Introduction
The making of table tortillas is established on the nixtamalization process, based on lime cooking of whole maize grains (Serna-Saldivar and Chuck-Hernandez, 2019). Nevertheless, this traditional process has several disadvantages, such as high-energy requirement, and abundant water consumption discarded as high alkaline effluent with a high soluble and suspended solids concentration. These wastewater solids, called nejayote, contain some nutrients associated with the maize pericarp tissue, lost during this alkaline treatment (Cuevas-Rodríguez et al., 2009; Gutiérrez-Uribe et al., 2010). Therefore, alternative processes to overcome disadvantages of the nixtamalization process are needed, especially in more environmental-friendly maize processing methods. The lime-cooking extrusion process has been studied with a technological alternative to make dough flours applicable to developing table tortillas with comparable attributes as those elaborated employing conventional processes (Milán-Carrillo et al., 2006). These emerging technologies offer benefits such as energy-saving, null generation of wastewater production, and better-quality control (Cuevas-Rodríguez et al., 2009). Hence, the nutrients and bioactive compounds associated with the outermost layers of maize are preserved, and the developing foodstuffs are similar to whole grain (Serna-Saldivar and Chuck-Hernandez, 2019).
Mexican native pigmented maize and its products have gained prominence due to a relevant source of diverse phytochemicals such as phenolics, carotenoids, dietary fiber, triglycerides, and phytosterols (Acosta-Estrada et al., 2014; Corrales-Bañuelos et al., 2016; López-Martínez et al., 2009; Mora-Rochín et al., 2016). Most of these bioactive compounds exert antioxidant effects preventing oxidative stress related to chronic diseases and cancer (Reynoso-Camacho et al., 2015; Urias-Lugo et al., 2015). The current interest has been expressed concerning the health benefits of consuming phospholipids and phytosterols. These phytochemicals inhibit the absorption of cholesterol from the small intestine, thus effectively lowering total blood cholesterol and low-density lipoprotein (LDL), a potential risk factor for cardiovascular diseases (Barrera-Arellano et al., 2019).
Fatty acids and phytosterols are unstable at high temperatures and, being unsaturated lipophilic compounds, are susceptible to oxidation. However, phytosterols are considered more stable than mono-unsaturated fatty acids (e.g., oleic acid) because of the steric hindrance in the ring structure. The extrusion process can prevent the release and oxidation of fatty acid by denaturing hydrolytic enzymes; on the other hand, phytosterols content can increase with a 90 to 110 °C extrusion temperature (Camire et al., 1990; Hu et al., 2018; Tolve et al., 2020). However, little has been reported about maize extrusion process conditions(to prevent a decreased in these compounds.
Response surface methodology (RSM) is a collection of techniques used in empirical studies to determine the relationship between a response variable and several input variables. It is a useful mathematical statistics method for establishing models evaluating the relative significance of variables and determining optimal conditions of desirable responses (Khuri and Cornell, 1987). Desirability optimization (DOM) is an analytical technique for optimizing multiple response variables employing mathematical transformations, developed by Harrington (1965), and later modified and extended by Derringer and Suich (1980). Some authors have reported that the extrusion cooking process could be resolved using RSM and DOM (Escalante-Aburto et al., 2014; Ramos-Enriquez et al., 2018; Ortiz-Cruz et al., 2020). Therefore, it could be used to optimize process conditions that lead to minimal changes (or maximization) in flours and tortillas lipophilic compounds, considering that no studies have been reported on fatty acids and phytosterols throughout the different steps of maize tortilla processing. Thus, this study’s purpose was to use RSM and DOM to optimize lime cooking extrusion conditions from mexican native blue maize, that would maximize lipophilic compounds to obtain a high-quality flour to make tortillas.
Materials and methods
Material
The study was performed on blue maize kernels from the Elotero Sinaloa landrace of Mexico’s northwestern region. These pigmented maize accessions were collected during 2018 from open-pollinated maintained by traditional farmers at their villages in the municipality of Concordia located at (23° 17′ 18″ N, 106° 4′ 3″ W), in the state of Sinaloa México. Maize samples were stored at -4 °C until use.
Production of extruded maize flour
The extrusion cooking experiments of pigmented maize grits (moisture content of 28%) were carried out on a single screw laboratory extruder Model 20 DN (CW Brabender Instruments, Inc., NJ, USA), with a length-to-diameter rate of 20:1, an internal diameter of 19 mm, nominal compression ratio 1:1 and die opening of 2.4 mm. A screw-operated feed hopper-fed the extruders at 30 rpm. The extruder feed was manual at 70 g/min. Extrusion temperature (ET) was defined as the temperature at the die end of the barrel. Table 1 shows the combinations of extrusion temperature (ET) and screw speed (SS) used for producing tortillas from extruded pigmented maize flours. Extrudates were cooled and dried, using ambient conditions (25 °C, RH = 65%), for one day and milled (UD Cyclone Sample Mill, UD Corp, Boulder, CO, USA) to pass through an 80-US mesh (0.180 mm) screen, packed in plastic bags, and stored at 4 °C.
Assay2 | Process variables3 [ET, (X1)] | Response variables4 | |||||
---|---|---|---|---|---|---|---|
Fatty acids | Phytosterols | ||||||
[SS, (X2)] | Linoleic Acid (LA) YLA | Oleic Acid (OA) YOA | Campesterol (CP) YCF | Stigmasterol (SP) YSF | β-sitosterol (βSP) Y βSF | ||
1 | 65 (-1) | 78 (-1) | 2,127.0 | 1,098.9 | 3,257.3 | 2,392.9 | 20,393.5 |
2 | 135 (+1) | 78 (-1) | 2,883.6 | 1,556.5 | 4,353.9 | 3,108.7 | 22,465.1 |
3 | 65 (-1) | 212 (+1) | 2,139.5 | 868.6 | 3,157.2 | 2,990.2 | 17,945.3 |
4 | 135 (+1) | 212 (+1) | 2,193.2 | 1,092.6 | 2,578.6 | 2,855.4 | 16,504.7 |
5 | 51 (-1.414) | 145 (0) | 2,362.2 | 945.0 | 3,193.6 | 2,374.0 | 15,465.7 |
6 | 150 (+1-414) | 145 (0) | 3,005.8 | 1,130.9 | 3,824.6 | 2,631.2 | 18,222.6 |
7 | 100 (0) | 50 (-1.414) | 3,020.9 | 1,480.1 | 4,839.9 | 3,208.3 | 24,447.6 |
8 | 100 (0) | 240(+1.414) | 2,123.7 | 1,161.2 | 3,268.1 | 2,954.7 | 18,471.1 |
9 | 100 (0) | 145 (0) | 3,066.3 | 1,491.9 | 3,627.8 | 3,299.5 | 20,211.4 |
10 | 100 (0) | 145 (0) | 3,198.3 | 1,492.2 | 3,832.2 | 3,292.9 | 19,991.4 |
11 | 100 (0) | 145 (0) | 3,190.4 | 1,512.3 | 3,775.9 | 3,408.9 | 21,224.1 |
12 | 100 (0) | 145 (0) | 3,295.6 | 1,462.1 | 3,007.6 | 3,170.9 | 19,789.1 |
13 | 100 (0) | 145 (0) | 3,025.6 | 1,578.3 | 2,990.2 | 3,154.7 | 20,849.4 |
1Central composite design with two factors and five levels; 13 assays. 2Does not correspond to order of experiments. 3ET = Extruded temperature (ºC), SS = Screw speed (rpm); values in parentheses are coded levels. 4LA = Linoleic acid (μg/100 g DW), OA = Oleic acid (μg/100 g DW), CP = Campesterol (μg/100 g DW), SP = Stigmasterol (μg/100 g DW), βSP = β-sitosterol (μg/100 g DW).
Tortilla preparation from extruded maize flours
Tortillas were made by mixing 200 g of extruded maize flours with 200 mL of water to achieve an adequate masa consistency to produce table tortillas. According to Cuevas-Rodríguez et al., (2009), the fresh dough was divided into 30 g pieces and flattened using a manual machine. The resulting disks were baked on a hot griddle at 220 ± 5 °C for 10 s on each side, until the puffing of the tortilla occurred. The fresh tortillas were dried and milled to pass through an 80-US mesh (0.180 mm) sieve and packed in plastic bags. Tortillas made by lime extrusion cooking flours were stored at -20 °C until use.
Lipid extraction
Lipids were extracted by mixing 4 g of maize samples with n-hexane/dichloromethane (40 mL, 1:1, v/v) under stirring conditions for 1 h at room temperature as previously described (Esche et al., 2012).
Lipophilic biocompounds identification and quantification
The identification and quantification of lipophilic biocompounds were performed in a HPLC-ELSD-UV (Model G1969A Agilent 1100 Santa Clara, CA, USA) as previously reported (Chávez-Santoscoy et al., 2014). The separation was performed in a Luna C8 (250 mm x 4.6 mm, 5 μm; Phenomenex, Torrance, CA, USA) column, set at 40 ºC, and an injection volume of 20 μL. The mobile phase consisted of acetonitrile (A) and 55% methanol, with 1% formic acid in water (solvent B). The gradient elution was: 0-7 min 0% B (flow rate 0.6 mL/ min), 7-15 min 0-15% B (flow rate 0.6-1.2 mL/min), 15-20 min 15-80% B (flow rate 1.2-1.5 mL/min) and 20-50 min 80-100% (flow rate 1.5 mL/min). The quantification was compared with linoleic and oleic acids, β-sitosterol, campesterol, and stigmasterol standards (Sigma-Aldrich Co). Results expressed in µg of fatty acids or phytosterols/100 g of dry sample.
Regression analysis and optimization
In this research, the RSM was used to determine the optimal experimental lime cooking extrusion process conditions (extrusion temperature and speed screw) to obtain extruded flour from mexican blue maize. The optimal response variables: fatty acids [linoleic (LA) and oleic (OA) acids], and phytosterols [campesterol (CP), stigmasterol (SP) and β-sitosterol (βSP)], in a central composite design (CCD) with two control variables, extruded temperature (ET) and speed screw (SS) were used. Maximum and minimum values were chosen according to previous data from preliminary trials. Table 1 shows the experimental design and the response data of tortillas made with extruded blue maize flours. The quadratic model applied to predict the response variables is given below.
Data were subjected to the stepwise regression analysis; the significant terms (P<0.05) were used to fit the predictive model for each response variable (Khuri and Cornell, 1987). According to the software program, multiple response optimization was implemented throughout the desirability function (Derringer and Suich, 1980). To establish the desirability of several arrangements of the experimental process variables (ET and SS) were set as ‘‘in the range”, whereas that of response variables (LA, OA, CP, SP, and (SP) was a goal set to obtain maximum, and the desirability value was calculated. The arrangement of experimental factors yielding the uppermost desirability was nominated as the optimal lime extrusion cooking condition.
Results and discussion
Appropriate models by response variables
The quadratic polynomial equations for each response variable (LA, OA, CP, SP, and (SP), efficiently fitted to the experimental values of lime extrusion cooking conditions (ET and SS) using multiple regression analysis, are shown in Table 1. The regression coefficients and analysis of variance of the quadratic models, showing the relationships among response variables and process variables for tortillas produced from extruded blue maize, are shown in Table 2. According to Vera-Candioti et al. (2014), a good predictive model should have the following statistical parameters: coefficient of determination (R2) and adjusted-R2 high (> 0.80), very small P-value (<0.05), coefficient of variation (CV < 10%), lack of fit test (P>0.05), and adequate precision > 4. The models were adequate for predicting the five response variables at different lime extrusion cooking conditions based on these criteria. In this study, the results reveal that the regression response models could depict the experimental region(s responses (Table 2).
Parameter | Regression parameter coefficients | ||||
---|---|---|---|---|---|
Fatty acids | Phytosterols | ||||
Linoleic (LA) | Oleic (OA) | Campesterol (CP) | Stigmasterol (SP) | β-Sitosterol (βSP) | |
Coded values | Coded values | Coded values | Coded values | Coded values | |
Intercept β0 |
3,155.2 | 1,495.4 | 3,743.3 | 3,285.3 | 20,393.1 |
Linear | 566.2** | ||||
β1 β2 |
215.1** -243.3** |
119.2** -144.3** |
213.8** -549.8** |
118.1** -1.8NS |
-2,107.6** |
Quadratic | |||||
β11 β22 |
-308.7** -364.6** |
235.6** -94.2** |
-209.6** 62.9NS |
-380.2** -90.7NS |
-1,740.6** 567.0NS |
Interactive | |||||
β12 | -175.7NS | -56.2NS | -493.8** | -212.7** | -878.1** |
Model F-values | 12.62 | 22.19 | 16.36 | 14.24 | 32.75 |
p-value | 0.002 | 0,0004 | 0.001 | 0.002 | 0.0001 |
R2 | 0.900 | 0.941 | 0.921 | 0.911 | 0.959 |
R2 ajust | 0.829 | 0.898 | 0.865 | 0.847 | 0.825 |
Lack of Fit | 0.06NS | 0.123NS | 0.085NS | 0.31NS | 0.36NS |
CV | 7.1 | 6.2 | 6,2 | 4.5 | 3.2 |
Ade Pre | 8.15 | 12.1 | 13.7 | 10.1 | 19.4 |
NS = Not significant (p>0.05); ** Significant (p<0.05)
Response surface model for linoleic acids
The regression analysis showed that Linoleic acid (LA) was significantly dependent on linear terms of extruded temperature (ET) and screw speed (SS), and quadratic terms [(ET)2, (SS)2] (Table 2). The following equation can be described as the predicted model for LA in terms of coded values:
The response regression model exhibitions, show lower (P< 0.002) and CV = 7.1% values, a satisfactory level by the correlation coefficients (R2 = 0.900; R2 adjust = 0.829), lack of fit (P > 0.06), and PRESS > 8.2 (Table 2). These statistical parameters demonstrated that the fitted model was suitable and reproducible. The surface response plots are revealed in Fig. 1A. The highest amounts of LA (3,280.9 μg/100 g DM) were observed at ET = 115-120 ºC/SS = 110-120 rpm.
Response surface model for oleic acid
The oleic acid (OA) content in the blue maize tortillas prepared from extruded flours was dependent on linear terms ET, SS, and quadratic terms [(ET)2, (SS)2] (Table 2), the following regression equation confirmed the estimate of the effects of independent variables on OA:
The response regression model displays the lower (P<0.0004) and CV = 6.2%, PRESS > 12.1, and coefficient of correlation (R2) from the developed model, which described 94.1% of the total variability to OA on extruded blue maize tortillas, suggesting the selected model adequately illustrates the information for this response. Response surface contours for OA as functions of the independent process variables are represented in Fig. 1B. The maximum (1,561.3 μg/100 g DM) values of OA were observed at ET = 110-120 ºC/SS = 78-85 rpm.
The behavior of fatty acids as a function of the lime cooking extrusion variables is shown in the surface graphs of Fig 1A and Fig 1B. In general, LA and OA’s desirable values in tortillas made with extruded flours were observed in the ET range (100-125 ºC) values and lowest SS values. Some research indicated that during cooking extrusion, the only processing variable that had a significant effect on the fat loss was temperature profile. Its increase caused an augment of fat loss, which could be attributed to forming a bond between starch and lipid fraction. Moreover, the highest barrel temperature values favor the migration of fat fraction outside from mass extruded (De Pilli et al., 2011).
Response surface model for campesterol
Changes in campesterol (CP) of tortillas prepared with extruded blue maize flours were affected by ET, SS, the quadratic term (ET)2, and interaction (ET)(SS) (Table 2); the following regression equation was used to estimate of the effects of independent variables on CP using coded variables.
The predictive model explained 92.1% of the total variability (P<0.001) in CP values (Table 2). The response surface contour plot for CP as a function of the lime extrusion cooking process variables is shown in Fig. 1C. The maximum (4,500.0 μg/100 g DM) value of CP were observed at ET = 120-135 ºC/SS = 78-85 °C.
Response surface model for stigmasterol
The stigmasterol (SP) of tortillas prepared from extruded native pigmented maize flours were influenced significantly by linear terms of ET, the quadratic term (ET)2, and interaction (ET)(SS) (Table 2). The regression equation relating the response function SP, measured as independent variables, was given in terms of uncoded variables by the following equation:
The significance (P<0.002) of the developed model for SP is given in Table 2. The lack of fit (P≥ 0.05), CV, and PRESS had satisfactory levels, indicating that the experimental data were satisfactorily explained in a 91.1%. The response surface contour plot for SP is shown in Fig. 1D. The maximum SP (3,296.3 μg/100 g DM) value was observed at ET = 105-128ºC/SS = 78-85 rpm.
Response surface model for β-Sitosterol
The regression analysis showed in linear terms (ET and SS), quadratic terms (ET)2, and interaction (ET)(SS), a significant effect on β-Sitosterol (βSP) of extruded blue maize tortillas (Table 2). The following equation can describe the predicted model for (SP in terms of coded values:
The predictive model explained 95.9% of the total variability (P≤0.0001) in (SP values (Table 2). The response surface contour plot for (SP is shown in Fig. 1E. Increases in extrusion temperature result in increases in (SP content, reaching a maximum (22,824.2 (g/100 g DM) at 100 to 135ºC ET and SS= 79 to 90 rpm. The remaining CP, SP, and (SP amounts in tortillas from extruded mexican blue maize varied from 25.8 to 48.4, 23.7 to 34.1, and 139.5 to 244.5 μg/100 g, respectively (Table 1). Our results showed that phytosterol levels tested in the present study were within the range of 25 varieties of raw maize reported by Esche et al. (2013). Overall, these significant losses of pigmented maize samples’ phytosterols during the tortillas elaboration could be attributed to the higher temperature during processing, facilitating thermo-oxidative degradation of phytosterols, including oxidized phytosterols, fragmented phytosterol molecules, volatile compounds, and oligomers (Rudzinska et al., 2009).
Optimization and validation of extrusion conditions
Fig. 2 exhibited the desirability functions response surface to attain optimum conditions in optimized tortillas made with extruded mexican blue maize. By applying the desirability function method, the optimum lime extrusion cooking conditions for the development of extruded native blue maize tortillas correspond to extrusion temperature (ET, 119 °C) and screw speed (SS, 79 rpm) with global desirability value (D = 0.906).
The fitness of the established model for the estimate was confirmed by relating the estimated and experimental values. The optimum condition (ET = 119 ºC and SS = 78 rpm) was experimentally tested to confirm the accuracy of the model equations, using the average values obtained in LA and OA experiments. CP, SP, and (SP under these optimal conditions were 3,019.9, 1,625.6, 4,484.9, 3,101.8, and 22,152.9 μg/100 g DW, respectively, which was in agreement with the predicted information (Table 3). The experimental results denote the correctness of the established quadratic models. It is noteworthy to mention that these best values are acceptable inside the specified array of process factors.
Response variables a | Predicted values b | Experimental values c | Percent relative error d |
---|---|---|---|
Linoleic acid (LA) | 3,155.6 | 3,019.9 ( 184.2 | 4.30 |
Oleic acid (OA) | 1,570.6 | 1,625.6 ( 87.8 | 3.50 |
Campesterol (CP) | 4,682.0 | 4,484.9 ( 331.9 | 4.21 |
Stigmasterol (SP) | 3,263.3 | 3,101.8 ( 151.9 | 4.95 |
β-Sitosterol ((SP) | 23,348.3 | 22,152.9 ( 1,572.8 | 5.12 |
aμg/100 g DW
bPredicted using response surface, quadratic model.
cMean ± standard deviation of triplicate determinations from experiments. d Percent error (percentage error) is the difference between an experimental and predicted value, divided by the predicted value, multiplied by 100 to give a percent.
Conclusions
Mexican pigmented maize used in this study was suitable for tortillas preparation throughout the lime cooking extrusion process. RSM and DOM were useful in producing predictive models and establishing relationships between processing factors and key responses for tortillas production from blue maize landrace. The optimum combination of lime extrusion cooking process variables, to produce optimized tortillas from extruded native blue maize, correspond to extrusion temperature of 119 °C and screw velocity of 79 rpm. Additional studies are needed to determine the nutraceutical potential of tortillas using optimized extruded blue maize landrace.