SET DOMAIN GROUP 721 protein functions in saline-alkaline stress tolerance in the model rice variety Kitaake.

To isolate the genetic locus responsible for saline-alkaline stress tolerance, we developed a high-throughput activation tagging-based T-DNA insertion mutagenesis method using the model rice (Oryza sativa L.) variety Kitaake. One of the activation-tagged insertion lines, activation tagging 7 (AC7), showed increased tolerance to saline-alkaline stress. This phenotype resulted from the overexpression of a gene that encodes a SET DOMAIN GROUP 721 protein with H3K4 methyltransferase activity. Transgenic plants overexpressing OsSDG721 showed saline-alkaline stress-tolerant phenotypes, along with increased leaf angle, advanced heading and ripening dates. By contrast, ossdg721 loss-of-function mutants showed increased sensitivity to saline-alkaline stress characterized by decreased survival rates and reduction in plant height, grain size, grain weight and leaf angle. RNA sequencing (RNA-seq) analysis of wild-type Kitaake and ossdg721 mutants indicated that OsSDG721 positively regulates the expression level of HIGH-AFFINITY POTASSIUM (K+ ) TRANSPORTER1;5 (OsHKT1;5), which encodes a Na+ -selective transporter that maintains K+ /Na+ homeostasis under salt stress. Furthermore, we showed that OsSDG721 binds to and deposits the H3K4me3 mark in the promoter and coding region of OsHKT1;5, thereby upregulating OsHKT1;5 expression under saline-alkaline stress. Overall, by generating Kitaake activation-tagging pools, we established that the H3K4 methyltransferase OsSDG721 enhances saline-alkaline stress tolerance in rice.

Introduction

Millions of hectares of irrigated and unirrigated agricultural land are affected by salinization and alkalization (Shahid et al., 2018; Shrivastava and Kumar, 2015). Saline–alkaline soils are characterized by both high sodium ion (Na+) concentration and high pH, which cause more complex stress effects on plants than pH‐neutral saline soils (Tang et al., 2014). To survive, plants growing in saline–alkaline soils have to cope with both physiological drought and Na+ toxicity, in addition to the cellular damage induced by high pH (Cheng et al., 2020; Liu et al., 2010). Rice (Oryza sativa L.) serves as a staple food crop for more than half of the world population (Gross and Zhao, 2014) and as a model monocot for bioenergy research (Izawa and Shimamoto, 1996). To date, multiple mutant rice collections have been generated in different cultivars including Dong Jin, Hwayoung, Nipponbare and Zhonghua 11 (Wang et al., 2013b; Wei et al., 2013). Most of these mutant collections have been generated artificially through EMS mutagenesis (Henry et al., 2014), irradiation (Li et al., 2017; Wang et al., 2013b; Wei et al., 2013), T‐DNA insertion (Chen et al., 2003; Hsing et al., 2007; Jeon et al., 2000; Sallaud et al., 2003; Wu et al., 2003), transposon/retrotransposon insertion (van Enckevort et al., 2005; Kolesnik et al., 2004; Miyao et al., 2003; Wang et al., 2013b), RNAi (Wang et al., 2013a), TALEN‐based gene editing (Li et al., 2012; Moscou and Bogdanove, 2009) and CRISPR/Cas9‐based genome editing (Jiang et al., 2013; Miao et al., 2013; Xie et al., 2015). Although insertion mutagenesis is an effective approach for determining the function of a genetically redundant gene, loss‐of‐function mutations often fail to reveal the function of a specific member of a gene family (Alonso et al., 2003). Activation tagging is a powerful gain‐of‐function approach used to elucidate the functions of genes, especially those that exhibit high sequence similarity to other genes, a feature recalcitrant to loss‐of‐function genetic analyses (Gou and Li, 2011; Wan et al., 2009; Weigel et al., 2000). T‐DNA activation tagging generates dominant mutations through the insertion of a T‐DNA‐carrying constitutive enhancer elements at a random position in the genome, resulting in the transcriptional activation of flanking genes (Jeong et al., 2002; Memelink, 2003). This method involves the generation of a large number of transformed plants using a specialized T‐DNA construct, followed by the selection of plants with the desired phenotype (Jeong et al., 2006). We recently established a mutagenized population in Kitaake, a model japonica rice variety with a short life cycle of 9 weeks (Li et al., 2016), using the activation tagging approach.

The SET domain proteins are known to methylate histone H3 at several lysine (K) residues, including H3K4, H3K9, H3K27, H3K36 and H4K20. The trithorax group (TrxG) protein family comprises a large number of functionally diverse regulatory proteins (Kingston and Tamkun, 2014), including the Su(var)3‐9, Enhancer‐of‐zeste and Trithorax (SET) domain‐containing proteins with H3K4 methyltransferase activity, ATP‐dependent chromatin‐remodelling factors and other associated proteins (Avramova, 2009; Krajewski et al., 2005). The rice genome harbours three genes encoding TrxG family proteins, including the SET domain group protein 723 (OsSDG723; related to ATX1 and ATX2), OsSDG721 (closely related to ATX3) and OsSDG705 (related to ATX4 and ATX5) (Jiang et al., 2018). Loss‐of‐function mutations of OsSDG721 and OsSDG705 in Oryza sativa ssp. japonica cv. Nipponbare caused abnormal phenotypes including semi‐dwarfism, reduced cell length and reduced panicle branching (Jiang et al., 2018). Moreover, genome‐wide H3K4me3 levels and H3K4me3 enrichment at the gene body regions of gibberellin (GA) metabolic and signalling genes, including SLR1, GID1 and GA20 oxidases, were reduced in the ossdg721 ossdg705 double mutant, indicating that OsSDG721 participates in the regulation of H3K4me3 levels to control gene expression (Jiang et al., 2018).

In this study, by screening the activation‐tagging transgenic line collection established in Kitaake background, we identified a gain‐of‐function mutant, activation tagging 7 (AC7), which showed increased tolerance to saline–alkaline stress. Thermal asymmetric interlaced PCR (TAIL‐PCR) revealed that the increased tolerance to saline–alkaline stress resulted from the upregulation of the gene encoding SDG721 with H3K4 methyltransferase activity. Transgenic lines overexpressing OsSDG721 showed saline–alkaline stress‐tolerant phenotypes characterized by an increased leaf angle and advanced heading and ripening dates, whereas ossdg721 loss‐of‐function mutants showed increased sensitivity to saline–alkaline stress, as indicated by the reduction in the survival rate, plant height, grain size, grain weight and leaf angle. Furthermore, ossdg721 mutants exhibited reduced expression levels of HIGH‐AFFINITY POTASSIUM (K) TRANSPORTER1;5 (OsHKT1;5), which was correlated with reduced H3K4me3 levels. We further showed that OsSDG721 binds to and modulates H3K4me3 levels in the promoter and coding region of OsHKT1;5 under saline–alkaline stress.

Results

Identification of an activation‐tagged rice mutant with enhanced saline–alkaline stress tolerance

To generate an activation‐tagging vector, we took advantage of the pCAMBIA1301 binary vector, which can be used for promoter trapping and activation tagging in Kitaake (Ozawa, 2009). We inserted the green fluorescent protein (GFP) gene driven by the maize ubiquitin (Ubi1) promoter between four repeats of the cauliflower mosaic virus 35S promoter (4×35S) and the 35S cassette (Figure 1a). This experimental design greatly facilitates the positive selection of transgenic callus generated by Agrobacterium‐mediated transformation (Figure S1a, b). In the T2 generation, line AC7 showed increased survival rates under saline–alkaline conditions (25 mm Na2CO3, pH = 10.0) (Figure 1b, c). The AC7 line was backcrossed with wild‐type Kitaake, followed by selfing, to obtain the BC1F2 progeny, in which approximately 76% lines showed GFP signals and saline–alkaline stress tolerance (Figure S1c, d), indicating that AC7 carries a dominant mutation. To further purify the mutant genotype and to remove unwanted mutations caused by the random T‐DNA insertion process (Østergaard and Yanofsky, 2004), we backcrossed line AC7 with wild‐type Kitaake three more times, followed by phenotyping the BC4F2 progeny. Genomic sequences flanking the T‐DNA insertion in line AC7 were obtained by TAIL‐PCR analysis (Figure 1d). The 4×35S sequence was detected in an intergenic region on chromosome 1 containing four genes within a 10‐kb sequence flanking the insertion site: OsKitaake01g083100, OsKitaake01g083200, OsKitaake01g083300 and OsKitaake01g083400 (Figure 1e). Transcript profiling of 4‐week‐old AC7 plants showed that the expression of OsKitaake01g083200 was increased by 8‐fold compared with wild‐type Kitaake, whereas that of OsKitaake01g083100, OsKitaake01g083300 and OsKitaake01g083400 was unaffected (Figure 1f). Moreover, AC7 mutant plants showed increased leaf angle and advanced heading date compared with wild‐type Kitaake (Figure 1g‐i).

Figure 1

Isolation of saline–alkaline stress‐tolerant activation tagging line, AC7. (a) Schematic representation of the vector used to generate activation tagging lines in Kitaake. RB, right boarder; 4×35S, four repeats of the cauliflower mosaic virus (CaMV) 35S promoter; P, CaMV 35S promoter; EGFP, enhanced green fluorescent protein; P, maize Ubiquitin1 promoter; Ter, nopaline synthase (NOS) terminator; Hyg, hygromycin resistance gene; LB, left boarder. (b, c) Images (b) and survival rates (c) of Kitaake, AC5, AC6 and AC7 plants before and after recovery from saline–alkaline stress treatment (25 mM Na2CO3, pH = 10.0). In (b), scale bar = 4 cm. Data in (c) represent mean ± standard deviation (SD) of three biological replicates, each containing 10 plants. Black circles represent the values of individual plants. Significant differences between wild‐type Kitaake and AC plants were evaluated by two‐way analysis of variance (ANOVA), followed by Tukey’s multiple comparison test. (d) Identification of the T‐DNA insertion site by thermal asymmetric interlaced PCR (TAIL‐PCR). AD, arbitrary degenerate primer; 2, secondary PCR products; 3, tertiary PCR products. (e) Schematic representation of the genomic DNA flanking the T‐DNA insertion site. (f) Semi‐quantitative RT‐PCR analyses using different primer sets. OsACT2 served as a loading control. (g) Photographs of 90‐d‐old Kitaake and AC7 plants grown in field. Scale bar = 15 cm. (h, i) Quantification of agronomic traits including leaf angle (h) and flowering time (i) in Kitaake and AC7 plants. Data represent mean ± SD of three biological replicates, each containing 12 plants. Significant differences between Kitaake and AC7 plants were evaluated by one‐way ANOVA, followed by Tukey’s multiple comparison test.

Ectopic expression of OsSDG721 exhibits saline–alkaline stress tolerant phenotypes

To further investigate the role of OsSDG721 in saline–alkaline stress tolerance, we generated three independent OsSDG721 lines (OsSDG721OX‐1, OsSDG721OX‐2 and OsSDG721OX‐3) by cloning the FLAG epitope at the 3' end of the OsSDG721 gene (Figure 2a). Real‐time quantitative PCR (RT‐qPCR) revealed that transcript levels of OsSDG721 were dramatically increased in all three transgenic OsSDG721OX lines compared with the wild‐type Kitaake (Figure 2b). Additionally, OsSDG721OX lines displayed saline–alkaline stress tolerant phenotypes compared with wild‐type Kitaake (Figure 2c) and showed higher survival rates than wild‐type Kitaake plants (Figure 2d). To examine the production of reactive oxygen species (ROS), including hydrogen peroxide (H2O2) and superoxide anion (O2 ‐), we stained the leaves of transgenic and wild‐type plants with diaminobenzidine (DAB) and nitrotetrazolium blue chloride (NBT). The results showed lower levels of ROS in OsSDG721OX lines than in wild‐type Kitaake (Figure 2e). Additionally, OsSDG721OX lines showed an increased leaf angle and advanced heading dates compared with wild‐type Kitaake (Figure S2). Taken together, these results indicate that OsSDG721 is required for saline–alkaline stress tolerance and plays an important role in plant growth and development.

Figure 2

Characterization of OsSDG721 overexpression (OsSDG721OX) lines. (a) Immunoblot analysis of the OsSDG721‐FLAG protein in Kitaake, OsSDG721OX‐1, OsSDG721OX‐2 and OsSDG721OX‐3 plants using anti‐FLAG antibody. Histone H3 was used as a loading control. (b) RT‐qPCR analysis of OsSDG721 transcript levels in Kitaake, OsSDG721OX‐1, OsSDG721OX‐2 and OsSDG721OX‐3 plants. Significant differences between Kitaake and OX lines were evaluated by one‐way ANOVA, followed by Tukey’s multiple comparison test. (c, d) Images (c) and survival rates (d) of Kitaake, OsSDG721OX‐1, OsSDG721OX‐2 and OsSDG721OX‐3 plants before and after recovery from the saline–alkaline stress treatment (25 mm Na2CO3, pH = 10.0). In (c), scale bar = 6 cm. Data in (d) represent mean ± SD of three biological replicates, each containing 10 plants. Black circles represent the values of individual plants. Significant differences between Kitaake and OX lines were evaluated by two‐way ANOVA, followed by Tukey’s multiple comparison test. (e) Accumulation of reactive oxygen species (ROS) in the leaves of rice seedlings treated with or without 25 mm Na2CO3 (pH = 10.0). After 24 h, seedlings were stained with nitrotetrazolium blue chloride (NBT) and diaminobenzidine (DAB) staining to assess the levels of O2 ‐ and H2O2, respectively. Three biological replicates were performed, each containing 20 seedlings.

Tissue‐specific expression patterns of OsSDG721, and subcellular localization analysis of OsSDG721

To examine the spatial and temporal expression patterns of OsSDG721, we generated transgenic plants expressing the β‐glucuronidase (GUS) gene under the control of the OsSDG721 promoter (OsSDG721) and examined the activity of the OsSDG721 promoter in different tissues at distinct developmental stages. GUS signals were detected in the leaf, leaf sheath, panicle, glume, root and stem tissues (Figure S3a). These results were also confirmed by RT‐qPCR (Figure S3b). Additionally, levels of OsSDG721 transcripts and the encoded protein rapidly increased upon exposure to saline–alkaline conditions (25 mm Na2CO3, pH = 10.0) for 2 and 4 h but decreased after recovery (Figure S3c, d). The stress‐responsive gene OsKitaakee01g425100 gene was used as a positive control (Figure S3c) in the RT‐qPCR analysis, and H3 was used as a loading control in the immunoblotting analyses (Figure S3d).

To examine the subcellular localization of OsSDG721, we cloned the GFP gene at the 3' end of the OsSDG721 coding sequence (CDS). The OsSDG721‐GFP construct or GFP alone (empty vector) was co‐transfected into rice protoplasts along with the NLS‐RFP construct, a nucleus localization maker. Green fluorescence signals co‐localized with red fluorescence in protoplasts co‐transformed with OsSDG721‐GFP and NLS‐RFP but were exclusively localized in the cytosol in protoplasts co‐transformed with GFP and NLS‐RFP (Figure S3e). These results indicate that OsSDG721 localizes to the nucleus.

Loss‐of‐function ossdg721 mutants exhibit saline–alkaline stress sensitive phenotypes

To confirm the role of OsSDG721 in saline–alkaline tolerance, we generated three independent ossdg721 loss‐of‐function mutants (ossdg721‐1, ossdg721‐2 and ossdg721‐3) using the CRISPR/Cas9 system. Single guide RNA (sgRNA) target sites for OsSDG721 were cloned into the CRISPR/Cas9 vector (Figure S4a), in which Cas9 was driven by the UBQ10 promoter (Ma et al., 2015). The vectors were then transformed into Kitaake plants, and homozygous ossdg721‐1, ossdg721‐2 and ossdg721‐3 mutant lines were identified via Sanger sequencing (Figure S4b). The ossdg721‐1 and ossdg721‐2 mutants carried a 1‐bp insertion of either A or T located 60 bp downstream of the transcription start site (ATG), which resulted in a frameshift mutation and consequently a premature stop codon before the core SET domain, and the ossdg721‐3 mutant contained a 2‐bp insertion at the same location, causing a frameshift (Figure S4b). To exclude the potential confounding effect of the Cas9 gene per se on the plant phenotype, we selected ossdg721‐1, ossdg721‐2 and ossdg721‐3 mutant plants for hygromycin sensitivity (Figure S4c).

To further confirm whether the loss‐of‐function mutation of OsSDG721 was responsible for saline–alkaline stress sensitivity, we generated complementation lines by expressing FLAG epitope‐tagged OsSDG721 under the control of the OsSDG721 promoter (OsSDG721) in the ossdg721‐1 mutant. Two independent complementation lines (Com#1 and Com#2) were chosen for further analysis, and the level of OsSDG721‐FLAG protein in these lines was detected using Western blot analysis (Figure S4d). Both Com#1 and Com#2 lines showed similar survival rates compared with wild‐type Kitaake plants under saline–alkaline stress (Figure 3a, b). Furthermore, staining leaves with DAB and NBT revealed increased ROS accumulation in ossdg721 mutants (Figure 3c, d), probably because high salt and pH reduce the photosynthetic rates, leading to increased production of ROS production including H2O2 and O2 ‐ (Sharma et al., 2012). Notably, ossdg721 mutants showed no significant changes under normal conditions. Next, we examined the contents of Na+ and K+ in shoots and roots of Kitaake, ossdg721 mutants and complementation lines under saline–alkaline stress. K+ and Na+ contents in shoots or roots did not differ significantly among the genotypes tested under normal conditions (Figure 3e, f, Figure S5). However, after treatment with saline–alkaline stress (25 mm Na2CO3, pH = 10.0) for 5 d, the shoot K+ content of ossdg721 mutants decreased dramatically compared with that of wild‐type Kitaake and complementation lines (Figure 3e), whereas the shoot Na+ content of ossdg721 mutants increased significantly compared with that of wild‐type Kitaake and complementation lines (Figure 3f). By contrast, in roots, the Na+ content of ossdg721 mutant lines was only slightly higher than that of wild‐type Kitaake and complementation lines, and the K+ content showed no significant differences among the ossdg721 mutants, wild‐type Kitaake and complementation lines (Figure S5). This result implies that the loss‐of‐function mutation of OsSDG721 perturbs the K+/Na+ ratio in shoots.

Figure 3

Characterization of loss‐of‐function ossdg721 mutants under saline–alkaline stress conditions. (a, b) Images (a) and survival rates (b) of Kitaake, loss‐of‐function mutants (ossdg721‐1, ossdg721‐2, ossdg721‐3) and complementation lines (Com#1 and Com#2) before and after recovery from the saline–alkaline treatment (25 mm Na2CO3, pH = 10.0). In (a), scale bar = 8 cm. In (b), data represent mean ± SD of three biological replicates, each containing 10 plants. Black circles represent the values of individual plants. Significant differences between Kitaake and other genotypes were evaluated by two‐way ANOVA, followed by Tukey’s multiple comparison test. (c, d) DAB (c) and NBT (d) staining of seedlings treated with or without 25 mm Na2CO3 (pH = 10.0). Three biological repeats were performed for each treatment, with 20 plants per treatment. (e, f) Plots showing K+ (e) and Na+ (f) contents of shoots of 4‐week‐old seedlings of the indicated genotypes treated with or without 25 mm Na2CO3 (pH = 10.0) for 5 d (n = 10 plants per genotype). Significant differences between Kitaake and other genotypes were evaluated by two‐way ANOVA, followed by Tukey’s multiple comparison test. DW, dry weight.

We further isolated nuclei from wild‐type Kitaake and ossdg721 mutants to determine the histone modification status in these genotypes. H3K4me3 levels were reduced in the three ossdg721 mutants compared with that in Kitaake, whereas H3K4me1 and H3K4me2 levels were not altered (Figure S6), which is consistent with previous reports (Jiang et al., 2018). In addition to the saline–alkaline stress tolerant phenotype, ossdg721 mutants also showed a slight reduction in plant height, grain size, grain weight and leaf angle (Figure 4a‐h). Together, these results indicate that OsSDG721 is not only involved in saline–alkaline stress tolerance but also controls grain yield and crop architecture.

Figure 4

Phenotypes of ossdg721 mutants and complementation lines. (a) Photographs of 90‐d‐old Kitaake, ossdg721‐1, ossdg721‐2, ossdg721‐3, Com#1 and Com#2 plants. Scale bar = 15 cm. (b) Plant height of the each indicated genotypes. Data represent mean ± SD (n = 25 plants per genotype). (c) Images of grains of the indicated genotypes. Scale bar = 7 mm. (d, e) Measurements of grain length (d) and grain width (e) of the indicated genotypes. Data represent mean ± SD of three biological experiments, each containing 25 grains. (f) Measurement of 500‐grain weight of each genotype. Data represent mean ± SD of 12 biological experiments, each containing 500 grains. (g, h) Images (g) and quantification (h) of leaf angles of 20‐d‐old plants. In (g), scale bar = 3 cm. In (h), data represent mean ± SD of three biological experiments, each containing 25 plants. Different lowercase letters represent statistically significant differences (P < 0.05; one‐way ANOVA, followed by Tukey’s multiple comparison test).

OsSDG721 enhances saline–alkaline stress response by upregulating OsHKT1;5 expression

To explore the effect of OsSDG721 on the genome‐wide transcriptional landscape, we conducted RNA sequencing (RNA‐seq) analysis of wild‐type Kitaake and ossdg721‐1 plants. Genes showing differential expression were identified in shoots and roots of rice seedlings between 0‐h vs. 6‐h saline–alkaline stress treatments via stringent statistical analysis of RNA‐seq data (see Experimental Procedures for details). Data shown in Figure S7a, b indicate that OsSDG721 dictates transcriptional reprograming under normal and saline–alkaline stress conditions (Tables S1–S4). Venn diagram analysis revealed that 10,112 and 10,150 genes were regulated under saline–alkaline conditions in Kitaake shoots and roots, respectively (Figure S7c, Table S5). Comparison of the differentially expressed genes (DEGs) between Kitaake and ossdg721‐1 showed that 1213 and 1705 overlapping genes responded to saline–alkaline stress in shoots and roots, respectively (Figure S7c, Table S6), which implies that the expression of these genes is affected by OsSDG721 under saline–alkaline stress. These saline–alkaline stress‐responsive genes included OsHKT1;5, OsHKT2;1 and OsHKT1;4. The HKT family proteins are involved in Na+ and K+ transport and homeostasis maintenance in many plant species (Hauser and Horie, 2010; Horie et al., 2009; Munns and Tester, 2008). Intriguingly, among the DEGs identified in this study, the expression of OsHKT1;5, located at the salt tolerance quantitative trait locus (QTL) SHOOT K (SKC1), was dramatically reduced in mutant roots, whereas that of OsHKT2;1, which encodes a plasma membrane‐localized subfamily II protein involved in Na+ uptake, was dramatically induced in mutant roots (Figure S7d). Given that OsSDG721 exhibits H3K4me3 methyltransferase activity and thus acts as a transcriptional activator, we deduced that the loss‐of‐function mutation of OsSDG721 reduces the expression of OsHKT1;5. To further examine the genetic interaction between OsSDG721 and OsHKT1;5, we generated ossdg721 oshkt1;5 double mutants using the CRISPR/Cas9 technology. Because OsSDG721 and OsHKT1;5 are genetically linked, we introduced CRISPR/Cas9 constructs harbouring two sgRNA target sites for OsHKT1;5 in the ossdg721‐1 mutant background (Figure 5a). Two double mutant lines were generated: ossdg721‐1 oshkt1;5‐c1 and ossdg721‐1 oshkt1;5‐c2; the former carried a 2‐bp deletion located 83 bp downstream of ATG, and the latter contained a 7‐bp deletion located 194 bp downstream of the ATG (Figure 5b). Sequence deletion in both double mutants resulted in a frameshift and, consequently, a premature stop codon (Figure 5b). We also generated oshkt1;5 single mutants in wild‐type Kitaake background using a CRISPR/Cas9 construct harbouring one of the sgRNA target sites. The oshkt1;5‐c3 single mutant harboured a 2‐bp deletion located 194 bp downstream of ATG, causing a frameshift and, consequently, a premature stop codon (Figure 5b). The ossdg721‐1 oshkt1;5‐c1, ossdg721‐1 oshkt1;5‐c2 and oshkt1;5‐c3 mutants were screened for non‐hygromycin resistance (Figure 5c). The saline–alkaline stress sensitivity of both double mutants was comparable with that of the single mutant, based on survival rates (Figure 5d, e). Moreover, no significant differences in shoot K+ and Na+ contents were detected among the three mutants (Figure 5f, g). These results imply that OsHKT1;5 acts downstream of OsSDG721, and reduced OsHKT1;5 expression in the ossdg721 mutant background contributes to saline–alkaline stress sensitivity.

Figure 5

OsHKT1;5 acts downstream of OsSDG721 under saline–alkaline stress conditions. (a, b) Sequence chromatograms showing mutations induced in the OsHKT1;5 gene in the ossdg721‐1 oshkt1;5‐c1, ossdg721‐1 oshkt1;5‐c2 and oshkt1;5‐c3 mutants using the CRISPR/Cas9 technology (a), as revealed by Sanger sequencing (b). (c) Isolation of Cas9‐free mutants. Plants of indicated genotypes were selected via hygromycin resistance. (d, e) Images (d) and survival rates (e) of Kitaake, ossdg721‐1, oshkt1;5‐c3, ossdg721‐1 oshkt1;5‐c1 and ossdg721‐1 oshkt1;5‐c2 plants before and after recovery from the saline–alkaline stress treatment (25 mm Na2CO3, pH = 10.0). In (d), scale bar = 12 cm. In (e), data represent mean ± SD of three biological experiments, each containing 10 plants. Black circles represent the values of individual plants. (f, g) Plots showing K+ (f) and Na+ (g) contents of the shoots of 4‐week‐old seedlings of the indicate genotypes treated with or without 25 mm Na2CO3 (pH = 10.0) for 5 d (n = 10 plants per genotype). DW, dry weight. Different lowercase letters represent statistically significant differences (P < 0.05; two‐way ANOVA, followed by Tukey’s multiple comparison test).

OsSDG721 regulates the saline–alkaline stress response via the regulation of H3K4me3 levels at the OsHKT1;5 chromatin

To test whether OsSDG721 impacts the H3K4me3 level under the saline–alkaline stress condition, we treated 2‐week‐old wild‐type Kitaake and ossdg721 seedlings with 25 mm Na2CO3 (pH = 10.0) for 0 or 6 h and examined the H3K4me3 levels by quantitative PCR (ChIP‐qPCR). The result showed no significant difference in H3K4me3 levels between Kitaake and ossdg721 seedlings under normal conditions; however, under saline–alkaline stress conditions, the increase in H3K4me3 levels was dramatically impaired in ossdg721 seedlings compared with wild‐type Kitaake seedlings (Figure 6a). Comparable changes were not detected in chromatin at the OsUBQ10 locus, a negative control (Figure 6b). To further test whether OsSDG721 associates with the OsHKT1;5 chromatin, we took advantage of the complementation lines (OsSDG721) to perform ChIP‐qPCR analysis using anti‐FLAG antibody. The results showed that OsSDG721‐FLAG specifically associated with the promoter and gene body of OsHKT1;5 (Figure 6c, d). Furthermore, after 25 mm Na2CO3 (pH = 10.0) treatment, the association of OsSDG721‐FLAG with OsHKT1;5 chromatin increased in the complementation lines compared with Kitaake seedlings. Taken together, these results indicate that OsSDG721 binds to the OsHKT1;5 chromatin and modulates its H3K4me3 levels. OsSDG721 established a peak level of H3K4me3‐modified nucleosomes nearly 300 bp downstream of the transcriptional start site (TSS) of OsHKT1;5. To further evaluate whether OsSDG721 directly binds to DNA, we performed EMSA using synthetic probes representing six different loci located in close proximity to this region. However, OsSDG721 did not bind to any of these probes (Figure S8), indicating that OsSDG721 does not exhibit DNA‐binding activity. GST and OsMYB106 proteins were used as negative and positive controls for the OsHKT1;5 promoter, based on the previous report (Wang et al., 2020).

Figure 6

OsSDG721 regulates saline–alkaline stress tolerance via the regulation of histone H3 lysine 4 trimethylation (H3K4me3) status of OsHKT1;5. (a, b) OsSDG721 regulates H3K4me3 levels at OsHKT1;5 (a) but not at OsUBQ10 (b) under saline–alkaline stress conditions, as shown by chromatin immunoprecipitation, followed by quantitative PCR (ChIP‐qPCR) using anti‐H3K4me3 antibody. The X‐axis denotes the genomic regions of OsHKT1;5 (a) and OsUBQ10 (b). Data represent mean ± SD (n = 3). (c, d) OsSDG721 associates with OsHKT1;5 (c) but not with OsUBQ10 (d), as shown by ChIP‐qPCR using anti‐FLAG antibody. The X‐axis denotes the genomic regions of OsHKT1;5 (c) and OsUBQ10 (d). C, control conditions; N, Na2CO3 (pH = 10.0) treatment conditions. Black boxes and grey boxes represent exon and untranslated regions (UTR), respectively. Data represent mean ± SD (n = 3). Different lowercase letters represent statistically significant differences (P < 0.05; two‐way ANOVA, followed by Tukey’s multiple comparison test).

Discussion

In this study, we generated an activation‐tagged mutant pool in Kitaake, a model rice variety. This experimental design greatly facilitates the positive selection of transformed calli and is also very useful for determining whether the activation‐tagged mutants are gain‐of‐function or loss‐of‐function mutants. Approximately 500 activation‐tagged insertion lines were screened in this study, of which approximately 8% showed saline–alkaline stress‐sensitive or ‐resistant phenotypes. Using this system, we identified a gain‐of‐function mutant, AC7, which showed enhanced saline–alkaline stress tolerance compared with the wild type. Genetic and physiological analyses of line AC7 confirmed that OsSDG721 positively impacts saline–alkaline stress tolerance in rice. Moreover, agronomic traits were negatively impacted in loss‐of‐function ossdg721 mutants, as evident from the reduction in plant height, grain size, grain weight and leaf angle but enhanced in OsSDG721OX lines, which showed an increased leaf angle and advanced heading and ripening dates compared with wild‐type Kitaake. Previously, in the japonica rice cultivar Nipponbare, loss‐of‐function mutations in SDG721 and SDG705 genes resulted in gibberellic acid (GA)‐deficient phenotypes, including reduced cell length, semi‐dwarfism and reduced panicle branching (Jiang et al., 2018). Thus, although we used a different background (Kitaake) in this study, our results were similar to those obtained by Jiang et al. (2018) in Nipponbare. RNA‐seq data analysis showed that OsKitaake07g154900 (OsSPL13) and OsKitaake09g157800 (OsSPL18) were downregulated in ossdg721‐1 shoots. In previous studies, osspl18 and osspl13 knockout mutants exhibited reduced grain width and thickness (Si et al., 2016; Yuan et al., 2019). Furthermore, OsKitaake03g264900 (OsIAA12) was downregulated in the ossdg721‐1 mutant, in which the overexpression of OsIAA12 increased the leaf inclination angle (Chen et al., 2018). It is possible that altering the impact of OsSDG721 on these genes explains the changes in agronomic traits to some extent. Previous studies indicate that the SET DOMAIN GROUP proteins affect different agronomic traits in rice. For instance, OsSDG708 impacts flowering time by controlling the H3K36 methylation status of key flowering regulatory genes including Heading date 3a (Hd3a), RICE FLOWERING LOCUS T1 (RFT1) and Early heading date 1 (Ehd1) (Liu et al., 2016); OsSDG701, which catalyses H3K4 methylation and plays an important role in sporophyte development and sporophyte‐to‐gametophyte transition (Liu et al., 2017). OsSDG724 modulates H3K36 methylation levels and regulates Ehd1, OsMADS50 and RTF1 expression (Sun et al., 2012); OsSDG714, which governs the H3K9 methylation status of chromatin, plays important roles in regulating trichome formation and CpG and CNG cytosine methylation at the Tos17 locus (Ding et al., 2007). Although SDG proteins exhibit different histone modification specificities, some of them play overlapping functions. Thus, further studies are needed to understand how different SDG proteins coordinate to regulate plant growth and development in rice.

Intriguingly, we found that the shoot K+ content of ossdg721 mutant lines was reduced, whereas the shoot Na+ content was elevated, indicating that OsSDG721 is involved in the maintenance of K+ and Na+ homeostasis. Furthermore, genome‐wide transcriptomic analysis revealed that OsSDG721 impacts the expression of OsHKT1;5, which was identified as the SKC1 locus. ChIP‐qPCR analysis revealed that OsSDG721 binds to the OsHKT1;5 chromatin and controls the H3K4me3 level at this locus. According to previous reports (Lu et al., 2013), OsSDG721 contains highly conserved domains, including PWWP, FYRN/C, plant homeodomain (PHD) and SET. Among these domains, only PWWP exhibits nucleosome‐binding capability (Lu et al., 2013). Moreover, a recent study demonstrated that the PWWP domain of mouse Dnmt3a exhibits little DNA‐binding activity, in contrast to Dnmt3b, which binds to DNA in a non‐sequence‐specific manner (Chen et al., 2004). As OsSDG721 also harbours the PWWP domain, we tested whether OsSDG721 also binds to DNA in a sequence‐specific or non‐sequence‐specific manner. However, our results showed that OsSDG721 does not bind to the OsHKT1;5 promoter or region downstream of the TSS. Recent studies revealed that a transcriptional regulatory complex composed of BCL2‐ASSOCIATED ATHANOGENE4 (OsBAG4), OsMYB106 and OsSUVH7 regulates OsHKT1;5 expression in response to salt stress. OsMYB106 and OsSUVH7 bind to the MYB binding cis‐element (MYBE) in the OsHKT1;5 promoter and to a miniature inverted repeat transposable element (MITE) located upstream of the MYBE (Wang et al., 2020). OsBAG4 functions as a bridge between OsSUVH7 and OsMYB106 to facilitate the binding of OsMYB106 to the consensus MYBE sequence in the OsHKT1;5 promoter, thereby activating OsHKT1;5 expression. Thus, a novel transcriptional complex, consisting of a DNA methylation reader, a chaperone regulator and a transcription factor, regulates OsHKT1;5 expression under salt stress. In the current study, we observed that the OsHKT1;5 gene was downregulated, and H3K4me3 levels in the OsHKT1;5 chromatin were decreased in ossdg721 mutants. Further ChIP‐qPCR analysis revealed that OsSDG721 binds to the OsHKT1;5 chromatin. It is possible that OsSDG721 combines with the OsSUVH7–OsBAG4–OsMYB106 transcriptional module to participate in the transcriptional regulation of the OsHKT1;5 gene. Further studies are needed to elucidate how OsSDG721 is recruited to the OsHKT1;5 chromatin for the deposition of H3K4me3 mark.

Experimental procedures

Plant material, growth conditions and saline–alkaline treatment

Whole genome‐scale activation tagging pools were established in Oryza sativa L. ssp. japonica cv. Kitaake through co‐cultivation of calli with Agrobacterium, as described previously (Ozawa, 2009). Seeds of wild‐type Kitaake, AC lines, ossdg721 loss‐of‐function mutant lines, ossdg721 osshkt1;5 double mutant lines and transgenic OsSDG721OX lines were sterilized, washed and cultured, as previously described (Yoshida, 1976). The hydroponic experiment was carried out in a growth chamber maintained at 28 °C day/25 °C night temperature, 14‐h light/10‐h dark photoperiod, approximately 70% relative humidity and 200 μmol photons m−2 s−1 light intensity. To perform breeding experiments, seeds were sown in the field in Changchung (125°41'N, 43°82'E) under normal conditions. To perform the saline–alkaline stress treatment, seeds were grown hydroponically in Yoshida’s culture solution for 4 weeks and then transferred to culture solution supplemented with 25 mM Na2CO3 (pH = 10.0) for 5 days (Oosthuizen and Greyling, 2001). After the saline–alkaline treatment, seedlings were transferred to Na2CO3‐free Yoshida’s culture solution for recovery.

Plasmid construction

To generate ossdg721 loss‐of‐function mutant lines and ossdg721 oshkt1;5 double mutant lines, single guide RNAs (sgRNAs) were cloned into pYLsgRNA‐OsU6a and pYLsgRNA‐OsU6b plasmids using OsSDG721‐CRISPR‐F1/‐R1, OsSDG721‐CRISPR‐F2/‐R2, OsHKT1;5‐CRISPR‐F1/‐R1 and OsHKT1;5‐CRISPR‐F2/‐R2 primer pairs, as described previously (Ma and Liu, 2016). To generate the OsSDG721 construct, a 2000‐bp fragment upstream of the OsSDG721 start codon was amplified by PCR using the primer pair OsSDG721pro‐F/‐R and then cloned into the pCAMBIA3301 binary vector. To determine the subcellular localization of OsSDG721, full‐length OsSDG721 CDS (minus the stop codon) was amplified and cloned into plasmid 326‐GFP under the control of the CaMV 35S promoter using recombination via XbaI and BamHI restriction endonucleases sites (Liu et al., 2018). To construct the CsV plasmid for the generation of OsSDG721OX lines, full‐length OsSDG721 CDS was amplified and cloned into pCsV1300, containing a 3×FLAG tag, using recombination (Xu et al., 2012). To construct the OsSDG721 plasmid for the generation of complementation lines (Com#1 and Com#2), the OsSDG721 gene (including the promoter region) was cloned into the pCAMBIA1302 binary vector using the G‐OsSDG721‐F/‐R primer pair. To generate GST‐OsSDG217 constructs, full length OsSDG217 was amplified and recombined into the pGEX‐4T1 vector via BamHI and EcoRI sites. All primers used for plasmid construction are listed in Table S7.

Mutant isolation and generation of transgenic plants

The ossdg721 and oshkt1;5 single mutants and ossdg721 oshkt1;5 double mutant were generated using the CRISPR/Cas9 technology (Ma and Liu, 2016). Briefly, OsSDG721‐ and OsHKT1;5‐specific sgRNAs were designed online at http://skl.scau.edu.cn/targetdesign/, and each sgRNA cassette was separately cloned into pYLCRISPR/Cas9P. The resulting constructs were separately introduced into wild‐type Kitaake or ossdg721‐1 plants via Agrobacterium‐mediated transformation (Lu et al., 2017). The resulting plants were selected in Yoshida’s culture solution supplemented with 50 mg L‐1 hygromycin, and all mutations were confirmed by Sanger sequencing. T1 seeds lacking hygromycin resistance were used for subsequent experiments. To generate OsSDG721OX or complementation lines (Com#1 and Com#2), the CsV or OsSDG721 construct was introduced into Kitaake or ossdg721‐1 background, respectively. Homozygous transgenic lines were identified by selection on media containing 50 mg L‐1 hygromycin (Chen et al., 1998).

GUS staining and subcellular localization analysis

Histochemical GUS staining was performed with OsSDG721 transgenic plants, as previously described (Jefferson et al., 1987; Nan et al., 2020). To determine the subcellular localization of OsSDG721, protoplasts were isolated from the leaf sheath of 3‐week‐old seedlings grown under 12‐h light/12‐h dark photoperiod and co‐transfected with OsSDG721‐GFP and NLS‐RFP constructs by polyethylene glycol (PEG)‐mediated transfection (Zhang et al., 2011). The protoplasts were cultured at room temperature for 12–16 h and then observed under a fluorescence microscope (Olympus).

Measurement of Na+ and K+ contents of shoots

Seedlings were grown in hydroponic culture solution for 4 weeks and then transferred to the same culture solution supplemented with or without 25 mm Na2CO3 (pH = 10.0) for 5 days. After the saline–alkaline stress treatment, rice seedlings were washed twice with deionized water and then dried at 55 °C for 3 days. The dried shoot samples were weighed, ground, resuspended in 10 mL of ultrapure water and boiled at 100 °C for 8 h. The Na+ and K+ concentrations in the solution were determined by atomic absorption spectrophotometry (Rus et al., 2001; Zhang et al., 2017).

NBT and DAB staining

Four‐week‐old seedlings were transferred to hydroponic culture solution supplemented without or with 25 mM Na2CO3 (pH = 10.0) for 24 h. To determine ROS levels, leaves were vacuum‐infiltrated for 30 min and then submerged for 12 h in 10 mM potassium phosphate buffer (pH = 7.8) containing 0.05% NBT (w/v) and 10 mM NaN3 or for 24 h in 0.1% DAB (pH = 5.8). The stained leaves were cleared in the destaining buffer (ethanol:lactic acid:glycerol = 3:1:1) to eliminate the background green color (Nguyen et al., 2017).

RNA‐seq

Total RNA was isolated from the roots and shoots of wild‐type Kitaake and ossdg721 mutant seedlings grown in Yoshida’s culture solution supplemented with 25 mM Na2CO3 (pH = 10.0) for 0 or 6 h using the TRIzol reagent (Invitrogen, CA, USA). RNA‐seq was performed, as described previously (Wang et al., 2020). Briefly, approximately 3 mg of RNA isolated from each sample was used for library construction, and RNA‐seq was performed using the Illumina HiSeq 2500 platform (Novogene, Beijing, China) in three biological replicates. Each sample generated approximately 4.0 Gb clean reads. Reads from each sample were trimmed, mapped and analysed using FASTX‐Toolkit (version 0.0.13), TOPHAT v.2.1.0 (Trapnell et al., 2009) and CUFFLINKS (http://cole‐trapnell‐lab.github.io/cufflinks/cuffdiff/index.html), respectively. The selected RNA‐seq data were confirmed by RT‐qPCR.

ChIP‐qPCR assay

The ChIP assay was performed, as described previously (Li et al., 2018; Liu et al., 2019), with slight modifications. Briefly, 2‐week‐old ossdg721 and OsSDG721 seedlings grown in Yoshida’s culture solution treated with 25 mm Na2CO3 (pH = 10.0) for 0 h (mock treatment) or 6 h. After the saline–alkaline stress treatment, approximately 20 g of seedlings were harvested and ground to a fine powder. Subsequently, chromatin complexes were extracted from the samples and sheared to 500‐bp fragments using FB120 Sonic Dismembrator (Fisher Scientific). Then, anti‐H3K4me3 (ab8580; Abcam) or anti‐FLAG (F1804; Sigma‐Aldrich) antibody was added to the chromatin complexes, and the samples were incubated overnight at 4 °C on a rotatory shaker. Protein A–agarose beads (Merck Millipore) were added to the mixture, and the samples were incubated for another 60–90 min. The protein–DNA cross links were reversed by incubating the samples at 65 °C for 8 h, and the released DNA fragments were purified. The enrichment of immunoprecipitated DNA was evaluated by RT‐qPCR, with OsUBQ10 serving as a negative control. Primers used for RT‐qPCR are listed in Table S7.

EMSA

The EMSA experiment were performed, as described previously (Gao et al., 2020). GST‐tag proteins were induced via the Escherichia coli BL21 (DE3) cell line and immobilized onto glutathione–sepharose beads (GE Healthcare). The purified protein was confirmed by SDS–PAGE and prepared for EMSA. DNA probes were synthesized and biotin‐labelled at the 5’‐end by Sangon Biotechnology. Recombinant proteins were incubated with double‐stranded probes at 4 °C in binding buffer for 30 min and then transferred to a nylon membrane via wet transfer and detected according to the instructions provided with the Chemiluminescent EMSA Kit (GS009; Beyotime). Primers used for EMSA are listed in Table S7.

Conflict of Interest

The authors have no conflict of interest to declare.

Author contributions

Z.‐Y.X. conceptualized and supervised the research. Y.L. performed most of the biochemical and cell‐based experiments. X.C. performed the physiological analyses of plants treated with saline–alkaline stresses. S.X. performed the physiological analyses of plants. T.Q. performed the Agrobacterium‐mediated transformation experiment. D.C. generated mutant lines. W.C. performed RNA‐seq and data analysis. M.L. performed molecular cloning. L.H., D.‐J.Y. and B.L. provided helpful suggestions for improving the manuscript. Z.‐Y.X. wrote the manuscript. All authors reviewed, revised and approved the final manuscript.

Accession numbers

Data generated in this study have been deposited in the National Center for Biotechnology Information Sequence Read Archive (accession number: PRJNA698439).

Figure S1. Identification of activation‐tagged lines based on the detection of green fluorescent protein (GFP) signal.

Figure S2. Agronomic traits of OsSDG721 overexpression (OsSDG721OX) lines.

Figure S3. Tissue‐specific expression patterns of OsSDG721, and subcellular localization of OsSDG721.

Figure S4. Generation of loss‐of‐function ossdg721 mutants.

Figure S5. Ion contents in the roots of different genotypes.

Figure S6. Examination of H3K4 methyltransferase activity.

Figure S7. OsSDG721 affects the transcriptional landscape in rice under saline–alkaline stress conditions.

Figure S8. Image of EMSA.

Click here for additional data file.

Table S1. List of differentially expressed genes (DEGs) in shoot in ossdg721‐1 vs. Kitaake under normal condition.

Table S2. List of differentially expressed genes (DEGs) in root in ossdg721‐1 vs. Kitaake under normal condition.

Table S3. List of differentially expressed genes (DEGs) in shoot in ossdg721‐1 vs. Kitaake under saline–alkaline condition.

Table S4. List of differentially expressed genes (DEGs) in root in ossdg721‐1 vs. Kitaake under saline–alkaline condition.

Table S5. List of saline–alkaline stress responsive genes in Kitaake.

Table S6. List of saline–alkaline responsive genes affected by OsSDG721.

Table S7. List of primers used in this study.

Click here for additional data file.