, crumb hardness and elasticity increased with amaranth flour substitution. Elasticity also decreased in flat bread elaborated with quinoa and wheat flour (El-Sohaimy et al., 2019), due to poor extensible gluten network. The effect of amaranth flour addition on wheat tortilla extensibility could be associated with amaranth albumin interactions with gluten, according to Oszvald et al. (2009). On the other hand, with respect to the effect on toughness, storage stability could be associated with starch retrogradation. Srichuwong et al. (2017) study in quinoa and amaranth suggests that short amylopectin branch chains and high soluble dietary fiber also may lead to a low retrogradation tendency of starch.

Morphological properties
The scanning electron micrographs are illustrated in Figure 1. There was some difficulty in viewing the samples owed to their high fat content. However, a big difference was observed in the tortilla samples with amaranth, compared to the wheat flour tortilla samples without amaranth. Table 3. Changes in textural biaxial properties during storage for 6 days at room temperature of hot-press tortillas produced with composite flours.
The samples with amaranth appeared to be more cohesive, while wheat tortilla samples appeared to be crumbling. Previous studies using amaranth flour in wheat pasta showed a more heterogeneous structure of cooked samples (Martinez et al., 2016), which may be explained by a higher protein content. There was no significant difference regarding the storage conditions.

Dietary fiber content, digestibility and sensory acceptance
In terms of fiber content, a significant difference between control and T3 treatments was seen (Figure 2(a)). Soluble dietary fiber content was increased by 31.60% with the addition of amaranth flour. Substitution of wheat flour by amaranth in bread has had similar results, increased only insoluble dietary fiber (Miranda-Ramos et al., 2019;Sanz-Penella et al., 2013). Regarding starch digestibility, this was reduced by 5.16% in T3 with respect to Control. In another study on isolated amaranth starch, it was observed a highest digestibility (Xia et al., 2015), however in systems like bread has been observed an increase in resistant starch according to wheat flour is substituted by amaranth flour (Sanz-Penella et al., 2013) and quinoa (Xu et al., 2019). Protein digestibility was increased 1.35% in T3 with respect to Control (Figure 2(b)).
The sensory evaluation showed that the texture of the control and T3 treatment had the same acceptance by the panelists (Figure 2). Similar results in bread have been reported using amaranth (Banerji et al., 2018), quinoa (Xu et al., 2019) or oat . The parameter aroma had the best rating in T3. This could be associated with the nutty flavor that has been observed in sensory evaluations in bread with amaranth (Banerji et al., 2018) or quinoa (Stikic et al., 2012;Xu et al., 2019). In general, the T3 has a better acceptance in terms of color, flavor and general acceptance (Figure 2(c)).
The production of hot-press wheat tortilla, fortified with amaranth, quinoa and oat flour, is an excellent way to improve the protein and dietary fiber consumption among consumers. This product is an excellent alternative for wheat flour tortillas because they provide higher nutritional value maintaining the similar physical characteristics and organoleptic acceptability.

Conclusion
The treatment with 30% substitution had superior characteristics than the rest of the treatments, including the control treatment. It had similar organoleptic characteristics than the control treatment, except in odor aspect, but higher protein content (27.1% increase) and dietary fiber content    2. Contenido de fibra dietaria (a), digestibilidad "in vitro" (b) y aceptación sensorial (c) de tortillas control y de T3. Columnas y líneas sólidas representan los resultados del control, en tanto que el las columnas con patrón y líneas punteadas representan los resultados de T3.
(31.4% increase), while total starch and crude fat values reduced to 9.5% and 35.6%, respectively. Rollability values were statistically similar among treatments. The addition of composite flours to cereal-based products is a great opportunity to the market for its nutrimental added value.

Disclosure statement
There are no relevant financial or non-financial competing interests to report.