Screening elite B. napus germplasms tolerance to Pb stress
To investigate the tolerance to Pb stress of different B. napus genotypes, the radicle lengths (RL) of 472 accessions grown under 0 or 100 mg/L Pb stress condition for seven days were compared. Although the RL varied significantly among all the accessions under both normal and Pb stress conditions (with a range from 31.15 to 130.50 millimeter (mm), and 8.67 to 80.60 mm, respectively), the RL of all accessions under Pb stress condition were shorter than that under normal condition (Figure 1A, Additional file 1: Figure S1). The average of RL under normal growth condition was 85.18±0.08 mm, whereas the average of RL under Pb stress condition was 39.77±0.05 mm (Figure 1A). This is consistent with previous reports [23, 45].
To eliminate the genetic variations in RLs under normal condition, the relative radicle lengths (RRL) was employed to evaluate the tolerance to Pb stress of B. napus as reported previously [23, 46]. We found that the RRL was ranged from 12.94 to 98.88, 12.17 to 99.84, 20.34 to 98.42 in three replications, respectively (Figure 1B, Additional file 5: Table S1). And the coefficient of variation ranged from 26.37% to 28.57% in three replications (Additional file 5: Table S1). These results indicate that this B. napus population exhibited a broad variation of Pb tolerance.
To select stable Pb-tolerant genotypes for potentially used in phytoremediation or new cultivar breeding, we performed correlation analyses, and found that the RRLs of three replications were significantly correlated with each other with a correlation coefficient value over 0.85 (Figure 1B). Based on the values of RRLs of all the accessions, six Pb-tolerant genotypes (RRL>80%) were selected (Additional file 6: Table S2).
Detection of QTLs associated with Pb tolerance
To select a most suitable model for GWAS analysis of Pb tolerance in the population, the native, population structure (Q), principal component analysis (P), kinship (K), Q+K and P+K models were tested. As shown in quantile-quantile plots (Q-Q) plot, the distribution of observed −log10(p) from Q+K model provided the best fit with the expected distribution (Additional file 2: Figure S2). Therefore, to decrease the rate of false-positive, Q+K model was chosen for subsequent analysis.
Six significantly associated single nucleotide polymorphisms (SNPs) (−log10(p) > 4.3) and three moderately associated SNPs (3.5<−log10(p)<4.3) located on chromosome A09, C03 and C04 were detected (Figure 2). Almost all of them (except for Bn-scaff_16614_1-p658026 and Bn-scaff_18559_1-p175628) were identified in more than two replications, and four out of the nine SNPs were detected in all replications (Table 1). In addition, the significant difference of RRLs between alleles in all nine SNPs were confirmed by t-test (Figure 3).
Further studies with linkage disequilibrium (LD) analyses indicated that these nine associated signals were located in four QTLs. QTL Pb-C03-1 (204.55 kb, position from 1,241,778 bp to 1,446,328 bp on chromosome C03) contained six SNPs, with a peak SNP Bn-scaff_16614_1-p721297 which gave a 5.61% contribution to the phenotypic variance (Figure 4, Table 1). Whereas, QTL Pb-A09 (265.76 kb, position from 8,148,958 bp to 8,414,720 bp on chromosome A09, Additional file 3: Figure S3A), QTL Pb-C03-2 (18.14kb, position from 58,079,114 bp to 58,097,249 bp on chromosome C03, Figure 4) and QTL Pb-CO4 (186.37 kb, position from 14,028,410 bp to 14,214,776 bp on chromosome C04, Additional file 3: Figure S3B) all contained only one associated SNP, and respectively gave a 3.81%, 4.43% and 4.31% contribution to the phenotypic variance (Table 1).
Identification of candidate genes related to Pb tolerance
For the identification of candidate genes related to Pb tolerance, all the 115 genes located in the QTL regions (29, 41, 24 and 21 genes in QTL regions Pb-A09, Pb-C03-1, Pb-C03-2 and Pb-C04, respectively) were annotated, by nucleic acid basic local alignment search tool (BLASTN) with A. thaliana genome and Kyoto Encyclopedia of Genes and Genomes (KEGG) databases. The top 20 enriched metabolic pathways were shown in Additional file 4: Figure S4 and Additional file 7: Table S3. Based on the criterion qvalue ≤ 0.05, three genes, BnaA09g14510D, BnaA09g14520D and BnaA09g14540D, enriched in glutathione metabolism pathway, and three genes, BnaC03g68440D, BnaC03g68450D, and BnaC03g68460D, enriched in the biosynthesis pathway of pantothenate and CoA, as well as in the biosynthesis degradation pathways of valine, leucine and isoleucine were selected for further analyses (Additional file 7: Table S3). The other three candidate genes, BnaC03g02630D, BnaC03g02690D and BnaC04g16200D, which were homologous with AtUBP13 (ubiquitin-specific protease 13), AtTBR (Trichome birefringence) and AtHIPP01 (heavy metal-associated isoprenylated plant protein) respectively, were also selected for further analyses. All these nine candidate genes may contribute to Pb tolerance in B. napus by regulating glutathione metabolism, cell wall development, ubiquitination and amino acid metabolism, respectively (Table 2).
Exploring the expression level of candidate genes
To investigate the expression levels of these candidate genes under both normal and Pb stress condition in both Pb-tolerant and Pb-sensitive accessions, we performed quantitative real time polymerase chain reaction (qRT-PCR) assay. We observed that the expression level of BnaA09g14520D, BnaA09g14520D and BnaA09g14540D located in QTL Pb-A09, and BnaC03g02630D and BnaC03g02690D located in QTL Pb-C03-1, were extremely higher in Pb-tolerant genotypes than in Pb-sensitive genotypes (Figure 5A, B, C, D, E). BnaA09g14520D and BnaC03g02690D were significantly induced by Pb stress only in two Pb-tolerant accessions (Figure 5B, E). BnaA09g14540D and BnaC03g02630D were significantly up-regulated in a Pb-tolerant accession III-229 and only slightly up-regulated in the other accessions under Pb stress (Figure 5C, D).
BnaC03g68440D, BnaC03g68450D and BnaC03g68460D located in QTL Pb-C03-2 were enriched in the same pathways. We found that BnaC03g68440D and BnaC03g68450D were significantly induced by Pb stress in III-229 (Figure 5F, G), and the expression levels of BnaC03g68440D and BnaC03g68450D in Pb-sensitive genotype EH3143 were extensively lower in comparison to Pb-tolerant genotypes (Figure 5F, G). Similarly, a higher expression level of BnaC03g68460D was also observed in the two Pb-tolerant genotypes than in two Pb-sensitive genotypes (Figure 5H). Under Pb stress condition, BnaC04g16200D, located in QTL Pb-C04, was remarkably up-regulated in Pb-tolerant genotype III-229 and down-regulated in Pb-sensitive genotype 6024-1 (Figure 5I).