Hybrid sequencing reveals insight into heat sensing and signaling of bread wheat

Abstract

Wheat (Triticum aestivum L.), a globally important crop, is challenged by increasing temperatures (heat stress, HS). However its polyploid nature, the incompleteness of its genome sequences and annotation, the lack of comprehensive HS-responsive transcriptomes and the unexplored heat sensing and signaling of wheat hinder our full understanding of its adaptations to HS. The recently released genome sequences of wheat, as well as emerging single-molecular sequencing technologies, provide an opportunity to thoroughly investigate the molecular mechanisms of the wheat response to HS. We generated a high-resolution spatio-temporal transcriptome map of wheat flag leaves and filling grain under HS at 0 min, 5 min, 10 min, 30 min, 1 h and 4 h by combining full-length single-molecular sequencing and Illumina short reads sequencing. This hybrid sequencing newly discovered 4947 loci and 70 285 transcripts, generating the comprehensive and dynamic list of HS-responsive full-length transcripts and complementing the recently released wheat reference genome. Large-scale analysis revealed a global landscape of heat adaptations, uncovering unexpected rapid heat sensing and signaling, significant changes of more than half of HS-responsive genes within 30 min, heat shock factor-dependent and -independent heat signaling, and metabolic alterations in early HS-responses. Integrated analysis also demonstrated the differential responses and partitioned functions between organs and subgenomes, and suggested a differential pattern of transcriptional and alternative splicing regulation in the HS response. This study provided comprehensive data for dissecting molecular mechanisms of early HS responses in wheat and highlighted the genomic plasticity and evolutionary divergence of polyploidy wheat.

Keywords: alternative splicing regulation; early heat stress; heat sensing and signaling; hybrid sequencing; spatio-temporal transcriptome; transcriptional regulation; wheat (Triticum aestivum L.).

Figure 1

Experimental workflow. (a) Wheat plants (T. aestivum cv. Chinese Spring) at 15 days after anthesis were subjected to heat stress (HS) (37°C). The flag leaves and filling grain were sampled at 0 min, 5 min, 10 min, 30 min, 1 h and 4 h under HS. In total, 36 samples (six time points for each of the two organs, three biological replicates per time point) were sequenced using second‐generation sequencing, and two mixed samples (the RNAs of 18 samples from each organ mixed in equal volume) were sequenced using third‐generation sequencing. (b) Bioinformatics pipeline for analyzing the hybrid sequencing data. Sequencing errors in full‐length and non‐chimeric (FLNC) reads were corrected with short reads. FLNC reads before and after error correction were mapped in parallel to IWGSC RefSeq v1.0, and the read with the best genomic match (see Experimental procedures) was retained in the downstream analysis. (c) Comparison of the transcript length between the IWGSC RefSeq v1.0 annotation and the PacBio data. (d) Comparison of the isoform number between the IWGSC RefSeq v1.0 annotation and the PacBio data. (e) Structure comparison of the IWGSC RefSeq v1.0 and PacBio transcripts.

Figure 2

CIRCOS visualization of different data at the genome‐wide level. The density was calculated in a 10‐Mb sliding window. (a) Karyotype of the wheat genome. (b) Comparison of transcript density between the IWGSC RefSeq v1.0 annotation and the PacBio data. From the upper to lower tracks: transcripts in IWGSC RefSeq v1.0, transcripts in grain and transcripts in leaves. (c–g) Distribution of heat stress (HS)‐responsive genes following HS treatment for 4 h, 1 h, 30 min, 10 min and 5 min. From the upper to lower tracks in each part: the HS‐responsive genes in grain with transcriptional regulation, the HS‐responsive genes in leaves with transcriptional regulation, the HS‐responsive genes in grain with alternative splicing (AS) regulation and the HS‐responsive genes in leaves with AS regulation. (h, i) Distribution of lncRNAs in grain (h) and leaves (i). (j) Linkage of fusion transcripts: intra‐chromosome (green), inter‐chromosome in the same subgenome (red) and inter‐chromosome in different subgenomes (blue).

Figure 3

Identification of differentially expressed genes (DEGs) and differentially spliced genes (DSGs) at each time point in leaves and grain. Number of DEGs (a) and DSGs (c). The x‐axis represents the heat stress (HS) treatment time points and the y‐axis represents the HS‐responsive gene number. Light blue and light red represent the number of DEGs and DSGs that were newly discovered loci from the PacBio data, respectively. (b) Number of different HS‐response genes between leaves and grain. Light blue represents newly discovered loci from the PacBio data. (d) Venn diagram of DSGs in leaves and grain.

Figure 4

Rapid changes of differentially expressed genes (DEGs) and differentially spliced genes (DSGs) in response to heat stress (HS). (a, b) Histograms plots of the time points at which the DEGs (a) and DSGs (b) first showed a significant difference in leaves and grain. Each gene is represented only once in each histogram. (c) Frequency over time of isoform switches (where the relative abundance of different isoforms is reversed in response to HS) in the time‐course transcriptomes. (d, e) Expression profiles of pre‐mRNA‐splicing factor SF2 (d) and HSFA6 (e). The isoform switch events were marked with black circles. For clarity, only the transcripts that were involved in isoform switches were plotted. The isoforms whose names start with ‘chr’ were novel isoforms identified from the PacBio data.

Figure 5

Differences in timing of diverse transcription factors (TFs) in the heat signaling and early heat response. (a, b) Heatmaps showing the fold enrichment of TFs in response to heat stress (HS) with transcriptional regulation (a) and alternative splicing regulation (b). Only significantly enriched families (q < 0.05) are indicated. The x‐axis represents HS‐treated samples and the y‐axis represents the TF families. ‘L’ and ‘G’ in the sample names represent leaf and grain, respectively.

Figure 6

Heat signaling and early heat response processes. (a, b) Heatmaps showing the fold enrichment of enriched KEGG pathways for differentially expressed genes (DEGs) (a) and differentially spliced genes (DSGs) (b). Only significantly enriched pathways (q < 0.05) are indicated. The full list of enriched pathways is presented in Tables S16 and S18. The x‐axis represents HS treatment time points and the y‐axis represents enriched KEGG pathways. ‘L’ and ‘G’ in the sample names represent leaf and grain, respectively.

Figure 7

Overlap of differentially expressed genes (DEGs) and differentially spliced genes (DSGs), and a heat shock factors coding gene that responds to heat stress with both transcriptional regulation and alternative splicing regulation. (a) Overlap of DEGs and DSGs in leaves. (b) Overlap of DEGs and DSGs in grain. (c) Schematic representation of the isoforms produced by TraesCS5D01G393200 (a member of the HSFA2 subfamily). Exons are represented as blue boxes and introns as lines. The isoforms whose names start with ‘Chr’ were novel isoforms identified from the PacBio data. The red triangle indicates the location of an in‐frame premature termination codon in the intron. For the right part, the numbers in rectangles represent average FPKM values of three replicates at each time point in leaves. The heatmap shows the fold change of each isoform at different time points (0 min time point was used as a control).

Figure 8

Subgenome bias in the heat stress response. (a, b) Heatmaps showing the fold change in the expression of each gene in each homologous triplet at the 30 min time point in leaves (a) and at the 1 h time point in grain (b). Only the triplets that contain differentially expressed genes are displayed. For the symbols on the x‐axis, ‘A’, ‘B’ and ‘D’ represent the A‐, B‐ and D‐homeologues in the triplets, respectively. ‘L’ and ‘G’ represent the leaves and grain, respectively. In the right part, the heatmaps display the pairwise ratio of the fold changes between the A‐, B‐ and D‐homeologues. The green, purple and orange represent the different responses of the A‐, B‐ and D‐homeologues in each pairwise comparison. A versus B: the comparison between the A‐ and B‐homeologues, A versus D: the comparison between the A‐ and D‐homeologues, and B versus D: the comparison between the B‐ and D‐homeologues. (c, d) Heatmaps display the differentially spliced genes (DSGs) in each homologous triplet at the 30 min time point in leaves (c) and at the 1 h time point in grain (d). Only the triplets that contain DSGs are displayed. The symbol ‘1’ indicates that the gene is a DSG and the symbol ‘0’ indicates that the gene is not a DSG.