CORRELATION BETWEEN HETEROSIS AND GENETIC DISTANCE BASED ON SSR MARKERS IN SUNFLOWER ( Helianthus annuus L . )

The objective of this study was to determine the correlation between SSR based genetic distance (GD) of new NS sunflower inbred lines and heterosis for six agronomic traits. Twenty three sunflower inbred lines (20 restorer lines and three female lines used as testers) and their hybrids were examined for plant height, head diameter, thousand seed mass, oil content, seed yield per plant and oil yield per plant. Significant heterosis was observed in hybrid combinations for all examined traits except oil content. Genetic distance between pairs of tested sunflower inbred lines ranged from 0.13 to 0.8. There was no significant positive correlation between genetic distance and midand better-parent heterosis, specific combining ability and mean value in any of the examined traits for all 60 hybrids. A highly significant negative correlation was found between GD and mean oil percentage (r=-0.33 p<0.01). Although GD was generally a poor predictor of heterosis, better results are obtained if hybrid combinations for each tester and each trait are analyzed separately.

term process, as it is necessary to cross a lot of inbred lines and evaluate hybrids in field trials.Therefore, only a limited number of hybrids among all possible crosses can be tested.Utilization of genetic distance for predicting hybrid heterosis has been of great interest to breeders.The efficiency of hybrid breeding programs could be increased if the inbred lines per se could be screened and the superior crosses predicted before field evaluation (M e l c h i n g e r et al. 1990).
Studies of genetic diversity in relation to hybrid performance have been undertaken in several crops.Investigations in corn, Zea mays L., have shown that the genetic diversity of parents was significantly correlated with hybrid performance and that yield heterosis could be predicted using  2000) used AFLP markers and found a significant correlation between GD and seed yield, but genetic distance was generally a poor predictor of hybrid performance.The objective of this study was to determine the association between SSR based genetic distance of new NS sunflower inbred lines for most important agronomic traits and heterosis.

Materials and Methods
Twenty three sunflower inbred lines (20 restorer lines and three female lines used as testers) were selected based on their diverse genetic background for examined agronomic traits.The selected restorer lines (labeled R-1 through R-20) are new inbred lines developed in the breeding program of the Oil Crops Department, of the Institute of Field and Vegetable Crops, in Novi Sad, Serbia.Female lines used as testers (HA-48, HA-26 and HA-19) are commercial lines with good combining abilities.
Female lines were crossed with restorer lines to produce all possible combinations of F 1 hybrids using the line x tester method (S i n g h and C h o u d h a r y 1976).Seeds of the 60 F 1 hybrids were produced and their parents were sown in a breeding nursery of the Oil Crops Department, of the Institute of Field and Vegetable Crops.The experimental design was a randomized block system with four replications.
Plant height (PH), head diameter (HD), thousand seed mass (TSM), oil content (OC), seed yield per plant (SY) and oil yield per plant (OY) were used for quantitative characterization of 23 parental lines and their 60 F 1 hybrids.Plant height and head diameter were measured at the end of flowering.Seed yield was measured by harvesting the middle row of each plot by hand.Seed samples from each plot were analyzed for oil content by nuclear magnetic resonance.
Analysis of variance and specific combining abilities (SCA) for quantitative traits were performed using the line x tester method (S i n g h and C h o u d h a r y 1976).Heterosis was determined as follows: Mid-parent heterosis (MPH) (%) = ((F 1 -MP)/MP)*100 Better-parent heterosis (BPH) (%) = ((F 1 -BP)/BP)*100 where, F 1 is the F 1 performance, MP = (P 1 +P 2 )/2 in which P 1 and P 2 are the performances of inbred parents and BP is the better parent value (G e l e t a et al.Values of genetic distance as measured by SSR markers were correlated with MPH and BPH to estimate their relationship using Pearson's coefficient of correlation.Correlations were done for F 1 combination from each tester line separately and all tester lines.Significance of correlation was determined using table of S n e d e c o r (1959).

Results and Discussion
Parental lines and 60 F 1 hybrids were evaluated in field trials for plant height, head diameter, thousand seed mass, oil content, seed yield per plant and oil yield per plant.There were great variation among inbred lines and hybrids, respectively (Table 1).The mean values of the hybrids were significantly higher than the parental lines for plant height, head diameter, thousand seed mass, seed and oil yield per plant.The heterotic effect was observed in all examined trait, except oil content (Table 2).The mean values of hybrids were between parental means for oil content and both parental lines were selected for high oil quantity.The highest effect of heterosis (MPH) was observed for oil yield per plant (143.77%) followed by seed yield per plant (142.04%).Analysis of fifteen SSR markers detected 44 alleles, with an average polymorphism PIC= 45.3%.The number of alleles per locus ranged between 2 and 5, with a mean of 2.93.Genetic distance between pairs of tested sunflower inbred lines ranged from 0.13 (HA-19 vs. HA-48 and R-12 vs. R-18) to 0.8 (HA-19 vs. R-18) (data not presented).
The relationship between genetic diversity based on SSR markers of all inbred lines and their hybrid performance depended on the trait examined.Correlation coefficients between GD and parental means, SCA and heterosis were not significant for the most examined traits (Table 3).The only significant correlation was a negative one, between GD and mean oil content (r=-0.33 p<0.01).For plant height, correlation between GD and heterosis was positive but not significant (r=0.232 and 0.172).Similar results were obtained for thousand seed mass (0.226 and 0.245).Correlation between genetic distance and heterosis was not significant for the most examined traits.The poor correlation might be due to the several causes.SSR markers used in this study were chosen solely for their high PIC values.C h a r c o s s e t et al. (1991) and B e r n a r d o et al. (1992) suggested that genetic distance can not accurately predict hybrid performance unless the DNA markers used in the analysis were linked to the genes affecting the trait.Therefore, the 60 F 1 hybrids were divided into three groups according to the parental tester line and correlation of the GD with hybrid performance and heterosis within the groups was examined for all six traits.Only significant correlations, but not highly, were found between GD and better-parent heterosis for thousand seed mass for hybrids with the tester line HA-19 (r=0.43 p<0.05) and between GD and mid-parent heterosis for plant height for hybrids with the tester line HA-26 (r=0.47 p<0.05) (Fig. 1).In these two cases hybrid heterosis increased linearly with increased GD between parental lines.However, the correlations obtained were too low to be of predictive value.T e r s a c et al. (1994) described relationships between heterosis and enzymatic polymorphism of 39 sunflower populations.The correlation coefficients for all enzyme systems were too low to be used as predictors of the general combining ability, but when enzymatic systems were analyzed separately, four of them turned out to be useful markers for breeding purposes.C h e r e s et al. (2000) have used 360 AFLP markers and found that although genetic distances were significantly correlated with hybrid seed yield and percent of heterosis for seed yield (r=0.79 and 0.76), hybrid performance varied greatly among hybrids of inbreds with similar genetic distance (GD).Z e i d et al. ( 2004) pointed out that the lack of association between heterosis and genetic dissimilarities for inter group hybrids might be explained by absence of crosses between related parents i.e. by the absence of variation for parental relatedness: all crosses have unrelated parents.
The results of this study confirm that GD generally correlates poorly with heterosis and specific combining abilities.Previous studies in various crop species such as corn, pepper, alfalfa, wheat, and rapeseed also showed

C o n c l u s i o n
Although genetic distance is a poor predictor of hybrid performance, our results indicate that better results are obtained if hybrid combinations for each tester and each trait are analyzed separately.Our further field trials for identification of sunflower heterotic performance will be planned on prior information on genetic distance of inbreds, obtained by more molecular markers, involving the ones associated with QTLs for examined traits.
molecular markers (B e t r a n et al. 2003, R i e f et al. 2003, S c h r a g et al. 2006).Conversely, weak correlations have been reported between genetic distance and hybrid performance and heterosis in oilseed rape, Brassica napus L. (D i e r s et al. 1996), pepper, Capsicum annuum L. (G e l e t a et al. 2004), faba bean, Vicia faba L. (Z e i d et al. 2004) and alfalfa, Medicago sativa L. (R i d a y et al. 2003).Different sunflower gene pools have been studied for its genetic diversity with enzymes (T e r s a c et al. 1993), RFLP markers (G e n t z b i t t e l et al. 1994, B e r r y et al. 1994, Z ha n g 1995, H o n g t r a k u l 1997) and SSR markers (Yu et al. 2002, T a n g and K n a p p 2003, P a n k o v i c et al. 2004, S o l o d e n k o et al. 2005).However, the literature data on the predication of sunflower heterosis and hybrid performance by marker based genetic distance of the parental lines is scarce (T e r s a c et al. 1994, C h e r e s et al. 2000).C h e r e s et al. ( 2004).Significance of heterosis was determined by the t-test (K r a l j e v i c -B a l a l i c et al. 1991).Genomic DNA of 23 parental lines was extracted following the modified method of D e l l a p o r t a et al. (1983).The 15 SSR sunflower primers were used in the study (T a n g et al., 2002).The selected primers have previously revealed DNA polymorphism of sunflower NS breeding material (P a n k o v i c et al. 2004, T e r z i c et al. 2006).Fragments were first separated on 2% agarose (voltage 220 V) and then on 6% denaturing polyacrylamide gels (power voltage is constant 50W), for the fine separation.DNA polymorphism between two inbred lines was estimated by comparison of amplified fragments.Jaccard coefficient (J) of similarity was calculated according to S t a u b et al. (2000).Genetic distances (GD) among the 23 parental lines were estimated according to S p o o n e r et al. (1996) as GD = 1-J.
low correlation of GD with heterosis (M e l c h i n g e r et al. 1990, D i e r s et al. 1996, G e l e t a et al. 2004, Z e i d et al. 2004, Ri d a y et al. 2003).
Tab. 1. -Mean values, standard error of the means and coefficient of variation (V)for the sunflower parental lines and their F 1 hybrids Tab. 3. -Correlation between genetic distance (GD) and mid-(MPH) and better-parent heterosis (BPH), specific combining ability (SCA) and mean values (MV) for each trait in sunflower hybrids (PH=plant height, HD=head diameter, TSM=thousand seed mass, OC=oil content, SY=seed yield per plant and OY=oil yield per plant)