There are 308 bHLH loci in the soybean genome. Amino acid sequences of all GmbHLH genes were downloaded from the database of phytozome². A comparative tree was constructed based on the amino acid sequence alignments of the 308 bHLH soybean genes (Supplementary Table S1). The clustering showed that there are 12 major subfamilies of related sequences that were strongly supported by bootstrap analysis; these subfamilies were titled I to XII (Figure 1). GmORG3 (Glyma03g28630), the red arrow-marked position, was classified as subfamily “II.” In a previous study, AtbHLH TFs were also divided into 12 subfamilies (Heim et al., 2003); the AtbHLH39, which is the orthologous gene of GmORG3, was classified as subfamily “Ib.” In Arabidopsis AtbHLH TFs were divided into 21 subfamilies in another study (Toledo-Ortiz et al., 2003); the AtbHLH39 was in the “2” subfamily.

The GmORG3 protein consists of 241 amino acids and contains one 61-residue-long conserved bHLH domain spanning amino acids 58–118 (Figure 2). The basic helix1-loop-helix2 domain sequence is PAMVKKLSHNASERDRR-KKVNDLVSSLRSLLP-GPDQTKKMS-IPATVSRVLKYIPELQH QVQ. Basic and helix domains are more conservative than are loop domains. There are conserved amino acids in the soybean bHLH transcription factors: the 4th residue in the basic part is L (Leucine)/V (Valine); the 3rd, 6th, and 13th residues in the helix1 part are V, L, and L, respectively; and 10th residue in the helix2 part is L. Those conserved amino acids should have important functions.

A GmORG3-green fluorescent protein (GFP) fusion construct and a control construct containing only GFP were created using the pJIT166-GFP vector. Transcription was driven by the CaMV 35S promoter in both constructs. Constructs were introduced into Arabidopsis protoplasts in accordance with the PEG4000-mediated method. The cells were inspected with a confocal laser-scanning microscope to visualize the subcellular localization of the proteins. The GmORG3-GFP protein was detected only in the nucleus; the control protein was observed throughout the cell (Figure 3). These results indicated that GmORG3 is a nuclear protein.

Transcript abundance of GmORG3 was determined in soybean plants. Total RNA was isolated from the roots, stems, leaves, flowers, and pods of soybean at the first-trifoliate, full-bloom, and full-pod stages. RNA was reverse-transcribed to obtain cDNA for semi-quantitative PCR analysis. The results showed that the transcripts of GmORG3 were detected only in the roots at all selected stages, and the highest transcript abundance in the root tissue was detected at the full-pod stage (Figure 4). GmActin was used as an internal control.

The transcript abundance response of GmORG3 was also detected in soybean plants growing under Cd stress. qRT-PCR analyses were carried out to determine the transcript abundance patterns of the GmORG3 gene in the roots of soybean plants. As shown in Figure 5, the expression of GmORG3 was clearly up-regulated in the roots of soybean seedlings after an overnight treatment with 100 μM CdCl₂. The expression level of GmORG3 under CdCl₂ was 6.13-fold higher than that under normal conditions.

The function of GmORG3 was investigated by its construction into the plant overexpression vector pCXSN and subsequent transformation to generate soybean mosaic seedlings via Agrobacterium rhizogenes mediation. Soybean mosaic seedlings were tested under Cd stress to determine Cd contents in roots and leaves.

Soybean mosaic seedlings having transgenic roots and non-transgenic shoots were generated in accordance with procedures explained elsewhere. PCR-validated transgenic seedlings of GmORG3 and those with empty vectors were grown in 100-ml glass tubes containing half-strength Hoagland solution for a few weeks to develop enough transgenic roots. Afterward, the mosaic seedlings were transferred to fresh ½-strength Hoagland solution that contained 20 μM CdCl₂ and grown for 9 days. The leaves and roots of both seedlings with empty vector and GmORG3 mosaic seedlings were harvested separately at 0, 1, 2, 4, 6, and 9 days after treatment. The Cd content in the samples was determined using an atomic absorption spectrometer with a continuous light source following the manufacturer’s instructions/manual (ContrAA^®300, Analytik Jena AG).

Soybean mosaic plants with GmORG3 transgenic hairy roots did not significantly differ from those with empty vector transgenic hairy roots after 9 days of growth under 20 μM CdCl₂ stress (Figure 7A). However, the Cd content was lower in GmORG3 transgenic roots than in those of plants with empty vectors during the first 2 days, but this finding was opposite at 4 and 6 days. No Cd was detected immediately after treatment (0 days) (Figure 7Ba). It is interesting that the translocation ratio of Cd from roots to shoots (Cd amount in shoots/total Cd amount) was consistent at every sample collection time. The GmORG3 transgenic root seedlings had a lower ratio at all sample collection times (Figure 7Bc). Therefore, when roots absorb the same amount of Cd, GmORG3 transgenic roots will translocate less Cd to shoots.

Considering that the absorption and transport of Fe is related to transporters that can also take up Cd and other metals, Fe accumulations in the roots and shoots of mosaic soybean seedlings were measured and analyzed (Figure 7Bb,d). The Fe content was higher in the roots than in the shoots, and a subtle increase in Fe uptake was evident with increasing time of Cd stress. GmORG3 showed slightly lower Fe content than did empty vector seedlings during the first 2 days, whereas at 4, 6, and 9 days of treatment, GmORG3 seedlings showed slightly higher or equal Fe content than did those with empty vectors (Figure 7Bc). Additionally, the Fe translocation ratio from roots to shoots (Cd amount in shoots/total Cd amount in plants) was also calculated. Interestingly, the GmORG3 transgenic mosaic seedlings had a higher ratio at all time points (Figure 7Bd).

Since GmORG3 significantly responded to cadmium stress and its overexpression conferred tolerance to Cd stress in yeast, further characterization of this gene was carried out. The heterologous ectopic over expression of GmORG3 in tobacco plants showed enhanced tolerance to cadmium stress. The GmORG3 CDS was used to prepare the construct of pCXSN-GmORG3, which have a 2X 35S promoter and a NOS terminator. In addition, the correct recombinant plasmid was transformed into tobacco by Agrobacterium-mediated leaf disk explants methods. The regenerated transformed seedlings were screened on medium containing hygromycin (100 mg/L). The putative tobacco transformants were identified using GmORG3 gene-specific PCR analysis. The T3 seeds of transgenic plant line 1 (L-1), line 2 (L-2), and line 3 (L-3) were used in subsequent experiments.

Wild-type (WT) and transgenic tobacco L-1, L-2, and L-3 seeds were sown on ¹/₂ MS medium (MS), ¹/₂ MS medium supplemented with 20 μM CdCl₂ (MS/+Cd), ¹/₂ MS medium without Fe (MS/-Fe) and ¹/₂ MS medium with +Cd/-Fe (MS/+Cd-Fe) (Figure 8). The growth of transgenic seedlings on MS control medium was similar to that of WT seedlings (Figure 8A): the roots of transgenic plants were slightly shorter than those of WT. The root length assays indicated that the average root length of WT plants was 1.10 cm and that of transgenic plants L-1, L-2, and L-3 were 0.98, 0.99, and 1.11 cm, respectively (Figure 8Ba). The leaf color of both WT and transgenic plants was also similar. The content of chlorophyll was 1.30 mg/g of fresh weight (FW) in the leaves of WT and was 1.24, 1.30, and 1.30 mg/g (FW) in the leaves of transgenic lines L-1, L-2, and L-3, respectively (Figure 8Bb).

The growth of transgenic seedlings on MS medium supplemented with 20 μM CaCl₂ showed obvious differences with that of WT. The roots of transgenic seedlings were longer than those of WT; there were significant differences between root lengths of transgenic L-1 and WT seedlings but no significant differences between those of transgenic L-2 and L-3 with WT seedlings (Figure 8Ba). The chlorophyll contents varied significantly between all transgenic lines and WT (Figure 8Bb). On Fe-deficient medium, the root lengths of transgenic seedlings were significantly longer than those of WT, while chlorophyll contents in transgenic lines were higher than those of WT (Figure 8B). The responses of all transgenic and WT seedlings were worse on +Cd/-Fe medium than on other media; all seedlings grown on +Cd/-Fe medium were short. However, the chlorophyll content of all the transgenic lines was significantly (P < 0.05) higher than that of WT. Similarly, the root length of transgenic seedlings was also significantly (P < 0.01) longer than that of WT.

The above results demonstrated that overexpressing GmORG3 improves the tolerance of transgenic tobacco plants to cadmium stress.

Plants of the Chinese cultivar Dongnong 690 of soybean (Glycine max L. Merr.), the Columbia ecotype of Arabidopsis (Arabidopsis thaliana) and tobacco (Nicotiana tabacum L.) were used in this study. Dongnong 690 plants was sown in 20-cm pots filled with a 1:1 mixture of vermiculite and nutrient-rich soil that were placed in the greenhouse. Arabidopsis seeds were sown in small pots containing vermiculite with sufficient water; the pots were covered with plastic wrap and placed into an incubator. Photoperiod and temperature were 12 h of light at 22°C and 12 h of darkness at 20°C. Seeds of tobacco and transgenic lines were sterilized with 10% NaClO for 10 min, then washed more than three times with double distilled water. After vernalization at 4°C in the dark for 2 days, the seeds were sown onto plates with 1/2 MS medium at pH 5.8. The plates were incubated at 22°C under a 12-h photoperiod. All the seedlings were then used for further analysis.

Bioinformatics analyses were performed by various web-based software programs. The amino acid sequences of the soybean bHLH transcription factors including GmORG3 were downloaded from the Phytozome database² and aligned using Clustal Omega web software³. Other bioinformatics software programs utilized included MEGA 5 and PowerPoint 10 for general sequence homology analysis, drawing of figures, domain analyses and the prediction of similarity and identity among different members of the subfamilies.

Total RNA was isolated from the tissues of roots, stems, leaves, flowers, and pods of Dongnong 690 at the first-trifoliate, full-bloom and full-pod stages using an SV total RNA kit (Promega, United States). RNA samples were quantified using a Nanodrop spectrophotometer (United States). cDNA was synthesized from 5 μg of total RNA with a SuperScript RT III first-strand cDNA synthesis kit (Invitrogen, San Diego, CA, United States) according to the manufacturer’s instructions. The synthesized cDNA was used as PCR template. GmActin was used as an internal control for semi-quantitative RT-PCR analysis. The GmORG3-specific primer pair, ORG3-F (5′-CGAGATCATCATAGCATTATCT-3′) and ORG3-R (5′-CTCCATTCTAATTTCCGAACT-3′), was utilized for expression analysis (semi-qPCR). The GmActin (accession number: NM_001250673) primer pair, GmActin-F (5′-CGGTGGTTCTATCTTGGCATC-3′) and GmActin-R (5′-GTCTTTCGCTTCAATAACCCTA-3′) was used as an internal control. PCR was carried out in 1x PCR buffer supplemented with dNTPs at 150 μM, 1.5 U of Taq DNA polymerase and 5 pmol of each gene-specific primer according to the following program: initial denaturation at 95°C for 5 min; 27 cycles of 94°C for 30 s, 55°C for 30 s and 72°C for 30 s; and a final extension step at 72°C for 5 min. Semi-quantitative RT-PCR experiments were repeated three times. The amplification products were separated via agarose gel electrophoresis and analyzed using a fluorescence imaging system (Peiqing CN).

Total RNA from the roots of 3-week-old soybean seedlings grown under control and 50 μM CdCl₂ stress conditions for 12 h was extracted, reverse-transcribed and used for real-time quantitative PCR analysis. Sample preparation and PCR analysis were performed using the SYBR Premix Ex Taq in a Roche Light Cycler 2.0 with Light Cycler software (build 4.1.1.21) (LightCycler Carousel-based system, F. Hoffmann-La Roche Ltd, Germany). A 10-μl reaction consisted of 5 μl of SYBR Premix Ex Taq, 0.8 μl of each primer (forward and reverse, 10 mM), 1 μl of template and 2.4 μl of ddH₂O. The GmORG3-specific primer pair, qORG-F (5′-TCGTTCACTTCTTCCTGGGC-3′) and qORG-R (5′-ATAGTGGAGCCTTGTGAGCC-3′), was utilized for its expression study under control and Cd stress conditions. Forward (5′-CGGTGGTTCTATCTTGGCATC-3′) and reverse (5′-GTCTTTCGCTTCAATAACCCTA-3′) primers of the soybean GmActin gene were used as an internal control for the expression analysis of GmORG3. An equal amount of cDNA template was used for each sample reaction, including the internal control. qPCR amplification conditions included an initial denaturation for 30 s at 95°C followed by 35 quantification cycles consisting of denaturation for 5 s at 94°C, annealing for 10 s at 58°C, and extension for 30 s at 72°C. A melting curve analysis was then performed to confirm the specificity of the PCR products. The real-time quantitative PCR analysis was repeated in three independent experiments.

An Agrobacterium rhizogenes-mediated soybean genetic transformation system was used to introduce the GmORG3 overexpression cassette into the hairy roots of soybean variety Dongnong 690. Soybean seeds were sown into pots containing vermiculite and nutrient-rich soil in a 1:1 ratio. Upon the emergence of cotyledons, healthy and robust seedlings were selected for transformation. Agrobacteria carrying the GmORG3 overexpression cassette/plant expression vector pCXSN-GmOGR3 were streaked onto an LB plate containing kanamycin and precultured at 28°C for 2 days. A single colony was used to inoculate liquid LB medium that contained kanamycin as a selectable marker. The culture was grown at 28°C for 12 h with shaking at 200 rpm. The culture was centrifuged at 4500 rpm for two min at room temperature to harvest the bacteria. The pellet was twice suspended gently in a 10 mM MgCl₂ solution. The OD₆₀₀ of the final bacterial suspension was adjusted to 0.4–0.6. Young seedlings with unfolded cotyledons were infected at the cotyledonary node and/or the hypocotyl with Agrobacterium rhizogenes carrying the pCXSN-GmORG3 construct. The injection point was covered immediately after inoculation with a wet vermiculite/soil mixture. The infection sites were kept in a high-humidity environment. Water was applied as required. The seedlings were grown in a greenhouse at 25°C with a 12-h photoperiod and approximately 60% relative humidity. Hairy roots were induced at the injection point after 1 week; the roots grew to 2–3 cm in length 1 week after inoculation. As such, the hairy roots developed 3 weeks after sowing. The main roots were removed when the emerged hairy roots could support the plants. Mosaic seedlings with one hairy root were transferred to 100-ml glass tubes containing ½-strength Hoagland solution. The glass tubes containing mosaic soybean seedlings were placed in a growth chamber with 25/22°C day/night temperature and a 12-h photoperiod for approximately 1 week. The Hoagland solution in the glass tubes was replaced every other day. Once the hairy roots reached bottom of the tube, the transformation of hairy roots was verified using PCR analysis. PCR was performed using a CaMV35S promoter-specific forward primer (5′-CAATCCCACTATCCTTCGCAAGACC-3′) and a GmORG3 gene-specific reverse primer. Genomic DNA from the hairy roots was used as a template. One set of soybean mosaic seedlings carrying a pCXSN empty vector was prepared for use as a control. Transgenic hairy roots of mosaic seedlings were immersed in ½-strength Hoagland solution containing 20 μM CdCl₂ and grown for 9 days in a growth chamber. The Hoagland solution containing Cd was replaced every other day. At 0, 1, 2, 4, 6, and 9 days, roots and shoots tissues were harvested for analysis of Cd and Fe content. The experiment was repeated three times.

For the heterologous expression of the GmORG3 in tobacco, A. tumefaciens strain EHA105 containing the pCXSN:2xCaMV35S::GmORG3-NOS construct was used to get transgenic tobacco plant according to a reported protocol, with a little modification. The hygromycin were used as selection maker. The positive transformed tobacco seedlings were grown in a contained net house. When the plantlets were large enough, DNA was extracted from the leaves, and positive plantlets were screened using PCR. T1 seeds were then harvested, and selection continued until the T3-generation seeds were collected and cultured for further study.

Arabidopsis protoplasts were isolated as previously described (Yoo et al., 2007). The CDS of GmORG3 was inserted into the pJIT166-GFP vector without a termination codon to create an in-frame fusion between GmORG3 and the green fluorescent protein. The fusion construct (p35S::GmORG3-GFP) and a control construct (p35S::GFP) were introduced into Arabidopsis protoplasts using the PEG4000 method as described by Abel and Theologis (1994). After the incubation of transformed Arabidopsis protoplasts in culture solution for 18–24 h without light at room temperature, imaging was performed using a confocal microscope (Zeiss, LSM510 Meta, Carl Zeiss AG) and a 40X lens. The fluorescence of GFP was measured at 495–545 nm, and chloroplast autofluorescence was measured at 480–685 nm.

Regarding the experiments to determine the function of GmORG3 under Cd toxicity stress, mosaic soybean seedlings were generated following the procedure described in the section “Development and treatment of GmORG3 mosaic soybean seedlings” above. Root and shoot tissue was harvested, and the Cd and Fe contents of the samples were then determined. Samples were kept at 105°C for 20 min and dried at 80°C to constant weight. Dry tissues were ground into a powder, and 1 ml of 100 mM acetic acid was added. Samples were incubated at 90°C for 6 h. Cadmium and iron contents were then determined using an atomic absorption spectrometer with a continuous light source (ContrAA^®300, Analytik Jena AG). The analysis was repeated three times.

Tobacco seeds of the GmORG3 transgenic lines and wild-type plants were grown on square plastic Petri dishes filled with 1/2 MS medium with 20 μM CdCl₂, without Fe, and 1/2 MS medium supplemented with 20 μM Cd but without Fe. The dishes were placed into a growth chamber. Three weeks later, the seedlings were imaged, chlorophyll contents were determined, root lengths were measured and samples were harvested. The estimation of chlorophyll from seedlings was performed in accordance with the method of Arnon (1949): 100 mg of tissue was homogenized thoroughly in 1 ml of 80% acetone and centrifuged at 3000 rpm for 2–3 min. The supernatant was retained, and absorbance was measured using a spectrophotometer at 663 and 645 nm. Total chlorophyll was estimated using the following equation: total chlorophyll = [(20.2 × A₆₄₅ + 8.02 × A₆₆₃)/(1000 × W)] × V. The experiment was repeated three times.