Inheritance and genetic variation of shoot elongation before winter in oilseed rape (Brassica napus L.)

Oilseed rape (Brassica napus L.) is a major oilseed in many parts of the world with well adaptation to cold, dry and moist growing conditions. Despite large adaptation, winter survival is a major limiting factor to success of the crop in the regions with extreme freezing temperatures. Selection for improved winter hardiness is a difficult task, because firstly, efficient selection can only be performed in extreme winters, which in principle are not accessible and predictable in all regions. Secondly, winter hardiness is a complex trait, which consists of other interacting stresses. Cessation of shoot elongation before and during winter, has been considered one of the main requirements for successful overwintering in winter plants. Therefore, shoot length before and during winter is particularly considered by oilseed rape breeders as a contributing trait for selection of winter-hardy genotypes. Vernalization requirement, also, is a main winter survival mechanism that reduces the rate of shoot development and extend the vegetative phase to prevent frost damage by prohibiting floral transition during long warm periods before and during winter. To analyze the genetic variation for shoot elongation before winter and vernalization requirement, field experiments were performed with a set of 19 European winter oilseed rape genotypes. Moreover, to understand the genetic architecture of shoot elongation before winter and its correlation with vernalization requirement in winter oilseed rape, QTL mapping was implemented in two bi-parental doubled haploid (DH) populations of Brassica napus (DHLE and DHSO). The DHLE population consists of 151 doubled haploid lines, derived from a cross between the resynthesized line L16 and the winter oilseed rape line Express617. The DHSO population consists of 226 doubled haploid lines, derived from a cross between the two winter oilseed rape cultivars Sansibar and Oase. The two DH populations and the parental lines were phenotypically characterized in series field and greenhouse experiments. The two field environments included the normal sowing time at end of August/beginning of September, called autumn sown environment, and the spring sowing time at end of March/ beginning of April, called spring sown environment. Since the greenhouse, autumn and spring sown experiments represented very different environments; they were called three mega environments. Shoot length, shoot diameter were measured around three months after sowing time in the all three environments. In addition, data from flowering time, plant height, and seed quality traits were used to study their phenotypic and genetic association with the studied traits in the two DH populations. In the set of 19 European winter oilseed rape genotypes, results showed large phenotypic variation with significant genotypic variance components for shoot length in the autumn sown and spring sown environment. Broad sense heritability was high for shoot length in the spring sown environment (h2=97%) while medium heritability (h2=62%) was observed for shoot length before winter. Spearman’s rank correlation revealed a medium positive correlation (rs= 0.48*) between shoot length before winter and shoot length in the spring sown environment. Scatter plot of spearmen’s rank correlation distributed the 19 genotypes mainly in two clusters; first cluster consisted of genotypes with short shoot length in the autumn sown and spring sown environment, including R53 and L16, Mohican, Lorenz and Sollux, Zenith, Apex and Akela. Second cluster consisted of cultivars which were short before winter but long in the spring sown environment including Montego, Tenor, Adriana, Sansibar, Oase, Express 617, SGEDH13, SGDH14, King 10 and hybrid cultivar Visby. Gaoyou was the only cultivar with the longest shoot length in the both field environments. Low vernalization requirement was characterized in genotypes with significant elongation of shoot in the spring sown environment. In the DHLE population, large and significant genotypic variance components were found for shoot length and shoot diameter in the all three mega environments. Express617 was significantly longer than L16 for shoot length before winter (61 vs. 34) and begin of flowering was around five days earlier in the autumn sown environment. In the spring sown environment, parent L16 was with 218 mm length significantly shorter than Express617 with 855 mm shoot length. A near bimodal frequency distribution was found for shoot length in the spring sown environment with large phenotypic variation that ranged from 87 mm to 1255 mm. In addition, comparison of frequency distribution of shoot length and percentage of visible buds in the spring sown environment indicated that genotypes shorter than L16 almost did not show any tendency to form an inflorescence, whereas genotypes longer than Express617 showed almost 100% tendency to form an inflorescence. Correlation analysis in the DHLE population indicated no significant correlation between shoot length before winter and shoot length in the spring sown environment. Shoot length before winter was also not correlated with begin of flowering neither in the autumn sown nor in the greenhouse environment upon eight weeks vernalization treatment. Significant negative correlation of rs =-0.54** was found between begin of flowering in the autumn sown environment and shoot length, visible buds in the spring sown environment. Also, significant correlation was found between shoot length in non-vernalized plants in the greenhouse and shoot length in the autumn (rs = 0.31**) and spring sown environment (rs = 0.22**). Multiple interval mapping in the DHLE population, detected six main QTL for shoot length before winter that contributed 49.2% to the phenotypic variance. For shoot length in spring sown environment a major QTL with R2=35.5% was localized on linkage group C09 which along with two QTL mapped to A02 and A07 explained 68% of the phenotypic variance. In the non-vernalized greenhouse environment, two QTL were found for shoot length on A06 and A09 which together explained 16.2% of the phenotypic variance. No collocation of QTL was found between shoot length before winter, shoot length in the spring sown and non-vernalized greenhouse environment. Two candidate genes BnA2.FT and BnFLC9b were identified in the genomic regions with overlapping confidence intervals of QTL on A02 and C06. The candidate gene BnA2.FT is a paralogue of gene FT that collocated with major QTL on A02 for begin of flowering and plant height in the autumn sown environment. The candidate gene BnFLC9b is a paralogue of gene FLC found within a hotspot of linkage group C09 in which major QTL for shoot length and visible buds were localized in the spring sown environment. In the DHSO population large and significant genotypic effects were found for shoot length in the three mega environments. High heritability was found for shoot length in the spring sown (h2=90%) and greenhouse environment with non-vernalized plants (h2=69%), in contrast shoot length before winter had rather medium heritability (h2=46%). Parent Oase showed significant higher value than Sansibar for shoot length before winter (40 ver. 29), while Sansibar was significantly longer than Oase in the spring sown field experiments (721 ver. 442). Tendency to form inflorescence in the spring sown environments was higher than 80% in Sansibar and Oase, showing low vernalization requirement in the parental lines. Spearman’s rank correlation reveled negative correlation between shoot length in the spring environment and begin of flowering, (rs=-0.52**), plant height (rs=-0.25**) in the autumn sown environment. Shoot length in the spring sown environment was positively correlated with the appearance of visible buds (rs=0.90**) and shoot length in non-vernalized greenhouse environment (rs=0.42**). Significant correlation was found between shoot length before winter and vernalization requirement determined by shoot length (rs=0.31**) and visible buds (rs=0.26*) in the spring sown environment. Likewise, shoot length before winter was negatively correlated with begin of flowering (rs=-0.29**). In the DHSO population, multiple interval mapping detected five main QTL for shoot length before winter, explaining collectively 23.5% of the phenotypic variance. Three QTL were found for shoot length in the spring sown environment. The major QTL on A02 explained 70% of the phenotypic variance for shoot length in the spring environment. Three QTL for shoot length in non-vernalized plants in the greenhouse contributed 26% to the phenotypic variance. QTL mapping for begin of flowering identified one major QTL on C06 which along with 5 additional QTL accounted for 71.6% of the phenotypic variance. On linkage group A02 a hotspot was found with the QTL for shoot length before winter, begin of flowering, end of flowering, duration of flowering and plant height in the autumn sown environment and the major QTL for shoot length and visible buds in the spring sown environment. Surprisingly, all the QTL were the biggest or almost the biggest QTL of respective traits. A copy of candidate gene FLC (BnFLC2) was found on the genomic region from 15.3 to 15.9 cM on A02. C06 was the second linkage group with overlapping confidence intervals of the biggest QTL for begin and end of flowering in the autumn sown environment and a QTL for visible buds in the spring sown environment with opposite additive effects in the DHSO population. A copy of FT (Bn.C6.FTb) was identified in the interval 25.6 to 29.7 cM within the QTL hotspot on C06. In the both populations, different copies of candidate genes FLC were identified in the vicinity of major QTL for shoot length in the spring sown environment, indicating that elongation of shoot is significantly influenced by vernalization induced genes. Furthermore, different copies of candidate gene FT were identified in the vicinity of the major QTL for flowering time, showing key role of FT in regulation of flowering time. Identification of QTL mapped for some seed quality trait in the extent of FLC in the DHSO population suggests pleiotropic effect of FLC on different traits or tight linkage between genes involved in respective traits.