Field planting

Field planting was carried out in the Campo Experimental Valle de México of the Instituto Nacional de Investigaciones Forestales, Agrícolas y Pecuarias (CEVAMEX, INIFAP) located in Chapingo, México (19° 29’ N and 98° 53’ W), altitude 2240 m. The climate is temperate with 643 mm of rainfall in summer and average annual temperature of 15.1 °C, and the soil is sand clay.

Lines and F1 crosses were sown on April 27, 2012, and both crops werere managed as described below.

The experimental design was complete randomized blocks with three replications. The genetic material was planted in plots comprising two furrows measuring 5.0x0.80 m, with 26 plants in each row 0.20 m apart. Plant density was 65 thousand plants ha^-1 and the dosage of fertilizer (kg ha^-1) was N (120), P (60), K (30). Auxiliary irrigation was applied 40, 90 and 120 d after planting (DAS).

Around physiological maturity of the crop (150 DAS), ten plants with complete competition per replication were selected at random. Ears were harvested and shelled manually; the kernels of the replications were mixed and two 100 g samples were taken to be used in the chemical determinations. The kernels, with 10 % moisture, were stored until analysis.

Genetic material

The genetic material was of the following types.

1) Six inbred lines of opaque2 (o2) maize, with white grain and normal textured endosperm (M1 to M6), derived from crossing lines CML 244 (S4), CML 352 (S5), CML 242 (S4), CML 246 (S3), CML 349 (S5), and CML 354 (S3) with the CIMMYT line CML 176 [(P63-12-2-1/P67-5-1-1)-1-2-B-B], which was the donor of the o2 gene. They were then backcrossed toward normal material and self-pollinated. Massive increases were then done to recombine and stabilize them (Table 1).

Table 1 Quality protein maize lines used in the study (^{CEVAMEX, 2011}).  

° = Self-pollinations after backcrosses.

2) The 15 single direct crosses formed with the above lines (S1, S2 or S3 plants, depending on their self-pollinations after backcrossing, with F₁ embryos in the crop field).

3) The F₂ generation of the above crosses (F₂ plants with F₂ embryos, in the crop field). H70, a commercial hybrid with normal protein and endosperm, was included as the regional control.

The genotypes studied are adapted to prosper in the Central Highland Plateau of Mexico (with altitudes of 2200 to 2600 m), and H70 predominates in this region (^{Arellano et al., 2011}), the reason it was chosen as the regional control.

Determination of tryptophan, lysine and total protein contents and quality protein index

Chemical determinations were done during 2013 in the Molecular Biotechnology Laboratory of the Biotechnology Interdisciplinary Professional Unit of the Instituto Politécnico Nacional (UPIBI, IPN) in Mexico City. The lysine (Lys), tryptophan (Trp) and total protein (TP) contents and quality protein index (ICP) were quantified for each genotype (six lines, 15 F₁ crosses and their respective F₂), and the two replications of 100 grains mentioned above were used.

The grains were ground and defatted with hexane for 6 h. One hundred milligrams of this flour was used for spectrophotometric determination (Zuzi 4201/20) of the amino acids (^{Galicia et al. 2009}). For Lys, the protocol was based on the reaction of 2-chloro, 3, 5-dinitropyridine (390 nm) and for Trp it was based on that of glyoxylic acid (560 nm).

Total protein was calculated by multiplying the content of total nitrogen, determined following ^{Galicia et al. (2009)}, by the factor 6.25 (^{Salinas and Vázquez, 2006}). The quality protein index (ICP) was calculated with the formula ICP = [tryptophan (%) x 100] [protein (%)]^-1 (^{Galicia et al. 2009}). The results were expressed as percentages, dry weight base, and were averages of the two replications.

Variation in quality protein between generations F ₁ and F ₂

To estimate the stability of quality protein variables between generations F₁ and F₂, the differences (%) in the values between the two generations were calculated: F₁-F₂ = (F₁-F₂) (100) (F₁)^-1; where F₁ = value of the single cross and F₂ = value of the corresponding segregating cross multiplied by (-1) to indicate reduction of the variable (adapted from ^{Hernández-Leal et al., 2013}).

Statistical analysis

With the data, an ANOVA was performed. Lysine, Trp and protein contents, and the ICP, between and within genotypes, were compared using the Tukey test (p≤0.05). To evaluate quality protein stability of each cross between their F₁ and F₂ generations (inbreeding depression), the t test for paired data was used (p≤0.05). All the statistical analyses were carried out with SAS software (^{SAS Institute, ver. 9.2}).

Chemical quantifications

Differences (p≤0.01) were found in Trp, Lys, and ICP for F₁ and F₂ lines and crosses and the regional control (H70) (Table 2). Thus, the genotypes expressed significant variation in the quantified characteristics of protein quality and, therefore, the means obtained in each case were compared (Table 3).

Table 2 Mean squares and statistical significance of the chemical traits evaluated in maize inbred lines, F₁ and F₂ crosses, and regional control. 

**p ≤ 0.01; CV = Coefficient of variation.

Table 3 Averages (% dry base) of tryptophan, lysine and protein contents, and quality protein index in maize inbred lines. 

Means with different letters in a column are statistically different (Tukey; p ≤0.05).

Diversity in expression of the evaluated characteristics may be because the genealogy of the six parent lines was different, although all of them shared line CML 176, the donor of the o2 gene. That is, the lines conserved their individuality and high level of inbreeding since only the o2 gene was incorporated and later backcrossed toward their respective unmodified, or normal, material (Table 1).

The lines (Table 3) and the F₁ crosses (Table 4) significantly surpassed the regional control (H70) in all of the measurements of protein quality evaluated, especially in essential amino acids and ICP, which are priority chemical traits for quality protein. Therefore, it may be pointed out that the evaluated materials showed large potential for producing hybrids with high protein quality.

Table 4 Averages (% dry base) of tryptophan, lysine and protein contents, and quality protein index in F₁ and F₂ generations of maize crosses. 

Means with different letters in a column are statistically different (Tukey; p≤0.05).

Lines M2 and M3 had the highest contents (p≤0.05) of Trp and PT and the highest ICP, followed by line M6, while M1, M4 and M5 had the lowest performance (Table 3). That is, there was genetic variation among the lines for Trp, PT and ICP, but not for LIS since the content of this amino acid did not vary among lines nor between F₁ and F₂ crosses. It was thus the experimental variable with the best response in all the study.

The F₁ crosses with the highest, most consistent values over measurements were obtained by crossing the lines situated in the first two significant levels (M2, M3, M4 and M6) (Table 3), but the crosses between less prominent lines generated hybrids with fewer quality protein attributes. In both cases, with hybridization, the intrinsic protein quality of the lines was conserved.

The crosses M2 x M5 and M5 x M6 had intermediate, but consistent, performance in generations F₁ and F₂ (Table 4). For this reason, they are a breeding option for formation of QPM, that is, hybrids with fewer, but more stable, quality protein attributes in F₂, which is the generation used for food and that in which expression of the beneficial maize quality protein is sought. The lines described above as superior (M2, M3 and M6) were parents of seven outstanding crosses for PT (p≤0.05), of which only three stood out also in Trp, Lys and ICP. Thus, the crosses consistently superior in all of the variables of the study were M2 x M4, M2 x M6 and M3 x M6 (Table 4).

The crosses mentioned as more favorable for PT than for preponderant factors of high quality protein, such as the essential amino acids and ICP, does not suggest that the proteins of the experimental materials had a good balance of amino acids. This argument coincides with the absence of correlation between protein content and quality protein index, which was reported by ^{Gutiérrez et al. (2014)}.

Variation in protein quality between generations F ₁ and F ₂

Examination of quality protein of the crosses passing from generation F₁ (controlled pollination) to F₂ (random pollination) showed that inbreeding depression was not significant in 70 % of the cases (Table 5). This indicated that, despite random recombination, most of the crosses, in generations F₁ and F₂, maintained similar contents of the studied variables, mainly Lys and ICP, followed by Trp and finally PT (Table 5). Thus, PT was the chemical component that the crosses maintained with greatest difficulty in the food grain generation (F₂).

Table 5 Inbreeding depression (%) of the chemical traits evaluated between the F₁ and F₂ generations. 

°=Student test of paired data. In columns: *p ≤ 0.05; *p ≤ 0.01; ns=Not significant.

The stability of the crosses between the F₁ and F₂ generations could also be explained by the high level of inbreeding of the parent lines with respect to o2 (as mentioned above), so that recombination of F₂ occurred among alleles of the same gene and, consequently, in most of the crosses there was no inbreeding depression in the variables that are determined by the o2 gene, which are Lys, Trp and ICP. For ICP, determination is indirect since it is the proportion of the quantity of Trp present in the grain proteins.

Another outstanding aspect of Lys is that its F₁ - F₂ differences were positive in 73 % of the crosses (Table 5), meaning that its content increased in the F₂ generation and thus there was no inbreeding depression, which did occur with Trp, PT and ICP.

The increases in Lys in F₂ would be attributable to the recombination between homozygous lines for o2, with high amino acid content, with which the F₂ crosses were obtained. They were also outstanding in the same variable and random pollination favored their Lys accumulation.

The crosses M2 x M4, M2 x M6 and M3 x M6 mentioned above as outstanding for their high significant levels of Trp, Lys, TP and ICP had reduced responses because of inbreeding depression. The exception was M3 x M6 which maintained its ICP with no changes in F₂. This cross was the best of the study because it showed high significant values in quality protein in F₁ and did not exhibit significant inbreeding depression.

The crosses M2 x M5 and M5 x M6, with intermediate response, were highly stable in F₂. They did not exhibit inbreeding depression (p≤0.05) in primordial aspects of quality protein (essential amino acids and ICP). For this reason, they are also valuable hybrids.

Some crosses maintained the level of quality of their lines and others did not, possibly because of their genetic constitution since expression of o2 is regulated by the genotype of the inbred line into which it was introduced (^{Gutiérrez-Rojas et al., 2008}). Moreover, there are genes that modify development of the endosperm (^{Holding et al., 2008}) and should vary with genotype.

Thus, there were different genotypic responses for inbreeding depression. Crosses with superior quality protein were more susceptible than allele combinations with lower quality attributes. Random genetic recombination had a favorable repercussion on the Lys variable and diminished amino acid content and ICP.

Obtaining o2 hybrids is a complex, but viable, process for which it is desirable to evaluate and characterize experimental germplasm based on the amount of essential amino acids and especially on ICP and stability of the manifestation of the virtues of o2 into generation F₂. In addition, the environment and the production technology level it requires should be defined. It is also possible to integrate the above elements using mathematical models that contribute to knowledge of the process.