Rice cultivar Dongjin (Oryza sativa ssp. Japonica cv. Dongjin) was used in this study. Transgenic rice plants were grown in soil in the greenhouse or on MS [Murashige and Skoog (Duchefa, Netherland)] agar medium [per liter: 4.4 g MS salt, 30 g sucrose, 0.5 g 2-(N-Morpholino) ethanesulfonic acid (MES), 8 g plant agar, pH 5.8] in a growth chamber (EYELA, Japan) maintained at 28°C and 60% relative humidity under long-day conditions (16 h light/8 h dark cycle).

To investigate the expression levels of OsDREB1G under various abiotic stress conditions, rice seeds were germinated and grown in soil in a greenhouse. For abiotic stress phenotype analysis, 3-week-old seedlings were treated with osmotic [150 mM mannitol (Sigma-Aldrich, United States)], salt [150 mM NaCl (Sigma-Aldrich, United States)], 5 μM ABA (Sigma-Aldrich, United States) and cold stress (4 C) were treated in MS media. Two-week-old T2 transgenic rice plants were subjected to cold stress (4°C) for 3 days and then recovered for 7 days in a growth chamber maintained at 28°C under long-day conditions (16 h light/8 h dark cycle). The experiments were performed with three biological repetitions. For RNA preparation, 2 week-old seedlings were transferred to a growth chamber at 4°C for cold stress treatment or sprayed with 0.1 mM ABA solution followed by sampling at the designated time points with three biological repetitions.

Total RNA was extracted from seedlings and T2 transgenic rice plants with TRIzol Reagent (MRC, United States). Fifteen micrograms of total RNA per lane was electrophoresed on a 1.2% formaldehyde agarose gel and transferred to nylon membranes (Amersham, United Kingdom) by capillary blotting, followed by UV-cross-linking. Pre-hybridization was performed for 30 min at 65°C in Church buffer (1% BAS, 1 mM EDTA, 0.5 M NaHPO4, 7% SDS, pH 7.2), followed by hybridization overnight at 65°C using a OsDREB1G probe. The probe was labeled using the random oligonucleotide priming method (Amersham, United Kingdom). The membranes were washed with 2 × SSC with 0.1% SDS for 10 min, followed by 1 × SSC and 0.5 × SSC (with 0.1% SDS) for 10 min each time at 65°C. The washed membrane was exposed to a BAS cassette (Fuji Film, Japan) for 1 day. For quantitative RT-qPCR analysis, first-strand cDNA was synthesized from 4 μg of total RNA using SuperScript III reverse transcriptase (Invitrogen, United States) and then we used the 1/20 volume of total cDNA per RT-qPCR reaction. Amplified signals were detected with a MyiQ real-time PCR system (Bio-Rad, United States) using SYBR Premix Ex Taq^TM (Takara, Japan). The amplification parameters were as follows: 5 min of denaturation and enzyme activation at 95°C followed by 40 cycles at 95°C for 5 s, 58°C for 15 s, and 72°C for 20 s. A final step was performed at 65–95°C (1°C/sec) for melting curve analysis. Data were normalized based on the expression of rice Ubi5, and relative gene expression was analyzed via the ΔΔCT method. RT-qPCR analysis was performed with at least three biological repetitions. The primer sequences used for RT-qPCR analysis are listed in Supplementary Table S2.

For subcellular localization, the full-length ORF of OsDREB1G without the termination codon was amplified by PCR with specific primers (forward, 5′-TCTAGAATGGACGTTTCTGCTGCGCTC-3′; reverse, 5′-TTGGATCCGTAGCTCCATAGCTGGACCTC-3′). After it was confirmed by sequencing, the amplified fragment was fused to the N-terminus of vector p326GFP under the control of the CaMV35S promoter to create 35S-OsDREB1G:GFP. The fusion construct (35S-DREB1G:GFP) was transformed into rice protoplasts, singularly or co-transformed with NLS-RFP, the protoplasts were prepared from young etiolated seedlings by PEG-mediated transformation (Kim et al., 2015). To stain the protoplast with DAPI (4′,6′-diamidino-2-phenylindol), after incubating the transformed protoplasts for 24 h at 28°C, protoplasts were fixed by exchanging buffer with W6 solution (154 mM NaCl, 125 mM CaCl₂, 5 mM KCl, and 10 mM PIPES adjusted to pH 6.8) and 4% paraformaldehyde treatment. After 2 h incubation, the buffer was exchanged by DAPI staining buffer (W6 and 5 ug/ml DAPI). After 10 min incubation and several wash out with W6 solution, DAPI and GFP signals were captured using a Leica TCS SP8 laser scanning confocal microscope. The combinations of excitation wavelength/detection range of emission on the confocal microscopy were 488 nm (solid state laser)/ 493 to 530 bandpass for GFP and 405 nm (solid state laser)/ 410 to 493 bandpass for DAPI. At the co-transformed protoplasts with OsDREB1G-GFP and NLS-RFP, GFP and RFP signals were captured lively by Leica TCS SP8 laser scanning confocal microscope. The excitation wavelength/detection range of emission for RFP was 530 nm (solid state laser)/ 557 to 640 bandpass. The detected images are presented in pseudocolor.

To construct the reporter plasmid vector including the promoter region of Remorin, Os03g63870 was fused with the firefly luciferase gene (fLUC), and the 1.5 Kb promoter regions of these genes were amplified from Dongjin genomic DNA with specific primers (Supplementary Table S2). The reporter plasmid was constructed using a 135 bp DNA fragment of the Rab16A promoter containing the DRE (GCCGAC) sequence. To construct the effector plasmids, the coding sequences of OsDREB1A–OsDREB1G and OsDREB1G were cloned into transient expression vector pGEM-3HA containing the maize ubiquitin promoter and a sequence encoding a 3 × HA tag. Freshly isolated rice sheath and leaf protoplasts were co-transfected with the designated reporter (4 μg per transfection) and effector (5 μg per transfection) plasmids. The UBQ10 promoter::Renilla luciferase plasmid (0.5 μg per transfection) was added to each sample as an internal control (Kim et al., 2015). Luciferase assays were performed using a dual luciferase assay kit according to the manufacturer’s instructions (Promega, United States). Luciferase assays were performed with at least three biological repetitions.

To generate OsDREB1G overexpression rice plants, the coding region of OsDREB1G was cloned into the plant expression vector pGA2897 under the control of the maize UBIQUITIN promoter (Kim et al., 2012). Recombinant DNA containing OsDREB1G was introduced into Agrobacterium tumefaciens LBA4404 by electroporation using a MicroPulser Electroporator (BioRad, United States). Transgenic rice plants were obtained by Agrobacterium-mediated transformation as previously reported (Han et al., 2012). Transgenic T0 plants were selected on medium containing hygromycin, transferred to soil in pots, and grown in a greenhouse. Transgenic plants overexpressing OsDREB1G were chosen for further study via RNA gel blot and RT-qPCR analysis. To investigate the effect of OsDREB1G on cold stress tolerance, 7 OsDREB1G-overexpressing plants (including independent lines #7, #8, #12, #42) and wild-type plants were grown together in the same pots filled with soil for 2 weeks in the greenhouse. The plants were transferred to a cold chamber (EYELA, Japan) at 4°C for 3 days, after which they were grown for 7 days under normal conditions (28°C, 16 h light/8 h dark cycle). Cold stress tolerance was assessed by determining the survival rate during the recovery period. And then four independent lines (including independent lines #7, #8, #12, #42) were selected to cold stress analysis in T2 generation using same method.

For the post-germination assay, sterilized non-transgenic and transgenic rice seeds were planted on 1/2MS medium with or without 40 mg/L hygromycin. Three days after planting, the plants were transferred to 1/2MS medium with or without 5 μM ABA (2-cis-4-trans-abscisic acid, 98% synthetic). Seedling growth was measured at 7 days after transfer.

Among OsDREB1s, only six members have been characterized functionally in terms of their effects on phenotype and genetics. Thus, to functionally characterize the other members of the OsDREB1 gene family, we searched the microarray database¹ to identify OsDREBs that are responsive to cold stress. We found that Os02g45450 was specifically expressed in response to cold stress, with an expression profile similar to those of OsDREB1A and OsDREB1B (Figure 1B). Os02g45450 was closest to OsDREB1E and OsDREB1F in a phylogenic tree constructed based on amino acid sequences (Figure 1A). This gene was identified as OsDREB1G and studied only about gene expression profiles under abiotic stress conditions. We summarized OsDREB1 subgroup genes published (Supplementary Table S1) We confirmed its expression pattern by RNA gel blot analysis in plants under several stress conditions (Figure 1C). Cold treatment increased OsDREB1G expression at 3 h, with expression levels reaching a maximum at 24 h and remaining quite high until 48 h after treatment. However, the expression of OsDREB1G did not change in response to other stress treatments including NaCl, mannitol, and abscisic acid (ABA) treatment (Figure 1C). We also compared the OsDREB1G expression pattern with those of other OsDREB1 genes using reverse transcription quantitative polymerase chain reaction (RT-qPCR) (Bustin et al., 2009). Cold stress induction of OsDREB1G was weaker than that of other OsDREB1 genes, although OsDREB1G had a longer induction period. OsDREB1G maintained high expression until 24 h after cold treatment (Figure 1D).

We compared the expression of this gene in several tissues to that of other OsDREB1 genes using RT-qPCR. OsDREB1G was strongly expressed in root, sheath, node, and leaf blade tissue and expressed in tiller and flower at low level. However, OsDREB1G was almost never expressed in seeds. This expression pattern was quite similar to those of OsDREB1A and OsDREB1B (Figure 1E).

The coding region of OsDREB1G is 675 bp and encodes a putative 224 amino-acid protein with a predicted molecular mass of 23.9 kDa (GenBank Accession No. XP_015624757, TIGR ID LOC_Os02g45450, RAP ID Os02g0677300). OsDREB1G contains a putative nuclear localization signal (NLS), a conserved AP2/ERF DNA binding domain, and an LWSY motif (Figure 2A). To investigate whether OsDREB1G is a functional transcription factor, we confirmed the subcellular localization of OsDREB1G protein. A fusion construct containing OsDREB1-GFP driven by the CaMV35S promoter was introduced into rice protoplast cells. The OsDREB1-GFP fusion protein was co-localized to the nucleus with NLS-RFP and DAPI stain, indicating that OsDREB1G is a nuclear protein (Figure 2B). DREB1s are well-known transcription factors that regulate gene expression by binding to the CRT/DRE sequences of stress-responsive downstream genes (Sakuma et al., 2002). We performed a transactivation assay of OsDREB1G compared to other OsDREB1s in rice protoplasts. The luciferase gene (LUC) fused to 2 × DRE (GCCGAC) and the 35S minimal promoter was used as a reporter plasmid. Relative LUC activity increased in the presence of all OsDREB1s in rice protoplasts. OsDREB1G, OsDREB1A, and OsDREB1B showed similar transactivation activities, while OsDREB1C had the highest transactivation activity (Figure 2C). These results indicate that OsDREB1G is a typical member of the OsDREB1 subgroup.

To determine whether the overexpression of OsDREB1G, like other OsDREB1s, would improve abiotic stress tolerance, we constructed and transformed the vector for overexpression into rice (Figure 3A) and we examined cold, drought, salt, and ABA tolerance in OsDREB1G overexpression (OsDREB1G-Ox) plants. We generated 42 transgenic rice plants, investigated the expression of OsDREB1G in these plants, and selected four OsDREB1G-Ox lines (#8, #12, #42) for further studies based on the gene expression level and confirmed the over-expression of selected lines using RT-qPCR (Figure 3C). Among the various stress treatments, OsDREB1G-OX plants showed clearly enhanced cold-stress tolerance compared to the control in the T2 generation (Figure 3B). Under cold-stress treatment, the control plants (Dongjin) showed a survival rate of 20%, whereas all four OsDREB1G-OX lines showed survival rates >60% (Figure 3D). However, OsDREB1G-OX plants were not sensitive to ABA in terms of growth (Supplementary Figure 1), and they did not exhibit clear and reproducible changes in drought or salt tolerance (Supplementary Figure 2). Finally, under normal growth conditions, the OsDREB1G-OX plants had shorter stems and lower seed productivity compared to the control (data not shown).

We performed expression analysis of cold stress-responsive genes in OsDREB1G-OX plants. Firstly we selected 16 cold responsive genes published by two research groups and unpublished 3 genes we have information in my laboratory (Rabbani et al., 2003; Byun et al., 2015). And then we examined the expression levels of them by RT-qPCR in OsDREB1G-Ox plants and we found out that 10 genes of them were upregulated in three different OsDREB1G overexpression lines and different time points of cold stress treatment compared to control. These genes were expressed at levels up to ∼3-times higher in OsDREB1G-Ox plants compared to the control under cold treatment, and Os03g60580 was expressed at higher levels in OsDREB1G-Ox plants, even at normal temperatures (Figure 4A–J). Thus, cold stress-tolerance genes were more highly expressed in OsDREB1G-OX plants than the control, suggesting that these plants have enhanced cold tolerance.

To investigate whether OsDREB1G directly induces the expression of these marker genes, we produced reporter constructs in which the promoters of the cold-stress marker gene Remorin and of Os03g60580 were individually fused with the luciferase reporter gene. We co-transformed rice protoplasts with an OsDREB1G effector construct and the reporter construct, finding that the Remorin promoter was induced at a level approximately 5-fold higher than the control, whereas the Os03g60580 promoter was induced at a level >50-fold higher than the control. Indeed, gene expression analysis showed that Os03g60580 was more highly induced in OsDREB1G-OX plants in response to cold treatment than Remorin (Figure 5). These results indicate that OsDREB1G directly induces cold-responsive gene expression in rice.