Dynamic regulation of genome-wide pre-mRNA splicing and stress tolerance by the Sm-like protein LSm5 in Arabidopsis

Abstract

Background: Sm-like proteins are highly conserved proteins that form the core of the U6 ribonucleoprotein and function in several mRNA metabolism processes, including pre-mRNA splicing. Despite their wide occurrence in all eukaryotes, little is known about the roles of Sm-like proteins in the regulation of splicing.

Results: Here, through comprehensive transcriptome analyses, we demonstrate that depletion of the Arabidopsis supersensitive to abscisic acid and drought 1 gene (SAD1), which encodes Sm-like protein 5 (LSm5), promotes an inaccurate selection of splice sites that leads to a genome-wide increase in alternative splicing. In contrast, overexpression of SAD1 strengthens the precision of splice-site recognition and globally inhibits alternative splicing. Further, SAD1 modulates the splicing of stress-responsive genes, particularly under salt-stress conditions. Finally, we find that overexpression of SAD1 in Arabidopsis improves salt tolerance in transgenic plants, which correlates with an increase in splicing accuracy and efficiency for stress-responsive genes.

Conclusions: We conclude that SAD1 dynamically controls splicing efficiency and splice-site recognition in Arabidopsis, and propose that this may contribute to SAD1-mediated stress tolerance through the metabolism of transcripts expressed from stress-responsive genes. Our study not only provides novel insights into the function of Sm-like proteins in splicing, but also uncovers new means to improve splicing efficiency and to enhance stress tolerance in a higher eukaryote.

Figure 1

Generation of the SAD1-overexpressing transgenic plants (SAD1-OE) and the splicing variants of SAD1 in the wild type, sad1 and SAD1-OE. (A) Morphology of the wild type, sad1 and SAD1-OE seedlings in soil. (B) Genotype analysis of plants shown in (A). The upper and lower bands of PCR products represent the endogenous SAD1 gene and the transgenic cDNA, respectively. (C) RNA-seq reads were visualized by the Integrative Genomics Viewer (IGV) browser across the SAD1 gene. Exon-intron structure was given at the bottom of each panel. The arcs generated by IGV browser indicate splice junction reads that support the splice junctions. The grey peaks indicate RNA-seq read density across the gene. The upper panel depicts the mutation of sad1 that changed the wild-type invariant dinucleotide AG to AA at the 3′ splicing acceptor recognition site of the first intron. The middle panel shows transcripts with two aberrant 3′ splice sites (3′SSs) that respectively occurred at the 20 bp (enlarged and marked by 3) and 2 bp (enlarged and marked by 2) downstream of the mutated splice site and transcripts with the retention of the first intron (marked by 1) in sad1. Also shown are SAD1 transcripts in the wild type where they were normally spliced. (D) Three variants of SAD1 transcripts discovered in sad1 by RNA-seq were validated by RT-PCR using junction-flanking primers. The three bands in the sad1 mutant from top to bottom represent transcripts with the first intron retained, the first aberrant 3′SSs and the second aberrant 3′SSs, respectively. Note in the wild-type and SAD1-OE plants, only one wild-type SAD1 band was detected. (E)SAD1 expression levels were shown using the reads per kilobase per million value and quantitative RT-PCR. bp, base pairs; RPKM, reads per kilobase per million; sad1, sad1 mutant; SAD1-OE, plants over-expressing wild-type SAD1 in the sad1 mutant background; WT, wild type.

Figure 2

Comparison of global alternative splicing between the wild type and the sad1 mutant. (A) The counts of each type of AS events in the wild type and sad1. The green/blue bars represent forward and reverse sequencing reads. (B) The total counts of the splice junction reads from each type of AS in the wild type and sad1. The P values were calculated by Fisher’s exact test comparing the junction read counts and the uniquely mapped reads between the wild type and sad1. (C) Three representative AS events validated by RT-PCR and visualized by IGV browser. For the validation of alternative 5′SSs and 3′SSs, there was only one band that represented the alternative-splice isoform, which was obviously detected in sad1 mutants, but not in the wild type and SAD1-OE. For exon-skipping events, the grey asterisk (*) to the right side denotes the alternative splice form. For the IGV visualization, exon-intron structure of each gene was given at the bottom of each panel. The arcs generated by IGV browser indicate splice junction reads that support the junctions. The grey peaks indicate RNA-seq read-density across the gene. The upper, middle and lower panels show the indicated genes with alternative 5′SSs, alternative 3′SSs and exon-skipping, respectively. These events were marked by red arrows and highlighted by red arcs. (D) The sequences around the alternative 5′SSs and 3′SSs that were over-represented in the mutant are shown by Weblogo. (E) Distribution of activated alternative 5′SSs and 3′SSs around the dominant ones. These alternative 5′SSs and 3′SSs were enriched in the downstream or upstream 10 bp region of the dominant 5′SSs and 3′SSs (position 0 on the x- axis), respectively. AS, alternative splice; sad1, sad1 mutant; SAD1-OE, plants over-expressing wild-type SAD1 in the sad1 mutant background; WT, wild type.

Figure 3

Comparison of intron retention between the wild type and sad1. The RPKM values for the exons and introns were plotted between the wild type and sad1. The expression of introns in the sad1 mutant shows a global up-regulation, but not that of exons. RPKM, reads per kilobase per million; sad1, sad1 mutant; WT, wild type.

Figure 4

Genes with abnormal splicing in sad1 are closely associated with stress response and transcriptional activation. (A) A two-dimensional view of the relationship between the genes with abnormal splicing and their functional annotations generated by the DAVID software. The top 50 functional annotations that were ordered by the enrichment scores were selected for the two-dimensional view, which indicates that genes with abnormal splicing were strikingly enriched (colored green) in the response-to-abiotic-stress category. (B) A heatmap was generated by mapping the genes enriched at the response-to-abiotic-stress pathways to the microarray database using Genevestigator. The heatmap indicates that genes with abnormal splicing in sad1 are mostly up-regulated (colored red) by ABA, cold, drought and salt stress but less regulated by biotic stress (bacteria infection). (C) A network generated by Mapman indicates that genes with aberrant splicing in sad1 are involved in various stress response pathways, including hormone-signaling pathways, MAPK-signaling pathways and transcription regulation. (D) Validation of the intron retention in 10 stress-responsive genes by RT-PCR using the intron-flanking primers. The grey asterisks (*) denote the intron-retained splicing variants. ABA, abscisic acid; SA, salicylic acid; JA, jasmonic acid; sad1, sad1 mutant; WT, wild type; HSP, heat shock protein; MAPK, mitogen-activated protein kinase; ERF, ethylene response factor; bZIP, basic-leucine zipper; WRKY, WRKY transcription factor; DOF, DNA-binding with one finger; PR-proteins, pathogenesis-related proteins; R genes, (plant disease) resistance genes.

Figure 5

Comparison of alternative splicing between the wild-type and SAD1-OE plants. (A) Profiling the normalized (by total uniquely mapped reads) read coverage of the splice junctions that were over-represented in the sad1 mutant relative to the wild type and SAD1-OE. The profiles indicate that the AS patterns in sad1 were completely or largely restored by overexpressing SAD1. (B) The counts of each type of AS event in the wild type and SAD1-OE. The green/blue bars represent forward and reverse sequencing reads. (C) The total counts of the splice junction reads from each type of AS in the wild type and SAD1-OE. The P values were calculated by Fisher’s exact test comparing the junction read counts and the uniquely mapped reads between the wild type and SAD1-OE. (D) The sequences around the alternative 5′SSs and 3′SSs that were absent in the SAD1-OE were shown by Weblogo. (E) Distribution of activated alternative 5′SSs and 3′SSs around the dominant ones are shown. These alternative 5′SSs and 3′SSs were enriched in the downstream or upstream 10 bp region of the dominant 5′SSs and 3′SSs (position 0 on the x-axis), respectively. (F) Profiling the normalized (by total uniquely mapped reads) read coverage of the introns that were over-represented in the sad1 mutant relative to the wild type and SAD1-OE. The profiles indicate that the intron retention in sad1 was largely restored by overexpressing SAD1. AS, alternative splice; sad1, sad1 mutant; SAD1-OE, plants over-expressing wild-type SAD1 in the sad1 mutant background; WT, wild type.

Figure 6

Genes with increased splicing efficiency in SAD1-OE plants are related to stress response and overexpression of SAD1 improves salt stress tolerance. (A) A two-dimensional view of the functional annotation of genes with increased splicing efficiency in SAD1-OE. The top 50 functional annotations that were ordered by the enrichment scores were selected for two-dimensional view, which indicates that genes with increased splicing efficiency were strikingly enriched (green) in the response-to-abiotic-stress pathways. (B) A heatmap was generated by mapping the genes enriched at the response-to-abiotic-stress pathways to the microarray database using Genevestigator. The heatmap indicates that genes with abnormal splicing are closely associated with stress responses and are up-regulated (red) by the indicated stresses. (C) Increased salt tolerance in seedlings overexpressing SAD1. Twelve-day-old seedlings on the regular ½ Murashige and Skoog (MS) medium were transferred to ½ MS media supplemented with the indicated concentrations of NaCl. The pictures were taken four days after the transfer. (D) Percentage of green leaves of seedlings on 200 mM NaCl media. Two-week-old seedlings grown on ½ MS media were transferred to ½ MS medium plates supplemented with 200 mM NaCl and incubated for five days before counting the number of green leaves or yellow and bleached leaves. A total 36 seedlings for each genotype were counted. Data are averages and standard deviations. Averages with different letters are statistically different (P <0.01, t-test). (E) Morphology of 28-day-old wild-type, sad1 and transgenic plants (SAD1-OE) that were subjected to four-day treatment with 400 mM NaCl solution. Also shown at the bottom are pictures of the damaged inflorescent stem and leaf seen in the wild type compared to undamaged ones in the SAD1-OE. ABA, abscisic acid; sad1, sad1 mutant; SAD1-OE, plants over-expressing wild-type SAD1 in the sad1 mutant background; WT, wild type.