Tartary buckwheat, called “Xiqiao No. 2,” was cultivated in a farm at Sichuan Agriculture University, and samples were collected to clone and detect the expression profiles of FtbHLH3 after multiple abiotic stresses. Arabidopsis thaliana (ecotype Columbia-0) was grown in an artificial climate chamber at 25°C and approximately 60% humidity (16 h light/8 h dark). Seven-day-old seedlings of tartary buckwheat were treated with the following conditions: 100 μM abscisic acid (ABA), 1 mM salicylic acid (SA), 302 nm UV-B light (302 nm, 0.1 mW/cm2), 4°C, 150 mM NaCl, and 30% PEG6000 (Su et al., 2014; Zhou et al., 2015), separately. For all of the treatments, the seedlings were collected at 0, 6, 12, 24, and 48 h (0 h was a non-treated control), and frozen in liquid nitrogen for further study. Three independent biological replicates were measured for each sample.

Total RNA was extracted from the samples using an RNAout kit (Tiandz, Beijing, China), and cDNA was synthesized with a RevertAid First Strand cDNA Synthesis kit (MBI, USA). Based on the flowering tartary buckwheat transcriptome database constructed by our laboratory (data not shown), a novel bHLH gene (designated as FtbHLH3) was isolated using specific primers. The gene sequence was analyzed via the NCBI database. Multiple amino acid sequence alignments were performed using ClustalX, and a phylogenetic tree was constructed using the neighbor-joining (NJ) method. The primers are listed in Supplementary Table S1.

To determine its exact subcellular location, the coding region of FtbHLH3 was amplified using primers incorporating the BamHI and SmaI sites at their terminus and ligated into the HBT95-GFP vectors with the same polyclonal site to produce FtbHLH3-GFP fusion proteins. Recombinant plasmids FtbHLH3-GFP and HBT95-GFP (used as a negative control) were separately transferred into the Arabidopsis protoplasts as previously described (Schirawski et al., 2000). The transformed cells were incubated at 22°C for 24 h in darkness and photographed using a laser confocal scanning microscope (Leica DM IRBE).

For the transactivation assay, the coding region of FtbHLH3 was PCR-amplified using the following primers 5′-CGCGGATCCATGGAGGTAAATGAAGATGGGTT-3′ (BamHI site underlined) and 5′-AACTGCAGCATATACTTTCCTCCATAACCTGCATT-3′ (PstI site underlined) and cloned into the pBridge vector to create the pBridge-FtbHLH3 constructs. The pBridge-FtbHLH3, the negative control pBridge (pBridge is the empty vector), and the positive control pBridge-GmMYBJ6 (GmMYBJ6 has transcriptional activation activities) plasmids were transformed into the AH109 yeast strain. The transformed cells were screened in SD/-His-Trp medium. Then, their LacZ activities were identified using the galactosidase filter lift assay.

The coding region of the FtbHLH3 gene was PCR-amplified from tartary buckwheat using the following primers 5′-GGGGTACCATGGAGGTAAATGAAGATGGGTTT-3′ (KpnI site underlined) and 5′-CGGGATCCCATATACTTTCCTCCATAACCTGC-3′ (BamHI site underlined), and the sequence was attached to the vector pCAMBIA1301. The recombinant vector pCAMBIA1301-FtbHLH3 was transformed into Arabidopsis Col-0 through the Agrobacterium tumefaciens strain GV3101 (Clough and Bent, 1998). The first generation (T0) seeds of transgenic plants were collected and screened in 1/2 MS medium with 50 mg L⁻¹ hygromycin. Hygromycin-resistant plants were transferred into soil pots and grown in a growth chamber with a 16 h photoperiod cycle at 25°C. Finally, transgenic Arabidopsis plantlets were identified by PCR amplification. The T3 homozygous positive lines were used for all further experiments.

For the osmotic stress tolerance assay, transgenic A. thaliana T3 and WT were used in vitro. Sterilized seeds were germinated in 1/2 MS medium under a 16 h photoperiod cycle at 25°C, and 4-day-old seedlings were transplanted to 1/2 MS medium supplemented with 150 and 250 mM mannitol. The germination rate, root length, content of hydrogen peroxide (H₂O₂), and SOD, CAT and POD activities were measured as previously described (Maehly and Chance, 1954; Beyer and Fridovich, 1987). For the drought stress tolerance assay, positive lines and WT plants were sown in pots and regularly watered for 3 weeks. At 4 weeks, plants with a similar growth status were subjected to water withholding for 30 days. The survival rate was measured after 30 days. Meanwhile, proline (Pro), malondialdehyde (MDA), ion leakage (IL), hydrogen peroxide (H₂O₂), superoxide anion radicals (O2−), peroxidase (POD), catalase (CAT), and superoxide dismutase (SOD) activities were detected for 15 or 30 days in the transgenic A. thaliana and WT plants (Heath and Packer, 1968; Pan et al., 2012). For the oxidative stress assay, the leaves of transgenic and WT plants were incubated on liquid MS medium and MS medium containing different concentrations of methyl viologen (MV) (10, 20, and 30 μM) for 32 h in the light. The chlorophyll and H₂O₂ content and antisuperoxide anion activity (an index indicating superoxide anion levels) were determined after treatment. Simultaneously, 5-week-old transgenic and WT lines were subjected to 20 μM MV for 15 days. Afterward, leaves were collected to detect chlorophyll fluorescence, DAB and NBT staining, and SOD, CAT, and POD activities (Beyer and Fridovich, 1987; Maehly and Chance, 1954). Furthermore, the expression levels of AtSOD, AtCAT and AtPOD were detected under the same treatment condition. Each sample contained three replicates.

The 5′-upstream promoter region of FtbHLH3 was isolated according to the Genome Walking Kit (TaKaRa, Japan) with nested specific primers sp1, sp2, sp3, and AP primers provided in the kit. The PCR product was subcloned into a pMD™19-T (TaKaRa, Japan) and sequenced. The possible putative regulatory elements of the FtbHLH3 promoter region were identified using the PLACE/Signal Scan database.

The obtained promoter region, designated FtbHLH3P, was amplified from F. tataricum genomic DNA using the specific primers 3P_f and 3P_r and cloned into the plant expression vector pBI101-GUS. The recombinant plasmids FtbHLH3P-GUS and positive cauliflower mosaic virus (CaMV) 35S-GUS were introduced into Arabidopsis using the Agrobacterium tumefaciens strain GV3101, and positive lines were selected using kanamycin (50 mg L⁻¹). The positive lines with FtbHLH3P:GUS and WT plants were grown in pots for 3 weeks, and 4-week-old seedlings were subjected to water withholding for a week. Histochemical analysis of GUS activity was performed.

Total RNA was extracted using the RNeasy Plant Mini Kit (Qiagen, Valencia, CA, USA), and the purity was determined by spectrophotometry. The first-strand cDNA was synthesized using the Revert Aid™ First Strand cDNA Synthesis Kit (MBI, US). Each reaction contained 5 μL of SYBR Green II Mix, 1 μL of cDNA sample, 1 μL of primer, and 3 μL of double distilled water. The PCR protocol was as follows: 95 °C for 30 s and then 39 cycles of 95 ° C for 5 s, and 60 °C for 30 s. The housekeeping genes FtH3 (ID: HM628903) and β-actin (ID: NM_112764) were used as internal standard genes in tartary buckwheat and Arabidopsis thaliana, respectively. The data were evaluated using the 2 ^−ΔΔCT method (Livak and Schmittgen, 2001). Three independent biological replicates were measured for each sample.

Statistical analyses were performed with SPSS software (SPSS 13.0). Analysis of variance was used to compare the significant difference based on Student's t-test (n = 4).

To identify putative bHLH genes that may be closely involved in the abiotic stresses, two bHLH genes AtbHLH116 and AtbHLH61 as probes were screened from the tartary buckwheat transcriptome database constructed by our laboratory. According to a set of unigenes with annotation as bHLH TFs, more than 4 full length bHLH genes were cloned. Of them, the gene (FtbHLH3) encoded a bHLH factor which was clustered with the PebHLH35 factor involved in drought stress in Populus euphratica (Dong et al., 2014). The FtbHLH3 gene encodes a protein of 318 amino acid residues with a predicted molecular mass of 36.12 kDa and a calculated pI of 4.89. FtbHLH3 has an 957 bp open reading frame (GenBank Accession No. KU296217). The predicted FtbHLH3 protein contained a putative NLS sequence. The multiple sequence alignment results implied that a sequence rich in acidic amino acids is present at its C-terminal region (Figure 1A). Phylogenetic analysis showed that FtbHLH3 was classified with stress-related bHLH TFs from other plants into the same cluster belonging to the subgroup III of Arabidopsis bHLH (Zhao et al., 2013). That means FtbHLH3 might have the similar functions involved in stress responses in plant (Figure 1B).

To further verify the characterization of FtHLH3 as a transcription factor, 35S:FtbHLH3:GFP was transiently expressed in Arabidopsis protoplast cells. Figure 2 showed that the 35S:GFP control displayed fluorescence throughout the whole cell. By contrast, the strong green fluorescence was very localized in FtbHLH3:GFP transformed cell, suggesting that FtbHLH3 could be located in the nuclear. Moreover, the transcriptional activity of FtbHLH3 was analyzed using a yeast assay system. Yeast cells containing pBridge-FtbHLH3 and pBridge-GmMYBJ6 (positive control) grew well in SD/-Trp/-His medium, whereas cells containing empty pBridge (negative control) did not grow (Figure 3A). The result indicates that FtbHLH3 activated the transcription of reporter gene His in the yeast AH109 strain. Meanwhile, there was no color change in the negative control, but the positive control and pBridge-FtWD40 had a significant blue reaction after the treatment with X-gal (Figure 3B). The above results indicate that FtbHLH3 is a transcriptional activator.

To determine the expression patterns of FtbHLH3 in different tissue, qRT-PCR was carried out with mRNA from different tissues. FtbHLH3 was expressed in all tissues examined including root, stem, leaf and flower. Its transcription level in the flower was 6.09-, 1.41-, and 1.89-fold compared to that in the root, stem and leaf, respectively (Supplemental Figure S1). To investigate the responses of FtbHLH3 to multiple abiotic stresses, 4°C, ABA, NaCl, PEG6000, SA, and UV-B treatments were applied. As a whole, the FtbHLH3 gene showed different expression patterns under various abiotic stress treatments. Under PEG treatment, the expression of the FtbHLH3 gene rapidly increased and reached a maximum value (14.56-fold) at 12 h, then tended to be high with little alteration (Supplemental Figure S2). Meanwhile, expression of FtbHLH3 was also induced by 4°C and ABA treatment, but the highest levels of transcription occurred at 24 h (Supplemental Figure S2). Under UV-B treatment, the FtbHLH3 expression increased slowly at the early stages (within 12 h), reached peak at 24 h, and then remained relatively stable (Supplemental Figure S2). In addition, The NaCl and SA treatments had no significant effect on the expression of FtbHLH3 gene (Supplemental Figure S2). In general, PEG had more pronounced effects than the other treatments.

To further characterize the function of FtbHLH3 under drought conditions, transgenic Arabidopsis plants with oxFtbHLH3 were generated. A total of 10 transgenic lines (T1) were obtained with hygromycin-resistance (50 mg L⁻¹) and PCR identification (Supplemental Figure S3A). The results indicated that FtbHLH3 was expressed in all of the transgenic plants but not in WT, and we selected three higher FtbHLH3 expression lines (#3, #5, and #9) for subsequent analysis (Supplemental Figure S3B). The selected transgenic lines produced T3 homozygous generation, which was used for the following experiments.

To confirm whether FtbHLH3 was involved in osmotic stress, the germination rate was measured on mannitol. When the seeds were sown on the 1/2 MS medium containing mannitol, the positive plants' germination rates were clearly higher than that of the WT plants, and the difference was more significant with an increase in the mannitol concentration (Figure 4A). However, there was almost no difference under normal growth conditions. In the follow-up experiment, the length of the root was measured under the same previous treatment condition. After mannitol treatment, transgenic lines showed a significantly longer root compared to WT plants (Figure 4B). In addition, the difference was more obvious with an increased concentration. Furthermore, to understand the role of FtbHLH3 in scavenging ROS, the H₂O₂ content and activities of antioxidant enzymes in transgenic lines and WT plants were detected after 7 days of mannitol treatment. The results showed that transgenic lines contained lower H₂O₂ accumulation and higher SOD and CAT activities than WT plants (Figure 5).

For drought tolerance analysis of FtbHLH3 transgenic lines in soil, 4-week-old transgenic lines (#3, #5, and #9) and WT were tested. After withholding water for 15 days, all of the leaves of the WT plants were heavily curled while only a few FtbHLH3 transgenic plants were affected by the water stress and most of them were still green and fully expanded (Figure 6A). After 30 days of water stress, all of the leaves of the WT plants wilted and their leaves became white, whereas the transgenic plants remained turgid and became purple. After 7 days of re-watering, most of transgenic plant leaves became green and grew buds again. The survival rate of the positive lines was significantly higher than WT (Figure 6B).

To clarify the effect of FtbHLH3 on physiological indexes of transgenic plants after withholding water, we also measured proline, MDA, and ion leakage (IL). The results showed that proline, MDA, and IL displayed a similar pattern in all of the plants under normal conditions. However, the content of proline was significantly higher than WT plants (Figure 6C). Meanwhile, after drought stress treatment, MDA and IL were significantly lower in the transgenic lines than WT after prolonged stress (Figures 6D,7A). At the beginning of drought stress, the content of O2− and H₂O₂ increased in both the experimental and control groups. Over time, the accumulation of O2− and H₂O₂ were significantly lower in the transgenic plants than that in WT (Figure 7A). We also detected the activities of three important antioxidant enzymes (POD, CAT, and SOD). Under adequate water conditions, the three antioxidant enzyme activities were not different between the transgenic lines and WT. After drought treatment, the transgenic plants showed higher SOD and CAT activities (Figure 7B). There was no obvious change in POD activity after treatment.

To more deeply study the molecular mechanism of FtbHLH3 in drought stress, the expression levels of several genes were detected in the transgenic and WT plants under normal and drought conditions (Figure 8). Genes selected for this analysis included AtSOD, AtCAT, and AtPOD involved in ROS detoxification, AtABA2, AtAAO, AtNCED, and AtZEP involved in ABA biosynthesis, AtP5CS and AtP5CR involved in proline biosynthesis, and ERD4, ATDR4, MDAR, ALDH3H1, PLC1, and AtMYC2 related to drought stress defense (Wang et al., 2016a). The qPCR results indicated that the expression of ATDR4 and AtNCED was higher in transgenic plants than in the WT plants under the normal conditions, while the other genes had no obvious change. After the drought treatment for 30 days, the tested genes, including MDAR, ALDH3H, AtMYC2, ERD4, ATDR4, AtCAT, AtSOD, AtNCED, AtZEP, and AtP5CS, showed remarkably increased expression in transgenic Arabidopsis compared to WT plants. The data could suggest that FtbHLH3 play an important role in the regulatory mechanism in some ways.

Drought conditions are often accompanied by oxidative stress. MV is a herbicide and is commonly used to simulate oxidative stress, which generates ROS and causes membrane damage and chlorophyll degradation when used to treat the plants (Slade, 1966). As shown in Supplemental Figure S4A, most of the wild-type leaves were bleached while the color of the transgenic lines retained more green under the same concentration. Furthermore, with the increase in MV, the leaves of wild-type plants were injured more seriously, but the transgenic plants showed less injury. The results of the chlorophyll content determination indicated that the positive plants had significantly higher amounts than the wild type (Supplemental Figures S4B,C). In addition, transgenic plants also displayed lower H₂O₂ contents than WT (Supplemental Figure S4D). In contrast, higher antisuperoxide anion activity was detected in the transgenic lines (Supplemental Figure S4E).

According to the above results, we selected 20 μM MV to treat the plants. Most of the WT plants died, whereas the transgenic lines exhibited better growth phenotypes when 30-day-old A. thaliana were treated with MV for 15 days (Figure 9A). The survival rates of the #3, #5, and #9 lines were 80, 90, and 60%, respectively, significantly higher than the 30% of WT (Figure 9B). DAB and NBT staining indicated that the WT plants accumulated more H₂O₂ and O2− than the three positive lines (Figures 9C,D). Meanwhile, the activity of three enzymes (POD, CAT, and SOD) increased in the WT and positive lines after oxidative stress; however, the increase in the positive plants was significantly higher (Figure 10A). We also examined the expression of AtPOD, AtCAT, and AtSOD, and the qRT-PCR analysis results showed that the expression of the three genes in the transgenic plants was significantly higher than in wild type under oxidative stress (Figure 10B).

Many studies have indicated that photosystem II (PSII) is the most relevant to osmotic stress. To analyze the effects of oxidative stress on PSII photochemistry in transgenic and WT plants, chlorophyll fluorescence images and values were detected using a modulated imaging fluorimeter. There was no significant difference in the color or value of the maximum photochemical efficiency of PSII in the dark-adapted state (Fv/Fm), the photochemical quenching (qP), the quantum yield of photosystem II (PSII) photochemistry (ΦPSII), and the non-photochemical quenching coefficient (NPQ) before treatment (Figure 11). Compared with transgenic plants, WT presented a significant decrease in the Fv/Fm, ΦPSII, and qP under oxidative stress for 1 and 3 days (Figure 11). Meanwhile, oxidative stress resulted in a significant increase in the level of NPQ in WT compared with transgenic plants.

In this study, we cloned a 989 bp promoter fragment upstream of the start codon of FtbHLH3. Sequence analysis indicated that multiple cis-acting elements involved in abiotic and biotic stress responses were contained in the FtbHLH3P region. For example, there are three copies of the ACGT element, which was reported to be involved in up-regulation of the erd1 gene in Arabidopsis thaliana by drought stress (Simpson et al., 2003). The GATA-box, T-box, and I-box participated in light responsiveness. In addition, many other cis-acting elements involved in the regulation of genes under abiotic stresses, such as ABA-response-element (ABRE), cold (DRE2 element), water (MYB), and salicylic acid signaling (W-box), were also observed (Supplemental Figure S5).

The activity of FtbHLH3P was tested in A. thaliana plants transformed with FtbHLH3P:GUS fusion constructs; non-transformed plants were used as negative controls, and plants transformed with the CaMV 35S promoter were used as positive controls. According to the results of the GUS activity assay, the positive control and transgenic lines with FtbHLH3P stained different degrees of blue (Figure 12). Meanwhile, the GUS histochemical assays were more obvious when transgenic Arabidopsis lines were treated with drought treatment.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.