1. Introduction
Rapeseed is one of the main oil crops in the world, and yield improvement is an important goal of rapeseed breeding. The utilization of heterosis is an important measure to improve yield in rapeseed and other crops [
1,
2,
3,
4]. Studies have demonstrated that heterosis also increases plant height [
5,
6,
7,
8], and plant height has a significant positive correlation with the lodging index [
9]. Therefore, the overgrowth of plants may also lead to increased lodging risk in rapeseed. Lodging not only affects rapeseed yield and seed quality but also increases the challenge of mechanized harvesting [
10,
11,
12]. The “Green Revolution” in the 1960s led to a significant increase in the yields of major staple grain crops (wheat, corn, and rice) through the widespread adoption of semi–dwarf plant varieties [
13,
14]. Considering the significant achievements of the “Green Revolution”, the excavation and utilization of dwarf and semi–dwarf plant resources have attracted the attention of breeders.
The key to rapeseed semi–dwarf breeding lies in the acquisition of germplasm resources and the identification and utilization of dwarfing genes. In recent years, a series of dwarf and semi–dwarf rapeseed germplasm resources have been created by natural mutation, interspecific hybridization, and physical mutagenesis, such as YA2016–12 [
15], ds–3 [
16], ds–4 [
17], sca [
18], G7 [
19], and H2M–1 [
20]. Studying the genetic mechanism of dwarf material and identifying its dwarfing genes can contribute to the reduction of plant height and the improvement of lodging resistance. At present, considerable progress has been made in the study of the rapeseed dwarfing mechanism. Some genes related to dwarfing have been mapped and identified, including
BnaDwf.C9 [
21],
Bzh [
22,
23],
DS–1 [
24],
DS–3 [
16],
DS–4 [
17],
G7 [
19],
BND2 [
25],
BnaC04.BIL1 [
26],
BnaA03.BP [
27], etc. It has been reported that the underlying mechanisms of these genes are primarily associated with plant hormone signaling transduction and biosynthesis. Although there have been numerous achievements in the exploration of dwarf breeding and the mechanism of dwarfism in rapeseed, it is still necessary to explore new dwarfing materials and identify new dwarfing genes. This will help to increase the genetic diversity of dwarfing materials and provide an effective strategy for cultivating “ideotype” rapeseed and increasing rapeseed yield.
With the advancement of next–generation sequencing technology in recent years, whole–genome sequencing and bulked segregant analysis (BSA–seq) have been widely used in the gene mapping and molecular marker development of crops, such as maize (
Zea mays L.) [
28,
29], watermelon (
Citrullus lanatus) [
30], squash (
Cucurbita pepo) [
31], rice (
Oryza sativa) [
32], cotton (
Gossypium hirsutum L.) [
33], rapeseed (
Brassica napus L.) [
25], peanut (
Arachis) [
34], and
Brassica rapa ssp.
chinensis Makino [
35]. For instance, using BSA–seq,
AhRt, a candidate gene controlling peanut red shell, was identified on chromosome 22 and further narrowed the mapping interval to the 530 kb region by linkage mapping with 220 recombinant lines [
34]. Through BSA–seq analysis of 638 F
2:3 individuals, a new height–related quantitative trait loci (QTL) (
qPH9) was identified, which was mapped to a 2.02 Mb region of chromosome 9. Using 13 Kompetitive allele–specific PCR (KASP) markers combined with traditional linkage analysis facilitated the localization of
qPH9 to a 126 kb region [
32]. In order to map the dwarf gene
BND2, BSA–seq was performed on 25 extremely dwarf and 23 extremely high individuals selected from the F
2:3 population, and the candidate interval was rapidly identified to locate in the range of 13.77–18.08 Mb on chromosome A08 with a length of 4.31 Mb [
25]. BSA–seq improves the accuracy and efficiency of locating major QTLs and mining target genes when combined with traditional gene mapping methods.
In a previous study, we identified a dwarf mutant in rapeseed, LSW2018. The average plant height of LSW2018 was 136.24 cm after transplanting. In addition, LSW2018 has a dark leaf color, thick wrinkled leaves, short petiole, short internode, and low branch position, which we call the “dwarf” phenotype of LSW2018 [
36]. The dwarf trait of LSW2018 is controlled by a major dominant gene, without cytoplasmic effects [
37]. In order to identify the chromosome position of the dwarf gene, LSW2018 and 389 were hybridized to construct an F
2 segregation population for genetic research in this study. Then, BSA–seq, combined with QTL mapping, was used to identify the dwarf gene. Based on the above research, we aimed to investigate the genetic basis of the dwarf trait in LSW2018, thereby providing a theoretical foundation for the utilization of the dwarf gene and breeding of dwarf rapeseed.
3. Discussion
In this study, we identified a dwarf rapeseed with potential for application in production. Research on crop dwarfing has shown that dwarf plants can reduce the incidence of lodging, increase the grain number per spike, and improve the harvest index, thereby resulting in amplified crop yield and heightened grain quality [
13,
32,
38]. However, the excessive dwarfing of plants results in the overlapping of canopy leaves, a reduction in biomass and yield, and decreased disease resistance [
39]. Thus, extremely dwarfed plants are not conducive to production and utilization. Therefore, rapeseed breeders commonly believe that the semi–dwarf (1.2–1.5 m) and compact type may be the “ideotype”, which can ensure sufficient biomass, lodging resistance, and harvest efficiency [
27,
40]. The dwarf material LSW2018 developed in this study has an average plant height of 136.24 cm in the field, belongs to semi–dwarf rapeseed, and is resistant to dense planting and suitable for mechanized production. Therefore, LSW2018 is an excellent rapeseed germplasm resource with potential production application value.
BSA–seq, combined with traditional gene mapping methods for genetic analyses, is a simple and rapid approach that can effectively narrow the major QTL interval. Compared to BSA–seq, traditional QTL analysis is a time–consuming and laborious process. The genome–wide development of molecular markers is necessary, and multiple segregating populations need to be constructed for candidate region screening and gene fine mapping [
41]. In genetic studies, BSA only requires the selection of extreme phenotypic individuals or representative individuals in the population to construct mixing pools for QTL identification and gene mapping [
42]. In different crops, BSA–seq technology combined with traditional gene mapping methods has been widely used to identify genes controlling the accumulation of epicuticular waxes [
28], sex determination [
30], disease resistance [
31], plant height [
32], branch [
33], and dwarfing [
25,
29]. In our study, through BSA–seq and traditional mapping methods, the genes related to dwarf traits in LSW2018 were accurately delimited to chromosome A03 within two years. Due to the BSA–seq method, only an F
2 population containing 548 individuals was used in the gene mapping, and the genetic region was narrowed to a 0.71 Mb region. These results indicate that BSA–seq is an effective and efficient method of genetic analysis.
A total of 42 genes were identified by QTL mapping and BSA–seq in the candidate region, of which 17 genes were differentially expressed in this region between high and dwarf individuals. Among them, BnaA03G0384600ZS had a higher expression level in the LSW2018 than HSW2018 (about 106–fold), and is predicted to encode long–chain acyl–CoA synthetase 8 (
LACS8). In
Arabidopsis, both lacs4 and lacs8 have an impact on cuticular lipid metabolism. After the
lacs4lacs8 double mutation, the overall size of the plant was reduced compared to the wild type, and the leaves exhibited characteristics such as being short, rounded, curled, and disordered [
36,
43]. In this study, the LSW2018 mutant also displayed a similar phenotype. Subsequently, we conducted chlorophyll leaching analysis and observed that the LSW2018 mutant exhibited higher cuticle permeability than HSW2018 (
Figure S1). Based on these findings, we hypothesized that the loss of function of BnaA03G0384600ZS may result from defects in cutin or wax deposition, consequently leading to corresponding phenotypic changes. However, this study found that the expression level of BnaA03G0384600ZS was increased in dwarf individuals. According to the report, after point mutations or deletion derivatives, one function of a protein will be inactivated, but the ability to interact with other large molecules will still be preserved, which might cause mutant phenotype by competition [
44,
45]. Therefore, we hypothesized that BnaA03G0384600ZS may have similar functions after mutation, and the enhanced expression makes LSW2018 exhibit a distinct mutant phenotype.
Brassica napus is an allotetraploid crop with a complex genome, which leads to potentially complex interactions between genes [
46]. These results cannot fully confirm that functional alterations in this gene are responsible for the dwarfing phenotype of LSW2018. We will further explore the role and mechanism of
LACS8 in the formation of the LSW2018 dwarfing phenotype in subsequent studies.
Numerous studies have demonstrated that plant hormones, including auxin [
47], gibberellin [
48,
49], brassinolide [
50], and strigolactone [
51] play a key role in regulating plant height. Based on gene function annotation, we found 16 genes have lower expression levels in dwarf individuals, including 2 hormone–related genes, BnaA03G0386300ZS and BnaA03G0388800ZS. BnaA03G0386300ZS is predicted to encode cytochrome b561, a member of the auxin–responsive family protein involved in auxin signaling. Mutation of the cytochrome b561–encoding gene
OsCYBDOMG1 exhibited a dwarfing phenotype in rice [
52]. BnaA03G0388800ZS is predicted to encode cycloartenol synthase, which plays a crucial role in the biosynthesis of brassinosteroids (BRs). BRs are a class of plant steroid hormones that play key roles in regulating plant growth and development. When BR synthesis and signal transduction pathways are blocked, plants will exhibit a dwarfing phenotype [
53]. The BR content was reduced in the pea mutant lk after mutation of the BR synthesis–related gene
DET2, resulting in plant dwarfing [
54]. Mutation of the BR signaling pathway–related genes
RLP44 and
RBI1 also led to a dwarfing phenomenon in
Arabidopsis thaliana plants [
55,
56]. Other genes may also be involved in regulating the dwarfing of LSW2018, but further experiments are necessary to confirm the functions of these genes.
4. Materials and Methods
4.1. Population Construction and Traits Investigation
LSW2018, 389, and two generations (F1 and F2) of their hybrid constructs were planted at the Guizhou Academy of Agricultural Sciences, with row spacing of 0.4 m and plant spacing of 0.33 m. The field planting and management of the experimental materials followed the farmers’ habits. The F2 segregating population was established, comprising approximately 548 individual plants, which exhibited a phenotypic ratio of 3:1 between the dwarf and high groups.
4.2. Whole–Genome Re–Sequencing
Bulked segregant analysis (BSA) [
57] was preliminarily used for the dwarf gene mapping. At the seedling stage, 10 plants of LSW2018 (dwarf parent, PD) and 389 (high parent, PH) were selected for constructing two parent DNA pools, and 30 extremely dwarf (bulk–D) and high (bulk–H) plants in the F
2 population were selected for constructing two bulked DNA pools. The tender leaves of the plants were sampled in 1.5 mL centrifuge tubes stored in dry ice, and then sent to the Genoseq Corporation (Wuhan, China) for database construction and sequencing. The specific methods and procedures are as follows:
The total DNA was extracted using the CTAB method [
58], and RNase was used to remove the RNA from the extracted DNA samples. The purity and integrity of the DNA samples were assessed using agarose gel electrophoresis, ensuring that the genomic DNA remained intact and without RNA contamination. The DNA concentration was detected using the NanoDrop 2000c (Thermo Scientific, Waltham, MA, USA), and the final mass concentration was adjusted to 40 ng/μL. The library was constructed according to a standard procedure, and the four DNA pools (PD, PH, bulk–D, and bulk–H) were extracted for whole–genome re–sequencing.
The bioruptorUCD–200 was used to process the qualified DNA samples through quality testing to obtain fragments with lengths between 200 bp and 500 bp. Then, the fragmented DNA was treated with terminal repair and A tail and then connected to the sequencing connector. Target fragments were selected according to their expected library size and purified to remove splice contamination. Sequencing libraries were constructed by PCR enrichment and purified using AMPure XP Beads. The quality of the library was tested, and the PCR products were subjected to paired–end 150 bp (PE150) sequencing on the Illumina HiSeq platform after passing the quality inspection. According to the above process, the four DNA pools (PD, PH, bulk–D, and bulk–H) were constructed and sequenced, respectively. The sequencing depth of the PD and PH pools was ~20x, and the sequencing depth of bulk–D and bulk–H pools was ~30x.
After sequencing, data quality control was performed based on the raw data. Cutadapt software (version 1.13) was used to remove the linker sequences in the reads, and trimmomatic software (version 0.36) was used to eliminate low–quality bases. The filtered clean reads were compared with the reference genome ZS11–v20200127 (download link:
http://cbi.hzau.edu.cn/cgi-bin/rape/download_ext) (accessed on 23 February 2021) using BWA software.
4.3. Polymorphic Site Detection and Variant Annotation
All samples were tested for variation using the GenotypeGVCFs module of GATK (version 3.7) software, including single–nucleotide polymorphisms (SNPs) and insertion/deletion mutations (InDels). ANNOVAR software (version 2016Feb1) was used to annotate the variants and predict the effects of the variants on gene function. According to the method of Takagi et al. [
41], the SNP frequency Δ(SNP–index) on each polymorphic locus was calculated using QTLseqr (R package). According to the distribution of the Δ(SNP–index) values throughout the genome, candidate gene intervals related to dwarfing traits were screened out.
4.4. Gene Localization
MISA software (version 2.1) was used to identify SSR loci in the reference genome and determine the location information of these loci. According to the results of the MISA software analysis, SSR markers were designed using Primer5 software. Combined with the preliminary localization results, the SSR markers were screened in the candidate region, and the primers were sent to Sangon Biotech (Shanghai, China) for synthesis. The specific primer sequences are shown in
Table S1 in Supplementary Materials.
SSR markers with polymorphisms between parents were selected to detect polymorphisms in the 30 F2 individuals, and non–polymorphic markers were removed. Finally, 548 individuals were genotyped using the selected polymorphic markers. SSR amplification was performed in a final volume of 25 µL with a PCR mixture including 2 μL template DNA (50 ng·µL), 12.5 μL2 × Taq Master Mix for PAGE (Vazyme, Nanjing, China), 1 μL forward and reverse primers (10 μM), and 8.5 μL ddH2O. PCR was performed in a T100™ thermal cycler (Bio–Rad, Hercules, CA, USA) with the following program: initial denaturation at 95 °C for 3 min; 32 cycles of denaturation at 95 °C for 15 s; annealing at 56–62 °C (for different markers) for 15 s and extension at 72 °C for 60 s; and a final extension step at 72 °C for 5 min. PCR–amplified products were detected by 7% acrylamide gel electrophoresis and observed after rapid silver staining.
According to the phenotype and genotype results, the chromosome positions of the dwarf genes were determined with genetic analysis. Based on BSA–seq, genes with large–effect variants within the localization interval were identified according to the sequence polymorphisms between the dwarf and high materials. The BnIR rapeseed multi–omics database (
https://yanglab.hzau.edu.cn/BnIR/) (accessed on 2 February 2024) and the Arabidopsis database (
https://www.arabidopsis.org/) (accessed on 22 February 2024) were used to perform functional annotations on the genes with large–effect variants, and dwarf candidate genes were predicted based on the functional annotations.
4.5. Candidate Gene Prediction
Quantitative real–time PCR (qRT–PCR) was performed to identify the candidate genes. Total RNA samples were isolated from the leaves of dwarf and high individuals in the separated population using GeneJET (Thermo Fisher Scientific, Waltham, MA, USA), and first–strand cDNA was synthesized using a RevertAid™ Master Mix with DNase I (Thermo Fisher Scientific, Waltham, MA, USA). The reverse transcription products were used as templates for PCR, using the designed primers (
Table S5) to examine the expression of the predicted genes. qRT–PCR was performed using the Genious 2X SYBR Green Fast qPCR Mix (ABclonal, Wuhan, China) and the Bio–Rad CF96 Real–Time system (Bio–Rad Laboratories, USA). Three biological replications were performed with three technical replicates for each sample. The comparative cycle threshold method was used to calculate the relative expression levels of the different genes. Rapeseed gene
BnACTIN7 was used to normalize the expression levels of the target genes, which were analyzed using the relative quantification method 2
−ΔΔCt [
59].