COMBINING ABILITY OF NITROGEN DEFICIENCY STRESS INDICES FOR PLANT HEIGHT IN RAPESEED VARIETIES

Although an average plant height is more preferable in most of agronomic crops, less reduction of this trait in stress condition makes yield components and seed yield sustainable in rapeseed varieties. Combining ability, heterosis and heritability of plant height at application and non-application of nitrogen environments (Hp and Hs, respectively) and its related stress tolerance indices were detected in half F2 diallel crosses of six spring rapeseed varieties. Significant mean squares of general and specific combining abilities (GCA and SCA) were detected for Hp, Hs, mean productivity (MP), geometric mean productivity (GMP) and stress tolerance index (STI), indicating the importance of additive and non-additive genetic effects for them. A high narrow-sense heritability estimate was exhibited for STI, emphasizing the prime importance of additive genetic effects for this stress tolerance index. A significant correlation among Hp, Hs, MP, GMP and STI showed the efficiency of these stress indices for improving plant height in plant breeding programme. In comparison to SCA effects, most of the crosses had significant high parent heterosis for Hp, Hs and all the stress indices. RGS003 with a significant positive GCA effect of Hs was considered as a suitable parent for improving this trait and most of the combinations of this genotype had significant negative SCA effects.


Introduction
Rapeseed (Brassica napus L.), also known as oilseed rape or canola, is very responsive to nitrogen (N) and in several studies it has been emphasized that N is a critical limiting factor for canola production (Jackson, 2000;Fageria and Baligar 2005;Danesh-Shahraki et al., 2008).The application of N fertilizers was shown to increase canola seed yield even under a variety of diverse challenging conditions (Sidlauskas and Tarakanovas, 2004;Kazemeini et al. 2010).The characteristics of canola such as plant height, number of branches per plant, number of pods per plant, seed yield and oil content are positively correlated with soil N level (Ahmadi and Bahrani, 2009).Canola yield is also indirectly affected by N as a result of increased stem length, higher number of flowering branches, total plant weight, seeds per pod, number and weight of pods and seeds per plant (Yasari and Patwardhan, 2007;Zhang et al., 2009).A number of investigations showed that nitrogen fertilizers caused substantial rapeseed yield increases even in diverse and contradicting conditions and also their timely applications of appropriate amounts have an important effect on yield associated traits (Fageria and Baligar, 2005;Rathke et al., 2005;Danesh-Shahraki et al., 2008).Although in rapeseed breeding based on ideotype form, dwarf genotypes are preferable (Downey and Rimer, 1993), a significant positive correlation of plant height with yield components and also with seed yield in rapeseed and other Brassica species was reported in several studies (Marjanović-Jeromela et al., 2008;Aytaç and Kınaci, 2009;Sabaghnia et al., 2010;Semahegn Belete, 2011;Rameeh, 2012).Therefore, decreasing of plant height in stress environments has an important role for decreasing of seed yield and hence it can be considered as a good indicator for predicting of seed yield reduction.Plant height is also more easily detectable than seed yield based on a single plant in segregating generation.Thus, genetic parameter estimation for this trait can be used as the indirect selection criterion for seed yield improvement.Diallel cross technique developed by Griffing (1956) was adopted for the estimation of heterosis and it was exploited in rapeseed (Brassica napus L.) and other Brassica species (Teklwold and Becker, 2005;Zhang and Zhu, 2006) for plant height and other yield associated traits.Several yield-based stress indices have been developed that may be more applicable to work on environmental stress tolerance such as drought tolerance (Saba et al., 2001;Cheema and Sadaqat, 2004;Golabadi et al., 2006;Moghani Nasri et al., 2006), salinity tolerance (Rameeh et al., 2004;Rezai and Saeidi, 2005;Rameeh et al., 2012) and temperature tolerance (Porch and Jahn, 2001).Rosielle and Hamblin (1981) defined stress tolerance (TOL) as the difference in yield between the stress (Ys) and non-stress (Yp) environments and mean productivity (MP) as the average yield of Ys and Yp.Fischer and Maurer (1978) proposed a stress susceptibility index (SSI) of the cultivar.Fernandez (1992) defined a new advanced index (STI=stress tolerance index), which can be used to identify genotypes that produce a high yield and geometric mean producitivity (GMP) under both stress and non-stress conditions.The geometric mean is often used by breeders interested in relative performance since drought stress can vary in severity in field environment over years (Ramirez and Kelly, 1998).Saba et al. (2001) reported that due to negligible narrow-sense heritability estimates of SSI and TOL for drought tolerance, these stress indices are not useful for the selection of drought tolerant genotypes in plant breeding programs.
Although diallel analyses are frequently used in rapeseed breeding to assess genetic parameters for yield associated traits and yield-based stress indices, they were focused on nitrogen deficiency stress effects and its stress indices only in a few studies (Saba et al., 2001).The objectives of this study were therefore to detect (i) the general and specific combining abilities and narrow-sense heritability for nitrogen deficiency stress indices of plant height among a set of adapted cultivars and (ii) the relationship among nitrogen stress tolerance indices and plant height of rapeseed cultivars and their F2 progenies at nitrogen application (Hp) and nonapplication of nitrogen (Hs) environments.

Material and Methods
Six spring rapeseed cultivars including RGS 003, Option 500, RW008911, RAS-3/99, 19H and PF7045/91 were crossed in half diallel method during 2004-2005.In order to produce F 2 progenies, fifteen F 1 s were selfed at Biekol Agriculture Research Station, located in Neka, Iran (53°, 13' E longitude and 36° 34' N latitude, 15 m above sea level) during winter 2005-2006.F 2 progenies and 6 parents were sown in a randomized complete block design with four replications at two experiments including N 0 : without nitrogen as stress condition and N + : 150 kg of nitrogen per hectare as non-stress condition during 2006-2007.The plots in each experiment consisted of four rows 5 m long and 40 cm apart.The distance between plants in each row was 5 cm.The soil was classified as a deep loam soil (Typic Xerofluents, USDA classification) containing an average of 280 g clay kg -1 , 560 g silt kg -1 , 160 g sand kg -1 , and 22.4 g organic matter kg -1 with a pH of 7.3.Soil samples were found to have 45 kg ha -1 (mineral N in the upper 30-cm profile).Fertilized variant of experiment (N + ) received 150 kg ha -1 N as Urea (in three equal applications: at planting time, at beginning of stem elongation, and at initial of flowering stage), while unfertilized variant (N 0 ) received no N.All the plant protection measures were adopted to make the crop free from insects.Plant height was recorded based on 10 randomly selected plants in each plot.
The stress tolerance indices were detected using the equations including stress intensity: SI=1-(µ s /µ p ), tolerance index: TOL=Hp-Hs, stress susceptibility index: SSI=[1-(Hs/Hp)]/SI, stress tolerance index: STI=(Hp•Hs)/(µ p ) 2 , mean productivity: MP=(Hs+Hp)/2 and geometric mean productivity: GMP=(Hs•Hp) 0.5 , respectively.Hs and Hp are numbers of plant height of all genotypes per trial under stress and nonstress conditions, respectively and also µs and µp are the means of this trait for all the genotypes per trial under stress and non-stress conditions, respectively.
Diallel analysis of variance for the parents and their crosses was based on Griffing's method 2, model 1 for fixed genotypes (Griffing, 1956).The analysis was performed using the diallel-SAS program written by Zhang and Kang (1997).The CORR procedure of SAS was used to estimate correlations among the traits.A t-test was used to test whether the GCA and SCA effects were different from 0. For each hybrid and each stress tolerance index, the difference between the hybrid and the mean of high parents was computed.A least significant difference (LSD) was used to test whether these differences were different from 0 (Mather and Jinks, 1982).

Diallel analysis of variance
Diallel analysis revealed significant mean squares of genotypes for plant height at N + (Hp) and N 0 (Hs) and also its nitrogen stress tolerance indices including mean productivity (MP), geometric mean productivity (GMP), and stress tolerance index (STI), indicating significant differences among parents and crosses for this trait and its respective nitrogen stress indices (Table 1).Similarly, Nassimi et al. ( 2006) and Sabaghnia et al. (2010) have reported the important role of GCA and SCA effects for plant height in rapeseed, and also this trait was a highly heritable character determined by genes that exhibit some degree of dominance.Significant mean squares of general and specific combining abilities (GCA and SCA, respectively) were estimated for Hp and Hs and also for related nitrogen stress tolerance indices of plant height including MP, GMP and STI indicated the importance of additive and non-additive genetic effects for them.High narrow-sense heritability estimates observed for STI showed the prime importance of additive genetic effects for this stress index.Therefore, the selection based on STI will result in the genotypes with higher stress tolerance and yield potential and this finding is well in accordance with earlier results (Fernandez, 1992).These results are also similar to the earlier finding of Saba et al. (2001), who reported that due to negligible narrow-sense heritability estimates of SSI and TOL for drought tolerance, these stress indices are not useful for the selection of drought tolerant genotypes in plant breeding programs.

Combining ability
Significant positive GCA effects of Hs were observed for RGS 003 (Table 2), so the plant height of this parent was less affected by nitrogen stress.The means of parents for Hs were varied from 115 to 150.25cm in Option500 and RGS003, respectively (Table 3).Due to significant positive correlations of plant height with yield components in non-stress environment and also a positive correlation with seed yield in nitrogen stress environment (the data was not shown), the parents with high plant height especially in stress environment will be preferable.Therefore, the parents including RGS003 and RAS-3/99 were considered as superior parents for this trait.A significant positive correlation of plant height with yield and with its components was reported in rapeseed and other Brassica species (Aytaç et al., 2009;Marjanović-Jeromela et al., 2008;Sabaghnia et al., 2010;Semahegn Belete, 2011).A significant negative GCA effect of Hs was detected for 19H and it showed that this parent had the reduction effect in a set of combinations for plant height and consequently seed yield in stress condition.The parents including RAS-3/99 and RGS003 with significant positive GCA effects of MP and GMP had a high amount of this trait in both non-stress and stress conditions.Non significant GCA effects of parents were determined for TOL and SSI, therefore, these stress indices did not have enough of efficiency for GCA effects in nitrogen deficiency.A significant positive GCA effect of STI for RGS003 again stressed of high efficiency of this parent at conditions without fertilization (N 0 ).The parents including 19H and Option500 had significant negative GCA effects of MP, GMP and STI, therefore, the plant height of these parents decreased in nitrogen stress environment.Significant positive SCA effects of Hp and Hs were observed for RW008911 x PF7045/91 in stress and non-stress environments, so this combination can be considered as a good one for increasing plant height in the both environments (Table 4).The high amounts of Hs, MP, and GMP means were related to RW008911 x PF7045/91 and 19H x Option 500.Significant negative SCA effects of MP, GMP and STI were determined for RAS-3/99 x RGS 003 and RGS 003 x PF7045/91.Therefore, these cross combinations displayed decreasing of plant height in stress and non-stress environments.Huang et al. (2010) reported the important role of GCA and SCA effects for plant height in spring cultivars of rapeseed.The crosses with significant negative SCA effects of Hs, MP, GMP and STI had low means for this trait and related nitrogen stress indices of plant height, therefore, the selection based on SCA effects of the crosses for these nitrogen stress indices improved their respective per se performances.

Heterosis
Data presented in Table 5 for heterosis of the crosses showed that out of 15 crosses, 8 crosses exhibited a significant heterosis for Hp.A significant negative heterosis for plant height indicated a significant negative dominance effect for this trait.RW008911 x PF7045/91 with the significant positive heterosis for Hs was considered as a superior combination.The crosses including RW008911 x Option 500 and RW008911 x PF7045/91 with the significant positive heterosis for MP, GMP and STI were considered as suitable cross combinations for these stress indices.All of the crosses did not have any significant heterosis for TOL and SSI, therefore, these two stress indices are not suitable for recognizing the preferable cross combinations in nitrogen stress environment.Heterosis for yield and other yield associated traits was reported in rapeseed (Brassica napus L.) and other Brassica species (Teklwold and Becker, 2005;Zhang and Zhu, 2006).In comparison to SCA effects, heterosis effects of Hp and Hs and their respective stress indices were more significant in the crosses, therefore, detecting of superior combinations based on heterosis effects will be more efficient than on SCA effects.Due to a significant positive correlation of Hs with MP, GMP and STI, any selection based on these stress indices will have considerable effects on Hs (Table 6).A significant negative correlation between Hs and SSI, indicating the selection based on a low amount of SSI will have an increasing effect on Hs.Non significant correlation between Hs and TOL indicated inefficiency of this stress tolerance index for improving Hs.Saba et al. (2001) reported that TOL and SSI were not strongly correlated to MP, GMP and STI.On the other hand, the selection based on STI and GMP will result in the genotypes with higher stress tolerance and yield potential (Fernandez, 1992).

Table 1 .
Mean squares from analysis of variance for Hp and Hs, MP, GMP, TOL, STI and SSI in six parents of rapeseed and their 15 F 2 progenies.

Table 2 .
Estimates of GCA effects for Hp and Hs, MP, GMP, TOL, STI and SSI in six parents of rapeseed.

Table 3 .
Means of rapeseed parents and their half diallel crosses for Hp and Hs and their respective stress indices.

Table 4 .
Estimates of SCA effects for Hp and Hs, MP, GMP, TOL, STI and SSI in the half diallel crosses of six parents of rapeseed.
* and ** : significant at 0.05 and 0.01 probability levels, respectively according to LSD test.Hp and Hs: plant height at N + and N 0 , respectively; MP: mean productivity; GMP: geometric mean productivity; TOL: tolerance index; STI: stress tolerance index; SSI: stress susceptibility index.

Table 5 .
High parent heterosis estimates for Hp and Hs, MP, GMP, TOL, STI and SSI in 15 F 2 progenies of rapeseed.

Table 6 .
Pearson correlation coefficient among Hp, Hs and the associated stress tolerance indices in six parents of rapeseed and their 15 F 2 progenies.Hp and Hs: plant height at N + and N 0 , respectively; MP: mean productivity; GMP: geometric mean productivity; TOL: tolerance index; STI: stress tolerance index; SSI: stress susceptibility index.