Genetic material

A corn compound of the Elotes Occidentales race was used, obtained from collections made in western Mexico in the states of Jalisco, Michoacán and Zacatecas. A 2 kg sample of the compound from the original population (OP) was divided into three subgroups by visual appreciation of the color intensity of the grain in clear (C), intermediate (I) and dark (O).

Location of the experiment

The crop to fix the color in grain, was established in the experimental field of the University Center of Biological and Agricultural Sciences of the University of Guadalajara that is located in Zapopan, Jalisco, Mexico. Its geographic coordinates are 20° 43’ North latitude and -103° 23’ West longitude at an altitude of 1,650 meters above sea level. With average annual temperature of 18 °C and precipitation of 950 mm. The soils are acid pH and sandy loam texture. In the cycles of S/S 2013 in open field under seasonal conditions, in autumn-winter (A-W) 2013-2014, O-I 2014-2015, spring-summer (S-S) 2015 and A-W 2016-2017, in greenhouse with irrigation. In the establishment of the crop, the genotypes are not evaluated through the environments, only the color was fixed.

Agronomic management

The first crop cycle was established in plots of two rows at a distance of 0.75 m between rows and 5 m in length, sowing with a planting density of 45 000 plants ha^-1. The formula 18 46 00 (N-P-K) was applied at a dose of 200 kg ha^-1 at sowing, in a first weeding, urea was applied with the formula 46-00-00 in a dose of 30 kg ha^-1. In the following greenhouse growing cycles, management was similar among them. They were established in plots of two rows of 20 m in length and 0.75 m between rows, 10 seeds of each family were planted (family by cobs), considering 20 families by intensity of coloration, directly in soil at a distance of 0.2 m between plants. The fertilizer was incorporated in the irrigation, making an application with the formula 12-61-00, the second application was with the formulas 12-00-46 and 11-00-00-23, a third application was made with the formula 32 -04-00. For the control of pests, weeds and diseases, incidents were monitored and, if applicable, chemical control was applied.

Method of obtaining the final population (FP)

From each cycle, 20 families were selected by visual appreciation by color intensity C, I and O. The separated seed was sown in the successor cycle without using a design. The first three cycles were performed self-pollinations (S₁, S₂ and S₃). In the next two cycles, fraternal crosses were made to obtain grain with uniform color intensity for each subpopulation (C, I and O) of the FP. A sample of 1 kg of grain from each subpopulation was used to determine the quality of the grain in terms of physical characteristics, technological variables for the production of pozole and nutraceutical properties, comparing them with subpopulations of the OP (Figure 1).

Figure 1 Classification of the populations of origin (OP) and final (FP) according to the intensity of coloration.  

Response variables

Physical characteristics

The color was measured with a Hunter-Lab team in the CIELab scale, in the raw grain (Gcr) and unprocessed grain modalities and in the processed grain for pozole that was denoted only as cooked grain (Gcd), which was dehydrated in a stove for 24 h at 40 °C. The values of the luminosity readings (Lum, %), a^* and b^*, were obtained, from these last two parameters the hue angle variables “hue” (h°) and color saturation index ‘chroma’ were calculated, by means of the expressions: h°= arctg (b^*/a^*), chroma= (a^*2 + b^*2)1/2 (^{McGuire, 1992}).

In the Gcr the variables were determined: length (L), width (A) and thickness (G) in a sample of 10 grains (mm), grain humidity (HG, %), hectolitre weight (PH, kg hL^-1), weight of one hundred grains (PCG, g). The hardness was determined indirectly by the flotation index (IF, %). Pericarp thickness (EP, μm) was measured in a block of five layers of pericarp from the opposite side to the location of the germ with a micrometer. The variables were performed in duplicate according to what was described by Salinas and Vázquez (2006).

Technological variables for the production of pozole

100 g of each of the samples were considered by color intensity (C, I and O) of both populations (OP and FP). For the preparation of pozole and the tests of nixtamalization and cooking for flowering of the grain: total solids in nejayote (STN) and in broth of cooking (STC) in g 100 g^-1, humidity of the grain nixtamalizado (HGN) and cooked (HGC) in g 100 g^-1 by the stove method; cooking time for grain flowering (TCF) in min; volume of expansion (VE) in mL and (%) of flowered grains (GF) were performed according to what was described by ^{Bonifacio et al. (2005)}. For the variable yield of grain to pozole (RGP) was calculated according to ^{Vázquez and Santiago (2013)} and was expressed as kg of product obtained per kg of processed corn.

Nutraceutical variables

Obtaining the extract

For the extraction, a sample of 25 grains was used from each of the subpopulations of the OP and FP in the Gcr and Gcd modalities. In both cases the grain was crushed with the help of a hammer and in the particular case of the germ, it was cut into small pieces with the help of a scalpel. The sample was degreased with petrol for eight hours in a Soxhlet equipment. It was placed in an oven at 40 °C to remove the solvent and milled in a cyclonic type mill (UDY, Tecator, City, State) with 0.5 mm mesh. For the extraction of anthocyanins 1 g of flour was considered, to which was added 20 mL of solvent (methanol acidified to 1% with trifluoroacetic acid, TFA). The samples were sonicated for 15 min, then cooled for 105 min. They were centrifuged at 10 000 rpm (11200 G) for 9 min in a centrifuge brand Hettich zentrifugen, model Universal 32, to recover supernatant and eliminate turbidity. The extract was filtered in Whatman No. 4 and its volume was measured. From this extract, the CAT, FST and CA were determined in duplicate.

Content of total anthocyanins (CAT)

The absorbance of the extract was read on a Perkin Elmer spectrophotometer (Lambda 25 UV/Vis) at 520 nm. A standard curve of chlorinated pelargonidin (Sigma Aldrich, MN) was developed to express the CAT content based on this compound, given that the predominant anthocyanins in the grain color analyzed are pelargonidin derivatives (^{Salinas et al., 1999}; ^{Paulsmeyer et al., 2017}). The results were expressed in mg equivalents of chlorinated pelargonidin (EPC) in 100 g of dry sample (MS).

Total soluble phenols (FST)

The Folin Ciocalteau method was used as described by ^{Singleton and Rossi (1965)}. Briefly, a 100 μL aliquot was contacted with Folin^® reagent (125 μL), 19% sodium carbonate (w/v) (1 250 μL) and distilled water (1525 μL). It was stirred and allowed to react for 75 min in the dark. It was centrifuged for 10 min at 4 000 rpm, recovering the supernatant and its absorbance was read at 760 nm. A ferulic acid standard curve was made (Sigma-Aldrich, St. Louis, MO, USA) to express the results based on this acid in mg equivalents of ferulic acid (EAF) per 100 g of MS.

Antioxidant capacity (CA) ABTS method

It was evaluated by the ABTS technique (2,2-azino-bis-3-ethylbenzotriazolino-6-sulfonic acid) as described by ^{Re et al. (1999)}. A reading at one minute and a second was considered 30 min after the reaction started at an absorbance of 734 nm. The percentage of reduced ABTS was calculated. The results were expressed in equivalent micromoles of Trolox (μmoles ET) (soluble form of α-tocopherol) per gram of MS (^{Salinas et al., 2017}).

Statistical analysis of information

The data of the quality variables were analyzed for the single-color intensity factor in three levels, by means of a completely random design with two repetitions. The analysis of variance was performed, the comparison of Tukey means (p≤ 0.05) when there was significance in the model and a correlation analysis between nutraceutical variables with the statistical package SAS, V.9.1 (^{SAS, 2004}).

Physical characteristics

The variables associated with the color of the grain, luminosity (Lum), hue angle or hue (h°) and saturation index of color or chroma, presented high significance (p≤ 0.01), so that there is variability in the color of the grain between the subpopulations of the OP and FP.

The light red grains (C and I) have a brighter appearance than those of blue grain (O) (^{Salinas et al., 2012}). Regarding the values of hue (h°) that describes the color tone, it is of the parameters that more association keeps with the visual appreciation and the values of chroma are related to the index of saturation of the color, with high values in colors cigars and low values in matt or less pure colors (^{Jha, 2010}).

For the variable h° it was observed that, in all the color subgroups, the values of the OP were higher than those of the FP, which means that the orange red tone of the OP changed to a red-purple tone in the FP. In the case of subpopulations C and I, the tendency was towards red-orange tones and in the subpopulations O towards purple tonalities. What coincides with that reported by ^{Salinas et al. (2013)}, where it separates the tone values between 0° and 25° which correspond to the purple-red color. The chroma values of subpopulations C and I of the OP were lower with respect to the corresponding ones in the FP, which indicates a higher color saturation index.

Regarding the color determined in Gcd, both populations (OP and FP) showed Lum loss of up to 62% (in subpopulation I of the OP) in relation to the Gcr. With the exception of subpopulation I of the FP, the Gcd showed greater luminosity (4.15%) with respect to the values obtained in the Gcr. ^{Vázquez and Santiago (2013)}, found that the nixtamal wash was brighter, this could be the reason why the Gcd of this subpopulation (I of the PF) has greater luminosity. The values of h° in Gcd that were presented in the subpopulations (C= 43.59, I= 54.64 and O= 63.66) of the FP were higher than those presented in the OP presenting an orange red tone.

The change of tone during nixtamalization is due to the fact that the alkali hydrolyze a large part of the pericarp, therefore the determination of the color in Gcd is carried out in the aleurone layer that acquires a yellowish hue (^{Bonifacio et al., 2005}). The GCD of the FP showed the highest chroma values in intensities C and I (18.47 and 18.28), related to reddish tones and presented a purer color, with respect to the OP. In the case of subpopulation O for both populations (OP and FP) in Gcd, they presented low chroma values (3.21 and 7.86, respectively), which indicates a matte color with lower purity.

Regarding the size of the grain, it was higher in the subpopulations of the OP that had higher L and A, since the subpopulation O had the highest value of L (12.72 mm), then subpopulations C and I and in the FP the subpopulation C had the lowest value of L (10.28 mm). While, for grain A, subpopulation I of the OP, presented the widest grains (14.17 mm) contrary to the FP than in subpopulation I, presented the narrowest grains (11.67 mm). However, for the thickness (G) they did not show significance, which shows that by means of the color fixation technique, the grain size of the FP that had shorter and narrower kernels compared to the OP was reduced and since it did not present variation in grain G, it is considered that this characteristic was not modified.

The average of the PCG was higher in the OP (63.24 g) over the FP (47.03 g). Considering the GWP as an indicator of grain size (>38 g) (^{Salinas et al., 2010}), the grains of this corn race are large. The subpopulation O of the OP was the one with the highest PCG (64.25 g). What coincides with ^{Bonifacio et al. (2005)} who reported for maize of the Cacahuacintle breed with a type of mealy endosperm, a PCG of 60 to 70 g, which compared with values of ‘hard’ grains (from 18 to 22 g in Palomero Toluqueño and Arrocillo, respectively) and coinciding with the values of L and A, ‘large’ size grains are considered, a desirable characteristic for their use in the production of pozole (^{Bonifacio et al., 2005}; ^{Vázquez and Santiago, 2013}; ^{Hernández et al., 2014}).

The FP values were lower with respect to the OP; subpopulations I and O of the FP (43.05 and 43.02 g, respectively), were the grains with the lowest PCG. The PCG suggests the size of the grain, and for the process of nixtamalization the grains of greater weight are more favorable (^{Salinas et al., 2013}). Grain size is a variable of interest in the nixtamalization process, due to its impact on cooking and water absorption processing. ^{Bonifacio et al. (2005)}, mention that the difference in grain weight is really due to the size of the grain and not to its density. The averages of the hectolitre weight (PH) were lower in the OP (67.75 kg hL^-1) than in the FP (70.05 kg hL^-1). the PH of 60, 59 and 58 kg hL^-1 was reported for the Cacahuacintle breed (^{Bonifacio et al., 2005}; ^{Vázquez and Santiago, 2013}; ^{Hernández et al., 2014}).

These values of PH indicate the amount of dry matter of the grain that there is in a certain volume and it is considered as an indirect measure of the hardness of the grain of corn so low values are characteristic of very soft grains (^{Vázquez and Santiago, 2013}). The HG was considered for the effect it has on the physical characteristics of the grain (^{Hernández et al., 2014}), the OP presented lower humidity than the FP with 10.6 and 11.27%, respectively. The flotation index (FI) showed little variation on the 6 subpopulations of both populations, subpopulation I of the FP presented the highest value of 100%, in the same population the subpopulation O presented the lowest value (92.67%), which indicates that the grains have the indispensable characteristic for the flowering of the grain that the pozole requires (Vázquez and Santiago, 2013).

In the pigmented maize the soft texture dominates, since they have a large proportion of floury endosperm. ^{Wellhausen, et al. (1951)}, found that in the selection by intensity of the color tends to maintain a pure race with respect to the gene of the floury endosperm. In floury endosperm grains, a pericarp thickness (PE) of 91.5 μm has been reported in the Bolita breed (^{Salinas-Moreno et al., 2003}) coinciding with the values obtained in light grains (C) of the OP and in intermediates (I) of the FP that presented values greater than 80 μm. In the meantime, for maize classified according to their final use as their own for the production of pozole, an average of 43.6 μm is reported (^{Narváez-González et al., 2007}), they have a pericarp content lower than the average of the normal creole corn, therefore, the maize with smaller thickness of pericarp is considered the most suitable for this purpose (^{Vázquez et al., 2014}), subpopulation C of the FP presented an EP of 48.6 μm, being the one that most resembles the one described previously.

Technological variables for the production of pozole

The total solid variables of the nejayote (STN), cooking time for flowering (TCF) and flowered grains (GF) showed a difference of 5% between the two populations. While the rest of the technological variables did not show significance, which indicates the little variation among populations for the process in the preparation of the pozole. The average STN content was lower in the OP (6.65 g 100 g^-1) than in the FP (7.46 g 100 g^-1), with subpopulation I of the OP having the lowest content (5.58 g 100 g^-1) and the subpopulation O of the FP, the highest content (8.42 g 100 g^-1). Contents of 4.9 g 100 g^-1 have been reported for Cacahuacintle corn. However, the values reach 8 to 12% when the grain is processed commercially (Pflugfelder et al., 1998), values similar to those presented in this study.

In the STC, there was an average content of 1.82 and 1.81 g 100 g^-1, in the OP and FP, respectively. In other works, 2.3 to 3% are reported (^{Hernandez et al., 2014}) and from 1.78 to 3.73%. In the case of the pozole cooking broth the pericarp is no longer present, so solids could be formed by solubilized proteins, starch and fragments of burst grains (^{Bonifacio et al., 2005}).

In the HGN, both subpopulation groups presented an average of 42.02 g 100 g^-1 in the OP and 41.61 g 100 g^-1 in the FP, so they showed little variation in water absorption during the process nixtamalization. For the HGC they presented an average of 56.89 and 59.5 g 100 g^-1 in the OP and FP, respectively. ^{Vázquez and Santiago (2013)}, indicate a positive correlation between grain moisture and the VE variables and the RGP, due to the resting time of the nixtamal.

The TCF of the grain was from 53 to 80 min in both populations, times well below that reported in previous studies, which indicate times of 103 to 218 min in maize grain Cacahuacintle (^{Bonifacio et al., 2005}; ^{Vázquez and Santiago, 2013}; ^{Hernández et al., 2014}). As for the reduction of the grain size in the FP, it could be considered that the TCF of the grain would increase, because ^{Bonifacio et al. (2005)}, confirm that rounded and globose grains are those that require less cooking time for trapping. What happened in subpopulation C (75 min), but in subpopulations I and O had shorter TCF with respect to the subpopulations of the OP (Table 2), taking into account that the reduction in grain size in the PF, therefore, was little the difference of the times reported in subpopulations of the OP.

In the yield of grain to pozole (RGP) both populations (OP and PF) presented very similar values on average (1.33 kg kg^-1 and 1.4 kg kg^-1, respectively), which indicates that despite the variations in to the PCG and the grain size of both populations, presented similar yields. The subpopulation O of the PO presented the highest value with 1.48 kg kg^-1, since it was the one that presented the grain with the highest L and PCG, while the subpopulations C and I presented the lowest RGP (1.26 kg kg^-1). In general, the FP presented higher performance on average coinciding with the higher HGC. ^{Vázquez and Santiago (2013)}, reported a yield of 1.5 kg of nixtamal kg^-1 of corn by the traditional method for the preparation of pozole in accordance with its moisture content.

Regarding the volume of expansion (VE) and percentage of flowered grains (GF), there was a difference between populations, since the FP presented the highest values above the OP with an average difference of 10.22 mL for the VE and 7.97% for the GF, presenting a proportional relation regarding the increase of the VE with respect to the greater percentage of GF. With the exception of the subpopulation O of the OP that had a VE of 58.67 mL (the highest value), surpassing that of the subpopulations of the FP (C= 53.33, I= 46.67 and O 50.67 mL) (^{Bonifacio et al., 2005}; ^{Narváez-González et al., 2007}). However, the percentage of GF of the subpopulation O of the OP was lower with respect to the VE, since the subpopulations of the FP had higher percentages of GF (Figure 2).

Figure 2 Technological variables expansion volume (VE) and flowered grain (GF) in the subpopulations of intensity C, I and O of the original (OP) and final (FP) populations.  

Nutraceutical variables

There was a difference in the CAT and CA, between the grain of corn with different intensity of coloration. In this study the Gcr of the subpopulation O of both populations, presented the highest average CAT (22.75 mg EPC 100 g-1 MS) followed by the subpopulations of intensity I and finally the subpopulations of intensity C (17.76 and 12.45 mg EPC 100 g^-1 MS, respectively).

Whereas, for the Gcd, the subpopulations of intensity I presented the highest average CAT (6.37 mg EPC 100 g^-1 MS), that of the subpopulations of intensity O and those of intensity C (4.95 and 4.01 mg EPC 100g^-1 MS). They presented the same behavior pattern for CAT both populations for Gcr (in order O, I and C) and Gcd (I, O and C).

In general, the OP had more CAT than the FP, both for the Gcr and the Gcd (Figure 3-A).

Figure 3 A) Total content of anthocyanins (CAT); B) total soluble phenols (FST); and C) antioxidant capacity (CA) in grain of the original (OP) and final (FP) populations in intensity C, I and O, in unprocessed (Cr) and cooked (Cd).  

The difference in CAT values can be attributed to the proportion of different aglycones (anthocyanidins), but also to the chemical structure in particular of each anthocyanin (^{Salinas et al.2012}). The size of the grain affects the concentration of anthocyanins, it influences by an effect of dilution of the anthocyanins in the grain flour used for the quantification (^{Salinas-Moreno et al., 2012}), so it is suggested that the lower CAT is due to the smaller size of the grain more than by effects of dilution in subpopulations of the FP.

Regarding the CAT of Gcr compared with Gcd, it has a relation with respect to the TCF, since through the process of elaboration of the pozole in the nixtamalization stage, a first loss of anthocyanins occurs due to the combination of alkalis and cooking, by effects of pH and temperature increase (^{Salinas-Moreno et al., 2012}).

In the extraction of anthocyanins other phenolic compounds and other flavonoids are also extracted (^{Salinas-Moreno et al., 2012}) so that the content of FST is greater than the CAT, the anthocyanins are the predominant phenols in the extract, therefore, they are the main responsible for the CA (^{Yang and Zhai, 2011}). The highest content of FST (136.3 mg EAF g^-1 MS) was presented by subpopulation I of the FP in Gcr; however, in subpopulations of intensity C and O, the FST contents were lower (108.8 and 119.6 mg EAF g^-1 MS, respectively) with respect to the subpopulations of the OP at the same intensities.

The CA is a reflection of the variables of phenolic composition (CAT and FST), a greater CA was present in grains of the OP than in Gcr, had a difference of 1.69 μmol ET g^-1 MS with respect to the PF in the subpopulations O and 1.91 μmol ET g^-1 MS in subpopulations C. While for subpopulations I, there was no greater variation with only a difference of 0.01 μmol ET g^-1 MS between populations. There are no significant differences of the Gcd of CA in the three subpopulations of the OP with respect to the FP. However, the subpopulations of the OP far outnumbered the FP (Figure 3-C).

A higher CAT is associated with higher CA (^{Salinas et al., 2012}) however, the type of pigment and differences in the pigment content are related to the antioxidant capacity in the corn grain, which are directly due to the pigments (^{Rodríguez et al., 2013}). Therefore, in the Elotes Occidentales maize population of this study, it showed variation in the CAT and in the CA, due to the intensity of color, which presented the same order in the content of CAT and CA in the subpopulations of the OP, being in Gcr O> I> C and in Gcd I> O> C.

In general, the OP presented higher CAT, content of FST and CA than FP, with the exception of subpopulation I of the FP in Gcr, which presented higher content of FST. The CA is not due only to the CAT, but to the content of FST. Therefore, there is a highly significant and positive correlation between the nutraceutical variables of CAT, FST and CA (p< 1%). Since in the extraction of anthocyanins phenolic acids and other flavonoids are extracted together, the value of FST is higher than that of CAT. However, anthocyanins are the predominant phenols in the extract, therefore, as they increase, the FST increases, which are the main responsible for CA (^{Salinas-Moreno et al., 2012}).

What explains the positive correlation (r= 0.96; p< 0.01) between the content of FST and CA, between CAT and CA (r= 0.91; p< 0.01) and between CAT and FST (r= 0.83; p< 0.01) (Figure 4).

Figure 4 Dispersion graphs of the correlations between the nutraceutical variables of the unprocessed and cooked grain, in the different color intensities of the original and final populations.  

Finally, by means of the method used of autobrothers it was possible to standardize the intensity of the grain in the different subpopulations (C, I and O) of the FP, mainly in the I, which presented color in pericarp and aleurone. However, there was no gain due to the additivity effect of the method on the physical and nutraceutical variables (it did not show higher CA), so the OP was not exceeded. It is considered that the method is not adequate for that purpose. As reported by ^{Rodriguez et al. (2013)}, in a single cycle of visual selection of the color of the grain, managed to increase the content of anthocyanins, but not in the subsequent ones, they explain that apparently the variability of the pigment content was exhausted in a single cycle and it was not possible to improve plus. In the case of ^{Landi et al. (2008)}, showed that recurrent selection for cob color among complete siblings found that a single selection cycle led to the fixation of the selected alleles.

It would be advisable to adapt it with another type of method in which the subpopulations obtained with an external genetic source could be included and obtain gain by dominance effects through hybridization. Authors such as ^{Lago et al. (2014)} report that through backcrosses between dominant homozygous inbred line as an anthocyanin donor and a variety of corn as a recurrent parent, they obtained lines that accumulated high levels of anthocyanins in the pericarp. And ^{Espinosa et al. (2009)}, who through direct diallel in direct and reciprocal crosses, found a significant effect in the direction of the cross over the content of anthocyanins in aleurone and pericarp.