-
In this study, two F2 populations were developed, including 'CJ-F2' (♀ZAASC4101 × ♂ZAASJ1401) and 'CR-F2' (♀ZAASC4101 × ♂ZAASR06). The DH line ZAASC4101, which was an early maturing cauliflower, took approximately 55 d from seedling transplanting to curd maturation. When a small curd became visible to the naked eye at the apex of the main stem, the plant's leaves exhibited a lotus-shaped appearance. The mean values for MSH and LSS in ZAASC4101 were notably short, measuring only 7.62 and 1.79 cm, respectively. ZAASJ1401, on the other hand, was an F8 inbred line of Chinese kale. The meristem at the apex of its main stem didn't form a curd but directly formed buds that led to flowering. In ZAASJ1401, the MSH and LSS reached average lengths of 28.61 and 5.47 cm, respectively. ZAASR06, a broccoli landrace had been inbred up to the F7 generation. Its shoot apical meristem (SAM) formed a loose curd composed of numerous small flower buds, resembling broccoli. The MSH and LSS of ZAASR06 were the longest among the three parental lines, measuring 35.22 and 6.21 cm, respectively (Table 1, Fig. 2).
Table 1. Statistical analysis of MSH and LSS for the F2 population and their parents.
Trait Populationa Parents F1 F2 populations Female Male Meanb ± SDc Ranged Mean ± SD Skewnesse Kurtosisf MSH CJ 7.62 ± 1.79 28.61 ± 3.03 19.37 ± 3.97 11.00−37.00 22.96 ± 5.46 0.067 −0.49 CR 7.62 ± 1.79 35.22 ± 2.92 21.81 ± 3.64 / / / / LSS CJ 1.79 ± 0.29 5.47 ± 0.49 4.23 ± 0.64 2.00−7.00 3.98 ± 0.94 0.56 0.21 CR 1.79 ± 0.29 6.21 ± 0.55 4.41 ± 0.62 2.10−5.90 3.82 ± 0.82 0.34 −0.43 a Population: CJ, ♀ZAASC4101 × ♂ZAASJ1401; CR, ♀ZAASC4101 × ♂ZAASCR06. b Mean (cm): The average value of phenotypic data from ten. c SD: Standard deviation of the phenotypic trait. d Range: The interval of phenotypic data of two F2 populations. e Skewness: Skewness of the phenotypic trait. f Kurtosis: Kurtosis of the phenotypic trait. Figure 2.
The MSH and LSS phenotypes of the parents and their F2 populations. (a) The MSH and LSS phenotypes of ZAASC4101 (female parent, cauliflower with short MSH), ZAASJ1401 (male parent, Chinese kale with high MSH), ZAASR06 (male parent, broccoli landraces with high MSH) and their F1 hybrids (CJ-F1 and CR-F1). (b) Frequency distribution of MSH and LSS of the F2 individuals in the CJ-F2 population and LSS of the F2 individuals in the CR-F2 population.
The F1 plants generated from the crosses CJ and CR exhibited intermediate values compared to their respective parents for MSH and LSS, measuring 19.37 and 21.81 cm, and 4.23 and 4.41 cm, respectively. In the CJ-F2 population, both MSH and LSS displayed a continuous and near normal distribution, ranging from 11.0 to 37.0 cm and 2.0 to 7.0 cm, respectively (Table 1, Fig. 2). Furthermore, the correlation coefficient between MSH and LSS in the CJ-F2 population was approximately 0.95, indicating a significant positive correlation. In the CR-F2 population, the range of LSS was slightly narrower, ranging from 2.1 to 5.9 cm, compared to the CJ-F2 population (Table 1, Fig. 2).
High-density SNP maps construction for the two F2 populations
-
The customized 10K MNP array of Brassica oleracea was employed for genotyping the parental lines and F2 individuals. All the SNP markers were assigned to nine linkage groups, corresponding to the nine chromosomes of cauliflower. In the CJ-F2 population, a total of 3,569 filtered SNPs were used to construct a high-density linkage map with a total length of 8,373.66 cM. The average distance between adjacent markers was 2.35 cM, corresponding to a physical distance of approximately 0.141 Mb on the reference genome (Fig. 3).
Figure 3.
High-density genetic maps constructed using a 10K chip for populations CJ-F2 and CR-F2. (a) The distribution of SNP markers on chromosomes in the CJ-F2 population. (b) The distribution of SNP markers on chromosomes in the CR-F2 population. The color represents the number of markers in the 10 cM interval.
Similarly, for the CR-F2 population, a total of 3685 filtered SNPs were used to construct the linkage map, covering a total length of 8,250.01 cM. The genetic distance and physical distances between adjacent markers were consistent with those in the CJ-F2 population, measuring 2.24 cM and 0.135 Mb, respectively (Fig. 3). A comprehensive summary of marker information for each chromosome in both F2 populations was provided in Supplemental Table S1. Notably, 95.32% of the SNPs in both maps had gaps of less than 3.0 cM, indicating a relatively uniform distribution of markers.
QTL identification for curd setting height in the two mapping populations
-
In the CJ-F2 population, the MSH and LSS data were used to detect QTLs associated with the curd setting height trait (Table 2, Fig. 4). A total of seven QTLs for MSH were identified on chromosomes C01, C02, C04, C05, C06, C07 and C09 with an LOD threshold of 4.28. These QTLs explained phenotypic variation (PVE) ranging from 3.05% to 33.0%, with additive effects ranging from −4.37 to 1.79. Among them, the QTL located on C01 (qCJMSH.C01-1) displayed the highest LOD value of 32.80 and explained the largest phenotypic variation of 32.99%. These values were significantly higher than those of the other six QTLs. In the CJ-F2 population, only two QTLs for LSS were detected on C01 and C05, respectively. Among these two loci, qCJLSS.C01-1 locus exhibited a relatively higher LOD value of 14.66 and explained a higher phenotypic variation of 23.48%. Furthermore, this locus overlapped significantly with the QTL qCJMSH.C01-1 controlling the MSH trait on chromosome C01.
Table 2. Summary of the identified QTLs detected.
Trait Population Identified QTL Chr.a Position CIb (cM) Physical interval (kb) LODc PVE (%)d Add.e MSH CJ-F2 qCJMSH.C01-1 C01 1,182 1,181.5−1,182.5 40,144−40,154 32.80 32.99 −4.37 qCJMSH.C02-1 C02 44 41.5−45.5 4,282−4,552 4.47 3.05 −1.35 qCJMSH.C04-1 C04 752 748.5−752.5 52,941−52,862 6.50 3.90 −1.43 qCJMSH.C05-1 C05 130 127.5−130.5 9,370−9,580 4.97 3.76 −1.51 qCJMSH.C06-1 C06 536 535.5−538.5 43,467−43,548 7.33 4.63 −1.73 qCJMSH.C07-1 C07 637 627.5−637.5 47,752−47,545 7.09 5.03 1.79 qCJMSH.C09-1 C09 360 355.5−360.5 19,188−19,188 5.21 4.37 −1.44 LSS CJ-F2 qCJLSS.C01-1 C01 1,205 1,181.5−1,183.5 40,144−40,277 14.66 23.48 −0.57 qCJLSS.C05-1 C05 110 203.5−211.5 6,824−7,020 6.13 8.96 −0.40 CR-F2 qCRLSS.C01-1 C01 777 774.5−779.5 38,823−40,682 5.52 10.02 −0.40 qCRLSS.C09-1 C09 669 664.5−674.5 49,244−49,346 5.52 9.97 −0.40 a Chromosome. b Confidence interval. c Logarithm of odds. d Phenotypic variation explained by the identified QTL. e Phenotypic variation explained by additive effect. Figure 4.
Quantitative trait locus (QTL) mapping and co-localization of the curd setting height-related traits. (a) Main QTL on C01 of MSH trait in CJ-F2 population. (b) Main QTL on C01 of LSS trait in CJ-F2 population. (c) Main QTL on C01 of LSS trait in CR-F2 population.
In the CR-F2 population, only the LSS trait was used for QTL detection (Table 2, Fig. 4). Two QTLs, qCRLSS.C01-1 on C01 and qCRLSS.C09-1 on C09, surpassed the LOD significance threshold of 4.45. Among these, the QTL on C01, which was consistently detected in the CJ-F2 population, was also identified in this population and overlapped with qCRLSS.C01-1 locus. The QTL located on C01 was repeatedly detected in both F2 populations, indicating it as the target region harboring a causal variant for the curd setting height.
Fine mapping of the qCRLSS.C01-1 locus for curd setting height
-
Among the three co-localized QTLs identified for curd setting height, qCJMSH.C01-1 locus had the smallest genetic distance, corresponding to a physical interval of approximately 10 kb (C01: 40,144–40,154 kb), with only two genes located within the interval. The qCJLSS.C01-1 locus had a larger genetic distance, with a corresponding physical interval of approximately 133 kb (C01: 40,144–40,277 kb), encompassing 11 genes within the interval. The qCRLSS.C01-1 locus exhibited the largest genetic distance, corresponding to a physical interval of approximately 1859 kb (C01: 38,823–40,682 kb), and contained 136 genes. To effectively identify candidate genes, qCRLSS.C01-1 locus with the largest physical interval was selected for the subsequent fine mapping, referred to as BocDWARF1 in the following paragraphs.
In the CJ-F2 population, a single plant (147-CJ-F2) with a heterozygous target locus (BocDWARF1) and homozygous other loci were selected for selfing to generate an F2:3 population consisting of 1,580 individuals. Within the target region, nine SNP loci were designed as stable KASP markers for fine mapping. A total of seven recombinants were identified (147-3-F3, 147-498-F3, 147-667-F3, 147-592-F3, 147-989-F3 147-1183-F3 and 147-1289-F3) with homozygous fragments originated from the biparental chromosome at one side of the exchange site and the other side being heterozygous. The offspring of each recombinant above were chosen for the selection of homozygote recombinants. Finally, plants 147-3-8-F4, 147-498-6-F4, 147-667-8-F4 and 147-1183-12-F4, which harbored ZAASJ1401 chromosomal segment within the interval of C01: 38,839−40,094 kb, displayed relatively short MSH (< 19 cm) and LSS (< 3.5 cm). On the other hand, lines 147-592-3-F4 and 147-989-11-F4, carrying ZAASJ1401 chromosomal segment within the interval of C01: 40,298−40,627 kb, exhibited significantly longer MSH (> 26.5 cm) and LSS (> 4.3 cm), indicating the presence of a curd setting height-related QTL in the later interval. Furthermore, the recombinant line 147-1289-2-F4, harboring ZAASJ1401 chromosomal segment within the interval of C01: 40,400−40,627 kb displayed a short MSH (12.7 cm) and LSS (2.8 cm), providing strong support for the region of interest (C1: 40,298−40,400 kb) as a causative region relevant to curd setting height (Fig. 5).
Figure 5.
The BocDWARF1 locus was fine-mapped in the 102 kb region between 40,298,271 bp and 40,400,283 bp. The green bars indicate homozygous segments from ZAASJ1401 with high MSH and red bars indicate homozygous segments from ZAASC4101 with short MSH. The ID indicates the name of recombinants from the F4 population.
Candidate gene analysis in the target region of C1: 40,298–40,400 kb
-
Based on the annotated information of the reference genome, a reduced target interval of ~102 kb (C1: 40,298−40,400) contained a total of 12 genes (Supplemental Table S2). To gain insights into the sequence variations between the parental lines within this target interval, we performed genome resequencing of ZAASC4101 and ZAASJ1401. Following alignment to the HDEM reference genome, we selected 13.36 and 12.15 Gb of clean reads for ZAASC4101 and ZAASJ1401, respectively, for subsequent analysis. These reads corresponded to a genome coverage of ~22× for ZAASC4101 and ~20× for ZAASJ1401. After filtration, a total of 1,139,237 high-quality SNPs and 127,915 InDels were identified between the two parents.
Within the target region of approximately 102 kb, a total of 65 SNPs and five indels were found distributed in the exon regions of the 12 genes. Notably, none of these variants resulted in termination or frameshift mutations. Strikingly, a particular gene (BolC1t04399H) caught our attention as it was a homolog of the classic plant height control gene DWARF1. Thirteen SNPs were identified within the exon region of BolC1t04399H, and four of these SNPs resulted in amino acid differences (Fig. 6, Supplemental Fig. S1). Furthermore, numerous SNPs and Indels were found between the two parents in the 2.0 kb regulatory region upstream and downstream of this gene (unpublished), suggesting potential effects on its transcription level. In summary, although further genetic functional validation is still required, the current results indicated that BolC1t04399H was the most likely candidate gene responsible for controlling the curd setting height trait of cauliflower.
Figure 6.
Amino acid sequence differences of BolC1t04399H between parental lines. The blue background indicates the conservative FAD binding domain. There are a total of four non synonymous SNPs causing amino acid differences.
Development of a KASP marker linked to CSH trait
-
To facilitate selection of high CSH carrying BocDWARF1 in future breeding programmes, a KASP marker for the SNP 1_40298271 that flanks BocDWARF1 was developed. The female line of ZAASC4101 with short MSH and LSS showed the genotype of TT at this locus, while the male line of ZAASJ1401 with long MSH and LSS presented the genotype of CC at this locus. TC was a heterozygous genotype with intermediate MSH and LSS. Among 218 individuals from CJ-F2 population, the MSH and LSS both displayed marked ascending patterns of CC > TC > TT (Fig. 7a, b).
Figure 7.
Association between the genotypes at SNP40298271 locus and curd setting height (CSH) variation. (a), (b) Association analysis between genotypes at SNP40298271 locus (CC, TC and TT) and leaf scar spacing (LSS)/main stem height (MSH) in 218 individuals of CJ-F2 population. (c) Association analysis between genotypes at SNP40298271 locus (CC and TT) and MSH in 38 germplasm accessions.
In addition, to explore the allelic distributions of the SNP 1_40298271 in other germplasm accessions, 38 breeding lines including F7-8 generation inbred lines and double haploid (DH) lines were analysed using KASP. From Fig. 7c, the MSH of the CC genotype was significantly higher than that of TT, indicating the usefulness of this KASP marker in the future marker-assisted selection. However, several accessions with CC genotype at this locus displayed lower MSH as that of TT, suggesting that there might be other loci controlling the CSH of cauliflower.
-
All data generated or analyzed during this study are included in this published article and its supplementary information files.
-
About this article
Cite this article
Sheng X, Cai S, Shen Y, Yu H, Wang J, et al. 2024. QTL analysis and fine mapping of a major QTL and identification of candidate genes controlling curd setting height in cauliflower. Vegetable Research 4: e008 doi: 10.48130/vegres-0024-0002
QTL analysis and fine mapping of a major QTL and identification of candidate genes controlling curd setting height in cauliflower
- Received: 16 October 2023
- Accepted: 19 December 2023
- Published online: 15 March 2024
Abstract: Curd setting height, a critical component of plant height, plays an important role in ideotype construction and yield improvement in cauliflower. However, the underlying mechanisms governing this trait in Brassica crops remain unclear. In this study, we developed two F2 populations by crossing a cauliflower DH line, ZAASC4101, having a short main stem, with two Brassica oleracea inbred lines possessing a long main stem. Subsequently, two high-density linkage maps were constructed for each F2 population using a 10K MNP (multiple nucleotide polymorphism) array developed for Brassica oleracea species. These maps covered genetic distances of 8,373.66 and 8,250.01 cM, respectively. Notably, a major QTL related to curd setting height (BocDWARF1) was consistently identified on chromosome C01 with the highest LOD value in both F2 populations. Through the analysis of the recombinants identified from the F2:3 and F3:4 populations at the BocDWARF1 locus, we successfully fine-mapped this locus to a narrow interval delimited by SNP markers 1_40298271 and 1_40400283. This interval corresponds to an approximately 102-kb region harboring 12 predicted genes. Combined with parental resequencing data with HDEM reference genome annotation information, we proposed BolC1t04399H as the most promising candidate gene. This gene exhibits homology to the well-known plant height-controlling gene DWARF1 and shows a total of four nonsynonymous SNPs in the exon region between the parents. This study lays the foundation for map-based cloning of BocDWARF1 and provides valuable insights for breeding improved cauliflower varieties with the ideal curd setting height.
-
Key words:
- Cauliflower /
- Curd setting height /
- QTL analysis /
- Candidate genes