STABILIZED1 Modulates Pre-mRNA Splicing for Thermotolerance^,

Heat-inducible STA1 activity is involved in the pre-mRNA splicing of heat stress response genes and contributes to the establishment of heat stress tolerance in Arabidopsis.

Abstract

High-temperature stress often leads to differential RNA splicing, thus accumulating different types and/or amounts of mature mRNAs in eukaryotic cells. However, regulatory mechanisms underlying plant precursor mRNA (pre-mRNA) splicing in the environmental stress conditions remain elusive. Herein, we describe that a U5-snRNP-interacting protein homolog STABILIZED1 (STA1) has pre-mRNA splicing activity for heat-inducible transcripts including HEAT STRESS TRANSCRIPTION FACTORs and various HEAT SHOCK PROTEINs for the establishment of heat stress tolerance in Arabidopsis (Arabidopsis thaliana). Our cell-based splicing reporter assay demonstrated STA1 acts on pre-mRNA splicing for specific subsets of stress-related genes. Cellular reconstitution of heat-inducible transcription cascades supported the view that STA1-dependent pre-mRNA splicing plays a role in DREB2A-dependent HSFA3 expression for heat-responsive gene expression. Further genetic analysis with a loss-of-function mutant sta1-1, STA1-expressing transgenic plants in Col background, and STA1-expressing transgenic plants in the sta1-1 background verified that STA1 is essential in expression of necessary genes including HSFA3 for two-step heat stress tolerance in plants. However, constitutive overexpression of the cDNA version of HSFA3 in the sta1-1 background is unable to execute plant heat stress tolerance in sta1-1. Consistently our global target analysis of STA1 showed that its splicing activity modulates a rather broad range of gene expression in response to heat treatment. The findings of this study reveal that heat-inducible STA1 activity for pre-mRNA splicing serves as a molecular regulatory mechanism underlying the plant stress tolerance to high-temperature stress.

The splicing of precursor mRNA (pre-mRNA) is a necessary step for intron-containing gene expression in eukaryotic cells to produce mature transcripts for protein translation (Wahl et al., 2009). This process is highly ordered and tightly controlled by multisubunit spliceosome activity to mix and match introns and exons of pre-mRNAs. The high molecular weight spliceosome complex comprises small nuclear ribonucleoprotein particles (snRNPs) called U1, U2, U4/U6, and U5 snRNPs. For splicing of pre-mRNA introns, U1 snRNP recognizes the 5′-splicing site, and U2 snRNP binds to the adenosine at the branch point of introns with the assistance of U2 auxiliary factors. U4/U6 and U5 trimeric snRNPs associate with each other and undergo a stepwise 3′-splicing site cleavage process. Eventually, U5 snRNP dissociates from the complex along with a lariat form of the intron. In this process, U5 snRNP accurately and dynamically swaps interacting partners with other snRNP subunits (Wahl et al., 2009).

The functions of ribonucleoprotein particles are evolutionarily conserved in most eukaryotes (Knowler and Wilks, 1980). Recent advances in functional genomic analysis platforms enabled the inference that alternative splicing of pre-mRNAs is a central regulatory module in extending the gene reservoir with a limited number of structural genes encoded in genomes for protein information (Reddy et al., 2013). Consequently, a few genomic resources address physiological and developmental complexities through transcript variants produced in cellular and environmental contexts (Reddy et al., 2013; Staiger and Brown, 2013).

The sessile lifestyle of plants facilitates the evolution of diverse adaptation processes at the multidimensional layers from cells to organisms under various environmental stress conditions (Godfray and Garnett, 2014; McClung, 2014). For example, in the cell autonomous manner, low temperature activates the ICE-CBF-COR pathway and establishes cold stress tolerance in Arabidopsis (Arabidopsis thaliana; Zhu et al., 2007). Such cellular adaptation processes also require posttranscriptional regulatory steps in gene expression. For example, COR15A expression was originally reported to be transcriptionally upregulated and further modulated posttranscriptionally upon cold treatment (Artus et al., 1996; Lee et al., 2006). In a recessive loss-of-function mutant sta1-1, pre-mRNA of COR15A accumulated in response to the cold stress and thus STA1 is implicated in controlling pre-mRNA splicing of COR15A (Lee et al., 2006).

STA1 is a stress-inducible U5 snRNP-interacting partner, of which cellular functions are involved in the establishment of cold stress and ABA tolerance (Lee et al., 2006). Recently, STA1 was also documented for heat stress responses without mechanistic details (Yu et al., 2016). High temperature (heat) is a detrimental environmental stress condition affecting plant biomass yields and its occurrence becomes more prevalent in the cropping fields under today’s climate change (Godfray and Garnett, 2014; McClung, 2014). Evolutionarily conserved heat stress transcription factors (HSFs) and heat shock proteins (HSPs) play key roles in establishing plant basal and/or acquired heat stress tolerance (Kotak et al., 2007; Scharf et al., 2012). For example, the gene regulatory modules of HSF and HSP expression have been well characterized in Arabidopsis. The key stress-related transcription factors DEHYDRATION-RESPONSIVE ELEMENT-BINDING PROTEIN (DREB) 2A and DREB2C, induce expression of heat-inducible transcription factor HEAT SHOCK TRANSCRIPTION FACTOR A3 (HSFA3), and its transcription activity induces downstream HSPs expression (Sakuma et al., 2006; Schramm et al., 2008; Chen et al., 2010). A similar transcription cascade is also found in the maize, and thus this transcriptional regulatory module is most likely conserved through the evolution of diverse plant species (Qin et al., 2007). Nevertheless, posttranscriptional regulatory steps of heat-inducible gene expression remain largely unknown in plants.

In this study, we used a (to our knowledge) novel splicing reporter in Arabidopsis leaf mesophyll protoplasts (LMPs) to examine STA1 activity on pre-mRNA splicing of stress-induced genes. STA1 was involved in the pre-mRNA splicing of heat-inducible HSFA3 as well as cold-inducible COR15A and IDD14. Clearly, STA1 splicing activity was indispensable for mature mRNA expression of HSFA3 and its downstream HSPs in a reconstituted DREB2A-dependent gene regulatory module. Further genetic analysis verified that STA1 involved in pre-mRNA splicing of essential genes, including HSFA3 and its target gene HSA32, are necessary for in the establishment of plant heat stress tolerance. However, a single heat stress transcription factor HSFA3 was unable to recapitulate the stress tolerance in the absence of STA1 activity. Our global target search of STA1 activity using differential display of RNA followed by sequencing (DDR-seq) revealed that STA1 played central roles in a broader range of heat-inducible gene expression. Our findings unraveled that heat-inducible STA1 activity secures appropriate gene expression under heat stress conditions, of which gene products contribute to plant tolerance.

RESULTS AND DISCUSSION

STA1 Function in Pre-mRNA Splicing

To evaluate the biochemical activity of STA1 in pre-mRNA splicing in a cellular system, a splicing reporter construct was generated using the translational fusion of the GUS reporter gene to the 3′-end of a genomic version of the structure gene under the regulation of a constitutive 35S promoter (Fig. 1A). This GUS reporter activity increases only when the intron of pre-mRNA is spliced out correctly. Otherwise, a premature termination codon of the intron retained in the pre-mRNA would interrupt the complete translation of the GUS reporter protein. Therefore, in principle, an increase or decrease in the cellular GUS activity may primarily reflect the amount of mature mRNAs serving as protein translation templates.

Figure 1.

STA1 induces splicing activity of specific pre-mRNA. A, Schematic diagram of the functional splicing assay is shown. The genomic version of splicing target gene was cloned into the GUS reporter construct with a translational fusion. B and C, Splicing activity was measured in the presence and absence of STA1 for gCOR15A-GUS in LMPs of Col (B) and sta1-1 (C). UBQ10-rLUC activity served as an internal control. STA1 protein expression was shown using protein-blot analysis with an anti-epitope-specific antibody. Rubisco small subunit proteins served as a protein loading control using Coomassie Blue staining. The means of three replicates are shown with se bars. Asterisks represent paired t test significance between samples (***P < 0.001, **P < 0.01, and *P < 0.05). D and E, Splicing activity was measured in the presence and absence of STA1 for gIDD14-GUS (D) and gMPK10-GUS (E) in LMPs of sta1-1. UBQ10-rLUC activity served as an internal control. The means of three replicates are shown with se bars. RBC, Rubisco small unit protein.

The design principle of the splicing reporter was validated by the splicing activity of STA1 with a gCOR15A-GUS reporter construct using the well-established Arabidopsis LMPs. The GUS-based splicing reporter and a UBQ10 promoter-driven renilla luciferase (rLUC) control reporter construct were cotransfected to either wild-type (Col) or sta1-1 protoplasts with or without a STA1-HA effector construct and then incubated for 6 h under light (Yoo et al., 2007). The splicing reporter activity of an intron-containing gCOR15A-GUS construct was clearly induced with the STA1 expression when compared to the basal activity obtained without the effector expression (Fig. 1, B and C). In the assay, the control reporter activity of intron-free rLUC construct was not altered in the presence or absence of STA1 expression, indicating that STA1 did not modulate transcription activity in this system. To further verify this notion, a control experiment was independently carried out with a 35S-driven intron-free fLUC reporter construct in sta1-1 LMPs. Again, STA1 did not affect the LUC reporter activity at all (Supplemental Fig. S1). Then, to reexamine whether the difference in the GUS reporter activity originates from the splicing efficiency depending on STA1 activity, both the intron-retained and intron-free forms of COR15A-GUS were detected using semiquantitative reverse transcriptase-dependent PCR (RT-PCR) using the RNA extracted from the transfected protoplasts of Col and sta1-1 (Supplemental Fig. S2). The mature mRNA of gCOR15A-GUS was substantially enriched in Col, whereas its pre-mRNA was instead enriched in sta1-1, indicating that a clear shift of pre-mRNA to mature mRNA was made for gCOR15A-GUS by STA1. Taken together, our cell-based assay demonstrated pre-mRNA splicing activity of STA1.

Then, to investigate whether STA1 activity has any preference in pre-mRNA splicing targets, two more splicing reporters were constructed with genomic versions of INDETERMINATE DOMAIN14 (gIDD14) and MAPK10 (gMPK10). IDD14 is another cold-inducible gene that produces a spliced variant and controls its own gene product activity in response to cold (Seo et al., 2011). The stress-related MPK10 encodes a gene for a signaling potent kinase, the function of which is largely unknown (Mao et al., 2011). The reporter activity of gIDD14-GUS, but not gMPK10-GUS, was increased in sta1-1 protoplasts by STA1 expression (Fig. 1, D and E). gIDD14 contains a single intron as gCOR15a but gMPK10 has multiple introns. However, the number of introns is seemingly irrelevant to the target preference of STA1 for pre-mRNA splicing based on the global analysis of STA1-dependent heat-inducible gene expression in this study (see below).

To characterize STA1 functions in abiotic stress-responsive gene regulation other than cold stress and ABA responses reported in the previous study (Lee et al., 2006), STA1 expression patterns were searched through the Arabidopsis eFP database, which integrates high-throughput transcriptome analyses (Winter et al., 2007). High-temperature stress induces STA1 expression in addition to cold stress (Supplemental Fig. S3). To verify the gene expression of STA1 in response to heat, semiquantitative RT-PCR was performed after mild heat treatment (37°C). STA1 expression increased within 1 h and maintained up to 3 h, and then decreased in 12 h after heat treatment (Fig. 2A). Because STA1 expression is under the regulation of heat stress, its gene product activity most likely plays a role in the stress-responsive gene expression.

Figure 2.

Heat-inducible STA1 involved in pre-mRNA splicing of HSFA3. A, STA1 expression was measured in response to heat treatment (37°C) using semiquantitative RT-PCR. PP2A served as a RNA control. B, A schematic diagram of genomic DNA of HSFA3 is shown with primer positions. C and D, Different types of HSFA3 transcripts were measured in Col (C) and sta1-1 (D) in response to heat treatment (37°C) using semiquantitative RT-PCR. Heat-inducible total mRNA (f1/r1), pre-mRNA (f1/f2), and mature mRNA (f1/f3) values of HSFA3 were measured. Experiments were triplicated with consistent results. Representative data are shown. E, Schematic diagrams (wild type), mDo (mutation on splice donor), and mAc (mutation on splice acceptor) of gHSFA3-nLUC are shown. F, Splicing reporter activities from WT gHSFA3-nLUC or mDo and mAC forms of gHSFA3-nLUC were measured in the presence and absence of STA1. Expression of HA-tagged STA1 variants was shown by protein-blot analysis using an anti-HA antibody. Rubisco small subunit proteins were used as a loading control. G, Splicing reporter activity from gHSFA3-nLUC was measured in the combination of STA1 and sta1-1 expression. Expression of HA-tagged STA1 and Flag-tagged sta1-1 was shown by protein-blot analysis using with an anti-epitope-specific antibody. Rubisco small subunit proteins served as a protein loading control using Coomassie Blue staining. 35S-fLUC activity served as an internal control. All of the experiments were repeated three times with consistent results. The means of three replicates are shown with se bars. Different letters indicate a significance difference by Tukey-Kramer test (P < 0.05). RBC, Rubisco small unit protein.

To investigate STA1 functions in heat stress-related gene expression, the gene expression of several HSFs including HSFA3, HSFB1, and HSFB2a were examined upon heat treatment. These HSFs are transcription factors that have important functions in plant heat stress adaptation (Schramm et al., 2008; Scharf et al., 2012). To precisely characterize the regulatory functions of STA1 on splicing of heat-inducible transcripts, a heat-inducible expression of the HSFs was specifically measured for the total mRNA (primer combination of f1 and r1), pre-mRNA (primer combination of f1 and r2), and mature mRNA (primer combination of f1 and r3) in Col and sta1-1 (Fig. 2B). Total mRNA expression of HSFA3 was induced and maintained in both Col and sta1-1 seedlings at 37°C, indicating that total transcript levels of HSFA3 increased with and without STA1 activity by heat stress (Fig. 2, C and D). Notably, pre-mRNA expression was marginal in Col; however, it was rather significantly induced in sta1-1, indicating that pre-mRNA transcripts of HSFA3 were not spliced appropriately in sta1-1 compared to those in Col. Mature HSFA3 mRNA expression was then induced in Col, but not in sta1-1. The expression patterns of HSFB1 and HSFB2a were similar to those of HSFA3 (Supplemental Fig. S4).

To examine STA1 activity for pre-mRNA splicing of heat stress-responsive genes, a new (to our knowledge) splicing reporter was constructed with the genomic version of HSFA3 (gHSFA3). This time the reporter was generated with Nano-luciferase (nLUC) that is smaller in size and thus more sensitive in cellular responses compared to firefly-luciferase (Hall et al., 2012). To substantiate this splicing assay, evolutionarily conserved splicing sites of the intron in gHSFA3 were mutated to be unspliced at splice donor (Do) or acceptor (Ac) site (Fig. 2E) and subjected to the cell-based assay with and without STA1 in sta1-1 protoplasts. The reporter activity from wild-type gHSFA3-nLUC increased with STA1 expression, but those from two mutated forms of gHSFA3-nLUC did not (Fig. 2F), confirming that the induction of splicing reporter activity resulted from authentic pre-mRNA splicing of the reporter genes by STA1.

In the loss-of-function allele sta1-1, STA1 transcript is deleted with six nucleotides in-frame and it produces sta1-1 protein omitting two amino acids (Lee et al., 2006). To investigate whether sta1-1 protein has any splicing activity, the reporter assay using gHSFA3-nLUC was carried out in a combination of STA1 and sta1-1 expression. Protein-blot analysis first showed that STA1 and sta1-1 accumulated to a similar level in Col protoplasts (Fig. 2G). The reporter activity was induced by STA1, but not by sta1-1, and its activity by both STA1 and sta1-1was similar to the activity obtained by STA1 alone (Fig. 2G), indicating sta1-1 did not carry any noticeable splicing activity of gHSFA3. However, sta1-1 alone reduced the reporter activity to some extent, and this might indicate that sta1-1 has a little dominant negative function, although such activity is competitively weaker than STA1. This null splicing activity of sta1-1 is coherent with the fact that sta1-1 is a recessive loss-of-function mutant (Lee et al., 2006).

Reconstitution of Heat-Inducible Transcription Cascades by DREB2A and STA1

To substantiate this finding, we reconstituted transcription regulatory circuits of heat-response gene expression in Arabidopsis protoplasts. Namely, DREB2A responsible for HSFA3 transcription under heat stress conditions was transiently expressed in Col or sta1-1 protoplasts with and without STA1 (Fig. 3A; Sakuma et al., 2006; Schramm et al., 2008). Before the measurement of HSFA3 pre-mRNA, the DREB2A and STA1 expressions were verified by RT-qPCR with the RNA extracted from the transfected protoplasts (Supplemental Fig. S5). With forward and reverse primers specifically designed to recognize the pre-mRNA transcripts of HSFA3 (Fig. 3B), the endogenous expression of pre-mRNA of HSFA3 exhibited a basal level and it was not induced by STA1 (Fig. 3, C and D). In contrast, the pre-mRNA expression was induced by DREB2A and its accumulation was rather pronounced in sta1-1 protoplasts when compared to that in Col protoplasts. Furthermore, the pre-mRNA of HSFA3 decreased in the presence of STA1, implicating its role in the splicing of HSFA3 pre-mRNA. This was consistent with our observation that pre-mRNA splicing of HSFA3 was dependent on the STA1 activity (Fig. 2, C and D).

Figure 3.

STA1 induces pre-mRNA splicing of HSFA3 driven by DREB2A. A, Working model is proposed for DREB2A-dependent heat-inducible gene expression. B, Schematic diagram of HSFA3 is shown with a set of primers used for detecting intron-retained pre-mRNA. C and D, Intron-retained HSFA3 pre-mRNA, gene expression of which was driven by DREBA2, was discerned in LMPs of Col and sta1-1 with/without STA1. The pre-mRNA accumulation was monitored using semiquantitative RT-PCR (C) and RT-qPCR (D). E, Expression of HSP23.6, HSA32, and HSP70T-2 in sta1-1 was monitored in combination of DREB2A and STA1 expression using RT-qPCR. Quantitative values were normalized with an internal control EIF4a and presented in comparison to those values in Col without effector transfection. All of the experiments were repeated three times and produced consistent results. The means of triplicates are shown with se bars. Asterisks represent paired t test significance between samples (***P < 0.001, **P < 0.01, and *P < 0.05).

To examine whether DREB2A-driven HSFA3 becomes a functional transcription activator in the presence of STA1 activity, mature mRNA expression of HSP23.6, HEAT STRESS-ASSOCIATED PROTEIN32 (HSA32), and HSP70T-2 was measured by RT-qPCR as potential target genes of HSFA3. Either STA1 or DREB2A alone could not induce mature mRNA of HSP23.6, HSA32, and HSP70T-2 in sta1-1 protoplasts that are disconnected in the process from pre-mRNA to mature-mRNA (Fig. 3E). The coexpression of STA1 and DREB2A did induce mature mRNA accumulation of the HSPs. Evidently, pre-mRNA of HSFA3 induced by DREB2A was processed by STA1, and then its mature mRNA products could induce the HSP gene expression appropriately. These results suggested that STA1 activity served an important regulatory step in mature mRNA expression of HSFA3, whose transcription was driven by DREB2A.

Pre-mRNA Splicing Function of STA1 for Plant Thermotolerance

The STA1 splicing activity was then examined for plant heat stress tolerance with wild type (Col), a loss-of-function sta1-1, STA1-expressing transgenic plants in Col background (STA1/Col), and STA1-expressing transgenic plants in sta1-1 background (STA1/sta1-1). Before heat-inducible gene expression analysis, endogenous and transgene expression of STA1 was confirmed by using semiquantitative RT-PCR (Supplemental Fig. S6). For a two-step acquired heat-stress tolerance assay, Arabidopsis seedlings were grown for 5 d at room temperature and exposed to heat at 38°C for 2 h and then at 45°C for another 2 h. Seedling viability was examined at 7 d after heat treatment. Col seedlings acclimated and survived from the stepwise heat treatment, whereas sta1-1 seedlings grew slowly, had shorter primary roots, displayed completely bleached shoots, and eventually died (Fig. 4A, upper panel). Both STA1/Col and STA1/sta1-1 seedlings showed their tolerance at a level similar to Col in response to the stepwise heat treatment, confirming that a genetic defect of STA1 led to the lack of heat stress tolerance in sta1-1. These seedling phenotypes were examined repeatedly with multiple independent STA1-expressing transgenic Col and transgenic sta1-1 lines (Supplemental Fig. S7). In contrast, after a one-step basal heat tolerance assay conducted with 5-d-old seedlings exposed to 45°C for 2 h, none of these genotypes including Col executed heat stress tolerance (Fig. 4A, lower panel). Seedling survival and lethal phenotypes of Col displayed for one-step basal and two-step acquired heat stress treatments, respectively, were consistent with a previous report (Silva-Correia et al., 2014). In summary, heat-inducible STA1 has a regulatory role in the establishment of the two-step acquired heat stress tolerance, but not that of the one-step basal heat stress, perhaps through its splicing function of heat-responsive gene expression.

Figure 4.

STA1 involved in the establishment of stress tolerance in response to two-step heat treatment in Arabidopsis. A, Seedling survival assay was carried out with Col, sta1-1, STA1-expressing Col (STA1/Col), and STA1-expressing sta1-1 (STA1/sta1-1) for two-step acquired and one-step basal heat treatments. B and C, Expression levels of tSTA1 (B) and STA1 (C) were quantitatively monitored before (Cont) and at 1 d after heat treatment (Heat) using RT-qPCR. D and E, Expression levels of total mRNA, pre-mRNA, and mature-mRNA of HSFA3 (D) and HSA32 (E) were quantitatively monitored using RT-qPCR in Col, sta1-1, and STA1/sta1-1 before (Cont) and at 1 d after heat treatment (Heat). Primer sets for total mRNA (f1/r1), pre-mRNA (f1/r2), and mature-mRNA (f1/r3) were described with schematic diagrams. Quantitative values were normalized with internal controls ELF4a. The means of triplicates are shown with se bars. Different letters indicate a significance difference by Tukey-Kramer test (P < 0.05).

To further understand STA1 function in heat-responsive gene expression, total RNA was extracted from whole seedlings before and at 1 d after the stepwise heat treatment, and gene expression was measured by RT-qPCR. To specifically measure the expression of wild-type STA1 and mutant sta1 transcripts we designed primers based on the 6-nucleotide in-frame deletion in sta1-1 (Lee et al., 2006). Total STA1 (tSTA1) expression combining STA1 and sta1 expression was comparatively high in STA1/sta1-1 seedlings before heat treatment (Fig. 4B), reflecting the fact that transgene expression was under the regulation of 35S constitutive promoter. In response to heat treatment, tSTA1 expression was highly induced in Col, sta1-1, and STA1/sta1-1. However, STA1 was highly induced only in Col by heat treatment, but not in sta1-1 and STA1/sta1-1 (Fig. 4C). The results suggested that the heat-induced tSTA1 expression resulted from the induction of sta1 expression in STA1/sta1-1, and further implicated that STA1 and/or STA1-dependent splicing products do not seem to be involved in the intron-free STA1 and/or sta1 expression.

Then, expression of our model transcript HSFA3 was monitored specifically for total mRNA (primer combination of f1 and r1), pre-mRNA (primer combination of f1 and r2), and mature mRNA (primer combination of f1 and r3) under heat stress conditions. HSFA3 expression in any type of mRNA was low in all genotypes before heat treatment (Fig. 4D). Total mRNA expression of HSFA3 was highly induced in all genotypes by heat treatment (Fig. 4D, upper panel). Pre-mRNA of HSFA3 was accumulated to a high level in sta1-1 by heat treatment, but not to the same level in Col and STA1/sta1-1 (Fig. 4D, middle panel). On the contrary, mature HSFA3 mRNA expression level was high in Col and STA1/sta1-1 by heat treatment, but such high expression level was not reached in sta1-1 (Fig. 4D, lower panel). These results indicated that HSFA3 transcription was driven by heat treatment regardless of STA1 activity, resulting in the synthesis and accumulation of its pre-mRNAs. However, HSFA3 splicing was carried out appropriately only in the presence of STA1 activity. To verify the RT-qPCR data at a technical point, STA1 and HSFA3 expression was monitored with total RNA in the absence of reverse-transcription as a negative control, resulting in null amplification of target genes (Supplemental Fig. S8). Taken together, mature mRNA expression of HSFA3 resulted from STA1-independent transcription and STA1-dependent pre-mRNA splicing in response to heat stress.

The higher induction HSFA3 expression in Col and STA1/sta1-1 correlated well with their seedling viability after the two-step heat treatment (Fig. 4A, upper panel). All these experiments thus clearly indicated that STA1-dependent splicing played an important regulatory step for HSFA3 expression in the establishment of Arabidopsis heat stress tolerance. To examine whether STA1-dependent splicing of HSFA3 pre-mRNA led to the expression of its downstream genes, HSA32 expression was monitored using RT-qPCR. Total mRNA expression of HSA32 was highly induced in Col-0 and STA1/sta1-1 by heat treatment, but unlike HSFA3, it was never induced to the same level in sta1-1 (Fig. 4E, upper panel), suggesting full induction of HSA32 expression requires STA1 activity. Pre-mRNA of HSA32 accumulated in sta1-1 to a certain level that was higher than those in Col and STA1/sta1-1 under the stress conditions (Fig. 4E, middle panel) and thus pre-mRNA splicing of HSA32 again required STA1 activity as HSFA3. Consequently, mature HSA32 mRNA expression level was higher in Col and STA1/sta1-1 than sta1-1 under the stress conditions (Fig. 4E, lower panel). These results indicated that heat-induced HSA32 expression is seemingly under the transcriptional regulation of HSFA3, splicing of which is also dependent on STA1 activity. In addition, its mature mRNA expression is also under the posttranscriptional regulation of STA1 activity. In the case of intron-free HSP18.2 that is another target gene of HSFA3 (Schramm et al., 2008), the gene expression was highly induced in Col and STA1/sta1-1 in response to heat treatment; however, such induction was clearly compromised in sta1-1 (Supplemental Fig. S9), again suggesting that HSP18.2 induction requires functional HSFA3. Taken all together, these results suggest that STA1 plays a necessary regulatory step in heat response gene expression, and its gene products have important roles in establishing plant two-step acquired heat stress tolerance.

The higher accumulation of HSFA3 pre-mRNA in sta1-1 (Fig. 4D, middle panel) could reflect the possibility that nonsense-mediated mRNA decay (NMD) activity is compromised in the absence of STA1 activity (Maquat, 2004). To address whether STA1 functions in NMD, the seedling survival assay in response to two-step heat treatment was carried out with well-characterized loss-of-function NMD mutants, upf1 and upf3 (Jeong et al., 2011), together with Col, sta1-1, and STA1/sta1-1. Consistent with the previous growth phenotype responses to heat stress treatment (Fig. 4A), Col and STA1/sta1-1 showed heat stress tolerance, but sta1-1 did not (Supplemental Fig. S10). upf1 and upf3 also showed a phenotype similar to Col before and after the stepwise heat treatment, indicating that STA1 most unlikely functions in NMD at least for the establishment of heat stress tolerance.

HSFA3 plays a key role in expression of many HSP genes, gene products of which contribute to either positive or negative feedback of heat stress responses and eventually provides efficient plant heat stress adaptation (Charng et al., 2006). Because mature mRNA expression of HSFA3 correlated well with plant heat tolerance in our two-step acquired heat assay, we further examined whether HSFA3 could drive plant heat stress tolerance for itself downstream of STA1. Two transgenic Arabidopsis lines in sta1-1 background were generated to constitutively express a cDNA version of HSFA3 that does not require splicing activity for its mature mRNA expression (HSFA3/sta1-1). The HSFA3/sta1-1 lines with single homozygous transgene insertion were selected and used for further analysis. The expression of HSFA3 was monitored in two transgenic HSFA3/sta1-1 lines by RT-qPCR (Supplemental Fig. S11). There was no drastic phenotypic difference among Col, sta1-1, STA1/sta1-1, and HSFA3/sta1-1, except seedlings in the sta1-1 background showed a relatively slower growth and simpler root architecture than Col (Supplemental Fig. S12). These transgenic lines were tested for heat stress tolerance under both mild and two-step acquired heat stress conditions together with Col and sta1-1. For mild heat stress tolerance assay, seedlings were exposed at 38°C for 1 d and seedling viability was monitored at 7 d after the heat treatment. Approximately two-thirds of seedlings of Col and transgenic STA1/sta1-1 complementation lines survived under mild heat stress conditions, but transgenic HSFA3/sta1-1 lines were unable to keep their viability as like Col and transgenic STA1/sta1-1 complementation lines (Fig. 5, A and B). In consistent with previous data (Fig. 4A), seedlings of Col and STA1/sta1-1 showed stress tolerance to the stepwise heat treatment, but transgenic HSFA3/sta1-1 lines again failed to survive under the stress conditions (Fig. 5C).

Figure 5.

STA1, but not HSFA3, establishes plant stress tolerance for mild heat and two-step acquired heat stresses. A, Seedling survival assay was carried out with Col, sta1-1, two lines of STA1-expressing sta1-1 (STA1/sta1-1), and two lines of HSFA3-expressing sta1-1 (HSFA3/sta1-1) for mild heat treatment. B, Viable seedlings were measured to calculate survival rate. The means of triplicates are shown with se bars. C, Seedling survival assay was carried out for two-step acquired heat treatment. D to G, Expression levels of mature mRNA of HSP18.2 (D), HSA32 (E), HSP23.6 (F), and HSC70-5 (G) were monitored in Col, sta1-1, two lines of STA1/sta1-1, and two lines of HSFA3/sta1-1 before and 1 d after the stepwise heat treatment using RT-qPCR. The means of triplicates are shown with se bars. Different letters indicate a significance difference by Tukey-Kramer test (P < 0.05).

To monitor protein functions of transgene HSFA3 in gene regulation, its target gene expression was measured by RT-qPCR. Expression of HSFA3 target genes HSP18.2 and HSA32 was highly induced in Col and STA1/sta1-1 under two-step acquired heat stress conditions (Fig. 5, D and E). Although HSFA3 expression was detected in transgenic HSFA3/sta1-1 lines before heat treatment because of the nature of 35S promoter used in the transgenic line construction (Supplemental Fig. S11), HSP18.2 and HSA32 expression was not any further induced in HSFA3/sta1-1 compared to sta1-1 after heat treatment. Furthermore, mature mRNA transcripts of two well-characterized heat stress-responsive genes HSP23.6 and HSC70-5 failed to accumulate in HSFA3/sta1-1 lines after heat treatment (Fig. 5, F and G). These results indicated that HSFA3 is expressed in response to heat stress and its gene products induce HSP expression, but its sole activity is not sufficient to drive plant heat tolerance. This implicated that STA1 is involved in pre-mRNA splicing of a broader range of genome responses to heat stress.

Global Target Analysis for STA1-Dependent Splicing Activity

To search for genome-wide global targets of STA1-dependent splicing activity in heat-inducible transcriptomes, differential display of RNA-followed by sequencing (DDR-seq) was conducted with Col and sta1-1. Arabidopsis seedlings were first grown on half-strength MS agar containing 0.5% Suc under a cycle of 16 h light and 8 h dark at 23 to 25°C for 5 d. For heat-induced DDR-seq analysis, seedlings were exposed to heat treatment at 38°C for 2 h and then immediately at 45°C for 2 h, and harvested 24 h after heat treatment. Because two genotypes showed some differences in growth phenotypes, we identified heat-inducible genes within each genotype and compared these identified sets of genes to learn similarities and differences in heat-responsive gene expression in Col and sta1-1. DDR-seq analysis was thus carried out with Col (driver) versus heat-treated Col (tester) and sta1-1 (driver) versus heat-treated sta1-1 (tester) and sequence reads were normalized and analyzed using CLC Genomics Workbench (Qiagen).

After mapping sequence reads to genome information, 7,762 and 8,529 genes were uniquely identified from 398,370 and 436,311 bp from Col and sta1-1, respectively (Supplemental Table S1, Supplemental Fig. S6A). These corresponded to ∼30 to 35% of protein coding genes in the Arabidopsis genome. In an ordinary process of differentially expressed gene analysis, transcripts of which expression was affected in the opposite direction in wild type and a loss-of function mutant would be recognized as primary responsive genes under the regulation of genetic factors of interest. However, we instead tried to analyze 5,745 genes that were commonly sequenced in Col and sta1-1 (Supplemental Fig. S13A) because genes under STA1 regulation seem to accumulate their pre-mRNAs in sta1-1 under induction conditions. Consistently twice more sequence reads matched up to introns in the raw data of sta1-1 compared to Col (Supplemental Table S1). In the results, roughly two-thirds of heat-inducible genes in Col-0 overlapped with those in sta1-1, suggesting that STA1 conveys pre-mRNA splicing activity for a significant portion of transcriptome responses to heat stress.

In the commonly identified DDR-seq data 268 genes overlapped with those found in microarray-based data for heat stress responses to 38°C for 3 h (Supplemental Data Set; http://Arabidopsis.org/info/expression/ATGenExpress.jsp). Because both DDR-seq and microarray experiments were carried out with different heat stress conditions, only a certain level of heat responses presented commonly in both data sets. Despite the limitation, Gene Ontology (GO) term analysis was carried out for this set of genes to understand representative cellular pathways under heat stress conditions. First genes for responses to heat (GO:0009408, 2.09e-19) and protein folding (GO:0006457, 6.02e-15) were recognized being enriched significantly (Fig. 6A). In addition, genes for responses to high light intensity, hydrogen peroxide, cold, and regulation of programmed cell death were enriched with equal or less than 1% of false discovery rates (FDR), which perhaps implicated that reactive oxygen species could play a role in heat-responsive gene expression.

Figure 6.

Heat-inducible genes enriched in DDR-seq analysis. A, Graphical view of GO term analysis for gene (Supplemental Data Set). B, Mature mRNA expression of HSP23.6, HSC70-5, and BOB1 was measured in Col and sta1-1 using RT-qPCR. C and D, Total mRNA (C) and premature mRNA (D) of HSP23.6, HSC70-5, and BOB1 were measured using RT-qPCR in Col and sta1-1. The means of three replicates are shown with se bars. Different letters indicate a significance difference by Tukey-Kramer test (P < 0.05).

Furthermore, 14 HSP genes were noticed in the category of genes for response to heat and those for protein folding that are known to be the major function of HSPs in stressed cells (Supplemental Fig. S13B). Twelve of them have more than one intron, of which splicing is most likely under STA1 regulation. HSP90-2 expression could not be measured by RT-qPCR with limited analysis and thus it was excluded from further analysis. We focused on these 11 intron-containing heat stress-related genes in the beginning.

When these intron-containing HSP genes were analyzed for their gene expression patterns under heat stress conditions, mature mRNA expression of HSP23.6, HSC70-5, BOB1, CPHSC70-1, and HSP60-3B was highly induced in Col, but much less in sta1-1 (Fig. 6B; Supplemental Fig. S14A). This pattern of mature mRNA accumulation was similar to that of HSFA3 (Fig. 4D). However, mature mRNA expression of HSP70-3, HSP70-4, HSP70-16, and HSP89.1 was induced similarly in both Col and sta1-1 or even higher in sta1-1 (Supplemental Fig. S14A), indicating that mature transcript accumulation of these genes was likely independent of STA1 activity. Taken together, mature mRNA expression of HSP23.6, HSC70-5, BOB1, CPHSC70-1, and HSP60-3B is most likely under the control of STA1-dependent pre-mRNA splicing.

To verify this notion, accumulation of total mRNA and pre-mRNA of HSP23.6, HSC70-5, BOB1, CPHSC70-1, and HSP60-3B was measured separately by RT-qPCR. Total mRNA expression of HSP23.6, HSC70-5, and CPHSC70-1 was induced in Col and sta1-1 by heat treatment (Fig. 6C; Supplemental Fig. S14B). However, their pre-mRNA transcripts were higher in sta1-1 than Col (Fig. 6D; Supplemental Fig. S14C). These results indicated that induction of HSP23.6, HSC70-5, and CPHSC70-1 expression by heat treatment was mainly regulated at the posttranscriptional level by STA1 splicing activity as shown with HSFA3 expression (Fig. 4D). In the cases of BOB1 and HSP60-3B, total mRNA accumulation was relatively less induced in sta1-1 compared to Col by heat treatment (Fig. 6C; Supplemental Fig. S14B), but their pre-mRNA levels were still higher in sta1-1 than Col (Fig. 6D; Supplemental Fig. S14C). These results implicated that STA1 not only modulated splicing of BOB1 and HSP60-3B at the posttranscriptional level, but also influenced their transcription perhaps by the function of STA1 splicing products as shown with HSA32 expression (Fig. 4E). Taken all together, STA1 splicing activity is indispensable for mature mRNA accumulation of a subset of HSPs under heat stress conditions.

In this study, STA1, a putative component of the U5 snRNP complex, was characterized as an essential regulatory factor involved in the splicing of a specific subset of pre-mRNAs including HSFA3 and its target gene HSA32 in response to high-temperature stresses. Phenome and molecular analyses in response to heat treatment demonstrated that STA1 functions in splicing, but not in NMD, as a regulator in heat stress tolerance. HSFA3-dependent transcript cascades that were reconstituted using transient expression of the stress-related DREB2A and STA1, unambiguously linked to the STA1 activity with pre-mRNA splicing of HSFA3, although it is not a sufficient factor for the establishment of plant heat stress tolerance. Consistently, STA1 controls pre-mRNA splicing of a broader range of heat-inducible transcriptomes. In conclusion, heat-inducible STA1-dependent posttranscriptional regulation of the stress-responsive gene expression plays a role in the establishment of the acquired heat stress tolerance in Arabidopsis.

More recently STA1 has been reported to play a role in miRNA biogenesis and RNA-directed DNA methylation based on analyses of relevant molecular and cellular phenotypes of sta1-1 (Ben Chaabane et al., 2013; Dou et al., 2013). Moreover, heat stress seemingly alters splicing of miRNA, perhaps linking STA1 activity to miRNA splicing (Yan et al., 2012). This, to our knowledge, new STA1 function in small RNA expression deserves more attention to elucidate specific STA1 functions in gene regulation.

Our findings unravel that STA1 has specific targets in pre-mRNA splicing for the posttranscriptional regulation of gene expression response to heat stress. Even so, how STA1 modulates heat stress tolerance remains unclear. Identification of the genetic components, and elucidation of the molecular mechanisms in the STA1-containing spliceosome complex responsible for its splicing specificity, will enlighten researchers on the molecular basis underlying stress-specific gene expression and, eventually, stress-adaptive responses in plants.

MATERIALS AND METHODS

Plant Materials and Growth Conditions

Plants were grown in soil for 22 to 24 d under a photoperiod of 13 h light/11 h dark (60 μmol/m²/s) at 25°C. Arabidopsis (Arabidopsis thaliana) Columbia-0 (Col) plants were used as the wild-type, and sta1-1 (Lee et al., 2006), upf1-5 (Jeong et al., 2011), and upf3-1 (Jeong et al., 2011) mutants were used for experiments. Plasmid constructs for transgenic plants were generated by inserting the cDNA of STA1 or HSFA3 between the 35SC4PPDK promoter (designated as HBT) and the NOS terminator in a mini-binary vector, pCB302 (Cho et al., 2012). The constructs were expressed in Col or sta1-1 plants. Transgenic lines with similar transgene expression levels were selected and used for further analyses. Multiple independent transgenic lines were generated and analyzed to identify consistent genetic effects. Transgenic plant phenotypes from at least two independent lines of the T3 generation were analyzed.

Heat Stress Survival Assay

To examine plant heat stress tolerance, surface-sterilized seeds were stratified for 4 d at 4°C in the dark, and plants were vertically grown on half-strength Murashige and Skoog (MS) agar medium containing 0.5% Suc for 5 d under a photoperiod of 16 h light/8 h dark (60 μmol/m²/s) at 25°C. For the acquired heat stress tolerance assay, seedlings were subjected to a stepwise heat treatment of 38°C for 2 h followed by 45°C for 2 h. For the basal heat stress tolerance assay, seedlings were subjected to 45°C for 2 h. For the mild heat stress tolerance assay, horizontally grown 3-d-old seedlings were subjected to 38°C for 24 h. Plant survival was observed 7 d after the heat treatment and scored based on the retention of green shoots.

Arabidopsis Mesophyll Protoplast Transient Expression Assay

Protoplast isolation and transient expression assays were carried out as previously described by Yoo et al. (2007) and Cho and Yoo (2010). The effector construct (STA1) was generated by inserting cDNA between the HBT promoter and the NOS terminator in a plant expression vector. All the reporter constructs (gCOR15A, gHSFA3, gIDD14, and gMPK10) were generated by inserting genomic DNA fused with GUS between the HBT (modified 35S) promoter and the NOS terminator in a plant expression vector. All the constructs were verified by DNA sequencing. The renilla luciferase driven by UBQ10 promoter (UBQ10-rLUC) was included as an internal control in the protoplast transient expression assay. The experiments were repeated, indicating consistent results among the replicates. The primers used for cloning are listed in Supplemental Table S2.

In the functional splicing assay, reporter activities were calculated based on the GUS/renilla-LUC ratio and normalized to the values obtained without the effector expression. To measure the GUS activity, the transfected protoplasts with designated constructs were lysed using a passive lysis buffer (Promega) containing 1% Triton-X (USB; Affymetrix) and briefly mixed by vortexing. The lysate was incubated at room temperature for 10 min and centrifuged at 13,000 rpm for 30 s. The protoplast lysate was mixed with 10 mm MUG (Gold Biotechnology) and incubated at 37°C for 90 min. The reaction mixture was frozen at −80°C to quench the reaction. The mixture was then diluted with 0.2 m Na₂CO₃, and the GUS activity was measured using the Glomax (Promega) single tube system with a UV module, following the manufacturer’s instructions.

RNA Isolation and Transcript Measurement

For the STA1 expression test, Col-RD29A-LUC was used as the wild type (Lee et al., 2006). Seeds were sown on 1× MS medium (2% Suc and 0.3% gelrite) after surface sterilization with sodium hypochlorite (4%). The seeds were stratified at 4°C for 2 d and grown at 22°C under continuous light. For heat treatment, the plants were placed in a 37°C incubator.

For gene expression analysis, total RNA was isolated by the Trizol method (Invitrogen), and 1 μg of total RNA was used for cDNA synthesis using M-MLV reverse transcriptase (Promega). Gene expression was quantitatively measured using real-time PCR (Bio-Rad) with the SYBR Green dye-added PCR mix (Bio-Rad). PROTEIN PHOSPHATASE 2A (PP2A; AT1G13320), TUBULIN4 (TUB4; At1g04820), and ELONGATION INITIATION FACTOR 4a (ELF4a; At3g13920) transcripts were used as the controls with gene-specific primers. Detailed primer sequences are listed in Supplemental Table S2. Each primer set was pretested by PCR for a single gene product. The experiments were repeated three times, and consistent results were obtained.

DDR-Seq

For sample preparation, plants were grown on half-strength MS agar medium containing 0.5% Suc for 5 d and subjected to an acquired heat stress as described in “Results and Discussion.” Heat-treated (tester) and control (driver) seedlings were harvested 1 d after heat treatment.

Total RNA was extracted based on the Trizol method (Invitrogen) and poly(A)-RNA was isolated using a Dynabead mRNA DIRECT Kit (Thermo Fisher Scientific). Fifty nanograms of poly(A)-RNA was used to synthesize double-stranded cDNA. After precipitation, double-stranded cDNA was digested with Dpn II enzyme (New England Biolabs) for 2 h, phenol/CHCl₃ was extracted, ethanol was precipitated, and the results were resuspended with water. cDNA was ligated with a preannealed R-24/12 adaptor (2 mg/mL) at 16°C for 12 h. R adaptor-ligated cDNA was phenol/CHCl₃ extracted, ethanol precipitated, and resuspended with water. To generate tester and driver amplicons, cDNA was amplified with multiple PCR reactions using high-fidelity DNA polymerase (Phusion High-Fidelity DNA Polymerase; New England Biolabs) and R-24 primer (72°C 5 min, 20 cycles of 94°C 1 min; 72°C 3 min, 72°C 10 min). The PCR products were phenol/CHCl₃ extracted and isopropanol was precipitated. To remove the R-24/12 adaptor, the PCR products were digested with Dpn II for 2 h, phenol/CHCl₃ was extracted twice, ethanol was precipitated, and the results were resuspended with water. The concentrations of the tester and driver products were quantified using a Quibit 2.0 fluorometer (Invitrogen). Five hundred nanograms of the tester was further ligated with preannealed J-24/12 adaptor (2 mg/mL) at 14°C or 12 h.

For the subtractive hybridization, 25 μg of driver and 250 ng of J-adaptor ligated tester cDNA were combined and phenol/CHCl₃ extracted, CHCl₃ extracted, and ethanol precipitated. The pellet was resuspended with 4 μL of 3× buffer solution containing 30 mm HEPES (USB; Affymetrix), pH 8.0, at 20°C and 3 mm EDTA (USB; Affymetrix). The solution was incubated at 98°C for 5 min for DNA denaturation. One microliter of 5 m NaCl was added to the solution and the mixture was further incubated at 67°C for 20 h to hybridize tester and driver. The hybridization solution was mixed well with 8 μL of 5 mg/mL of yeast RNA and further diluted with 367 μL water. For each subtraction, eight 100-μL PCR reactions were prepared including 20 μL of hybridization mix without the primer. The PCR mixtures were incubated at 72°C for 5 min and 0.8 μL of J-24 primer (1 mg/mL) was added. Ten cycles of PCR reaction were performed (94°C 1 min, 70°C 3 min) and incubated at 72°C for 10 min. PCR products were combined, phenol/CHCl₃ extracted twice, ethanol precipitated, and the results were resuspended with 40 μL water. Twenty microliters of PCR products were digested with Mung Bean Nuclease (New England Biolabs) at room temperature for 30 min. To stop the reaction, 160 μL of 50 mm Tris, pH 8.9, was added and incubated at 70°C for 10 min. For amplification, eight 100-μL PCR reactions were prepared including 20 μL of Mung Bean Nuclease-treated products and 0.8 μL of J-24 primer (1 mg/mL) without polymerase. The PCR mixtures were incubated at 95°C for 1 min and cooled down to 80°C. Polymerase (2.5 U) was added and 18 cycles of PCR reaction were performed (94°C 1 min, 70°C 3 min). After further incubation at 70°C for 10 min, PCR products were combined, phenol/CHCl₃ extracted twice, CHCl₃ extracted, ethanol precipitated, and the results were resuspended with 50 μL water, generating a difference product.

For next-generation sequencing, the library was prepared according to the Ion Xpress Plus Fragment Library Kit (Thermo Fisher Scientific) and sequencing was conducted according to the manufacturer’s instructions with Ion PGM (Thermo Fisher Scientific). Raw reads were normalized and analyzed using a trial version of the CLC Genomics Workbench (Qiagen) platform.

GO Analysis

GO analysis was carried out using the agriGO web-based tool (http://bioinfo.cau.edu.cn/agriGO/; Du et al., 2010). Complete_GO was used for identification of enriched GO with Fisher’s test. For multiple significance tests, a Benjamini and Yekutieli FDR correction was used and GO terms with FDR < 0.01 were visualized with graphical view.

Accession Numbers

Sequence data from this study can be found in the Arabidopsis Genome Initiative or GenBank/EMBL databases under the following accession numbers: STA1, AT4g03430; COR15A, AT2G42540; IDD14, AT1G68130; MPK10, AT3G59790; DREB2A, AT5G05410; HSFA3, AT5G05410; HSFB1, AT4G36990; HSFB2a, AT5G62020; UPF1, AT5G47010; UPF3, AT1G33980; HSA32, AT4G21320; HSP23.6, AT4G25200; HSP70T-2, AT2G32120; HSP70-16, AT1G11660; HSP70-4, AT3G12580; HSP60-3B, AT3G23990; HSP70-1, AT5G02500; BOB1, AT5G53400; HSP70-3, AT3G09440; CPHSC70-1, AT4G24280; HSP70-7, AT5G49910; HSC70-5, AT5G09590; and HSP89.1, AT3G07770.

Supplemental Data

The following supplemental materials are available.

• Supplemental Figure S1. Splicing activity was measured in the presence and absence of STA1 for 35S-fLUC activity in LMPs of sta1-1.

• Supplemental Figure S2. Intron-retained pre-mRNA and exon-only mature mRNA of gCOR15A-GUS were discerned using semiquantitative RT-PCR with RNA extracted from transfected LMPs of Col and sta1-1.

• Supplemental Figure S3. STA1 expression upon cold and heat treatment.

• Supplemental Figure S4. Different types of HSF transcripts were measured in Col and sta1-1 in response to heat treatment (37°C) using semiquantitative RT-PCR.

• Supplemental Figure S5. Analysis of STA1 and DREB2A transcripts.

• Supplemental Figure S6. Analysis of the endogenous and transgene expression of STA1 in transgenic plants.

• Supplemental Figure S7. Seedling survival assay was carried out with Col, sta1-1, two lines of STA1-expressing Col (STA1/Col), and two lines of STA1-expressing sta1-1 (STA1/sta1-1) for two-step acquired and one-step basal heat treatments.

• Supplemental Figure S8. The expression levels of STA1 and HSFA3 mRNA were monitored using RT-qPCR without reverse transcription as control.

• Supplemental Figure S9. The expression levels of HSP18.2 in Col, sta1-1, and STA1-expressing sta1-1 (STA1/sta1-1) were monitored using RT-qPCR.

• Supplemental Figure S10. STA1 involved in pre-mRNA splicing but not in nonsense-mediated mRNA decay for the establishment of plant heat stress tolerance.

• Supplemental Figure S11. Analysis of HSFA3 expression in two HSFA3/sta1-1 transgenic plants.

• Supplemental Figure S12. Non-heat-treated 12-d-old seedlings of Col, sta1-1, two lines of STA1/sta1-1 and two lines of HSFA3/sta1-1 are shown.

• Supplemental Figure S13. Analysis of DDR-seq of Col and sta1-1 in response to heat stress.

• Supplemental Figure S14. Analysis of heat-inducible genes enriched in DDR-seq analysis.

• Supplemental Table S1. DDR-seq information.

• Supplemental Table S2. Primers used in the study.

• Supplemental Data Set. GO analysis of commonly enriched 268 genes in DDR-seq data of Col and sta1-1.