QTL mapping for low temperature germination in rapeseed

Rapeseed, a major oil crop in the world, is easily affected by low-temperature stress. A low temperature delays seed germination and increases seedling mortality, adversely affecting rapeseed growth and production. In the present study, a tolerant cultivar (Huyou21) was crossed with a susceptible genotype (3429) to develop a mapping population consisting of 574 F2 progenies and elucidate the genetic mechanisms of seed germination under low temperatures. Two quantitative trait loci (QTL) for low-temperature germination (LTG) were detected, one on chromosome A09 (named qLTGA9-1) and the other on chromosome C01 (named qLTGC1-1), using the QTL-seq approach and confirmed via linkage analysis in the mapping population. Further, qLTGA9-1 was mapped to a 341.86 kb interval between the SSR markers Nys9A212 and Nys9A215. In this region, 69 genes including six specific genes with moderate or high effect function variants were identified based on the Ningyou7 genome sequence. Meanwhile, qLTGC1-1 was mapped onto a 1.31 Mb interval between SSR markers Nys1C96 and Nys1C117. In this region, 133 genes including five specific genes with moderate effect function variants were identified. These specific genes within the two QTL could be used for further studies on cold tolerance and as targets in rapeseed breeding programs.


Results
Analysis of LTG and construction of bulks. Based on a preliminary screening for LTG, the tolerant cultivar Huyou21 and the susceptible cultivar 3429 were selected for further study. At optimal temperature (20 °C), the two parents demonstrated similar germination rates (> 95%) within four days after imbibition (DAI). At low temperature (8 °C), Huyou21 exhibited excellent tolerance to cold stress, with a mean germination rate higher than 90% within four DAI. In contrast, 3429 showed high susceptibility at 8 °C, with a mean germination rate lower than 40% within four DAI (Fig. 1A). The mean germination rate of all F 1:2 seeds derived from the F 1 plants of 3429 × Huyou21 cross was 76.9%, and that derived from the F 1 ′ plants of Huyou21 × 3429 cross was 70.2%. The F 2 lines showed large germination rate variations, ranging from 17.5% to 100.0%, exhibiting a distribution skewed toward tolerance (Fig. 1A).
Regression analysis of parent-offspring revealed an LTG heritability estimate of 0.26 for the F 2:3 families. After cold stress, 344 plants out of F 2:3 families showed a germination rate larger than 90%, and 230 plants showed a germination rate equal to or less than 90%. Besides, the tolerant and susceptible plants fit well in the 9:7 ratio (χ 2 = 1.58, p = 0.21) (Table S1). These results indicate the role of two dominant genes in controlling LTG in Huyou21. The F 2:3 families showed pronounced variation and segregation in cold stress tolerance or sensitivity. The LTG tolerance distribution was continuous and approximately normal in the F 2:3 families, indicating a quantitative inheritance of LTG. Meanwhile, the phenotypic trait indices (PTI) of F 2 progeny ranged from 0.19 to 2.58 (Fig. 1B). To establish the low temperature tolerant (LT) bulk and low-temperature sensitive (LS) bulk, 30 LT and 30 LS F 2 individuals were selected. The PTI of the 30 plants in the LT bulk was more than 2.33, and that of the 30 plants in the LS bulk was less than 0.99.
Resequencing and mapping of reads. The  Candidate genomic regions identified for LTG by QTL-seq. Further, the SNP index was calculated for each bulk to identify the candidate genomic regions related to LTG. Then, Δ(SNP-index) was calculated and plotted against the genome positions by combining the information on the SNP-index in LT bulk and LS bulk   2). The Δ(SNP-index) was calculated by subtracting the SNP-index of LS bulk from the SNP-index of LT bulk, and plotted across the 19 rapeseed chromosomes to map the putative genomic regions associated with the phenotype for LTG. Two QTL regions on the two chromosomes A09 and C01 were detected based on Δ(SNPindex) plot ( Fig. 2C), at 95% significance. The peak of qLTGA9-1 was located between 40.00 Mb and 49.73 Mb on chromosome A09 (Fig. 2D), and that of qLTGC1-1 was located between 38.90 Mb and 55.38 Mb on chromosome C01 (Fig. 2E). The results revealed one QTL related to LTG at the 9.73 Mb region of chromosome A09, named qLTGA9-1, and another QTL at the 16.48 Mb region of chromosome C01, named qLTGC1-1.

Marker development and QTL fine mapping.
A total of 351 SSR markers, including 163 primers in the 9.73 Mb candidate region (qLTGA9-1) on chromosome A09 and 188 primers in the 16.48 Mb candidate region (qLTGC1-1) on C01, were developed and genotyped in Huyou21 and 3429 to analyze the result of QTL-seq. Among these, 17 primer pairs (eight on A09 and nine on C01) produced steady and clear polymorphic bands between the two parents ( Table 2), indicating that these primers could be used in QTL mapping for LTG. These polymorphic markers were selected for linkage map construction, and linkage analysis was performed in the 574 F 2 individuals with the LTG phenotype. The genetic map and the LTG phenotypic analysis demonstrated both QTL on the predicted regions (Fig. 3). The QTL qLTGA9-1 was mapped in a 1.78 cM interval between the tightly linked markers, Nys9A212 and Nys9A215, which only explained 5.93% of the phenotypic variation for LTG but had a higher Limit of detection (LOD) value of 8.43 ( Fig. 3 and Table 3). Meanwhile, qLTGC1-1 was mapped in a 16.40 cM interval between the tightly linked markers, Nys1C96 and Nys1C117, which explained a phenotypic variation of 5.39% and had a LOD value of 6.57 ( Fig. 3 and Table 3). Based on the physical position of the tightly linked markers, qLTGA9-1 was localized in a 341.86 kb interval between 44.72 Mb and 45.06 Mb on rapeseed chromosome A09, and qLTGC1-1 in the 1.31 Mb interval between 48.19 Mb and 49.50 Mb on chromosome C01. Further, annotation based on Ningyou7 genome sequence 22 predicted that 69 and 133 genes were located in the qLTGA9-1 region and qLTGC1-1 region, respectively. Further, 74 SNP and 25 InDel mutations related to 32 predicted genes in the qLTGA9-1 region, and 162 SNP and 26 InDel mutations related to 35 predicted genes in the qLTGC1-1 region were screened based on the sequencing results for each QTL (Table S3 and Table 4). Of these predicted genes, only one gene in the qLTGA9-1 region was www.nature.com/scientificreports/ a variant (frameshift insertion/deletion) with a high effect, and five genes in qLTGA9-1 and qLTGC1-1 regions were variants (missense mutation) with moderate effects (Tables 4 and 5). These gene variants with moderate or high effects may be related to low-temperature tolerance at the generation stage in rapeseed. The Ningyou7 annotation 22 identified that seven of the predicted genes (four genes in qLTGA9-1 and three genes in qLTGC1-1) encode proteins associated with plant growth or temperature stress response as follows: ChrA09g005501 encodes a formiminotransferase, ChrA09g005507 encodes a protein phosphatase 2C (PP2C), ChrA09g005509 encodes an aminotransferase, ChrA09g005523 and ChrA09g005524 encode DRG family regulatory proteins (DFRP), ChrC01g004357 encodes a SWEET sugar transporter, ChrC01g004359 encodes a PHDfinger, and ChrC01g004400 encodes a plant invertase/pectin methylesterase inhibitor (PMEI) ( Table 5).  TGT ATA ACT CCT TGA ATCC CCA CAA GAA ATT AGA GGT CG  A09  43716935-43717137   Nys9A207  TAC TCC TTT GGA AGG AAA CA  TCC CTC TCC ATC TGA AAA TA  A09  44059061-44059257   Nys9A208  ATT CTG TGT ATC CCA TTT CG  GAG GAG GTT TCT GAG GAG TT  A09  44230061-44230241   Nys9A145  TTC AGT AAA GCT GAC ACA CG  CGA GAG GTT ATT AGG GGT TT  A09  44611834-44612080   Nys9A212  CAA AAG AGG GAA TTT CAG TG  TGT CTC TAG TGA GAA AGC ATTG

Discussion
The late direct seeding area for rapeseed in China has continuously increased with intensive cropping development. However, low temperature occurs in late-autumn or early-winter quickly affects rapeseed germination with the delay of rapeseed seeding times 6 . Therefore, it is a breeding goal to select varieties with good germination under both optimal and stressed conditions to adapt to changing temperatures under late direct seeding cultivation. In the previous studies, a high degree of variability was observed in the seed germination rate among the rapeseed genotypes under low-temperature conditions. The genotypic variability helped LTG studies in rapeseed breeding to deal with cold stress under late direct-seeding conditions. Studies have also revealed variations in LTG among plant species, which are generally affected by inheritance and environmental factors and regulated by QTL and multi-genes [6][7][8] . QTL mapping studies for LTG have been conducted in various crops, such as rice, maize, soybean, and wheat [9][10][11][12][13][14][15] ; however, few QTL or genes for LTG are reported in rapeseed. QTL-seq combined with BSA and NGS is a practical approach used to identify QTL 18 . The method has been successfully used to rapidly identify QTL of different traits in rapeseed 23,24 . In this research, a segregating population was employed to detect the LTG QTL of rapeseed using QTL-seq. The analysis of the LTG in the population derived from the 3429 × Huyou21 cross revealed that two dominant nuclear genes or QTL controlled the LTG of these populations, with a heritability of 0.26. Further, QTL-seq revealed two QTL (namely qLTGA9-1 and qLTGC1-1) associated with LTG on chromosomes A09 and C01; both were verified with classical QTL analysis through map construction. Here, qLTGA9-1 was mapped between the flanking SSR markers Nys9A212 and Nys9A215, and qLTGC1-1 was mapped between Nys1C96 and Nys1C117. Besides, the study found that the QTL (qLTGA9-1) on chromosome A09 was mapped around 30.2 Mb apart from each other based on the physical position of linkage markers on ZS11 genome sequence 25 . Earlier, GWAS mapping by Luo et al. 7 showed that the candidate genes related to seed vigor under low temperature were localized around the physical    25 . As the physical distance between the locus and the QTL identified in this study is more than 27 Mb, the QTL, including qLTGC1-1, identified are novel in rapeseed for LTG. Furthermore, to recapitulate the physical position of the linkage markers (Nys9A212 and Nys9A215, Nys1C96 and Nys1C117) based on the Ningyou7 genome sequence 22 , 69 and 133 genes were predicted in qLTGA9-1 and qLTGC1-1, respectively, and 11 of these predicted genes were variants with moderate/high effect according to QTL-seq. Among these predicted genes were ChrA09g005507 in the qLTGA9-1 region encoding a PP2C and ChrC01g004357 in the qLTGC1-1 region encoding a SWEET protein. PP2C is a key player in ABA signal transduction, which plays an important role in seed germination under cold stress [26][27][28] , while SWEET protein is an important plant sugar transporter family and plays a crucial role in seed germination and stress response 29,30 . In addition, ChrA09g005523 and ChrA09g005524 in qLTGA9-1 region encode DFRPs, while ChrC01g004359 and ChrC01g004400 in qLTGC1-1 region encode a PHD-finger and a PMEI, respectively. Like PP2C or SWEET, the DFRP, PHD-finger, and PMEI proteins also play significant roles in defense response or under abiotic stress [31][32][33] . However, further research should analyze and functionally validate these candidate genes.

Materials and methods
Plant materials. The parents used to develop the mapping population were Huyou21 (tolerant to cold stress) and 3429 (susceptible to cold stress) (Fig. 4). Seeds of Huyou21 and 3429 cultivars were obtained from the Shanghai Academy of Agricultural Sciences. Huyou21, developed from a double-cross between rapeseed lines 9714/9711 and 84004/8920, is widely cultivated in the lower reaches of the Yangtze River Basin, China, while 3429, used as the female, is a new line derived from a self-cross plant of the hybrid variety Qinyou99. The 3429 × Huyou21 hybrids were advanced from the F 1 generation by selfing to yield the F 2:3 families for mapping of LTG. Besides, the seed germination of 500 seeds harvested at the same time from both parents and their generations was confirmed.
Phenotypic evaluation. Five hundred healthy and plump F 3 seeds obtained per F 2 plant, all together F 2 seeds per F 1 plant, and their parental controls were placed on two layers of filter paper with 15 mL distilled water in Petri dishes (9 cm inner diameter). The germination experiment was conducted in a low-temperature incubator set at 8 °C in a completely randomized design with three replications per treatment. The number of germinated seeds (N 1 ) was counted on the fourth day after imbibition (DAI). Seeds that did not germinate after four days under low temperature (8 °C) were removed to normal temperature (20 °C) to start the recovery process for three days, and the number of remaining germinated seeds (N 2 ) under normal temperature were counted to exclude the percentage of low-vigor seed influences. A seed was considered germinated once the radicle emerged and elongated 2-3 mm from the seed. The seed germination under low-temperature was calculated using the following formula: Plants with a germination rate larger than 90% were classified as cold tolerant, and those with a germination rate equal to or less than 90% were classified as cold susceptible. The broad-sense heritability of LTG expression Germination rate (%) = N 1 /(N 1 + N 2 ) × 100%. www.nature.com/scientificreports/ was estimated using parent-offspring regression methods based on the variance of parents, F 1 -derived F 2 population, and F 2 -derived F 3 families 34 . The data on germination rate were arcsine transformed to improve the homogeneity of variance 35,36 . Further, the phenotypic trait indices (PTI) of F 2:3 families for QTL mapping of LTG were calculated using the following formula to reduce the influence of microenvironment features during cold treatments: where X represents the grand mean of germination rate of the F 2:3 family, P 1 and P 2 represent the germination rate of their parental controls.

SSR marker analysis and QTL fine mapping. LTG-related QTL identified by QTL-seq were validated
and fine mapped through the traditional QTL mapping method 40 . A total of 351 SSR markers in the predicted regions were mined from the whole-genome sequence 22 . The SSR markers were used to survey the polymorphism between parents, which were designed with SSR Locator 41 based on the parameters as previously described 40 . The newly developed markers were named NysX(A/C)Y markers, where Nys represents the microsatellite from the physical sequence of Ningyou7 rapeseed, the number X indicates the chromosome in subgenome (A or C), and Y represents a numerical code for the newly designed marker. Further, PCR and amplicon detection for the F 2 individuals and their parents were performed as previously described 42 with minor modifications. Polymorphic markers were further selected to analyze F 2 population to construct a linkage map for QTL fine mapping. The genetic linkage map was drawn using the MAP functionality in QTL IciMapping v4.1 43 , and QTL was conducted using the BIP functionality 43 . The map distance (cM) was calculated using the Kosambi mapping function 44 , and the mapping method adopted was ICIM-ADD 45 . The LOD threshold and recombination frequency were set at 3.0 and 0.30, respectively. The QTL were designated by the term qLTG followed by the chromosome number.
Candidate gene annotation. The candidate genes within the detected QTL for LTG were obtained based on the Ningyou7 rapeseed genome annotation 22 , and the functional variant effects of SNPs or Indels in each gene were predicted using SnpEff 46 software.

Permission statement.
All the experiments on plants, including the collection of rapeseed materials, were performed in accordance with relevant guidelines and regulations.

Conclusions
A total of 574 F 2:3 families were constructed to elucidate the genetic mechanisms of seed germination under low temperature in rapeseed. Based on the QTL-seq and linkage analysis of the populations, two QTL were identified from 'Huyou21' . One QTL was mapped to a 341.86 kb interval between the SSR markers Nys9A212 and Nys9A215 on rapeseed chromosome A09, and another was mapped to a 1.31 Mb interval between the SSR markers Nys1C96 and Nys1C117 on chromosome C01. These findings provide a basis for further studies on genetic breeding and assist in cloning candidate genes for cold tolerance in rapeseed.

Data availability
This whole genome resequencing reads used in QTL-seq has been deposited in the National Center of Biotechnology Information Sequence Read Archive (SRA) under BioProject accession number PRJNA751740.