Jatropha curcas seeds were collected from Guizhou province, China and planted on farm land in Guangzhou, Guangdong province, China. For analysis of JcDREB2 gene expression in physic nut, roots, stem cortex, leaves, flowers, and seeds were sampled at 35 days after pollination and stored at -80°C until required for RNA isolation. For salinity treatment, plants at the six-leaf stage were irrigated with Hoagland solution plus 150 mM NaCl. For GA₃ and ABA treatments, plants at the six-leaf stage were sprayed with 100 μM GA₃, 100 μM ABA or distilled water (control), and the fourth leaves were collected after 1, 3, 6, and 12 h of GA₃ and ABA stresses.

The japonica rice (Oryza sativa L.) cv. Zhonghua 11 (ZH11) was used as the wild type in this study. Seeds were germinated and cultured in soil in basins in a greenhouse under natural sunlight.

All protein sequences of JcDREB2 orthologs were downloaded from GenBank¹. ClustalX was used to analyze multiple sequence alignment (Thompson et al., 1997). The phylogenetic relationship of JcDREB2 orthologs was constructed, using MEGA 5 software, by the Neighbor–Joining method with 1000 bootstrap replicates (Tamura et al., 2011).

The full length coding domain sequence of JcDREB2 (after removal of the termination codon) was amplified by RT-PCR from total RNA extracted from physic nut leaves using the primers given in Supplementary Table S1. The target sequence was cloned into the pSAT6-eYFP-N1 vector to construct the pSAT6-JcDREB2-eYFP fusion expression vector. The 35S::JcDREB2-YFP construct and a control (35S::YFP) plasmid were introduced into Arabidopsis protoplasts by the polyethylene glycol (PEG) mediated method. Next the transformed cells were incubated in the light for 16 h at 25°C. YFP fluorescence signal was detected by laser scanning confocal microscopy. Arabidopsis protoplasts was prepared according to Axelos (1992).

Total RNA was isolated from the leaves of 14-day-old physic nut seedlings using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions, and first-strand cDNA was obtained following Tang et al. (2016). JcDREB2 fragments including the complete coding sequences were amplified from physic nut by PCR with the primer pairs shown in Supplementary Table S1. The products were then cloned into the vector pMD 18-T (TaKaRa, Otsu, Japan) for use in sequencing and analysis. The target sequences were digested with the restriction enzymes Sac-I and Xba-I and then inserted into the corresponding restriction sites of the pCAMBIA1301 vector under the control of a CaMV35S promoter. Next, Agrobacterium harboring the constructs was used to transform and regenerate rice seedlings, as described by Chen et al. (2003). In total we obtained seven independent transgenic lines, and selected three single gene insertion homozygous transgenic lines for further study based on an approximately 3:1 segregation ratio being observed among T2 plants. The segregation ratios, which were confirmed by GUS staining, were 216:70, 198:63 and 237:76. T2 seeds were germinated and 80 roots were taken from 10-day-old seedlings of each line of transgenic rice for GUS staining. GUS expression was observed in all roots from each line of 10-day-old seedlings.

After germination, rice seedlings were cultured in Yoshida’s culture solution (Yoshida et al., 1976) at 25°C under 16/8 h (light/dark) conditions in a growth chamber. Two weeks later, seedlings were planted in soil in plastic pots and transferred to a greenhouse under natural sunlight. For GA₃ treatment, wild-type and JcDREB2 overexpressing rice seeds were kept in water for 2 days for germination, then the seeds were incubated in Yoshida’s culture solution containing 10 μM GA₃ for 10 days under natural sunlight. As an alternative GA treatment, plants grown in a greenhouse under natural sunlight were sprayed daily with 100 μM GA₃ for 6 days (10-day-old seedlings) or 10 days (4-week-old seedlings). For the salt tolerance assay, 2-week-old seedlings were incubated in Yoshida’s culture solution containing 150 mM NaCl for 5 days, followed by incubation in Yoshida’s culture solution for 10 days at 25°C under 16/8 h (light/dark) conditions in a growth chamber, and then the survival rates were calculated. The relative electrolyte leakage (REL) and proline content were determined 2 days after salt treatment. Leaf samples were collected 2 days after salt treatment. All tests were repeated three times with three biological replicates for each test.

About 0.2 g of leaf sample was used for each determination. Firstly, each sample was washed five times with deionized water and placed in a test tube containing 10 mL of deionized water. The leaf samples were immersed and vibrated continuously at 25°C for 2 h, then a conductivity meter was used to measure the electrical conductivity (B1) of the solution. Next the samples were boiled for 15 min, and after cooling the solutions to room temperature, their conductivity (B2) was measured again. The REL was calculated according to the following formula: REL (%) = B1/B2 × 100. The content of free proline in the rice leaves was determined as previously described (Bates et al., 1973). The content of malondialdehyde (MDA) in the leaves was determined as described by Cai et al. (2015). The activities of catalase (CAT) and superoxide dismutase (SOD) in the rice leaves were determined as previously described (Azooz et al., 2009).

Total RNA was isolated from the leaves of 14-day-old rice seedlings using TRIzol reagent (Invitrogen, USA) according to the manufacturer’s instructions. Total RNA was extracted from roots, stem cortexes, flowers, and seeds of physic nut plants using a HiPure Plant RNA Mini kit (Promega, Madison, WI, USA) following the manufacturer’s instructions. First strand cDNA was obtained by reverse transcription of total RNA using M-MLV reverse transcriptase (Promega, Madison, WI, USA) following the manufacturer’s instructions. Primers used in the experiment are listed in Supplementary Table S1. Semi-quantitative RT-PCR with gene-specific primer pairs was used to detect the level of JcDREB2 gene expression in wild-type and transgenic plants. The OsUbiquitin gene was used as a control.

Quantitative real-time PCR was performed in a reaction volume of 20 μL containing 2 μL of cDNA, 10 μL of 2 × SYBR Premix ExTaq, 0.4 μL forward primer (10 μmol), 0.4 μL reverse primer (10 μmol) and 7.2 μL of ddH2O using a LCS480 system (Roche²). Conditions for all quantitative real-time PCR amplifications were as follows: 95°C for 10 min for DNA polymerase activation, followed by 40 cycles of 95°C for 15 s and 60°C for 35 s. The relative expression level was calculated using the 2^-ΔΔCT method, and OsUbiquitin and JcActin were used as the reference genes for rice and physic nut, respectively. All experiments included three biological replicates, each with two technical replicates.

All experiments included three biological replicates, and the data collected were analyzed using the Duncan multiple range test (Duncan, 1955) with the SAS software package³.

The full-length JcDREB2 cDNA was isolated from physic nut by reverse transcription PCR. The JcDREB2 sequence so obtained contained an open reading frame of 1074 bp that encoded a protein of 358 amino acids with one conserved AP2/ERF domain. The JcDREB2 sequence was deposited in the NCBI GenBank database under accession number JCGZ_24071. The conserved Val (14th) and Glu (19th) residues in the AP2/ERF domain are crucial in regulating the binding activity of DREB proteins to the DRE element (Cao et al., 2001). Further analysis of the deduced amino acid sequence indicated that the JcDREB2 protein contained these conserved 14th valine and 19th glutamic acid residues in the AP2/ERF domain and that it was highly homologous to DREB proteins from other plants (Figure 1A).

To further determine the relationship between JcDREB2 and other DREB proteins, a phylogenetic tree was established using the complete amino acid sequence (Figure 1B). Among the proteins included, the JcDREB2 protein was most closely related to cassava MeDREB, with a 75% match at the amino acid level (Figures 1A,B).

To determine the expression pattern of JcDREB2 in physic nut, we analyzed its expression in five different tissues: roots, stem cortexes, leaves, flowers, and seeds by real-time quantitative PCR (qRT-PCR) using JcDREB2-specific primers. JcDREB2 was constitutively expressed in all of the tissues examined, but the highest level of expression was detected in leaves (Figure 2A).

Quantitative PCR analysis was performed to measure JcDREB2 expression under various abiotic stresses, including salt, GA₃ and ABA. As shown in Figure 2B, JcDREB2 expression was repressed in physic nut leaves 1, 3, and 6 h after treatment with salt and GA₃. However, under ABA stress, the level of JcDREB2 transcript was markedly up-regulated from 6 to 12 h after treatment.

In order to determine the subcellular localization of the JcDREB2 protein, we examined the localization of JcDREB2-YFP by laser scanning confocal microscopy (Figure 3). The YFP fluorescence produced by 35S::JcDREB2-YFP was localized in the nucleus, whereas YFP fluorescence produced by 35S::YFP was detected throughout the cell. These results indicate that JcDREB2 is a nuclear-localized protein.

To investigate the function of JcDREB2, we overexpressed the JcDREB2 gene in rice under the control of the CaMV 35S promoter. Changes in levels of JcDREB2 transcripts in leaves were analyzed by semi-quantitative RT-PCR (Figure 4A). We observed that all transgenic rice plants overexpressing JcDREB2 (OeJcDREB2) had a severe dwarf phenotype under normal growing conditions (Figure 4B). The shoot and root length was shorter in OeJcDREB2 rice plants, compared with those of wild-type plants (Figures 4C,D).

Gibberellin, the most important of the hormones regulating shoot elongation, flowering, and seed development, plays an important role in determining plant height (Sakamoto et al., 2004). To examine whether the dwarf phenotype of OeJcDREB2 rice plants was caused by GA deficiency, we cultivated wild-type and transgenic seedlings in Yoshida’s culture solution containing 10 μM GA₃ for 10 days at the germination stage. Under these conditions, shoots of transgenic as well as of wild-type plants showed rapid elongation (Figures 5A,D), but the elongation ratio (+GA₃/-GA₃) of OeJcDREB2 plants was significantly higher than that of the wild-type (Figure 5G). Furthermore, spraying OeJcDREB2 rice plants with 100 μM GA₃ rescued shoot elongation; in particular, the length of the leaf sheath was fully restored to a level similar to that in wild-type plants (Figures 5B,C,E,F). The elongation ratio (+GA₃/-GA₃) of OeJcDREB2 plants was also significantly higher than that of the wild-type (Figures 5H,I). These results suggest that OeJcDREB2 rice plants do indeed exhibit a classic reduced-GA phenotype.

To investigate whether overexpression of JcDREB2 altered the expression of GA biosynthesis genes, we used qRT-PCR assays to detect the expression of genes encoding two key enzymes [gibberellin 20-oxidase (GA20ox) and gibberellin 3β-hydroxylase (GA3ox)] as well as other genes responsible for the early steps in GA biosynthesis, including OsCPS1, OsKO2, and OsKAO. As shown in Figure 6, the expression levels of OsGA20ox1, OsGA20ox2, OsGA20ox4, OsGA3ox2, OsCPS1, OsKO2, and OsKAO exhibited a significant reduction in OeJcDREB2 seedlings compared with their levels in wild-type. GID2 (gibberellin-insensitive dwarf2) is involved in the GA signaling pathway in rice (Gomi et al., 2004). We therefore also examined the expression levels of the GID2 gene, but no significant difference in GID2 expression was seen between OeJcDREB2 and wild-type plants (Figure 6).

To examine whether JcDREB2 plays a role in salt stress tolerance, we compared the salt tolerance of OeJcDREB2 and wild-type plants at the vegetative growth stage. After 5 days of salt treatment with 150 mM NaCl, 73.3% of wild-type seedlings remained green, whereas OeJcDREB2 seedlings showed severe leaf aging and rolling (Figure 7A). The survival rates of wild-type and OeJcDREB2 seedlings after 10 days of recovery were statistically analyzed. Our data showed that approximately 38% of the wild-type seedlings survived, whereas no OeJcDREB2 seedlings survived after the recovery period (Figure 7B). However, overexpressing JcDREB2 had no significant effect on drought tolerance (data not shown) in rice. Proline accumulation is considered to be a plant adaptive response to high salinity and drought stresses (Xiong et al., 2012). Under normal growth conditions, the proline content did not differ between wild-type and OeJcDREB2 plants. However, under salt treatment, the proline content of OeJcDREB2 plants was significantly lower than that of the wild-type (Figure 7C). This finding indicated that salt stress was more damaging to the OeJcDREB2 plants than to the wild-type plants. Electrolyte leakage is an important index of cell membrane damage in plant stress responses. The relative level of electrolyte leakage from wild-type leaves was lower than that from OeJcDREB2 leaves after the salt stress treatments (Figure 7D).

Membranes abound in polyunsaturated fatty acids which serve as antioxidants by capturing reactive oxygen species, thereby protecting plants from being damaged under extreme conditions. This process generates MDA, a typical marker of the oxidation of polyunsaturated fatty acids. The MDA content of wild-type plants was lower than that of OeJcDREB2 plants under salt treatment (Figure 7E). These results suggested that there was more cell membrane damage in OeJcDREB2 leaf cells than in wild-type leaf cells under salt stress.

Superoxide dismutase is an antioxidant enzyme that catalyzes the conversion of the superoxide radical into hydrogen peroxide, and CAT further catalyzes the transformation of harmful hydrogen peroxide into harmless water, so protecting plants from damage by reactive oxygen species arising as a result of abiotic stress. We analyzed CAT and SOD activities under salinity stress and normal conditions. The results suggested that the activities of CAT and SOD from OeJcDREB2 leaves were lower than those from wild-type leaves under salinity stress, whereas no significant difference was found under normal growth conditions (Figures 7F,G).

Finally, we measured the expression levels of several genes which have previously been shown to respond to salt stress in rice seedlings, since overexpressing these genes could improve the salinity stress tolerance of transgenic plants. The products of the OsHKT1;1 gene are crucial for salt tolerance in rice through the exclusion of Na⁺ ions from sensitive cells (Ren et al., 2005). The OsGR3 gene products are involved in scavenging reactive oxygen species (Wu et al., 2015). Increases in the levels of expression of OsHKT1;1, OsP5CS and OsGR3 under salt treatment were strongly impaired in the leaves of two OeJcDREB2 plants (Figure 7H). In addition, the expression level of the salt-stress-related transcription factor gene SNAC1 (STRESS-RESPONSIVE NAC 1) (Hu et al., 2006) was significantly lower in OeJcDREB2 leaves than in wild-type leaves under salinity stress (Figure 7H).