DEAD-Box RNA Helicase 42 Plays a Critical Role in Pre-mRNA Splicing under Cold Stress^{1  [OPEN]}

Abstract

Low temperature is an important environmental stress that adversely affects rice (Oryza sativa) growth and productivity. Splicing of pre-mRNA is a crucial posttranscriptional regulatory step in gene expression in plants and is sensitive to temperature. DEAD-box RNA helicases belong to an RNA helicase family involved in the rearrangement of ribonucleoprotein complexes and the modification of RNA structure and are therefore involved in all aspects of RNA metabolism. In this study, we demonstrate that the rate of pre-mRNA splicing is reduced in rice at low temperatures and that the DEAD-box RNA Helicase42 (OsRH42) is necessary to support effective splicing of pre-mRNA during mRNA maturation at low temperatures. OsRH42 expression is tightly coupled to temperature fluctuation, and OsRH42 is localized in the splicing speckles and interacts directly with U2 small nuclear RNA. Retarded pre-mRNA splicing and plant growth defects were exhibited by OsRH42-knockdown transgenic lines at low temperatures, thus indicating that OsRH42 performs an essential role in ensuring accurate pre-mRNA splicing and normal plant growth under low ambient temperature. Unexpectedly, our results show that OsRH42 overexpression significantly disrupts the pre-mRNA splicing pathway, causing retarded plant growth and reducing plant cold tolerance. Combined, these results indicate that accurate control of OsRH42 homeostasis is essential for rice plants to respond to changes in ambient temperature. In addition, our study presents the molecular mechanism of DEAD-box RNA helicase function in pre-mRNA splicing, which is required for adaptation to cold stress in rice.

Rice (Oryza sativa) is among the most important crops in the world and is the main staple food for almost 50% of the world’s population (Fairhurst and Dobermann, 2002; Gross and Zhao, 2014). As a tropical and subtropical crop plant, rice is sensitive to cold stress (Lee et al., 1995). Low temperature impairs rice development at the vegetative and reproductive stages, affecting germination, seedling growth, plant height, photosynthesis, heading days, and fertility (Suh et al., 2010). Therefore, cold stress is a major limiting factor for rice growth and production (Jena et al., 2012; Xie et al., 2012; Zhang et al., 2014).

Cold response in plants, including rice, involves a complex network of pathways. Complex mechanisms underlying the molecular and physiological changes that enable adaptive response to cold stress in plants have been detected (Hashimoto and Komatsu, 2007). Previous studies of rice response to cold stress have mostly focused on the cold-induced transcriptional response, and many cold-related transcription factors, such as CBF/DREBs, MYB4, MYB3R-2, MYBS3, OsNAC5, OsbZIP52, and OsbZIP73, have been identified in rice (Vannini et al., 2004; Ito et al., 2006; Dai et al., 2007; Su et al., 2010; Takasaki et al., 2010; Liu et al., 2012, 2018). Overexpression of certain cold-related transcription factors can enhance the tolerance of rice to cold stress (Ito et al., 2006; Wang et al., 2008; Ma et al., 2009; Su et al., 2010; Takasaki et al., 2010). In addition to basic transcriptional regulation, posttranscriptional regulatory mechanisms, such as splicing, capping, polyadenylation, mRNA transport, mRNA stability, and translation of the functional mRNA, also influence the cold stress-response regulatory network in many different plants (Gong et al., 2005; Nishimura et al., 2005; Lee et al., 2006; Narsai et al., 2007; Adamo et al., 2008; Chinnusamy et al., 2008; Kim et al., 2017; Shi et al., 2019).

RNA molecules, including mRNA, rRNA, and tRNA, perform a specific function based on their well-defined structure during gene expression (de la Cruz et al., 1999; Rajkowitsch et al., 2007). Low temperature causes overstabilization of incorrectly folded RNA and therefore leads to RNA molecular inactivation (Sosnick and Pan, 2002; Melencion et al., 2017). RNA chaperones and RNA helicases function to ensure the formation of mature RNAs of the correct structure by means of their RNA-unwinding and RNA-unfolding activities (Rajkowitsch et al., 2007). RNA helicases are enzymes that can rearrange ribonucleoprotein (RNP) complexes and modify RNA structures and are therefore involved in all aspects of RNA metabolism (Linder, 2006; Hubstenberger et al., 2013; Russell et al., 2013). DEAD-box helicases, which constitute the largest family of RNA helicases, exhibit variable protein sizes and compositions of N- and C-terminal extension sequences. These extension sequences have been proposed to provide substrate-binding specificity, signals for subcellular localization, or interaction domains with accessory compartments (Korolev et al., 1998; Cordin et al., 2006; Byrd and Raney, 2012).

DEAD-box helicases are present in most prokaryotes and all eukaryotes, including plants (Okanami et al., 1998; Aubourg et al., 1999; Rocak and Linder, 2004; Owttrim, 2006, 2013; Umate et al., 2010; Xu et al., 2013). Certain DEAD-box RNA helicases in plants are associated with a variety of cellular functions, developmental regulation (Stonebloom et al., 2009; Huang et al., 2010a, 2010b; Liu et al., 2010; Nishimura et al., 2010; Burch-Smith et al., 2011; Asakura et al., 2012; Kanai et al., 2013; Hsu et al., 2014), and response to biotic and abiotic stresses (Gong et al., 2005; Kant et al., 2007; Huang et al., 2010c; Xu et al., 2011; Khan et al., 2014). Expression of five DEAD-box RNA helicases, namely AtRH7 (Huang et al., 2016a), AtRH9/PMH1 (Kim et al., 2008), AtRH22 (Tripurani et al., 2011), AtRH25/STRS2 (Kant et al., 2007), and AtRH53/PMH2 (Kim et al., 2008), is induced by cold stress. Overexpression of AtRH25 confers tolerance to cold stress in Arabidopsis (Arabidopsis thaliana; Kim et al., 2008). AtRH38/LOS4 (Gong et al., 2005) and AtRH42/RCF1 (Guan et al., 2013) perform crucial roles in the export of RNA molecules from the nucleus to the cytoplasm and in pre-mRNA splicing, respectively, and both proteins have an impact on the cold-stress response mechanism. The protein AtRH7 participates in rRNA biogenesis and is also involved in the cold-response mechanism in Arabidopsis (Huang et al., 2016a). Thus, expression of these DEAD-box RNA helicases is essential and may promote plant adaptation to cold stress.

To date, the majority of plant DEAD-box RNA helicases shown to function in the cold-response mechanism have been characterized in Arabidopsis. Although the physiological roles of a limited number of DEAD-box RNA helicases, such as OsABP (Macovei et al., 2012), OsRH53 (Nawaz et al., 2018), TOGR1 (Wang et al., 2016), and OsRH58 (Nawaz and Kang, 2019), have been reported in response to abiotic stresses in rice, most rice DEAD-box RNA helicases remain uncharacterized.

In this study, we identified and characterized the cold-induced expression of the DEAD-box RNA helicase OsRH42. OsRH42 interacted directly with U2 small nuclear RNA (snRNA). Gain- and loss-of-function analyses revealed that both OsRH42 overexpression and knockdown in transgenic rice seedlings confer hypersensitivity to cold stress. RNA sequencing (RNA-seq) analysis indicated that cold-induced pre-mRNA splicing is disrupted in both OsRH42-overexpressing and -knockdown transgenic lines. Combined, these results demonstrate the pivotal role of OsRH42 in the accurate regulation of cold-induced pre-mRNA splicing and cold-stress tolerance in rice.

RESULTS

Cold and Heat Stresses Induce the Expression of OsRH42

To identify DEAD-box RNA helicases involved in the rice response to cold stress, we first identified relevant low temperature-induced expression targets. A rice DEAD-box RNA helicase, namely OsRH42 (Os08g0159900/LOC_Os08g06344), was revealed to be a putative cold-induced gene following a search of a public gene expression profiling database (Genevestigator) in rice (Supplemental Fig. S1A). The cold-induced expression pattern of OsRH42 was verified in rice suspension-cultured cells by reverse transcription PCR (RT-PCR) analysis (Supplemental Fig. S1B) and in 2-week-old rice seedlings by reverse transcription quantitative PCR (RT-qPCR) analysis (Supplemental Fig. S1C) with gene-specific primers (Supplemental Table S1). Expression of OsRH42 under treatment with salt (150 mm NaCl), drought (10 min of air drying), polyethylene glycol (PEG; 20%, w/v), heat (42°C for 3 h), or cold (4°C for 48 h) was analyzed in 2-week-old rice seedlings. The OsRH42 mRNA abundance increased by about 10-fold in cold-stressed seedlings and about 7-fold in heat-stressed seedlings compared with control mock seedlings (Fig. 1A). The abundance of OsRH42 mRNA increased slightly, by a factor of 4, in drought-stressed seedlings (Fig. 1A). These results indicate that OsRH42 expression was highly responsive to cold and heat stresses and that OsRH42 may participate in the response mechanism to these stresses in rice.

The expression patterns of OsRH42 in various tissues and organs of rice were determined by RT-qPCR analysis. The OsRH42 transcripts were detected in all selected tissues and organs, including the root, stem, leaf, sheath, panicle, and seedling, and relatively high amounts of OsRH42 mRNAs were detected in vegetative leaves and flag leaves (Fig. 1B). Thus, OsRH42 expression was ubiquitous for all the plant developmental stages examined and was predominant in leaves. The temporal and spatial expression patterns of OsRH42 were further examined in three independent transgenic plants that harbored an OsRH42 promoter-driven GUS chimeric gene (Supplemental Fig. S2A). The GUS activity was weak and only detected in leaf tips of the seedling (Supplemental Fig. S2B). We also examined the mRNA accumulation pattern of OsRH42 mRNA in the leaf blade. Three samples of leaf blade, leaf apices, leaf midribs, and leaf bases were collected from 2-week-old, 1-month-old, and 3-month-old rice plants and their total RNAs were subjected to RT-qPCR analyses. Accumulation of OsRH42 mRNA in all three parts of the leaf blade was maintained at similar levels (Supplemental Fig. S2C). The studies described here suggest that although the expression of OsRH42 in rice is ubiquitous and occurs predominantly in leaves, OsRH42 expression may be regulated not only at the transcriptional level.

Expression Knockdown and Overexpression of OsRH42 Cause Defective Plant Development

Loss- and gain-of-function approaches were applied to explore the biological and cellular functions of OsRH42 in rice. The RNA interference (RNAi) target region comprised 290 bp of the OsRH42 cDNA 3ʹ untranslated region sequence, which was used as a query to search rice mRNA databases of the National Center of Biotechnology Information. No RNA region had an identical sequence of 16 or more nucleotides. In addition, a public web-based computational tool developed for the identification of potential off-targets, siRNA Design Software (Naito and Ui-Tei, 2012), was used to search rice mRNA databases. No potential off-target effects based on use of the RNAi target region were detected. Several independent transgenic plants that harbored the OsRH42 RNAi construct, which is controlled by a maize (Zea mays) ubiquitin gene (Ubi) promoter (Supplemental Fig. S3A), were generated, and OsRH42 mRNA abundance was determined by RT-qPCR analysis. The residual levels of OsRH42 mRNA detected in three independent T1 transgenic lines, RNAi-4, RNAi-17, and RNAi-20, were about 25% of those of wild-type lines (Supplemental Fig. S3B) and were therefore selected for further investigation. In addition, OsRH42-overexpression transgenic rice lines were generated using OsRH42 cDNA fused downstream of the maize Ubi promoter (Supplemental Fig. S3A). Three independent homozygous lines that showed high abundances of OsRH42 mRNA were obtained: OX-6, OX-17, and OX-18 (Supplemental Fig. S3C).

OsRH42-overexpression and -knockdown transgenic lines were generated up to the T3 to T5 generations, and these showed a number of morphological defects at vegetative and reproductive stages. The OsRH42-overexpression transgenic lines were more severely impacted than the OsRH42-knockdown lines. Reduced seedling growth was observed in the OsRH42-knockdown and -overexpression transgenic lines compared with the growth of wild-type seedlings; the height of 2-week-old transgenic seedlings was 12% and 20% shorter, respectively, than those of wild-type seedlings of identical age (Fig. 2, A and B). At the reproductive stage, the plant height of OsRH42-knockdown and -overexpression transgenic lines was less than that of wild-type plants (Table 1; Fig. 2, C and D). The culm and internode lengths of the transgenic lines were reduced compared with those of wild-type plants (Supplemental Fig. S4). The OsRH42-knockdown and -overexpression transgenic lines showed a 13% to 46% reduction in terms of the number of panicles compared with the wild-type plants (Table 1). After pollination, the OsRH42-knockdown and -overexpression transgenic lines exhibited a decrease in tiller number, panicle number, and panicle length, and produced fewer and lighter seeds, compared with wild-type plants (Table 1; Fig. 2E).

Comparison of agronomic traits in field-grown wild-type, OsRH42 knockdown (RNAi), and OsRH42 overexpression (OX) lines

OsRH42 Is Essential for Tolerance to Prolonged Cold Stress in Rice

To investigate the response of rice ‘Tainung 67’ to cold stress, 2-week-old wild-type seedlings were incubated in growth chambers at 4°C for various periods and then transferred to 28°C for 7 d. Survival percentage declined from 100% to 70% after 6 d of cold stress and then decreased progressively until almost no seedlings remained alive after 9 d (Fig. 3A). A 6-d cold treatment of seedlings was used to assess whether OsRH42 plays a role in cold-stress response. Under prolonged cold stress, the leaves of all examined rice seedlings showed dramatic rolling. After cold treatment, the three independent OsRH42-knockdown lines, RNAi-4, RNAi-17, and RNAi-20, did not grow and exhibited withered, dried leaves before eventually dying (Fig. 3B), whereas the wild-type seedlings continued to grow and showed a relatively green and healthy appearance (Fig. 3B). Survival percentages of the RNAi-4, RNAi-17, and RNAi-20 lines ranged from 26% to 42%, which were lower compared with 85% for the wild type (Fig. 3C). Unexpectedly, continued growth of seedlings of the OsRH42-overexpression lines was inhibited by cold treatment (Fig. 3B). Three OsRH42-overexpression lines, OX-6, OX-17, and OX-18, showed survival percentages of 3.6%, 4.2%, and 4.6%, respectively (Fig. 3C). Electrolyte leakage was measured in seedlings grown under 4°C for 6 d to assess the degree of cold stress-induced membrane injury. Electrolyte leakage during cold treatment was higher in both OsRH42-knockdown and -overexpression lines than in the wild-type seedlings (Fig. 3D). These results indicate that OsRH42 plays an essential role in the rice response to cold stress, and ectopic constitutive overexpression of OsRH42 is unable to improve rice tolerance to cold stress.

To investigate the mechanism by which OsRH42 acts in the response of rice to cold stress, OsDREB1A (DRE-binding protein 1A/CBF3) and OsDREB1B expression patterns were examined in wild-type, RNAi-4, and OX-6 seedlings treated at 4°C for 12 and 18 h. The wild-type seedlings exhibited increased OsDREB1A and OsDREB1B expression levels under cold stress for 12 and 18 h (Supplemental Fig. S5). Consistently, the expression levels of OsDREB1A and OsDREB1B induced by cold stress in the RNAi-4 and OX-6 seedlings were similar to those of the wild type (Supplemental Fig. S5). This result indicates that the function of OsRH42 in the response to cold stress is independent of the CBF-mediated system.

OsRH42 Is a Prp5-Homologous RNA Helicase That Interacts Directly with U2 snRNA

The OsRH42 gene was located on rice chromosome 8 and included two exons. The deduced amino acid sequence of OsRH42 cDNA consisted of nine conserved RNA helicase domains and the characteristic amino acid residues D-E-A-D in motif II (Fig. 4A). Comparison of the amino acid sequences and analysis of the phylogenetic relationships among plant, yeast, and mammalian OsRH42-like proteins showed that OsRH42 is closely related to Arabidopsis RCF1, yeast Prp5, and human DDX46 (Fig. 4A; Supplemental Fig. S6). All homologs contained a conserved motif, E_XDPLP^Y/_FM, unique to the family. Moreover, N-terminal external regions of all homologs, except in the case of Saccharomyces cerevisiae, included RD/RS dipeptide repeats (Fig. 4A).

Both Prp5 and DDX46 are involved in branch-point site selection of pre-mRNA splicing (Hozumi et al., 2012; Liu and Cheng, 2015; Tang et al., 2016) and are therefore located in splicing speckles. To determine the subcellular localization of OsRH42, an OsRH42-GFP fusion gene under the control of the cauliflower mosaic virus 35S promoter was generated and introduced into rice protoplasts. OsRH42-GFP-derived fluorescent signals were detected in the nucleus with specific foci, whereas the GFP control signals were observed in the nucleus and cytoplasm (Supplemental Fig. S7). In addition, rice protoplast cells were coexpressed with the fusion proteins OsRH42-GFP and RNPS1-mCherry, a splicing speckle marker. Both GFP and mCherry signals were detected in specific nuclear foci (Fig. 4B), which indicates that OsRH42 is localized in splicing speckles in rice.

In both yeast and humans, Prp5 has been demonstrated to bind directly to U2 snRNA (Xu et al., 2004; Liang and Cheng, 2015). Moreover, Prp5 comes into contact with U2 snRNA regions on and near the branch-point-interacting stem loop (Liang and Cheng, 2015). In addition, Prp5 also can interact with U1 snRNA in fission yeast Schizosaccharomyces pombe and humans, but the interaction was not found in S. cerevisiae (Xu et al., 2004). Therefore, an RNA immunoprecipitation (RIP) assay of OsRH42-GFP rice transgenic calli was performed to detect whether OsRH42 binds directly to U2 snRNA in vivo. Two transgenic lines independent of OsRH42-GFP, namely 42-GFP-1 and 42-GFP-2, were obtained and their expression of OsRH42-GFP was verified by RT-PCR and immunoblot analysis (Fig. 5A). The transgenic calli were subjected to RIP analysis using an anti-GFP antibody. Immunoprecipitated RNAs were examined by RT-PCR with primers specific to the U1, U2, U4, and U6 snRNAs (Supplemental Table S1). High amounts of U2 snRNA and low amounts of U1, but no U4 or U6, snRNA were detected (Fig. 5B). By contrast, the RIP assay showed no binding signal in the GFP control transgenic callus (Fig. 5B). This result indicates that OsRH42 binds strongly to U2 but only weakly to U1 snRNA. Our result is consistent that of S. pombe and humans, in that Prp5 interacts with U1 and U2 small nuclear ribonucleoproteins (snRNPs) to bridge between the two snRNPs in the formation of the prespliceosome (Xu et al., 2004).

Pre-mRNA splicing defect analysis of AtSK12, AtPRR5, AtEBF2, and AtPUB45 was carried out in the Arabidopsis rcf1 (an OsRH42 homolog) mutant (Guan et al., 2013). Rice homologs of these genes were identified and their pre-mRNA splicing was examined in OsRH42-knockdown rice. Two-week-old wild-type and RNAi-4 seedlings were cultured under nonstress conditions (28°C) and then treated with cold stress (4°C) for various periods. Total RNA was isolated from the seedlings and subjected to RT-PCR to evaluate the abundance of tested mRNA precursors using gene-specific primers (Fig. 6; Supplemental Table S1). The unspliced mRNAs of SHAGGY-LIKE KINASE12 (OsSK12; Os01g0252100), PSEUDO-RESPONSE REGULATOR59 (OsPRR5; Os09g0532400), EIN3-BINDING F-BOX PROTEIN2 (OsEBF2; Os02g0200900), and PLANT U-BOX PROTEIN45 (OsPUB45; Os01g0901000) were detected after cold-stress treatment (Fig. 6), which indicates that pre-mRNA splicing for these genes was affected in rice seedlings under cold stress. The relative abundance of unspliced mRNA fragments in cold-stressed seedlings was higher in the RNAi-4 seedlings than in the wild-type plants (Fig. 6), which indicates that OsRH42 participates in rice pre-mRNA splicing under cold stress.

Accurate Expression of OsRH42 Is Important for Genome-Wide RNA Splicing in Rice under Cold Stress

To determine the role of OsRH42 in genome-wide pre-mRNA splicing under cold stress, we performed RNA-seq analyses of seedlings of the wild type, an OsRH42-knockdown line (RNAi-4), and an OsRH42-overexpression line (OX-6) under nonstress and cold-stress conditions. Approximately 62.4, 70.75, 49.34, 60.81, 53.79, and 48.48 million paired-end sequence reads, with a read length of 150 bp, from wild type, RNAi-4, and OE-6 under nonstress and cold-stress conditions, respectively, were generated (Supplemental Table S2). More than 83% of the reads were perfectly aligned to the rice IRGSP-1 reference genome sequence (Supplemental Table S2). Comparison of mapping frequency in individual samples showed a similar distribution of nonsplice reads and splice reads (Supplemental Table S2), which indicates that the RNA-seq data were of high consistency. Thus, alternative splicing (AS) of pre-mRNA, including intron retention (IR), exon skipping (ES), and alternative 5′ or 3′ splice site (AltDA) events, was analyzed for these RNA samples.

Aberrant IR, ES, and AltDA in 268 (255 genes), five (five genes), and 71 (48 genes) events, respectively, were detected in the OsRH42-knockdown seedlings under nonstress conditions compared with the wild-type seedlings (Table 2; Supplemental Data Set S1). Interestingly, a large number of aberrant IR, ES, and AltDA events, specifically 3,822 (2,747 genes), 30 (30 genes), and 301 (239 genes), respectively, were detected in the OsRH42-knockdown seedlings under cold-stress conditions (Table 2; Supplemental Data Set S1). Comparison of aberrant IR, ES, and AltDA genes in the RNAi-4 seedlings, which were detected in 98, one, and 20 genes, respectively, overlapped in the nonstressed and cold-stressed seedlings (Supplemental Fig. S8). These results indicate that expression of OsRH42 is required for accurate splicing of pre-mRNAs in rice in response to cold stress. In the OsRH42-overexpression seedlings, aberrant AS, including IR, ES, and AltDA events, also accumulated under cold-stress conditions, but with reduced abundance, with 1,247 (1,048 genes), 24 (24 genes), and 69 (59 genes) detected, respectively, compared with those in the OsRH42-knockdown seedlings (Table 2; Supplemental Data Set S2). However, more than 1,200 aberrant AS events, especially IR events (1,194), were detected in the OX-6 seedlings under nonstress conditions (Table 2; Supplemental Data Set S2). Among aberrant IR genes, 225 genes were detected under both nonstress and cold-stress conditions (Supplemental Fig. S8). These results suggest that overexpression of OsRH42 causes pre-mRNA splicing defects of certain genes in rice seedlings, regardless of cold stress.

Differential AS (DAS) events and genes in the OsRH42-knockdown (RNAi-4) and OsRH42-overexpression (OX-6) seedlings under nonstress and cold-stress conditions

OsRH42 Affects AS of a Subset of Cold-Induced Transcripts

Differentially expressed gene (DEG) analysis between nonstress and cold-stress conditions revealed that 1,059 and 894 genes showed increased and reduced expression in response to cold stress, respectively (Supplemental Data Set S3). Among the genes up-regulated by cold treatment, 396 and 175 genes showed aberrant IR events in the RNAi-4 and OX-6 seedlings, respectively (Fig. 7A). Of these genes, aberrant IR events were induced for 134 genes by knockdown expression as well as by overexpression of OsRH42 (Fig. 7A). Two up-regulated genes that encode MULTIDRUG RESISTANCE7 (MR7) and CALMODULIN BINDING PROTEIN-LIKE (CBP-LIKE), which have been shown previously to be up-regulated by cold treatment (Su et al., 2010), were selected to assess representative IR splicing events by RT-PCR analysis with gene-specific primer sets (Fig. 7; Supplemental Table S1). The accumulation of various intron-containing mRNA isoforms of these two genes was highly increased in the RNAi-4 seedlings, and slightly increased in the OX-6 seedlings, incubated at 4°C for 18 h (Fig. 7, C and D). For an additional two cold-induced genes, which encode an initiator-binding protein (TRF-LIKE2) and a haloacid dehalogenase-like hydrolase (HAD-LIKE), we verified that IR was enhanced only in the RNAi-4 seedlings in our RNA-seq data (Fig. 7, E and F). The levels of intron-containing mRNA isoforms of these two genes were higher in the RNAi-4 seedlings but slightly reduced in the OX-6 seedlings (Fig. 7, E and F). In addition, some genes up-regulated by cold treatment showed aberrant IR and/or ES and/or AltDA events in the current RNA-seq data (Fig. 8). Two up-regulated genes, encoding ALKALINE INVERTASE6 (INV6) and DST COACTIVATOR1 (DST1), were analyzed using RT-PCR with gene-specific primers (Fig. 8; Supplemental Table S1). For INV6, cold-induced In4 retention mRNAs were enhanced in the RNAi-4 and OX-6 seedlings, but cold-induced In1 retention mRNAs were only enhanced in the RNAi-4 seedlings (Fig. 8A). An alternative 5′ splicing site at In4 of INV6 was detected in all cold-treated seedlings and was enhanced by expression knockdown or overexpression of OsRH42 (Fig. 8A). However, INV6 mRNA with an alternative 3′ splicing site was only detected in OX-6 seedlings. For DST1, IR at the third intron, alternative 5′ splicing site at the fourth intron, and ES at the fourth exon mRNAs were detected at increased frequencies in cold-treated RNAi-4 and OX-6 seedlings compared with wild-type seedlings (Fig. 8B). Two alternative 3′ splicing sites at the 10th intron were only enhanced by knockdown expression of OsRH42 (Fig. 8B). These results indicate that OsRH42 plays a critical role in the regulation of mRNA splicing of cold stress-responsive genes under cold stress.

Regulation of Cold-Induced AS Is Affected by Abnormal Expression of OsRH42

Cold stress induces pre-mRNA AS events in Arabidopsis (Calixto et al., 2018). Rapid and dynamic AS have impacts on Arabidopsis response to cold stress (Calixto et al., 2018). In this study, several DAS, including 9,720 differential IR (DIR), 719 differential ES (DES), and 571 differential AltDA (DAltDA) events, were detected in wild-type seedlings after cold stress for 18 h (Table 3; Supplemental Data Set S4). Among these DAS events, DIR represented the majority, which occurred in 6,000 rice genes (Table 3). Cold-induced DES and DAltDA events were less abundant, with transcripts of 628 and 138 genes, respectively (Table 3). Compared with cold-induced DEGs, 863 cold-responsive genes showed cold-induced DAS events, including DIR, DES, and DAltDA events (Supplemental Fig. S9). These results suggest that AS for a number of genes, including cold-induced DEGs, were regulated in the rice response to cold stress.

Cold-induced DAS events and genes in wild-type, OsRH42 knockdown (RNAi-4), and OsRH42 overexpression (OX-6) seedlings

Cold-induced DAS events were dramatically increased in cold-treated RNAi-4 seedlings, especially DIR, with 16,775 events in 8,371 genes detected in the RNAi-4 seedlings (Table 3; Supplemental Data Set S5). The number of cold-induced DES and DAltDA events was also increased in the RNAi-4 seedlings, but to a much lesser degree than cold-induced DIR events (Table 3; Supplemental Data Set S5). Comparison of cold-induced DAS genes between the RNAi-4 and wild-type seedlings reveals that almost 87% (5,216 of 6,000 genes) of genes showing DIR events in the wild type were observed in the RNAi-4 seedlings (Fig. 9A; Supplemental Data Set S6), and 3,153 unique cold-induced DIR genes were detected in the RNAi-4 seedlings (Fig. 9A; Supplemental Data Set S6). In addition, 77% (481 of 628 genes) of DES genes and 59% (267 of 450 genes) of DAltDA genes in the wild type were also detected in the RNAi-4 seedlings under cold stress (Fig. 9A; Supplemental Data Sets S7 and S8). These results suggest that expression knockdown of OsRH42 increased the number of cold-induced DAS events and genes. By contrast, overexpression of OsRH42 slightly reduced cold-induced DIR, DES, and DAltDA events and genes, respectively (Table 3; Supplemental Data Set S6). Only 68% (4,073 of 6,000 genes), 55% (344 of 628 genes), and 44% (199 of 450 genes) of cold-induced DIR, DES, and DAltDA genes, respectively, in the wild type were shared with those in the OX-6 seedlings (Fig. 9A; Supplemental Data Sets S7–S9). These results suggest that overexpression of OsRH42 reduced the frequency of cold-induced DIR, DES, and DAltDA transcripts observed in the wild type but led to the accumulation of an additional subset of abnormal AS transcripts.

Three cold-induced DIR events detected in three genes up-regulated by cold treatment, namely MR7, CBP-LIKE, and INV6, were selected to assess the effects of OsRH42 on the regulation of cold-induced DAS. Two-week-old seedlings of the wild type, RNAi-4, and OX-6 were incubated under cold stress for various periods. The accumulation of intron-containing mRNAs of MR7 was detected in RNAi-4 and OX-6 following 6 h of cold stress, whereas accumulation in the wild type was detected after 12 h of cold stress (Fig. 9C). Increased abundance of MR7 IR mRNAs in RNAi-4 and OX-6 seedlings was detected after 6 to 24 h of cold stress, respectively, compared with the wild type, which was detected from 12 to 24 h (Fig. 9C). Similar expression patterns were observed for CBP-LIKE (Fig. 9D) and INV6 (Fig. 9E); IR-containing pre-mRNAs of these two genes were detected earlier and showed greater abundance in RNAi-4 and OX-6 seedlings in comparison with the wild type. In addition, rice seedlings were incubated under cold stress for 24 h and then transferred to the recovery (nonstress) condition (28°C) for various periods. Accumulation of MR7 IR mRNAs was rapidly reduced in the wild type, RNAi-4, and OX-6 when the seedlings were moved to the nonstress condition for recovery after 2 h (Fig. 9C). The CBP-LIKE and INV6 IR mRNAs also rapidly decreased in abundance in the recovery phase in wild-type seedlings (Fig. 9, D and E). However, continued high abundance of CBP-LIKE and INV6 IR mRNAs was detected in RNAi-4 and OX-6 seedlings after 6 to 12 h of recovery. Taken together, these results indicate that abnormal expression of OsRH42 impairs the rapid regulation of certain genes involved in AS in the response of rice to cold stress.

DISCUSSION

Investigation of rice adaptation to a variety of abiotic stresses, together with the improvement of yield and quality, are the focus of current research to address the problem of climate change exacerbated by soil degradation. Low temperature is a major environmental stress, which adversely affects rice growth and productivity. Splicing of pre-mRNA is a crucial posttranscriptional regulatory stage in gene expression in plants and is closely connected to environmental changes (Dubrovina et al., 2013; Reddy et al., 2013). Pre-mRNA splicing is mediated by the spliceosome, which is composed of five uridine-rich snRNPs, namely U1, U2, U4, U5, and U6. These snRNPs function to interact with unspliced pre-mRNAs with the correct sequence at the 5′ splicing site, the 3′ splicing site, and a branch-point region to process intron removal and exon joining. A number of DEAD-box RNA helicases facilitate RNA folding and rearrange the snRNPs at various specific steps during the pre-mRNA splicing process (Liu and Cheng, 2015). Given that the RNA conformation is temperature sensitive (Hofinger and Zerbetto, 2010; Bevilacqua et al., 2016), DEAD-box RNA helicases are of increasing importance to maintain the efficiency of pre-mRNA splicing under a low-temperature environment (Urushiyama et al., 1997; Schaffert et al., 2004; Lee et al., 2006; Du et al., 2015; Kim et al., 2017). In this study, we demonstrate that a nuclear DEAD-box RNA helicase, OsRH42, plays a critical role in pre-mRNA splicing in rice exposed to cold stress and is essential for the cold-stress tolerance of rice growth.

OsRH42 has a similar structure in its RNA helicase domain, DPLD motif, and RD/RS domain to Arabidopsis RCF1. However, it is uncertain whether these two proteins perform similar or different biological functions. Here, our evidence shows that although OsRH42 and RCF1 are very similar proteins, and both are required in cold-induced pre-mRNA splicing pathways in rice and Arabidopsis, respectively, the precise involvement of OsRH42 in the pre-mRNA splicing mechanism in rice may differ from that of RCF1 in Arabidopsis. There are five important points to make, as follows. (1) The cold-induced expression pattern of OsRH42 was similar to that of RCF1. The expression of OsRH42 in rice is ubiquitous and is seen predominantly in leaves and may be regulated not only at the transcriptional level (Fig. 1A; Supplemental Fig. S2C). However, the equivalent findings for RCF1 are as yet unclear. (2) The OsRH42-GFP protein is localized to the splicing speckles (Fig. 4B). However, the RCF-GFP fusion protein was detected only in the nucleus (Guan et al., 2013). (3) Three rice OsRH42-knockdown transgenic lines showed hypersensitivity to cold stress (Fig. 3) as did the rcf1 mutants of Arabidopsis (Guan et al., 2013). However, RCF1 has a negative role in regulating CBF gene expression in Arabidopsis (Guan et al., 2013), and the function of OsRH42 in the response of rice to low temperature is independent for the CBF-mediated cold-response system (Supplemental Fig. S5). (4) Both OsRH42 and RCF1 were involved in pre-mRNA splicing in rice (Figs. 6–9; Tables 2 and 3) and Arabidopsis (Guan et al., 2013), respectively, but only for OsRH42 was there a direct interaction with U2 and U1 snRNA in vivo (Fig. 5). (5) Overexpression of OsRH42 in transgenic rice had a negative effect on rice plant growth at the vegetative and reproduction stages (Fig. 2). Moreover, the OsRH42-overexpression lines showed disruption in terms of pre-mRNA splicing and hypersensitivity to low temperatures (Figs. 6  9; Tables 2 and 3). By contrast, constitutive overexpression of RCF1 increased the tolerance of Arabidopsis to cold stress (Guan et al., 2013).

Both OsRH42-knockout and -overexpression transgenic rice lines exhibited defects in plant growth and development (Fig. 2; Table 1), and their pre-mRNA splicing was affected under normal growth conditions (Table 2). These results suggest that disorder in pre-mRNA splicing causes defects in plant growth and development, as reported previously in Arabidopsis (Syed et al., 2012; Loraine et al., 2013; Staiger and Brown, 2013). However, the OsRH42-overexpression lines showed a more severe phenotypic defect than the OsRH42-knockdown lines (Fig. 2; Table 1). Accumulation of aberrant pre-mRNA splicing genes in the OsRH42-overexpression lines, compared with the OsRH42-knockdown lines, under normal growth conditions (Table 2) is a reasonable explanation for this. More severe phenotypes were also observed in OsRH42-overexpression transgenic seedlings under cold-stress conditions. Survival rates of the OsRH42-overexpression transgenic seedlings were lower than those of the OsRH42-knockdown seedlings (Fig. 3). In this study, survival rates were determined 10 d after recovery from cold stress. It is possible that, in the recovery stage of plants, OsRH42-overexpression transgenic seedlings have more aberrant pre-mRNAs than OsRH42-knockdown lines, thus explaining their lower survival rates compared with the OsRH42-knockdown lines. However, our findings do not rule out the possibility that overexpression of OsRH42 has a particular effect on rice plant development and response to cold stress, which might not be related to its function in pre-mRNA splicing.

Rice OsRH42 is a Prp5-homologous RNA helicase. In yeast, Prp5 binds directly to U2 snRNA to function in the conformational change of U2 snRNP in the recognition of the branch-point region of pre-mRNA (O’Day et al., 1996; Liang and Cheng, 2015). After U2 snRNP binds to the branch-point site, Prp5 is released from the complex, which is essential for tri-snRNP (U2, U4, and U5) to be recruited into the spliceosome; the spliceosome then proceeds following the steps of the splicing pathway (Liang and Cheng, 2015). In consideration of the expression pattern, the subcellular localization, the snRNA-binding abilities of OsRH42, the pre-mRNA splicing phenomena in the OsRH42-knockdown and -overexpression transgenic lines, and the yeast pre-mRNA splicing mechanism, a hypothesis is proposed here to address the biological relevance of observations with pre-mRNA splicing in rice in this study. Under normal temperature conditions, the abundance and function of OsRH42 are irrelevant possibly due to the existence of other RNA helicases (Fig. 1); the U2 snRNP has a functional structure that can participate in pre-mRNA splicing and generate mature mRNAs for rice growth. By contrast, increased OsRH42 under low temperatures (Fig. 1) is important to process pre-mRNA splicing, and this ensures a sufficient quantity of mature mRNAs for adaptation to cold stress in rice. OsRH42 was shown to bind directly to U2 snRNA (Fig. 5). The OsRH42-knockdown transgenic lines have reduced amounts of OsRH42 to bind U2 snRNP in the splicing speckles, leading to less U2 snRNP to associate with the spliceosome in cold-stressed rice plants, and thus their seedlings exhibited the phenotype of cold-stress hypersensitivity (Fig. 3). Increased OsRH42 abundance caused a reduction in the number of aberrant pre-mRNA splicing genes in the transgenic rice seedlings under cold-stress conditions (Table 3). However, a high amount of OsRH42 may also increase the likelihood that OsRH42 still binds to U2 snRNP after U2 snRNA binds to the branch-point site. As speculation from the pre-mRNA splicing mechanism in yeast (Liang and Cheng, 2015), tri-snRNP cannot be recruited into the spliceosome without the release of OsRH42 from U2 snRNA. Thus, rice transgenic lines overexpressing OsRH42 showed disruption of the pre-mRNA splicing pathway (Figs. 7–9; Tables 2 and 3), growth retardation (Fig. 2), as well as intolerance of cold stress (Fig. 3). Taken together, rice plants need accurate control of OsRH42 homeostasis in order to respond to ambient temperatures.

Forward genetic studies have shown that several spliceosome-related proteins are important for pre-mRNA splicing and stress response in Arabidopsis (Filichkin et al., 2010; Staiger and Brown, 2013; Ding et al., 2014). For example, SAD1/LSM5 may be involved in stabilization of the U6 snRNP during pre-mRNA splicing. A mutation in the homolog of the LSM5 protein, SAD1/LSM5, leads to increased sensitivity to drought and abscisic acid (Xiong et al., 2001; Cui et al., 2014). RDM16, a component of the U4/U6 snRNP, is involved in pre-mRNA splicing and plant response to salt stress and abscisic acid (Huang et al., 2013). The spliceosomal protein U1A is involved in 5′ splicing site recognition and salt-stress tolerance (Gu et al., 2018). In addition, STABILIZED1, a Prp1p and Prp6p homolog, and PRPF31, which regulates the formation of the tri-snRNP, are required for the response to cold stress in Arabidopsis (Urushiyama et al., 1997; Schaffert et al., 2004; Lee et al., 2006; Du et al., 2015; Kim et al., 2017). Our results here suggest that OsRH42 facilitates pre-mRNA splicing mediated by U2 snRNP function at the branch-point site (Fig. 5). Indeed, RCF1, an OsRH42 homolog, performs an essential role in pre-mRNA splicing and is involved in cold tolerance in Arabidopsis, as reported previously (Guan et al., 2013). These studies suggest that the pre-mRNA splicing mechanism is generally conserved in eukaryotes (Lorković et al., 2000). However, different spliceosome-related proteins are indicated to be connected to a unique stress response in the plant. The branch-point site recognition and tri-snRNP formation may be cold stress-targeted steps in the pre-mRNA splicing pathway. A recent report indicated that the expression of U2B″-LIKE is required for cold acclimation and freezing tolerance in Arabidopsis (Calixto et al., 2018), which provides further support for this hypothesis.

AS involves different splicing events from the same pre-mRNA, leading to a variety of mature mRNAs that may code for different functional proteins. In many different plants, AS has been shown to contribute in the response to different stresses (Mastrangelo et al., 2012; Leviatan et al., 2013; Staiger and Brown, 2013; Hartmann et al., 2016; Klepikova et al., 2016; Laloum et al., 2018). In Arabidopsis, cold-induced rapid and dynamic AS has an effect on the cold response transcriptome, which in turn affects plant growth at low temperatures (Calixto et al., 2018). In our genome-wide study, we demonstrate that AS is also detectable in rice seedlings after 18 h of cold stress. Furthermore, analysis of three selected genes reveals that cold-induced AS was detectable at earlier time points. Accumulated IR-containing mRNAs were detected after 6 h of cold stress (Fig. 9). When cold-stressed rice seedlings were transferred to nonstress growth conditions for 2 h, the pre-mRNA splicing mechanism rapidly returned to normal. Accumulated IR-containing mRNA was undetectable in rice seedlings after 2 h under nonstress conditions (Fig. 9). Thus, cold-induced AS is an additional strategy in rice for adaptation to cold stress. Moreover, our results indicate that OsRH42 plays a role in the regulation of cold-induced AS. OsRH42 expression knockdown leads to the acceleration of cold-induced AS in cold-stressed rice and delays normal-type AS in the recovery phase (Fig. 9).

An increasing body of evidence indicates that overexpression of a certain spliceosome or other splicing factors can increase plant tolerance to various environmental stresses (Guan et al., 2013; Cui et al., 2014; Cui and Xiong, 2015; Gu et al., 2018). Thus, manipulation of the AS of specific genes may be an effective approach for plants to cope with abiotic stress. However, overexpression of these proteins has not yet been investigated in crop plants. Given that constitutive overexpression of OsRH42 in rice cannot protect seedlings from chilling injury and that such transgenic plants show severe stunting of growth in rice fields under normal growth conditions, accurate control of OsRH42 homeostasis is important for rice adaptation to cold stress. Use of stress-inducible promoters for the expression of OsRH42 may minimize its negative effects on rice plant growth in future applications.

MATERIALS AND METHODS

Plant Materials and Growth Conditions

Rice (Oryza sativa) ‘Tainung 67’ was used in this study. Transgenic rice plants were cultivated at the Agricultural Experiment Station, National Chung-Hsing University (Taichung, Taiwan). Sterilized seeds were placed on one-half-strength Murashige and Skoog agar medium supplemented with 3% (w/v) Suc and cultivated at 28°C under constant light for 10 d. Seedlings were then transferred to a hydroponic culture medium (Kimura B solution) for 4 d before being used in experiments. Stress treatments were as follows: salt, seedlings were transferred to a culture solution containing 150 mm NaCl for 2 d; drought, seedlings were air dried for 3 h; osmotic, seedlings were transferred to a culture solution containing 40% (w/v) PEG for 48 h; heat, seedlings were transferred to 42°C for 2 d; cold, seedlings were transferred to 4°C for 2 to 9 d.

Primers

The nucleotide sequences of all primers used for plasmid construction, PCR, RT-PCR, and RT-qPCR analyses are listed in Supplemental Table S1.

Plasmids

Plasmid pMDC43 (Curtis and Grossniklaus, 2003) was used for fusion of the OsRH42-GFP chimeric protein. The pCAMBIA vectors were obtained from CAMBIA.

Plasmid Construction

The OsRH42 coding regions were amplified with specific primers (Supplemental Table S1) using Phusion High-Fidelity DNA Polymerase (New England Biolabs) using the cDNA of seedlings at the three-leaf stage as templates. The PCR products were cloned into the yT&A cloning vector (Yeastern) to generate pOsRH42. For subcellular localization of OsRH42, full-length cDNA fragments were excised from pOsRH42 with AscI and NotI and then ligated into the corresponding sites of the pENTR-TOPO vector to generate the pOsRH42-ENTR vector. Using LR Clonase (Invitrogen), the OsRH42 DNA fragments were transferred from the entry vector to the destination vector, pMDC85, by recombination to generate the OsRH42-GFP plasmid.

To generate a construct for ectopic expression of OsRH42, pOsRH42 was digested with BamHI, and the full length of the OsRH42 cDNA fragment was isolated and then introduced into the pAHC18 vector between the maize (Zea mays) Ubi promoter and the Nos terminator to generate pUbi-OsRH42. The pUbi-OsRH42 construct was further digested with SspI and introduced into the pCAMBIA1301 binary vector to generate p1301OsRH42.

For the construction of the OsRH42 RNAi vector, a 290-bp DNA fragment located at the 3′ untranslated region of OsRH42 was amplified using specific primers (Supplemental Table S1). This DNA fragment was cloned into the yT&A cloning vector, generating pRH42Ri. The GFP cDNA was amplified by PCR using a forward primer and a reverse primer (Supplemental Table S1) and was then subcloned into the yT&A cloning vector to generate pGFPRI. The OsRH42 RNAi DNA fragment was isolated from pRH42Ri by digestion with EcoRI and BamHI, the GFP DNA fragment was isolated from pGFPRI by digestion with EcoRI, and these two fragments were ligated into the BamHI site of the pAHC18 expression vector, generating pAHC18-OsRH42-Ri. This RNA-silencing construct was linearized by digestion with HindIII and inserted into the HindIII site of the pCAMBIA1301 binary vector for Agrobacterium tumefaciens-mediated gene transformation.

RT-PCR and RT-qPCR Analyses

Total RNA was isolated from whole seedlings as previously described (Huang et al., 2016b). First-strand cDNA was synthesized using ReverTra Ace reverse transcriptase (Toyobo) with oligo(dT) primers. A 20-fold dilution of the resultant first-strand cDNA was subjected to PCR (22–35 reaction cycles) with gene-specific primers (Supplemental Table S1). For RT-qPCR, a 20-fold dilution of the first-strand cDNA was subjected to qPCR using a PikoReal Real-Time PCR System (Thermo Fisher), in accordance with the manufacturer’s instructions. The PCR procedure was repeated independently at least three times. The relative gene expression levels are expressed as ratios of the abundance of the target gene’s mRNA to that of Act1 mRNA. Data were analyzed using the PikoReal software provided by the manufacturer. The gene-specific primers used for RT-qPCR are listed in Supplemental Table S1.

Plant Transformation

Rice embryonic calli were induced from germinated seeds on N6 solid medium with 9 μm 2,4-dichlorophenoxy. A. tumefaciens strain EHA105 was used to perform rice transformation as previously described (Huang et al., 2010b). Transformed calli were selected on N6 medium containing 50 mg L⁻¹ hygromycin B.

Subcellular Localization Analysis

A protoplast transient expression system was used as previously described (Chou et al., 2017). Rice protoplasts were isolated from the sheaths of 10-d-old seedlings and then incubated with expression vectors and PEG for 20 min. After the PEG had been washed away, the rice protoplasts were incubated in WI solution for 6 h and observed using an Olympus IX71 (Olympus) inverted fluorescence microscope.

Electrolyte Leakage Assay

Examination of electrolyte leakage was performed as described by Yoon et al. (2016) with a slight modification. Briefly, 2-week-old seedlings were treated with 4°C for 6 d. Leaf samples were cut into 1-cm slices from the fourth leaf of seedlings and subjected to electrolyte leakage measurement. The cold-induced electrolyte leakage (%) was calculated as electrolyte leakage (cold)/electrolyte leakage total × 100.

RNA Immunoprecipitation Assays

RNA immunoprecipitation assays were performed using an RIP-Assay Kit (MBLI). Fresh calli were ground in liquid N₂ and then resuspended in a lysis buffer with protease inhibitor cocktail (Roche) and RNase inhibitor (Toyobo). The lysed supernatant was incubated with anti-GFP antibody (Abcam) coupled with Mag Sepharose-Protein G (GE Healthcare). After the preparation had been washed four times, the RNA was extracted from GFP antibody-immobilized protein G Sepharose beads-RNA-protein complex. First-strand cDNA was synthesized with random primers, and RT-PCR was performed with specific primers.

RNA-Seq Analysis

Total RNA was isolated from three biological replicates of wild-type, RNAi-4, and OX-6 seedlings under 4°C for 0 and 18 h using TRIzol Reagent (Thermo Fisher). The RNA quality was examined using a NanoPhotometer spectrophotometer (IMPLEN). Three biological replicates of RNAs were pooled, polyadenylated RNAs were isolated, and RNA-seq libraries were constructed using the NEBNext Ultra RNA Library Prep Kit for Illumina (New England Biolabs). Having assessed the library quality, the library preparations were sequenced on the NovaSeq 6000 system with 150-bp paired-end reads. Sequence reads were mapped to the rice IRGSP-1 reference genome sequence using TopHat v2.0.12. HTSeq v0.6.1 was used to count the read numbers mapped to each gene. The fragments per kilobase of transcript per million mapped reads were used for estimation of gene expression levels (Trapnell et al., 2010). The read counts were adjusted using the edgeR program package through one scaling-normalized factor before performing differential gene expression analysis. Differential expression analysis of two conditions was performed using the DEGSeq R package (1.20.0). The corrected P value (false discovery rate) of 0.005 and log₂ (fold change) of 1 were set as the threshold for significantly differential expression. DAS events of two conditions were analyzed using the RACKJ (http://rackj.sourceforge.net/) package as previously described (Chang et al., 2014; Kanno et al., 2017). All genes, the average depths of all exons and all introns, and read counts for all splicing junctions of read counts were computed by the RACKJ. In this study, events containing nonannotated splicing junctions or nonspliced sequences spanning annotated splicing junctions in representative gene models were considered AS events. For the preference of IR, a χ² for goodness of fit (Sasaki et al., 2015) was used, where read depths of an intron in two samples were compared with the background of read depths of neighboring exons. The null hypothesis was assumed that the probabilities of IR were the same in the two samples, and a significant false discovery rate (less than 0.001) indicates that the chance of IR was altered in one of the two samples. Using similar methods as for the IR events, the χ² test was applied for goodness of fit to measure the preference of ES events and AltDA events, in which those with a skipped exon and other junctions of the same intron were compared.

Statistical Analyses

The data presented here were statistically analyzed by Student’s t test, P < 0.05 as significant difference. Statistical analysis was performed with the Excel program.

Accession Numbers

Sequence data from this article can be found in the GenBank/EMBL data libraries under the following accession numbers: Os08g0159900 (OsRH42), Os09g0522200 (OsDREB1A), Os09g0522000 (OsDREB1B), Os01g0252100 (OsSK12), Os09g0532400 (OsPRR5), Os02g0200900 (OsEBF2), Os01g0901000 (OsPUB45), Os03g0718100 (ACT1), Os01g0695800 (MR7), Os01g0134700 ) CBP-LIKE), Os03g0274300 (TRF-LIKE2), Os03g0273800 (HAD-LIKE), Os11g0175400 (INV6), and Os10g0456800 (DST1).

Supplemental Data

The following supplemental materials are available.

• Supplemental Figure S1. Screening for rice cold-induced DEAD-box RNA helicase gene.

• Supplemental Figure S2. Expression pattern of the OsRH42p::GUS chimeric gene.

• Supplemental Figure S3. Characterization of OsRH42-knockdown and -overexpression transgenic lines.

• Supplemental Figure S4. Comparison of internode distance of 4-month-old rice plants among the wild type, RNAi-2, RNAi-17, RNAi-20, OX-6, and OX-17.

• Supplemental Figure S5. Comparison of OsDREB1A and OsDREB1B expression in wild-type, RNAi-4, and OX-6 seedlings.

• Supplemental Figure S6. Phylogenetic relationships of RH42 family members.

• Supplemental Figure S7. Subcellular localization of the OsRH42-GFP fusion protein.

• Supplemental Figure S8. Venn diagram of aberrant AS in RNAi-4 and OX-6 seedlings under nonstress and cold-stress conditions.

• Supplemental Figure S9. Venn diagram of cold-induced DAS genes and DEGs in wild-type, RNAi-4, and OX-6 seedlings.

• Supplemental Table S1. List of primers used in this study.

• Supplemental Table S2. Mapping statistics of RNA-seq.

• Supplemental Data Set S1. DAS genes between RNAi-4 and wild-type seedlings.

• Supplemental Data Set S2. DAS genes between OX-6 and wild-type seedlings.

• Supplemental Data Set S3.Cold-induced DEGs in wild-type seedlings.

• Supplemental Data Set S4. Cold-induced DAS genes in wild-type seedlings.

• Supplemental Data Set S5.Cold-induced DAS genes in RNAi-4 seedlings.

• Supplemental Data Set S6.Venn diagram of cold-induced DIR genes in wild-type, RNAi-4, and OX-6 seedlings.

• Supplemental Data Set S7.Venn diagram of cold-induced DES genes in wild-type, RNAi-4, and OX-6 seedlings.

• Supplemental Data Set S8.Venn diagram of cold-induced DAltDA genes in wild-type, RNAi-4, and OX-6 seedlings.

• Supplemental Data Set S9.Cold-induced DAS genes in OX-6 seedlings.

ACKNOWLEDGMENTS

We thank the Agricultural Experiment Station, National Chung-Hsing University, for providing space to plant transgenic rice.

LITERATURE CITED

Author notes