Identification of low temperature resistance-associated loci containing MIRNA genes in B. napus using Genome-wide association study
B. napus is closely related to B. rapa and both areBrassica crops. To identify MIRNA genes associated with low temperature stress in Brassica crops, we constructed a B. napus population with 220 accessions. The phenotype of low-temperature resistance was divided into 5 levels (Grade 1 indicates most cold-resistant, while Grade 5 indicates most cold-sensitive) and was measured in the field in January of 2017 and 2018, and the phenotypic data of rapeseed population showed a nearly skewed distribution (Xu et al., 2021). The genome re-sequencing of these 220 B. napusaccessions generated 2.13 Tb clean reads, and single nucleotide polymorphism (SNP) genotyping of B. napus population identified 3.80 million highly quality SNPs (MAF > 0.05, geno < 0.2) (Xu et al., 2021). Overall, the SNPs were not evenly distributed across the whole genome. The SNP density in the A sub-genome (5.25 SNP/Kb) was higher than that in the C sub-genome (4.74 SNP/Kb) (Figure S3A). A principal component analysis (PCA) was used to assess the genetic relationship in the association panel. Based on the PCA, theB. napus population was roughly divided into three groups corresponding to winter (W), semi-winter (SW), and spring (S) ecotypes (Figure S3B). We further analyzed the linkage disequilibrium (LD) throughout the genome. The LD was estimated asr2 (the squared Pearson correlation coefficient) between all pairs of SNP markers. The average distance over which LD decayed to r2 =0.2 of its maximum value was 100 kb, indicating that SNPs located in the 200-kb genomic region around each peak SNP represented a GWAS-quantitative trait locus (QTL) (Figure S3C).
The GWAS was performed using a mixed linear model, and the SNPs significantly associated with low-temperature resistance were identified at a threshold of p < 10-6. As a result, we identified 32 highly credible GWAS-QTLs (Figure 3A). To discover candidate MIRNA genes within these QTLs, we collected 163 pre-miRNA sequences specific to the Cruciferae family from four species (B. napus , B. rapa , B. oleracea, andArabidopsis ) according to previous reports (Jones-Rhoades and Bartel, 2004; Alves-Junior et al., 2009; Yu et al., 2011; He et al., 2018; Zhang et al., 2018; Li et al., 2019). All the pre-miRNA sequences were mapped to the B. napus reference to identify their location (Table S4). We found that four candidate MIRNA genes,bna-MIR166C , bna-MIR1885 , bna-MIR168A andbna-MIR845A, were in the QTLs associated with low-temperature resistance.
A previous study proved that knock-down of miR165/166 expression conferred a cold-resistant phenotype in A. thaliana (Yan et al., 2016). Conserved miR845a was known to be differentially expressed in leaves of winter turnip rape under cold stress (Zeng et al., 2018). To further determine whether these four miRNAs were responsive to low temperature, we measured the abundance of mature miRNAs in B. napus under cold stress by qRT-PCR. From this analysis, we found that all of them were cold-responsive. bna-miR1885 and bna-miR168a were up-regulated under cold stress compared to under normal condition at the same growth time points, while bna-miR845a and bna-miR166c were down-regulated (Figure 3B and Figure 4A).
To identify if any genetic variation in cis-regulatory element contribute to the accumulation of mature miRNAs in the population of rapeseed, we studied the LD of SNPs in bna-MIR1885, bna-MIR166C, bna-MIR168A and bna-MIR845A gene locus, respectively. We found that the SNPs in their corresponding genes exhibited significant linkage (Figure S3D, F, G, H). Importantly, we identified 22 SNPs and 7 INDELs inbna-MIR1885 promoter region, which were found to be significant associated with cold resistance. We further investigated the five variants located in the 200-bp genomic region around the low temperature-responsive (LTR: CCGAAA) element identified above. TheB. napus accessions were classified into three haplotype groups based on the genotypes of these variants (Figure 3C). We found that 179 accessions contained the H1 haplotype, which variants were identical with the genome reference, and this haplotype was associated with low temperature-sensitive phenotype. 24 accessions carried H2 haplotype, which contained mutated LTR cis-element and was associated with low temperature-resistant phenotype. Consistently, H3 haplotype had an average resistance degree, which indicated more resistant than H1 haplotype, and less resistant than H2. Next, we randomly selected 4 accessions with low temperature-resistant variant alleles and low temperature-sensitive variant alleles, respectively, to measure their abundance of miR1885 after cold treatment by qRT-PCR. The qRT-PCR data indicated that the abundance of miR1885 was significantly higher in low temperature-sensitive accessions with normal LTR in the promoter (H1 haplotype), as compared to low temperature-resistant accessions with mutated LTR in their promoter (H2 haplotype) (Figure S3E). In total, these results suggested that genetic variation in the LTR region atMIR1885 promoter in the rapeseed population potentially affected the miR1885 expression and were associated with cold resistance.