Earth climate changes have set bigger challenges for plant breeders searching for superior genotypes (Araus et al., 2008). Beyond genetic gains for high yield and grain quality, maintaining yield stability by the reducing of losses by environmental stresses has received great attention (Oliveira et al., 2011). Among the strategies, the introduction of stay-green (delayed leaf senescence at late grain filling stage) character has shown promising results in several species. Strong evidences of the greater adaptability of stay-green genotypes under drought stress have been reported for sorghum (Kassahun et al., 2010), maize (Costa et al., 2008) and wheat (Izanloo et al., 2008). An evaluation of 936 wheat lines detected a significant association between tolerance to stresses by high temperatures and stay-green, indicating the delayed leaf senescence as a criterion for heat tolerance selection (Kumari et al., 2007). Plants with the character stay-green showed a lower reduction of the photosynthetic active area close to the physiological maturity (Ahlawat et al., 2008). Thus, delaying leaf senescence closer to physiological maturity stages may promote higher translocation and grain filling rates with direct reflex on yield components (Kumar et al., 2010).
Strong G x E interaction effects on grain yield has been reported for the wheat plant, making it difficult to select superior performance genotypes (Sanchez-Garcia et al., 2012). The analysis of adaptability and stability is essential for the partitioning of this interaction, allowing the identification of cultivars with predictable behavior and responsive to changes in environment (Franceschi et al., 2010). Among the various methods available for this determination, those based on regression of phenotypic values in relation to an environmental index have been the most widely used. Among those, is the model proposed by Eberhart and Russell (1966), commonly reported for cereals (Crestani et al., 2010). From regression models, the termed traditional method has also been used to estimate the stability with the advantage of application to a limited number of environments (Cruz et al., 2004). It is important to combine the adaptability and stability analysis via regression with other methods such as traditional, strengthening the inferences to be drawn against the predictability of genotypes (Silva and Duarte, 2006).
The contribution of genes that promote delayed leaf senescence of the plant may represent an effective mechanism that favors grain filling, adding benefits related to the higher yield stability. Thus, the aim of this study was to elucidate if the delayed leaf senescence in wheat by the expression of stay-green character brings forward effective contributions to the parameters of adaptability and stability in grain yield and thousand grain weight, longing for the possibility of genetic gain in selection of more stable and productive genotypes.
The populations were formed by crossing the genetic constitutions of wheat TB438 (bearer of character "stay-green") and TB188 (synchronized maturation) comprising lines selected by the breeding program of Embrapa Clima Temperado (Pelotas/RS) obtained via recurrent selection with several promising lines that had the character stay-green and others revealing synchronized senescence. In 1998, crosses were made between these lines and the F1 generation was obtained and used to obtain the two backcrosses and the F2 population. These lines were advanced for many generations (Fn) in order to reach high grain yield and differences in maturity, that is, stay-green (SG) and synchronized (SZ). Moreover, the backcrosses, RC1F1 (P1 // P1/P2) and RC2F1 (P2 // P1/P2), respectively, were subjected to self-pollination and selected for the presence and absence of the stay-green character until highly homozygous.
It is noteworthy that in these populations the selection also involved the simultaneous analysis for high grain yield and presence/absence of stay-green character. In 2002, 14 stay-green (SG30, SG39, SG47, SG53, SG65, SG71, SG74, RC1SG32, RC2SG34, RC2SG40, RC2SG46, RC2SG54, RC2SG62, RC2SG67), 16 synchronized (SZ31, SZ37 SZ49, SZ57, SZ69, RC1SZ43, RC1SZ45, RC1SZ55, RC1SZ58, RC1SZ68, RC1SZ72, RC1SZ76, RC2SZ35, RC2SZ42, RC2SZ56, RC2SZ61) were obtained and compared to parental lines TB438 and TB188.
The lines that were conducted in 2003, 2004 and 2005 in the experimental field located in Capão do Leão County, Rio Grande do Sul State, Brazil. And the soil is classified as Red Yellow Podzolic unit in Mapping Pelotas, which its U.S. equivalent is Typic Hapludalf (USDA, 2010). The County is situated at 31°52'00'' S lat and 52°21'24'' W long at an altitude of 13.24 m, with an average annual rainfall of 1280.2 mm.
A randomized complete block design with three replications was used, and each experimental unit consisted of five lines with three meters in length and spacing of 0.2 m, in a seeding rate of 300 viable seeds m-2. The fertilization and liming were made based on recommendations for wheat to an expected crop yield around 2 t ha-1. Other cultural practices such as weed control, diseases and pests were performed according to the indications for the crop.
This study evaluated the grain yield (GY, kg ha-1) obtained by the yield value of each plot and scaled to a hectare, weight of a thousand grains (WTG, g), by the count of 250 grains in each plot with subsequent weighing and multiplied by four. Data were subjected to variance analysis (ANOVA) and comparison of means between wheat lines contrasting for the maturation in different growing seasons. After, it was determined by traditional phenotypic stability method based on environmental variation for each genotype, recommending that those who showed the smallest mean square values are considered stable (Cruz et al., 2004). Also, adaptability and stability parameters were obtained using the Eberhart and Russell (1966) method, based on linear regression using the mathematical model Yij = βoi + β1iIj + δij + , where Yij is the average of genotype i at environment j; βoi: overall average genotype i; β1i: linear regression coefficient which measures the response of the genotype to variation of the environment; Ij: environmental index coded (∑jIj = 0); δij: deviations from regression; and : average experimental error. The genotype predictability was obtained with the determination coefficient (R2). All analyzes were performed using the GENES software (Cruz, 2006).
The genotype versus year (GxY) analysis indicated a significant source of variation for grain yield (GY) therefore a comparison of means was performed seeking to decompose the interaction effects (Table 1). The GY average for the lines TB438, SG65, SG74 and RC1SZ72 did not differ between years, suggesting a trend towards stability. However, despite the stability shown by the genotype RC1SZ72, it expressed a lower average grain yield than the lines carrying the stay-green character in different years, showing superiority of genotypes with within the expected
average of 2 t ha-1. Accordingly, some of the highest averaging genotypes were also from the delayed maturation type. In 2005, the low average rainfall between the months of November and December were decisive for the lower production values, including restrictive conditions during grain filling (Table 2).
This condition enabled a grain yield of 1598 kg ha-1 (Table 1). The reduced availability of water towards the end of the grain filling stage changes the plant relationships between source and drain, the concentration of reactive oxygen (Chen et al., 2010) and accelerates senescence and photoassimilate accumulation (Samarah et al., 2009), causing yield losses.
In more restrictive conditions (2005), nine genetic constitutions showed superior performance in GY, seven carrying the stay-green character, including one parent and two expressing synchronized maturation (Table 1). Some delayed maturation genotypes showed, in these conditions, GY values greater than 2 t ha-1, strengthening the evidence that the stay-green character favors a greater ability to stress tolerance of reduced water availability in wheat, agreeing with other reports (Adu et al., 2011). A strong relationship in maize chlorophyll content and delayed leaf senescence with a higher yield stability under more restrictive water supply (Messmer et al., 2011). In the analysis involving three year average GY for each genotype (Table 1), near all the stay-green genotypes showed the better performance. On the other hand, in the synchronized group, only RC1SZ68 showed superior behavior. Wheat genotypes with delayed senescence indicate significant gains in GY, bringing great prospects for increasing production efficiency by modification of plant photosynthetic ability (Parry et al., 2011).
In the WTG analysis (Table 3), both simple effects such as interaction were statistically significant. Similar to GY, WTG also showed greater magnitude of the mean square for the year of cultivation, indicating the most significant source of variation in character changes (Table 3). Similar results were obtained by G x E analysis in oat (Crestani et al., 2010). A large number of lines showed unchanged average values in WTG over the years of assessment (Table 3), indicating stable behavior of delayed senescence stay-green genotypes. In 2003, a favorable year for wheat cultivation, there was a predominance of synchronized maturation genotypes and only three stay-green genotypes showing high WTG (Table 3). In this condition, a similar behavior was observed in the parents. In 2004, similar results were observed, with the highest WTG values predominantly displayed by lines of the synchronized maturation group. Of the seventeen lines of superior performance, six were stay-green and eleven synchronized.
Also, the synchronized displayed higher WTG expression when compared to the delayed senescence parent (Table 3). Thus, in favorable years (2003 and 2004) little contribution from the stay-green character to WTG was observed, prevailing a superiority of synchronized senescence genotypes. These results disagree with the literature regarding contributions of stay-green to grain filling (Ahlawat et al., 2008). In wheat, genetic stability for high WTG and the availability of photoassimilates next to anthesis may favor other components connected to yield (Silva et al., 2005). The low performance of stay-green lines for WTG may be results of negative relationship among weight average of grains and number of grains per ear, seen that in favorable environments the presence of stay-green character improve also the fertility of spikelets, specially of those presents in the base and apical of the ear, where the grains are smaller, reducing to WTG (Luche et al., 2013). Also, positive effects on the number of fertile flowers and spikelets and ear grains have been reported (Silva et al., 2005; Ahlawat et al., 2008). In 2005, the year with highest water stress conditions (Table 2), an opposite behavior was observed in lines with the different maturity groups (Table 3). The stay-green genotypes showed higher values for WTG. These results coupled with genotype performances in the overall average between years (Table 3) appear to
show the greatest effect of stay-green on WTG is more evident under conditions of environmental restrictions, enhancing the tolerance to abiotic stresses. Results were also observed in other species (Izanloo et al., 2008; Adu et al., 2011).
Seeking to strengthen the evidence and hypotheses, the results of adaptability and stability analyses through simple regression, proposed by Eberhart and Russell (1966) and stability of the traditional method were obtained (Table 4). The variable GY in the stay-green character genotypes showed higher values than on those with synchronized maturation. Moreover, WTG means were equivalent for both groups of maturation. This reinforces the hypotheses reported by other researchers, indicating the superiority of stay-green genotypes in yield of grain goes beyond the greater capacity of grain filling, having an influence on other characters also linked to the formation of yield components (Silva et al., 2005; Ahlawat et al., 2008). The lines studied for GY indicated that the synchronized genotypes accounted for the majority of individuals with wide adaptability (β1=0), showing greater responsiveness to environmental improvements. Moreover, the stay-green parent showed specific adaptability to the unfavorable environment, ranking on top in GY, unlike the parent synchronized, adjusted to favorable environments, but among the lowest in the character expression. However, TB438, SG65 and SG74 showed significantly lower adaptability values (β1 <1.0), which, associated with phenotypic stability, discloses independently of method used, lines of great potential as sources for alleles for high GY and stability in unfavorable environments.
For WTG, the numbers of individuals with wide adaptability from different maturation groups were similar. The stay-green parent showed specific adaptability to harsh environments and was also included in the group expressing the best averages. On the other hand, the synchronized parent showed general adaptation, was also represented the group that expressed the highest values in character. Therefore, the similarity in WTG mean between the parents and the proportions of lines between maturity groups possibly indicate that the contribution of stay-green also affects other important characters in the increase of grain yield (Table 4). Stability parameters in the method of Eberhart and Russell (S2d and R2), the majority of exhibit delayed senescence genotypes were in stable expression of GY. It is noteworthy stability observed for who shows more permanent green plant next harvest parent, unlike what happened to the parent of synchronized pattern. In the analysis involving the traditional method (NDE) for the same character, there was a reduction in the number of phenotypic stability genotypes, showing that all tested lines, four belonged to the stay-green group and only a trend with the synchronized maturation. In a general analysis, involving the two observation methods, TB438, SG65, SG74 and RC2SG54 genotypes were those that showed the stability and effectiveness fully belong to stay-green ripening group. Even were included that expressed the best medium and limited adaptability to harsh environments group. We highlight the RC2SG54 lineage, because in addition to expressing high average with stability, was the one who showed ability to adapt front to GY expression in different years of assessment. Thus, maintenance of the photosynthetic active machinery for favoring higher chlorophyll concentration in tissues, provides maintenance of grain filling, providing greater stability in yield and grain yield, particularly under stress conditions (Izanloo et al., 2008).
Regarding WTG (Table 4), the adaptability parameter (β1) from the Eberhart and Russell method was found to be similar between the two maturity groups, with the stay-green parent indicating adaptability to restricted environments and the synchronized with wide adaptation. It is highlighted by this method (S2d and R2), a slight tendency to stability in a larger number of lines representing the stay-green group. In the mean square stability analysis, six stay-green (SG30, SG39, SG71, RC2SG34, and RC2SG40 RC2SG46) and only one synchronized (RC1SZ72) line were stable. It is noteworthy that for WTG, the RC2SG46 genotype showed both high adaptability and stability by both methods.
The traditional method was more conservative than the Eberhart and Russell, detecting a lower number of stable genotypes. On the other hand, the study of correlations between methods for estimating stability and adaptability, indicated a high positive (0.83) correlation between the traditional method and proposed by Eberhart and Russell (1966) and Silva and Duarte (2006). However, the model in the Eberhart and Russell (1966) method takes into account a higher number of parameters for the analysis, defining an ideal genotype as one which displays high yield (β0), regression coefficient equal to 1.0 (β1= 1.0) and regression deviation equal to zero (S2d =0), in other words, high average values for the character, wide adaptability and stability, respectively (Cruz et al., 2004).
