Genetic control and combining ability of agronomic attributes and northern leaf blight-related attributes in popcorn

Abstract

The present study was conducted to investigate the genetic control and to estimate the general and specific combining abilities of popcorn for agronomic attributes and attributes related to resistance to northern leaf blight (NLB). The 56 hybrids (F1 and reciprocals), together with the eight parent lines and six controls, were evaluated in two harvests, in a randomized-block design with four replications. Dominance components were more expressive than the additive components for grain yield and expression of resistance, and hybridization was the most suitable option for obtaining resistant and productive genotypes. For grain yield, popping expansion, and resistance to NLB, there was no significance for reciprocal effects, which indicates that the direction in which the cross is performed does not interfere with the hybrid’s performance. Then, the superior hybrids recommended for more profitable growth were P8 x L61, L61 x L76, and L61 x L77.

Introduction

Popcorn (Zea mays L.) is a type of corn whose main characteristic is its hard and small kernels, which hold the ability to expand as a result of an internal pressure when heated (Hoseney et al., 1983; Silva et al., 1993). Typically, when compared with cultivars of common corn, popcorn plants display a greater susceptibility to the attack of pests and diseases (Hallauer, 2001; Arnhold, 2008; Leonello et al., 2009). Besides contributing to low yields, this fact elevates production risks. Therefore, producing genotypes resistant to the main leaf diseases should be considered a relevant aspect in breeding programs for this species (Arnhold, 2008).

Northern leaf blight (NLB), a disease caused by Exserohilum turcicum (Pass.) Leonard and Suggs (sin. Helminthosporium turcicum Pass.), is widespread all over the world and practically in all corn grown (Smith and White, 1988; Carson, 1995). In popcorn, in particular, it is considered one of the main leaf diseases (Fantin et al., 1991; Miranda et al., 2002; Sabato and Pinto, 2013). The pathogen causes losses that can exceed 40% of the grain yield under favorable climatic conditions and in susceptible genotypes. In Brazil, the greatest severities of diseases have occurred in off-season crops, when the pathogen infects the plants during the flowering period (Costa et al., 2009). Because of the growing participation of the off-season seasons for this species, it is even more important to control diseases of higher incidence during this growing season.

Several control measures are implemented to minimize the damage caused by NLB (spot blotch), e.g., the spraying of fungicides and the planting of resistant varieties. Furthermore, some practices are adopted, such as crop rotation, the use of the adequate planting density and spacing, and the use of balanced fertilization and elimination of crop residues (Bergamin Filho and Amorim, 2011). Of the recommended measures, the use of resistant cultivars is the most effective means of control (Ishfaq et al., 2014), as it decreases production costs and reduces the risks to the activity of man and the environment (Vieira et al., 2009).

According to Paterniani and Miranda Filho (1978) and Hallauer et al. (2010), the breeding of corn has two alternatives that can be implemented together: obtaining genetically improved populations or hybrids. In the first case, by using adequate selection methods, the frequency of favorable genes in the enriched population can be gradually increased. In the second scenario, the breeding strategy is aimed at producing inbred lines that, when in proper combinations, can produce hybrids superior to the populations of origin. In the conception of Cruz et al. (2012), in the case of hybrid production, the diallel analysis has been exploited successfully, as it provides an estimate of useful parameters in the selection of parents for hybridization in a simple manner, as well as information about promising combinations.

Despite the advantages of diallel-cross strategies, however, little research has been conducted with popcorn using this methodology. The studies undertaken by Larish and Brewbaker (1999), Pinto et al. (2007), Pajic et al. (2008), da Silva et al. (2010), Viana et al. (2011), Vieira et al. (2011), Moterle et al. (2012), and Cabral et al. (2015) are the few examples in which diallels were used in popcorn, in which lines are used as parents. In this group, Vieira et al. (2011) evaluated grain yield (GY), popping expansion (PE), and partial resistance to southern corn rust (Puccinia polysora Underw) in hybrids originating from crosses in a partial circulating diallel scheme among ten popcorn lines (IAC 112 line groups x ‘Zaeli’ line groups). The authors observed a significant effect of general combining ability (GCA) in the Zaelin group for GY, PE, and resistance to southern corn rust and concluded that there is a predominance of additive genes in the expression of resistance, and these recurrent selection methods are recommended for obtaining gains in these traits.

At the moment, there are no studies on resistance to NLB employing diallel analysis in popcorn, although different levels of resistance to the disease have been reported in evaluations of popcorn genotypes (Fantin et al., 1991; Miranda et al., 2002; Vieira et al., 2009). Given the considerations mentioned above, the present study aimed to evaluate the genetic effects and the combining ability of hybrids of popcorn lines for incidence and severity of NLB, as well as for GY and PE, via diallel analysis, in the first and second harvests.

Material and Methods

Single-cross hybrids were produced, and parents and F₁ hybrids with reciprocals were evaluated at the Antônio Sarlo State Agricultural College, located in Campos dos Goytacazes - RJ, Brazil. The parent lines and diallel hybrids were cultivated in the agricultural periods of the first harvest (October 2014 to January 2015) and the second harvest (May to August 2015). The trial consisted of 70 treatments, including 56 simple hybrids (F₁ and reciprocals), eight parents (S₇ lines), and six controls, which were selected according to their performance regarding agronomic attributes and resistance to spot blotch, aiming to compare these traits in the controls with the hybrids. The adopted controls were IAC 125, BRS Angela, UENF 14, UFV M2-Barão de Viçosa, and hybrids L70 x L54 and P8 x L54.

The trials were implemented in a randomized-block design with four replications. The lines were chosen at random and separated from the hybrids to prevent competition effects. Plots consisted of a 5.0-m planting row with 25 plants, with 0.90 m spacing between rows and 0.20 m between plants. Grains were seeded at a depth of 0.05 m, using three grains per furrow; 30 days later, the area was thinned, leaving one plant per furrow. Plots were fertilized at seeding with 60 kg/ha K2O, 30 kg/ha N, and 60 kg/ha P2O5, in addition to 100 kg/ha N. Irrigation was applied by a sprinkling system, and herbicides and insecticides were applied whenever necessary.

To evaluate the incidence and severity of NLB, caused by E. turcicum, five plants were analyzed per plot. The assessments took place during the flowering period and the grain’s dough stage. Three evaluations were carried out, once a week. To study the incidence of E. turcicum in the plant, the scores proposed by Agroceres (1996) were adopted, and then, the proportion of leaf area of the plant was evaluated, considering all leaves. The severity of the disease in the leaf was determined by using a diagrammatic scale proposed by Vieira et al. (2013). In the study of severity in the leaf, only the leaf of the uppermost cob of the plant was considered.

Further, GY and PE were evaluated. The former was determined by weighing the grains after eliminating the cob, relative to the area extrapolated to hectares, and was expressed as kg/ ha. To determine the PE (mL/g), the weight of 30 g grains was heated in a microwave oven inside a special bag for popping at the power of 1000 W for 2 min and 20 s. The popcorn volume was quantified in a 2000-mL beaker, by dividing the popped volume by 30 (weight of grains).

The data obtained from the experimental plots were initially subjected to an individual analysis of variance for each location, and combined analysis of variance in a completely randomized block design. Whenever a significant effect was detected, treatment means were grouped by the Scott-Knott algorithm at the 5% probability level.

In the diallel analysis, we used Griffing’s (1956) Method 1, Model B. The GCA effects (�?_i) of each parent and the specific combining ability (SCA) effect (�?_ij) from a set of parental (p) and [p (p - 1) / 2] F₁ hybrids with the reciprocals were estimated. All analyses were performed using the computer resources of the Genes software (Cruz, 2013).

Results and Discussion

For both incidence and severity of E. turcicum, there were significant effects for all tested sources of variation (data not shown); it thus follows that there is variation between the treatments, which allows us to anticipate the opportunity to identify lines and hybrids of interest for obtaining selective gains. A significant effect of growing harvest was already expected, given that the development of the disease is influenced by environmental variations, e.g., temperature and relative humidity of the air (Agrios, 1988). Costa et al. (2009) reported that the disease is more severe in the conditions of the second harvest, which is in line with the results obtained in the present study.

Considering the occurrence of a significant treatment x harvest interaction effect, individual analyses were performed for each growing harvest. A significant effect of treatment was observed in both harvests, indicating the existence of genetic variation, and consequently, the possibility of selection of superior genotypes. The variation of treatments as observed by the F-test was also detected by the Scott-Knott algorithm, and then, three (for severity in the first harvest) and six (for incidence in the second harvest) groups were formed (Table 1).

Table 1: Means clustering test for incidence and severity of northern leaf blight (NLB) evaluated in 8 parents, 56 hybrids, and 6 controls in the first harvest (October 2014 to January 2015) and the second harvest (May to August 2015) in Campos dos Goytacazes, RJ, Brazil.

The group of genotypes that showed the lowest mean values for incidence and severity in both growing harvests, and consequently, the highest resistance levels was formed by parent L76, control UFV M2-Barão de Viçosa, and the combinations L55 x L61, L55 x L70, L55 x L76, L55 x L88, L55 x P8, L61 x L76, L61 x L77, L70 x L55, L70 x L88, L76 x L77, L77 x L76, L88 x L77, L88 x L61, L88 x L70, L88 x L76, L88 x P8, P8 x L61, and P8 x L88 (Table 1). According to the genealogy of the hybrids mentioned above, it is observed that the 18 combinations included lines L88, L77, L55, and L76 in their constitution that were obtained from varieties Viçosa and Beija-Flor, and were considered potential sources of resistance to NLB caused by E. turcicum (Miranda et al., 2003). The same was true for lines L70 and L61, which originated from cultivar BRS Angela, which has been reported as moderately resistant to E. turcicum (Embrapa, 2008). In this case, it was inferred that the lines provided experimental hybrids with the genetic contribution that culminated in higher levels of resistance in hybrids originating from lines of variety Beija-Flor. Therefore, the use of these parents in crosses tends to benefit the introduction of resistance genes, thereby making them of interest for breeding.

For grain yield, the means of the genotypes in the first harvest were allocated into three groups, whereas four groups were formed in the second harvest. In the two growing harvests, good performance was shown by the hybrids, since the means expressed by the genotypes that made up the group with the highest yields exceeded the magnitude of 3000.0 kg/ha. Of the genotypes that showed, in the first harvest, the best performance for grain yield, hybrids P8 x L61, L61 x L88, L77 x L55, and L88 x L77 stood out with estimated yields of 4767.01, 4648.00, 4502.03, and 4495.34 kg/ha, respectively. We also observed that among the lowest magnitudes for grain yield were varieties BRS Angela and UFV-M2 Barão de Viçosa, with mean values of 2560.12 and 2685.74 kg/ha, respectively. Therefore, these controls did not adapt well to Campos dos Goytacazes in the first growing harvest. In the second harvest, 20 genotypes expressed mean values for grain yield between 4197.92 and 5411.23 kg/ha. The highest mean for this trait was obtained in the second harvest. The experimental hybrids with the highest means for grain yield were L77 x L55, L61 x L77, L61 x L55, L77 x P1, with estimated yields of 5411.23, 4969.94, 4946.64, and 4787.83 kg/ha, respectively.

For popping expansion in the first harvest, the genotypes were distributed into five groups, while four groups were formed in the second harvest. Within the groups formed in the first harvest, fifteen genotypes showed average estimates greater than 32.71 mL/g, composing the group with the best performance. Among these, parent P8 and controls L70 x L54 and IAC 125 had the best performance, with estimated popping expansion of 36.25, 34.79, and 34.71 mL/g, respectively. Regarding experimental hybrids, L76 x P8 expressed the highest popping expansion, 34.46 mL/g. However, the estimated grain yield was not sufficiently high: 3290.00 kg/ha; nevertheless, this magnitude exceeds by 9.67% the minimum recommended for a release, characterizing it as a hybrid that can be indicated for commercial crops.

In the second harvest, parents L61 and P8 had the best means composing the group of genotypes with the highest estimates for popping expansion: 33.71 and 31.09 mL/g, respectively. Controls IAC 125 and P8 x L54 had the best performance for this growing harvest, with estimates of 33.08 and 31.92 mL/g. These higher mean values of parent P8 in both growing harvests demonstrate its potential among the evaluated parents, characterizing it as promising to be used in breeding programs aimed at increasing the popping expansion. Experimental hybrids L61 x P8, P8 x P1, L77 x L61, and P1 x L77 had the best performances for popping expansion, with estimated values of 31.83, 31.59, 31.58, and 31.33 mL/g, respectively.

The mean squares for GCA and SCA, in the second harvest, were significant for incidence and severity of E. turcicum, as well as for GY and PE, which indicates the existence of variability resulting from the action of additive and non-additive effects in the control of the gene expressions of these traits (Table 2). By contrast, for the first harvest, a significant effect of GCA was only observed for severity of E. turcicum, GY, and PE; for SCA, only for the incidence of leaves with symptoms of NLB. As also shown in Table 2, there was no significant reciprocal effect for the evaluated traits, either relating to resistance to NLB or for agronomic attributes, which suggests that the performance of the parent as a pollen donor or receptor can be the same. Thus, resistance to NLB cannot be superior if the crosses are inverted, which is an important fact in breeding programs to prevent additional work and costs in potentiating variability through hybridization.

Table 2: Estimates of mean squares of treatments (parents, F₁, and reciprocals), general (GCA) and specific (SCA) combining abilities, reciprocal effects and residue, as well as of square means for the combining ability effects for six traits evaluated in a full diallel among eight parents, with the reciprocals, in the first harvest (October 2014 to January 2015) and the second harvest (May to August 2015) in Campos dos Goytacazes, RJ, Brazil.

Non-additive genetic effects predominated in the evaluated traits (Table 2), except popping expansion, which determines that hybridization is the best alternative to obtain gains in breeding aimed at resistance to E. turcicum and grain yield. Results concerning the prominence of gene effects on resistance to E. turcicum in conventional corn have been controversial. Paterniani et al. (2000) and Vivek et al. (2010), for instance, found that the additive genes were the most important in the expression of resistance to E. turcicum; the opposite occurred in the study developed by Nihei and Ferreira (2012). These findings indicate that genetic control for this trait may vary across the different sources of resistance. In this case, it is up to breeding programs to characterize their sources of resistance to achieve maximum efficiency during the selection process (Lopes et al., 2007).

For grain yield, the mean squares of the effects associated with GCA were lower than those associated with SCA, demonstrating the higher importance of non-additive concerning additive effects in the genetic control of this trait. Analogously to these results, superior significant SCA effects concerning GCA were found by Pereira and Amaral Júnior (2001) and Freitas Jr et al. (2006) in popcorn.

As for popping expansion, the additive genetic effect was the most important; therefore, intrapopulation methods are the most recommended for obtaining gains in this trait. These results agree with those obtained by da Silva et al. (2010), Vieira et al. (2011), and Cabral et al. (2015).

In the first harvest, parents L76, L61, and L70 presented negative GCA estimates for incidence and severity of E. turcicum (Table 3), and are thus recommended for the generation of superior hybrids. In the second harvest, negative GCA estimates for incidence and severity of E. turcicum were expressed by parents L88, P8, L77, and L76. Only L76 displayed a negative estimate of GCA for incidence and severity of E. turcicum in both growing harvests (Table 3); moreover, according to the Scott-Knott test, L76 was allocated in the group with the greatest resistance to both variables related to NLB (Table 1). Therefore, L76 can be considered a line of interest for participation in breeding programs aimed at the production of segregants with resistance to NLB, especially in the composition of hybrids, given the greater evidence of dominance in the expression of resistance.

Table 3: Estimates of general combining ability, evaluated in a full diallel with eight parents, in the first harvest (October 2014 to January 2015) and the second harvest (May to August 2015) in Campos dos Goytacazes, RJ, Brazil.

Contrary to what has been considered for disease-related traits, in the evaluation of grain yield, higher positive estimates of GCA are preferred since the effects of allelic complementation are predominant in the generation of superior genotypes. For this trait, in the first harvest, the highest GCA estimates were observed in parent L88, followed L76 and L77. Of these, L88 and L76, together with L70, were the lines that stood out, also in the second harvest, for revealing high positive GCA estimates. Thus, for both growing harvests, L88 and L76 can be considered the parents of interest for obtaining gains in grain yield.

Regarding popping expansion, in the first harvest, the lines L61, P8, P1, and L70 showed the most significant positive estimates for GCA, whereas in the second harvest the lines L77, P8, and L70 were superior. For popcorn breeding programs, genotypes that conciliate favorable genes for GY and PE should be prioritized, because these are the characteristics of highest economic importance, besides other traits that may make the genotype superior, of course. In this context, most of the parents that were highlighted for GY were not for PE. In popcorn breeding, the existence of a negative correlation between PE and GY complicates the selection of genotypes for these two traits of economic value (Daros et al., 2004). For this reason, none of the parents that were superior for PE and GY could be simultaneously highlighted.

In the first harvest, 17 hybrid combinations were noteworthy for their negative SCA estimates, both incidence and severity of NLB. In the second harvest, however, 17 hybrids displayed negative SCA estimates for both traits related to resistance to NLB (Table 4). The superior hybrids - for showing negative estimates for incidence and severity of NLB - common for both growing harvests were L88 x L61, L88 x L70, L88 x L55, P8 x L70, L61 x L77, L61 x P1, L70 x L77, L77 x P1, and L55 x L76, which characterizes them as of interest when aiming at reduced attack of NLB in crops.

Table 4: Estimates of �?_ij and �?_ii effects in a full diallel with eight parents in the first harvest (October 2014 to January 2015) and the second harvest (May to August 2015) in Campos dos Goytacazes, RJ, Brazil.

For grain yield, in both growing conditions, a high number of combinations with positive SCA estimates were obtained. The three combinations that most stood out in both harvests were L61 x L77, L61 x L76, and L77 x L76, whose means were higher than 3900.00 kg/ha.

The SCA for popping expansion in the first harvest revealed that 17 combinations presented positive magnitudes. However, combinations L76 x P8, P8 x P1, L70 x P8, P8 x L77, P8 x L55, P8 x L61, L76 x L70, and L77 x L76 can be considered promising for that trait, because in addition to revealing positive values for SCA they also showed satisfactory means for the trait (Table 1) and originated from at least one parent with the potential to increase popping expansion. This demonstrates that the desirable effect of the genetic accumulation of parents L70, L77, and P8 was translated into a satisfactory effect of genetic complementation in the combinations.

In the second harvest, 15 combinations expressed positive SCA estimates for popping expansion, in which P8 x L61, L77 x P8, L77 x L70, L61 x L77, L61 x L76, L70 x P8, and L61 x L70 were noteworthy. In this regard, P8 x L61 showed a higher SCA estimate than the other selected combinations, averaging 31.92 mL/g (Table 1), thus proving to be of great potential to increase PE. Despite the difficulty gathering high yields and good popping expansion in one hybrid, L61 x L77 managed to comprise quality with productivity and parent L61 had the highest GCA for popping expansion, while L77 expressed a good GCA estimate for grain yield. This denotes that the hybrid responded as expected based on the parental GCA. This was not true for the other combinations, whose hybrids, despite originating from GCA desirable for PE and GY, did not display good allelic complementation for both traits together.

Given the results, promising hybrids can be indicated to increase GY and PE and reduce incidence and severity of NLB. In this context, in the first harvest, combination P8 x L61 expressed the highest estimate for both traits together, whereas in the second harvest, combinations L61 x L76 and L61 x L77 stood out for the elevated negative SCA estimates for incidence and severity of NLB as well as high positive estimates for GY and PE.

Conflicts of interest

The authors declare no conflict of interest.

Acknowledgments

The authors thank the Foundation for Research Support of the State of Rio de Janeiro (FAPERJ) and the National Council for Scientific and Technological Development (CNPq) for the financial support to the experiment, and the Coordination of Improvement of Higher Education Personnel (CAPES), for granting a doctoral scholarship to J.S. Santos.