Cold-Dependent Expression and Alternative Splicing of Arabidopsis Long Non-coding RNAs

Abstract

Plants re-program their gene expression when responding to changing environmental conditions. Besides differential gene expression, extensive alternative splicing (AS) of pre-mRNAs and changes in expression of long non-coding RNAs (lncRNAs) are associated with stress responses. RNA-sequencing of a diel time-series of the initial response of Arabidopsis thaliana rosettes to low temperature showed massive and rapid waves of both transcriptional and AS activity in protein-coding genes. We exploited the high diversity of transcript isoforms in AtRTD2 to examine regulation and post-transcriptional regulation of lncRNA gene expression in response to cold stress. We identified 135 lncRNA genes with cold-dependent differential expression (DE) and/or differential alternative splicing (DAS) of lncRNAs including natural antisense RNAs, sORF lncRNAs, and precursors of microRNAs (miRNAs) and trans-acting small-interfering RNAs (tasiRNAs). The high resolution (HR) of the time-series allowed the dynamics of changes in transcription and AS to be determined and identified early and adaptive transcriptional and AS changes in the cold response. Some lncRNA genes were regulated only at the level of AS and using plants grown at different temperatures and a HR time-course of the first 3 h of temperature reduction, we demonstrated that the AS of some lncRNAs is highly sensitive to small temperature changes suggesting tight regulation of expression. In particular, a splicing event in TAS1a which removed an intron that contained the miR173 processing and phased siRNAs generation sites was differentially alternatively spliced in response to cold. The cold-induced reduction of the spliced form of TAS1a and of the tasiRNAs suggests that splicing may enhance production of the siRNAs. Our results identify candidate lncRNAs that may contribute to the regulation of expression that determines the physiological processes essential for acclimation and freezing tolerance.

Keywords: alternative splicing; cold transcriptome; diel time-course; high-resolution RNAseq; long non-coding RNA; primary microRNA.

Figure 1

Expression and alternative splicing of lncRNAs and pri-miRNAs in response to cold. The number of (A) lncRNA and (B) pri-miRNA genes in RTD2 (Zhang et al., 2017) according to their expression in the RNA-seq time course and identifying the number which are differentially regulated by cold at the expression (DE) and/or alternative splicing level (DAS).

Figure 2

Expression of lncRNA transcripts in response to cold. Transcripts below 1.85 TPM in all time-points are not shown (except in E, where we used a 0.5 TPM cut-off value). Transition to cold is represented by a vertical dashed blue line at time-point 9. (A) The single transcript of AT2G15128 is DE showing a significant upregulation upon long exposure to cold. (B) AT5G59662_ID1 transcript is DE showing a significant downregulation upon cold and loss of a high amplitude rhythm is also observed. (C) The AT1G68568_ID2 transcript is DE showing a significant upregulation throughout the cold treatment and a gain of a high amplitude rhythm is also observed. (D) AT1G22403 (DAS-only) undergoes splicing regulation rapidly in the cold and the relative abundance of the AT1G22403.2 transcript is maintained in day 4 at 4°C (adaptive) whereas the total gene level is not significantly affected by cold. (E) AT2G31751 and (F) AT1G25098 are DE and DAS showing a significant downregulation upon cold. (G) AT1G34418 (DAS-only) undergoes only significant splicing regulation showing a rapid decrease in abundance of AT1G34418.1 and increasing expression during the day. (H) Rhythmic expression of AT1G53233 (not affected by cold). Rapid and significant isoform switches detected by TSIS (Guo et al., 2017) are labeled with a red circle. NAT lncRNA genes (B,C,E,F,H); sORF (G); other RNA (A,D).

Figure 3

Expression of pri-miRNA transcripts in response to cold. Transcripts below 1.85 TPM in all time-points are not shown (except in H, where we used a 1 TPM cut-off value). Transition to cold is represented by a vertical dashed blue line at time-point 9. (A) pri-miR159a is DE showing a significant upregulation upon cold and a gain of a high amplitude rhythm is also observed. (B) pri-miR5014a, is an intronic miRNA in the GLUTAMATE DECARBOXYLASE 2 gene (AT1G65960); the profile of the pre-mRNA is DE showing a significant upregulation only in the first day of cold treatment. (C) pri-miR5640 is an intronic miRNA in CALLOSE SYNTHASE 1 (AT1G05570) which is DE in the cold, while upon longer exposure to cold (day 4) rhythmicity is lost/dampened. (D) pri-miR414 is rhythmically expressed and not significantly affected by cold. (E) pri-miR162a is DE and DAS showing a significant downregulation throughout the cold treatment. (F) pri-miR4245 is intronic in an AGENET DOMAIN-CONTAINING PROTEIN gene (AT5G52070) which undergoes splicing regulation with no significant expression change at the gene level. (G) pri-miR1888a is an intronic miRNA in an L-ASCORBATE OXIDASE gene (AT5G21100) that is downregulated at the transcriptional level and undergoes differential alternative splicing. (H) pri-miR3434 is an intronic miRNA in an ALPHA-ADAPTIN gene (AT5G22770) which undergoes only splicing regulation. Rapid and significant isoform switches detected by TSIS (Guo et al., 2017) are labeled with a red circle.

Figure 4

Cold-induced alternative splicing of Trans-acting siRNA 1a. (A) Transcript structures of TAS1a isoforms showing the intron and the position in the intron of the region containing the miR173 binding site and phased cleaved siRNAs. The up arrow shows the cleavage site of the TAS1a RNA by miR173 and the down arrows the cleavage sites which release the various siRNAs (Allen et al., 2005). (B) Expression profile of TAS1a transcript isoforms and gene; TAS1a is DAS-only (no significant change in gene level expression). (C) 5-week-old Arabidopsis rosettes harvested at dawn (arrow) after 12 h of variable reductions in temperature. Reductions (Δ) of 2, 4, 8, 12, and 16°C in temperature applied at dusk provides evidence of temperature-sensitive, long-term changes in AS. (D) Unspliced (intron retention – IR) of TAS1a (AT2G27400_ID1) is sensitive to reductions in temperature of 8°C. Student’s t-tests were performed to compare each temperature reduction results against 20°C control. (E) 5-week-old Arabidopsis rosettes harvested rapidly after transfer to cold. The temperature was gradually reduced from 20°C at 0 h to 11°C at 40 min and eventually 4°C at 120 min into the cold treatment; rosettes were harvested across the first 3 h of cold at the times shown allowing the measurement of the speed of transcriptional and AS changes due to temperature reduction. (F) The unspliced (IR) transcript of TAS1a (AT2G27400_ID1) responded rapidly to cold within 90 min, when the temperature reaches 5°C. RT-qPCR was used to measure relative expression levels for data presented in D and F, see Section “Materials and Methods.” Student’s t-tests were performed to compare each temperature reduction results against 20°C control. Significant differences are labeled with asterisks (^∗∗p < 0.01; ^∗∗∗p < 0.001).

Figure 5

Cold-induced reduction of miR173 and TAS1a-derived siRNAs. Small RNA gel blot analysis was performed using RNA from rosette leaves harvested at different time-points after application of cold. Samples were hybridized with radioactively labeled probes specific to miR173 and the different siRNAs and bands were quantified. (A) miR173, (B) siR752, (C) siR255, and (D) siR477. The temperature at the different time-points was 20°C (0 min), 5°C (90 min), and 4°C (120 and 180 min). The letters on the bars indicate significantly different expression levels of small RNAs using Student’s t-test (p < 0.05); different letters mean significant differences between the samples.

Figure 6

Rapid cold-induced alternative splicing of lncRNAs. (A) Transcript structures of AT1G34844 isoforms showing the fully spliced transcript (AT1G34844_ID2), AT1G34844.1 which retains introns 2 and 3, and AT1G34844_ID4 which retains intron 2. (B) Transcript structures of AT3G26612 isoforms showing the fully spliced transcript (AT3G26612.1), AT3G26612_c1 which has an alternative 5′ splice site in exon 3, and AT3G26612_ID4 which retains intron 3. (C) AT1G34844 is regulated by alternative splicing only. AT1G34844_ID2 and AT1G34844.1 show a transient decrease and increase, respectively, between 20°C and day 1 at 4°C while AT1G34844_ID4 remains unchanged throughout the experiment. (D) AT3G26612 is regulated by alternative splicing only. AT3G26612.1 significantly decreases in the first 6 h (day 1) and throughout the cold treatment, while the other two transcripts have similar levels of expression, with AT3G26612_c1 slightly increasing and AT3G26612_ID4 remaining unchanged. (E,F) High-resolution RT-PCR analysis of splicing ratios at 20°C (control – left panels) and decreasing to 4°C (right panels). (E) The fully spliced and I2R&I3R transcripts of AT1G34844 respond rapidly to changes in temperature within 40 min when the temperature reaches 11°C and onward as the temperature decreases further. (F) The fully spliced and Alt5′ssE3 transcripts of AT3G26612 respond rapidly to changes in temperature within 40 min when the temperature reaches 11°C and onward as the temperature decreases further, while I3R is mostly unresponsive. Student’s t-tests were performed to compare each temperature reduction results against 20°C control. Significant differences are labeled with asterisks (^∗∗p < 0.01; ^∗∗∗p < 0.001).