Integration of small RNAs, degradome and transcriptome sequencing in hyperaccumulator Sedum alfredii uncovers a complex regulatory network and provides insights into cadmium phytoremediation

Abstract

The hyperaccumulating ecotype of Sedum alfredii Hance is a cadmium (Cd)/zinc/lead co-hyperaccumulating species of Crassulaceae. It is a promising phytoremediation candidate accumulating substantial heavy metal ions without obvious signs of poisoning. However, few studies have focused on the regulatory roles of miRNAs and their targets in the hyperaccumulating ecotype of S. alfredii. Here, we combined analyses of the transcriptomics, sRNAs and the degradome to generate a comprehensive resource focused on identifying key regulatory miRNA-target circuits under Cd stress. A total of 87 721 unigenes and 356 miRNAs were identified by deep sequencing, and 79 miRNAs were differentially expressed under Cd stress. Furthermore, 754 target genes of 194 miRNAs were validated by degradome sequencing. A gene ontology (GO) enrichment analysis of differential miRNA targets revealed that auxin, redox-related secondary metabolism and metal transport pathways responded to Cd stress. An integrated analysis uncovered 39 pairs of miRNA targets that displayed negatively correlated expression profiles. Ten miRNA-target pairs also exhibited negative correlations according to a real-time quantitative PCR analysis. Moreover, a coexpression regulatory network was constructed based on profiles of differentially expressed genes. Two hub genes, ARF4 (auxin response factor 4) and AAP3 (amino acid permease 3), which might play central roles in the regulation of Cd-responsive genes, were uncovered. These results suggest that comprehensive analyses of the transcriptomics, sRNAs and the degradome provided a useful platform for investigating Cd hyperaccumulation in S. alfredii, and may provide new insights into the genetic engineering of phytoremediation.

Keywords: Sedum alfredii Hance; cadmium stress; coexpression network; integration analysis; phytoremediation.

Figure 1

Summary of the distribution of the four identified groups of miRNAs in Sedum alfredii. CountsMIRb: the counts of miRNAs from miRBase; Expression level: low indicates <10; middle indicates >10 but less than average; high indicates over average.

Figure 2

Cadmium (Cd)‐responsive miRNAs in Sedum alfredii. (a) Differential expressed miRNAs in eight different Cd treatment durations (0 h, 0.5 h, 6 h, 12 h, 24 h, 48 h, 72 h and 96 h) by hierarchical clustering. Red indicates higher levels of miRNAs and green indicates lower levels of miRNAs. The names of the samples are shown at the top. The original expression values of the miRNAs were normalized using Z‐score normalization. The absolute signal intensity ranges from −3.0 to +3.0, with corresponding colour changes from green to red. (b) The number of differentially expressed miRNAs under Cd stress compared with the control.

Figure 3

Gene ontology (GO) functional classification of identified target genes. (a) Gene ontology classification of target genes for the identified miRNAs, (b) GO enrichment of target genes. The arrows represent the important biological pathways.

Figure 4

A combined view of expressions levels between differentially expressed miRNAs (a) and their target genes (b) in Sedum alfredii at eight different Cd treatment durations. The original expression values of miRNAs and their target genes were normalized by Z‐score normalization.

Figure 5

Expression correlation between miRNAs and their targets at eight different Cd treatment durations. (a ) mtr‐miR166C‐2SS20TC21CA; (b) mtr‐miR172d‐3p_R+1; (c) PC‐3P‐3461203‐2; (d) mtr‐miR164a_1ss17GA; (e) cme‐miR858_L‐1R+1; (f) bra‐miR2111a‐5p; (g) bra‐miR162‐3p_1ss8AG; (h) cpa‐miR167d_R‐1; (i) han‐miR3630‐3p_L‐1_1ss2GA; (j) PC‐5p‐3760_2402; (k) rco‐miR535; (l) mtr‐miR2592bj‐p3_1ss12TC. The thick and thin lines indicate miRNAs and accordingly the target abundance, respectively, based on the RT‐qPCR results.

Figure 6

The coexpression subnetwork of ARF4 and AAP3.