Study material

Two populations of blue/purple corn from each of the following races were used: Chalqueño (CHAL), Conical cob (EC) and Bolita (BOL). The populations of CHAL were obtained from the Álvaro Obregón delegation in the “El Tejocote” area of San Bartolo Meyalco (19° 21’ 23.39” north latitude, 99° 14’ 10.26” west longitude, at 2 700 meters above sea level). Those of EC came from the localities of Mexicalzingo (19° 12’ 33” north latitude, 99°35’9” west longitude) and Tenango of Aire (19° 9’ 27”north latitude, 98° 51’ 29” west longitude ) State of Mexico, located at heights of 2 600 and 2 380 meters above sea level, respectively. The BOL breed was obtained in the towns of San Martín Tilcajete (16° 51’ 32” north latitude, 96° 41’ 42” west longitude) and Valde Flores (15° 51’ 5.04” north latitude, 96° 47’ 35.88” west longitude) in the state of Oaxaca, located at 1 500 and 1 447 meters above sea level, in that order.

Samples of 1-2 kg of ears of each stock were obtained directly from the producers and stored in plastic bags under refrigerated conditions until processed. The identity of the populations with their corresponding race was realized with the support of personnel of ample experience in the classification of maize germplasm. As a control was incorporated a white grain corn (H-40) that is destined in the region of Valles Altos for the elaboration of tortillas. In order to obtain information about the hardness of the grain in the study populations and to be able to give them adequate cooking time during the nixtamalization, the grain hardness was determined by the flotation index (^{Salinas et al., 1992}).

Nixtamalization and making tortillas

The nixtamalization was performed from 100 g of clean grain, with 200 mL of distilled water and 0.7 g of calcium oxide, as described by ^{Vázquez et al. (2014)}. The mixture was heated on an electric grill to boiling and kept under these conditions for 20 min for the CHAL and EC samples, while the BOL samples were held for 30 min. These times were adjusted according to the hardness of the grain. Samples were rested at room temperature for 12-14 h. The nixtamal was rinsed with tap water to remove excess lime and ground in a stone mill. The dough was kneaded and 20 g portions were molded in a manual tortilla and cooked on a metal plate (comal) with a time of 15 s on one side, one minute on the opposite side and 15 s on the start side, for the formation of the “blister”.

Color of the tortilla

It was measured in four freshly made tortillas on the opposite side of the ampoule with a Hunter Lab colorimeter (Mini Scan XE Plus 45/0-L). The CIELab values of L^*, a^* and b^* were obtained. From these two last ones the parameters of Hue (tone angle) and Chroma (color saturation) were calculated, by means of the expressions: Hue = arctg(b^*/a^*), chroma= (a^*2 + b^*2)^1/2 (^{McGuire, 1992}).

Preparation of tortilla samples for analysis

The tortillas, cut into small pieces, were placed in aluminum trays and dehydrated in an oven (Riossa, MX) at 40 °C for 24 h. They were then milled in a 0.5 mm mesh cyclone mill (UDY, USA) to obtain the flour, which was placed in amber vials inside a desiccator and analyzed immediately for anthocyanins, phenols, antioxidant activity, and proximal and mineral composition.

Analysis proximal and minerals

The proximal analysis was performed according to the methods 920.152, 940.26 and 970.20 for protein, ash and ethereal extract, respectively of the ^{AOAC (1998)}. The starch content was determined by the Megazyme^® method (^{McCleary et al., 1994}). The mineral composition of the tortillas was determined by digestion with a diacid mixture (HClO₄-HNO₃) and analyzed by flame emission spectrophotometry (^{Jones and Case, 1990}). All analyzes were performed in duplicate.

Phenolic composition of tortillas

Extraction of soluble phenols. 1 g of ground and dried tortilla were weighed and 20 mL of 1% acidified methanol was added with trifluoroacetic acid (TFA). The mixture was sonicated for 15 min at room temperature and stored at refrigeration temperature (4 °C) for 115 min. Centrifuged for 15 min at 4 000 rpm and filtered on Whatman No. 4 paper. The volume was adjusted to 20 mL with the same extraction solvent. This extract was used for the determinations of total anthocyanins (^{Salinas et al., 2005}), total soluble phenols (^{Singleton and Rossi, 1995}) and antioxidant activity (^{Re et al., 1999}; Soler et al., 2000).

Total anthocyanins (CAT). A standard curve was made with cyanidin-3-glucoside (Extrasynthese, FR). The results were expressed in mg equivalents of cyanidin 3-glucoside/100 g of dry sample (MS). Total soluble phenols (FST). Analysis was performed by the Folin-Ciocalteu assay (^{Singleton and Rossi, 1995}). The content of FST was expressed as a function of gallic acid, for which a standard curve of this acid was elaborated. The results are reported in mg equivalents of gallic acid (EAG) per 100 g of MS.

Antioxidant activity (AA)

DPPH method

The free radical DPPH (1,1-diphenyl-2-picryl-hydrazine) method was used as described by ^{Soler et al. (2000)}. Antioxidant activity was expressed as a percentage of reduced DPPH.

ABTS method

It was determined by the methodology of ABTS (2,2’-azinobis (3-ethylbenzthiazoline-6-sulfonic acid) described by ^{Re et al. (1999)}. The percentage of reduced ABTS was calculated and expressed as equivalent micromoles Trolox (soluble form of α-tocopherol).

Statistical analysis of the data

Data were analyzed under a completely randomized design, with two replicates. A statistical analysis of variance and mean comparison tests between treatments (tortillas of the three races and white-grain tortillas) was performed using the ANOVA statistical procedure, when significance was presented in the model. The analyzes were performed with the statistical package SAS for microcomputer (^{SAS, 2002}).

Color of tortillas

In the Figure 1A shows the hue (pitch angle) and chroma (color saturation index) data of the purple blue grain tortillas of the corn populations under study, in addition to the white corn tortilla. The blue corn tortillas of the BOL race were located between the yellow-green tones, with hue values of 114.9 ±2.2° and 133.7± 1.6°, for collections 89 and 120, respectively. One of the samples of the CHAL race (collection 16, hue= 167.3 ± 1.1°) was also located between these tones.

Figure 1 Color of the blue/purple corn tortillas and the whitegrain corn tortilla in the hue-chroma plane (A and B) and in the hue-luminosity plane (C).  

The other CHAL sample was placed between the green and blue tones, but very close to the green hue, with an hue value of 185.1 ±1.2°. On the other hand, the tortillas of the two collections (4 and 84) of the EC race were located in the quadrant between the blue and red tones, with hue values of 288.8 ±10.8 and 285.2 ± 1.3°, respectively. ^{Del Pozo et al. (2007)} reported hue values of 357.6° and 5.9° in national and American blue corn tortillas, respectively. These values are associated with purple-red (357.6°) and purple-purple (5.9°) tones. As expected, the white corn tortilla ranged from yellow to red, very close to yellow, with an hue value of 82.0 ±0.7.

The chroma values observed in the purple blue grain tortillas were low (between 3 and 5), relative to the value of the white corn tortilla, which presented a value of 25 (Figure 1A). By removing from the graph of Figure 1A, the data of the white corn tortilla, the color shades of the purple blue grain tortillas and the differences for this variable among the tortillas of the three races could be seen more clearly (Figure 1B ).

The brightness is a variable related to the brightness of the sample, varies between 0 for black and 100% for white. In the morello tortillas, this variable was between 37.2 and 50.3%, tortillas of the BOL breed were the ones with the highest values. As expected, the white-grain tortilla presented the highest luminosity, which was 72.6 ±1.5% (Figure 1C). In blue corn tortillas obtained from national and American grain, ^{Del Pozo et al. (2007)} have indicated brightness values of 32 to 40%.

The analysis of variance showed statistical difference (p≤ 0.05) of the color variables among the tortillas of the populations of the CHAL, EC and BOL maize breeds under study. The results of the comparison of means between the tortillas of the three races are presented in Table 1, in which the data of the variables a^* and b^* were also added. The most relevant of this table is that the tortillas of the three races presented a greenish tone, given by the negative values of a^*.

CHAL= Chalqueño, EC= Elotes Cónicos, BOL= Bolita. L^*= luminosidad (%). Los valores después del signo ± corresponden a la desviación estándar de n= 4. Valores seguidos por la misma letra, dentro de cada columna no son estadísticamente distintos (Tukey p≤ 0.05). DMS= diferencia mínima significativa

Table 1 Color tortillas made from corn populations of blue/purple three races grain. 

These results differ from those reported by ^{Del Pozo et al. (2007)} who reported positive values in this variable, indicative of a red dye, in tortillas obtained from blue corn of national and American origin, a similar result was reported by ^{Sánchez et al. (2015)} for blue corn flour obtained by extrusion. The greenish hue observed in the tortillas of this study is possibly due to the combination of the characteristic yellow color of the solubilized pericarp during the nixtamalization process and the purple blue tone that the anthocyanins of the grain acquire when they are in an alkaline environment (^{Salinas et al., 2005}; ^{Torskangerpoll and Andersen, 2005}).

The observed hue differences among tortillas of the three races could be attributed to differences in content and type of anthocyanins in the grain of each particular breed. In this regard, it has been reported that the grain of collections of the BOL race has a lower content of anthocyanins than that of races such as EC and CHAL, in addition to presenting differences in anthocyanin profiles, with absence of peonidin in the collections of the breed BOWL. The chemical structure of anthocyanins influences the color they acquire at alkaline pH, in addition to temperature and concentration (^{Torskangerpoll and Andersen, 2005}), which complicates the precise origin of color differences between tortillas.

Another fact that stands out in the information contained in Table 1 is that only tortillas of the breed EC, presented negative values in b^*, which is related to its blue tone.

Negative values of b^* similar to those observed in tortillas of the EC race have been reported when using as calcium source calcium hydroxide or calcium lactate and high values of L (63.68 to 67.94%) (^{Sánchez et al., 2015}), similar to those of the white corn tortilla.

There are no published works that analyze the effect of color in the purple blue grain tortilla on the acceptability by the consumers, although it is one of the attributes of the appearance of the foods that the consumers most consider when selecting them (^{Jha, 2010}). This is an area that must be addressed in order to generate information on this aspect of the blue/purple grain populations that occur in many of the Mexican races.

Proximal and mineral analysis of tortillas

Statistical differences (p≤ 0.05) are presented for oil, ash and protein between blue/purple grain tortillas and H-40 white grain maize (Table 2). The largest oil content was the tortilla of the CHAL breed. The oil content in the tortilla will be influenced by the grain content and the losses occurring during the nixtamalization process. ^{Martínez et al. (2002)} found losses of 10.3% of oil when transforming in commercial tortilla, by the traditional method. The losses are due to the saponification of the grain oils by the alkali of the nixtamalization (^{Yahuac et al., 2013}).

^†= los valores están expresados en (%) base seca; ^††= expresados en mg 100 g^-1 base seca; DHS= diferencia mínima significativa

Table 2 Chemical and mineral composition in tortillas made from corn populations with blue/purple grain of three races and a white grain control. 

The tortillas of the BOL breed were the ones with the lowest oil content, which was statistically the same as that of the white corn tortillas. This fact is relevant, if one takes into account that one of the traditional uses of this breed is for the elaboration of tlayudas, which are giant tortillas that are consumed in the Oaxacan kitchen and that require to have a shelf life longer than the tortilla common. The values obtained in oil content in the tortillas of the present work are similar to the value of 3.2% dry basis reported by ^{Vázquez et al. (2014)} for white corn tortillas, but are slightly lower than the 3.8% dry basis reported by ^{Hernández et al. (2007)} in blue corn tortilla.

The ash content ranged from 1.4 to 1.8% and did not present statistical differences (p> 0.05) between the tortillas of the three blue/purple grain races, but was higher than that observed in white corn tortillas. An important fraction of the ashes in the corn tortilla is the calcium that is incorporated during the nixtamalization process, which is concentrated in the pericarp and germ of the nixtamalized grain (^{Fernández et al., 2004}). The highest (p≤ 0.05) protein content was presented in BOL tortillas, the lowest presented tortillas of the CHAL and EC breeds.

The higher protein content in the tortillas of the analyzed collections of the BOL race can be attributed to the grain hardness, which according to ^{Salinas et al. (2012)}, is higher than in cereal grains such as EC and CHAL. There is a direct relationship between protein content in corn grain and its hardness (^{Fox and Manley, 2009}). The starch content in the blue tortillas varied from 65.3 to 69.7%, while in the white grain it was 69.7%. The values obtained for this component of the tortilla are slightly lower than the 72.92 ±0.44% reported by ^{Rendón et al. (2002)} in commercial white corn tortilla obtained under the traditional nixtamalization method.

In general, the values observed in the nutritional composition of blue/purple grain tortillas are similar to those reported by ^{Hernández et al. (2007)} for blue tortilla. Calcium content ranged from 130 ±20 to 170 ±10 mg 100 g^-1 dry basis (BS) and was statistically different (p≤ 0.05) between blue/purple grain tortillas and white corn tortilla, which presented the highest value, while the minor had the tortilla CHAL. Nixtamalization significantly increases the calcium content, relative to grain (7.7 mg 100 g^-1 MS in grain at 114 mg 100 g^-1 MS in tortilla) (^{Figueroa et al., 2001}).

During this process, the calcium is fixed differentially in the grain structures, in the pericarp> germen> endosperm order (^{González et al., 2005}). However, much of the pericarp is lost in the cooking and washing water and with it, calcium, so it is possible that maize that retain more pericarp may contain more calcium in their tortillas. In the nixtamalization process, factors such as amount of alkali, cooking time and rest, and rinse intensity affect the tortilla’s calcium content (^{Bressani et al., 2004}), making it difficult to objectively compare results with those of other authors.

The values obtained in blue / purple and white grain corn tortillas are lower than the value of 204.9 ±22.9 mg (%), reported for white corn tortillas by ^{Bressani et al. (2004)}, but similar to the 114.0 mg (%) reported by ^{Figueroa et al. (2001)} for a white-grain tortilla sample.

The iron content did not present statistical difference (p> 0.05) between the analyzed tortillas, while zinc showed differences (p≤ 0.05). The highest zinc content was the tortillas of the CHAL breed, the lowest of the EC breeds and the tortillas of the white corn.The iron values in the blue and white tortillas of this study are greater than the 1.55 mg reported by ^{Bressani et al. (2004)} for white corn tortilla.

The values of zinc obtained in the blue and white tortillas are similar to those indicated by ^{Figueroa et al. (2001)} for white corn tortillas. Although it has been reported a higher content of iron in blue-grain maize, in relation to white or yellow grain maize (^{Bodi et al., 2008}) in the blue-purple tortillas analyzed, the content of this mineral was equal (p> 0.05) than that of white-grain tortillas.

Since the iron content in the corn tortilla is not affected by the conditions of the nixtamalization process (^{Bressani et al., 2004}), the differences can be attributed to the environmental effect and the genetics of maize. The concentration of micronutrients (Fe and Zn) in maize grain is strongly influenced by the production environment, particularly soil type (^{Field et al., 2005}). Since this information is not available for the maize populations analyzed, it is difficult to determine the cause of the differences observed for the Zn content. An objective comparison of the effect of genetics on micronutrient content in corn tortilla requires grain of cultivars grown in the same environment. In the case of native populations, this requirement is made difficult by the limited adaptation they have.

Total soluble phenols (FST), anthocyanins and antioxidant activity in tortilla

According to Table 3, FST in blue/purple grain tortillas ranged from 68.8 ±4.27 to 82.78 ±0.93 mg gallic acid equivalent (EAG) 100 g^-1 dry sample (MS), while in the omelet of white maize the value of this variable was 59.32 ±2.53 mg EAG 100 g^-1 MS. These values are superior to those reported by De la Parra et al. (2007) for blue-grain corn tortilla (39.1 ±1.5 mg EAG 100 g^-1 MS) and white grain (47.2 ±1.8 mg EAG 100 g^-1 MS). Among the main phenolic compounds identified in the FST fraction are ferulic and p-coumaric acids (De la Parra et al., 2007), as well as cyanidin 3-glucoside (^{Salinas et al., 2003}).

DMS= diferencia mínima significativa; FST= fenoles solubles totales (mg equivalentes de ácido gálico 100 g^-1 de muestra seca); CAT= contenido de antocianinas totales (mg equivalentes de cianidina 3-glucósido 100 g^-1 de muestra seca); DPPH= porcentaje de DPPH reducido. ABTS= micromoles equivalentes de Trolox g^-1 de muestra seca.

Table 3 Total soluble phenol content, total anthocyanins and antioxidant activity in tortillas of blue/purple grain maize populations and a white-grain maize. 

The highest numerical value was presented in tortillas of the race EC, while the smaller one had the tortillas of the race BOL. Due to the high variability in FST content between the two populations of the CHAL breed, no significant differences were found for this variable among tortillas of the three races. The lowest value of FST was presented in tortillas of white corn.

Among the different cereal products to which the Mexican population has access, are bread and flour tortilla, as well as corn tortilla. The purple-blue corn tortilla has on average 50% more FST than white bread (37 mg ferulic acid equivalents 100 g^-1 MS, (^{Menga et al., 2010}) and that the wheat flour tortilla (16.88 ±0.3 mg ferulic acid equivalents 100 g^-1 MS, (^{Anton et al., 2008}). The CAT of the tortillas of the CHAL and EC breeds was higher than those of BOL, whereas the content of the white grain maize (H-40) was marginal. The CAT in the blue/purple grain corn tortilla is influenced by the anthocyanin content in the grain. The grain of CHAL and EC populations has higher anthocyanin content than BOL (^{Salinas et al., 2012}), and this pattern was conserved in tortillas obtained under the traditional nixtamalization method.

According to the results obtained with the DPPH method, the antioxidant capacity of the purple blue grain tortillas was in the order CHAL> EC> BOL> H-40; with the ABTS method, there was no difference in CA between tortillas of the CHAL, EC and BOL races, but between tortillas and white corn tortillas. The differences in the results of the two methods can be attributed to the differential sensitivity of the phenolics present in the extract to the free radicals of each method (^{Alam et al., 2013}).

The highest AA in blue/purple grain tortillas in relation to white grain, observed under the two methods used is attributed to anthocyanins, since phenolic acids are common in tortillas of both grain colors. During the nixtamalization a large quantity of the anthocyanins of the grain is lost by the action of the alkali and the high temperature (^{Salinas et al., 2003}; De la Parra et al., 2007). The predominant anthocyanin in the tortilla is cyanidin 3-glucoside, since the predominant acylated type (cyanidin 3- (6’-malonylglucoside and cyanidin 3- (3’6 ‘dimalonyl glucoside) lose their acyl radical which is esterified to sugar and converted to cyanidin 3-glucoside.