Isolation and characterization of GmMYBJ3, an R2R3-MYB transcription factor that affects isoflavonoids biosynthesis in soybean

Isoflavonoids are secondary metabolites that play a variety of roles in plant-microbe interactions and plant defenses against abiotic stresses. Here we report a new MYB transcription factor (TF) gene, GmMYBJ3, that is involved in the isoflavonoids biosynthesis. The GmMYBJ3 gene is 1,002 bp long and encodes a protein of 333 amino acids. Amino acid sequence analysis showed that GmMYBJ3 is a typical R2R3 MYB TF. Yeast expression experiment demonstrated that GmMYBJ3 has its transcription activity in the nucleus and is transiently expressed in onion epidermal cells. The GmMYBJ3 gene was transformed into soybean and the expression activity of the GmMYBJ3 gene was significantly positively correlated with total isoflavonoid accumulation in soybean. Transient expression assays indicated that GmMYBJ3 can activate CHS8 expression. Furthermore, we analyzed the expressions of several genes known involved in the isoflavonoid biosynthesis, including CHS8, CHI1A, PAL1, IFS2 and F3H, in the GmMYBJ3 transgenic plants. The results showed that the expression levels of CHS8 and CHI1A were significantly increased in the transgenic plants compared to wild-type plants, but those of PAL1, IFS2 and F3H remained similar between the transgenic and wild-type plants. These results suggest that GmMYBJ3 participates in the isoflavonoid biosynthesis through regulation of CHS8 and CHI1A in soybean.


Introduction
Isoflavonoids are a group of plant natural secondary metabolites. Isoflavonoids are mainly found in the Papilionideae family, and are abundant in soybean, Glycine max (L.) Merr., and other legume species. They play a variety of roles in plant-microbe interactions, such as functioning as signal molecules for symbiosis between soybean and Bradyrhizobium japonica [1,2]. They also serve as phytoalexins in response to pathogen attacks [3,4] and abiotic stresses such as UV irradiation and drought [5][6][7]. Interest in these compounds has recently increased PLOS  because they are associated with important preventive and therapeutic medicinal properties [8,9]. Isoflavonoids are synthesized through a branch of the phenylpropanoid pathway present in legumes. Chalcone establishes the first step in the branched pathway for the synthesis of flavonoids and isoflavonoids [10] and is generated in a reaction catalyzed by chalcone synthase (CHS). CHS is encoded by a single gene in Arabidopsis thaliana, but there are multiple copies in other plants, including petunia and soybean [11,12]. The soybean genome contains a gene family consisting of nine CHS gene members, designated from CHS1 to CHS9, with CHS1 having a duplicate [13]. Comparative gene expression analysis between soybean cultivars with contrasting seed isoflavonoid contents revealed a critical role of the CHS7 and CHS8 genes in isoflavonoid biosynthesis [14]. RNAi silencing of the CHS8 gene reduced the level of isoflavonoids in soybean hairy roots, providing a line of evidence that CHS8 is involved in regulation of isoflavonoid biosynthesis [15]. In the isoflavonoid biosynthesis, isoflavonoid synthase (IFS) is also necessary. The first key step of isoflavonoid biosynthesis is liquiritigenin and naringenin conversion to daidzein or genistein, which is catalyzed by IFS. Two IFS genes were identified in soybean, IFS1 and IFS2, and there are 14 amino acids that differ between their protein products. The expression of IFS1 in A. thaliana can induce the production of the isoflavone genistein in this non-legume plant [16].
The phenylpropanoid pathways are predominantly regulated at the transcriptional level by members of transcription factors, especially the MYB TFs [17][18][19]. Genetic markers linked to controlling corresponding quantitative traits may be used to select for favorable alleles effectively, and was verified in some agronomic traits [20,21]. Therefore, we focus in this study on the MYB TFs that are mapped in the soybean isoflavonoid QTL. Even though interaction exists between MYB TFs and other factors to regulate several branches of the phenylpropanoid pathways, MYB proteins are believed to be the key components [18]. For example, GmMYB176 affects isoflavonoid biosynthesis in soybean hairy roots through interaction with 14-3-3 protein in the nucleus [22]. The genes required for the synthesis of flavonols and anthocyanins are induced by maize MYB C1 in tomato, except for F3 0 H, F3 0 5 0 H and CHI [23]. Introduction of CRC (a chimeric transcription factor of maize C1 and R) into soybean resulted in a significant increase in seed isoflavonoid levels [24]. GmMYB39 down regulates the isoflavonoid biosynthesis in soybean [25]. MYB134 regulates proanthocyanidin synthesis in poplar [26]. Overexpression of AtMYB12 in Arabidopsis displays a flavonol accumulation in transgenic lines [27].
In this study, we compared the isoflavonoid QTLs previously mapped [28][29][30] and found a common isoflavonoid QTL mapped with different populations in different environments. Only one R2R3 MYB TF, GmMYBJ3, was found in this common QTL, suggesting that it is likely a candidate gene for isoflavonoid biosynthesis. We associated the expression activity of GmMYBJ3 with the isoflavonoid content in soybean. Subcellular localization assays in onion epidermal cells indicated that GmMYBJ3 encodes a nucleus-localized protein and yeast hybrid analysis showed that it has transcriptional activity. Transient expression assays showed that GmMYBJ3 could activate CHS8 expression. Overexpression of GmMYBJ3 in the soybean transgenic plants up-regulates the expression level of CHS8 and CHI1A and increases the total isoflavonoid content.

Isolation and sequence analysis of GmMYBJ3
Two adjacent QTLs, GLY1 and qGm06, were previously mapped to Chromosome 6 that control soybean isoflavonoid content [31,32]. GLY1 were mapped to a region closely linked with SNP marker BARC-031337-07051 using a population consisting of 274 F5:8 RILs (recombining inbred lines) derived from Essex × Williams 82 and grown at three locations (Knoxville, TN; Harrisburg, IL; and Stuttgart, AR). The SNP of BARC-031337-07051 was physically mapped to the 16679945 position of the chromosome 6 (http://soybase.org/). The QTL qGm06 was mapped to an interval between BARC-063661-18416 and BARC-066175-19800 using a population consisting of 188 F7 RILs derived from Magellan × PI 437654 and grown at two locations (the University of Missouri Bradford Research and Extension Center, and the University of Missouri Delta Research Center). The SNP of BARC-063661-18416 was physically mapped to the 17230167 position and that of BARC-066175-19800 to the 18736894 position of Chromosome 6 (http://soybase.org/). Therefore, the region between BARC-031337-07051 and BARC-066175-19800, physically spanning from the 16679945 position to the 18736894 position, contains at least one of the two QTLs controlling soybean isoflavonoid content. Therefore, we searched this region carefully and found only one MYB transcription factor (TF) and the gene ID of this MYB TF is Glyma.06g193600 (Version: Glyma2.0). We designated this gene as GmMYBJ3. GmMYBJ3 has a full-length ORF of 1002 bp and was predicted to encode a protein of 333 amino acid residues with a calculated mass of 37.7 kDa and a pI of 5.81. The CDD v3.15 (https://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi) was used to search for the conserved domains located at amino termini and responsible for binding to DNA sequence. As shown in Fig 1, there are two MYB imperfect repeats, R2 and R3, which consist of 50 (13-63) and 48 (66-114) amino acids, respectively. Phylogenetic analysis (Fig 2) showed that GmMYBJ3 clustered with Arobidopsis AtMYB60 that regulates anthocyanin biosynthesis in lettuce [33]. Both anthocyanin and isoflavonoid are products of the phenylpropanoid metabolic pathway. Because GmMYBJ3 was found from the soybean isoflavonoid QTL region, it is homologous to AtMYB60, and its R2R3 participates in regulating the phenylpropanoid pathway, we hypothesized that GmMYBJ3 is involved in the transcriptional regulation of isoflavonoid biosynthesis.

Subcellular localization of the GmMYBJ3 protein
The GmMYBJ3 protein was localized in the nucleus by online prediction (http://www.csbio. sjtu.edu.cn/bioinf/plant/). To further confirm this localization result of the GmMYBJ3 protein, we deleted its stop codon and fused its ORF to the N terminus of GFP that is under control of the CaMV 35S promoter. The pBI121 constructs containing GmMYBJ3:GFP and containing GFP alone were transformed into onion epidermal cells, respectively, through Agrobacterium EHA105-mediated transient expression. The results showed that the GmMYBJ3:GFP fusion protein resided in the nucleus of the cells, which was consistent with the online predicted result, but the control construct containing GFP alone was visualized throughout the cells (Fig 3).

Transcriptional activation of GmMYBJ3 in vivo
The transcriptional activity of GmMYBJ3 was tested using the pGBKT7 vector that expresses proteins fused to the GAL4 DNA binding domain from the constitutive ADH1 promoter in the yeast GAL4 system (PT3248-5, Clontech, USA). The transcriptional activator protein would activate the HIS and lacZ reporter genes, thus allowing yeast to survive on the histidine-deficient medium and showing color in the β-galactosidase assay. The fusion plasmid pGBKT7-GmMYBJ3 and the empty vector pGBKT7 were transformed into yeast AH109 cells, respectively, and screened on the solid medium SD/-Trp so that the positive transformants will survive (Fig 4A and 4B). To further characterize the positive transformants, we streaked the transformants on the solid medium SD/-Trp/-His/-Ade plus 3-AT. The fusion plasmid cells survived on the SD/-Trp/-His/-Ade plus 3-AT plate, but the pGBKT7 control plasmid cells did not. This result indicated that the fusion effectors were expressed and able to activate the expression of the HIS reporter gene (Fig 4C). For the 5-bromo-4-chloro-3-indolyl β-D-galactopyranoside activation assay, strong blue signals were observed for the pGBKT7-GmMYBJ3 cells, reflecting that the lacZ reporter gene was activated, whereas no colorful signal was detected for the pGBKT7 cells, indicating that no transcriptional activation occurred ( Fig 4D).

GmMYBJ3 activates the transcription of CHS8 in vivo
To find whether GmMYBJ3 activates CHS8 expression in soybean, we performed a transient expression experiment using tobacco leaves. A CHS8pro:GUS fusion reporter construct, an effector plasmid and a control plasmid were constructed ( Fig 5A). The GUS activity of the 35Spro:GUS construct without the effector was higher than the GUS activity of the transformed CHS8pro:GUS construct in tobacco leaves. However, the co-transfection experiment with the effector and CHS8pro:GUS reporter constructs showed that the GUS activity was 1.26-fold higher than that of the 35Spro:GUS control plasmid and was 1.59-fold higher than that of the CHS8pro:GUS reporter plasmid. The GUS activity was significantly different between the CHS8pro:GUS construct and the CHS8pro:GUS construct co-transfected with the effector plasmid. From the results of the transient expression assay (Fig 5B), we noted that GmMYBJ3 highly likely trans-activated CHS8pro:GUS expression.

GmMYBJ3 is expressed in the tissues where isoflavonoids are synthesized in soybean
To confirm that GmMYBJ3 expresses in tissues where isoflavonoids accumulate, we investigated the expression of GmMYBJ3 in various tissues of soybean ( Fig 6A) by quantitative RT-PCR and analyzed the correlation between its expression level and isoflavonoids content. The results showed that GmMYBJ3 expressed in all the tissues studied, but its expression level increased as embryos developed. Fig 6B shows the contents of isoflavonoids in tissues. Correlatioin analysis showed that the GmMYBJ3 expression level positively correlated with the total isoflavonoids accumulation in soybean (Pearson's r = 0.674, P 0.05) (S1 Table). These results indicated that the expression activity of GmMYBJ3 was associated with the synthesis of isoflavonoids.

Overexpression of GmMYBJ3 enhances isoflavonoid content in transgenic soybean
To demonstrate the function of GmMYBJ3 in isoflavonoid biosynthesis, the overexpression of GmMYBJ3 was performed using cv. Jilin 35 as the host plant through the Agrobacterium- https://doi.org/10.1371/journal.pone.0179990.g002 mediated transformation of soybean embryonic tips (unpublished). The T 0 transgenic plants were selected on the medium containing Basta. Southern blots were prepared from the leaves of independent T 2 plants and hybridized to confirm the transgenic plants. The results showed that GmMYBJ3 was transformed into Jilin 35 cultivar (Fig 7A). Quantitative RT-PCR with the positive T 2 transgenic plants showed that the expression level of GmMYBJ3 was significantly increased in the transgenic plants compared to the WT control ( Fig 7B) and the isoflavonoids content was measured by HPLC ( Fig 7C). HPLC analysis showed that the glycitin, genistin and total isoflavonoid contents were significantly increased in the positive transgenic plants compared to the WT control, but the contents of daidzin, genistein and daidzein largely remained unchanged.

Expression of the genes known to be involved in the isoflavonoid biosynthesis in the seeds of transgenic plants
The enzymes, PAL1 (phenylalanine ammonia-lyase), CHS8 (chalcone synthase), IFS2 (isoflavone synthase) and CHI1A (chalcone isomerase), are known to catalyze the biosynthesis of isoflavonoids in soybean seeds [14,15,35], and F3H (flavanone-3-hydroxylase) is a competitive enzyme for anthocyanin synthesis in a branched fashion. Therefore, we further analyzed the expressions of the genes encoding these enzymes in the T 2 transgenic plants to infer the role of GmMYBJ3 in the isoflavonoids biosynthesis in soybean. The results showed that the CHS8 and CHI1A genes were significantly up-regulated in the seeds of the T 2 transgenic plants. However, the expression levels of PAL1, IFS2 and F3H underwent little changes compared to the WT plants (Fig 8). These changes suggested that GmMYBJ3 may be a regulator of the isoflavonoid biosynthesis pathway in soybean seeds.  The data are presented as the averages of three independent assays ± SD. ** represents significant difference at P 0.01 between the CHS8pro:GUS construct and the CHS8pro:GUS construct co-transfected with the effector plasmid. +,-represent whether the reporter and/or effecter plasmids were co-transfected or not.

Discussion
The MYB TFs, especially the group of R2R3-MYB, are known to be involved in the regulation of the phenylpropanoid metabolic pathway [36]. In this study, we have identified a MYB transcription factor from soybean, designated as GmMYBJ3, and characterized it in its functions in the phenylpropanoid metabolic pathway. Analysis of the GmMYBJ3 gene indicates that it is a typical R2R3-MYB with conserved R2 and R3 motif, and gene phylogenetic analysis shows that it clusters with GmMYBJ2, and is also homologous to AtMYB60. Overexpression of GmMYBJ2 enhances stress tolerance in Arabidopsis thaliana [37]. AtMYB60 regulates stomatal movements, plant drought tolerance [38] and anthocyanin biosynthesis in lettuce [33]. Pleiotropy is very common in MYB TF [39], and GmMYBJ3 may play different roles in different physiological and biochemical processes. Because GmMYBJ3 was identified from a genomic region containing a soybean isoflavonoid QTL, we speculate that GmMYBJ3 may be involved in regulating the isoflavonoid biosynthesis. The sub-localization of GmMYBJ3 reveals that it is a nuclear localization protein, which is in agreement with its role as a transcription factor. Most R2R3 MYB TFs are presumed to be transcriptional activators with activation domains (ADs) in the C-terminal region [40], but it is known that the N-termini of MYB TFs are conserved. One yeast hybrid analysis has confirmed that the GmMYBJ3 gene has a transcription activation domain, even though further tests in the plant system remain.
Analysis of gene expressions during soybean embryo development revealed the crucial roles of the CHS7 and CHS8 genes in isoflavonoid biosynthesis [14]. Further studies of the CHS7 and CHS8 promoters have indicated that CHS8 is involved in isoflavonoid biosynthesis during seed development [15]. In the transient expression experiments of this study using tobacco leaves, the co-transfection of the CHS8 gene with the effector construct demonstrated that the CHS8pro:GUS construct was activated by GmMYBJ3, and the GUS activity of the co-transfectant was 1.59-fold higher than that of the CHS8pro:GUS construct. In previous studies, GmMYB176 was found to trans-activate the CHS8pro:GUS expression by 4.7-fold in Arabidopsis leaf protoplasts compared to the CHS8pro:GUS and the level of endogenous CHS8 transcripts was increased by 169-fold in soybean protoplast within 48 h [41]; and GmMYB12B2 trans-activated CHS8pro:GUS expression by 2-fold in soybean calli [42]. In this study, although GmMYBJ3 activateed a lower expression level of CHS8 than GmMYB176 and GmMYB12B2, the co-transfection results indicate that CHS8 is one of the target genes for GmMYBJ3.
This study has also shown that GmMYBJ3 expresses in the tissues where isoflavonoids are synthesized in soybean. The expression of GmMYBJ3 is most active in embryos and increases gradually as they develop, even it also expresses in other tissues including root, stem, leaf and floret. The correlation between GmMYBJ3 expression level and total isoflavonoids accumulation further demonstrates that GmMYBJ3 is involved in the isoflavonoid biosynthesis in soybean. Moreover, the roles of GmMYBJ3 in regulating isoflavonoid biosynthesis have been further confirmed by the overexpression analysis in the transgenic plants. The results of significantly increased total seed isoflavonoid content and individual isoflavones, glycitin and genistin, but not for daidzin, genistein and daizein in the transgenic plants compared to the wild RT-PCR analysis showing the expression level of GmMYBJ3 in the seeds of the positive T 2 transgenic plants compared to the wild type, Jilin 35 (control). Data represent the mean ± SD of three independent seeds, with each plant having three technical replicates. * and ** represent significant differences at P 0.05 and 0.01 between the transgenic lines and the WT. (C) Isoflavonoid content in the seeds of the positive T 2 transgenic compared to the WT control. The data of wild type (control) represent the mean ± SD of three independent plants, with each plant having three technical replicates. The data for each of the independent positive transgenic lines are the means of three technical replicates, and the error bars show SEM. * and ** represent significant difference at P 0.05 and 0.01 between the transgenic lines and the WT.
https://doi.org/10.1371/journal.pone.0179990.g007 type provide a strong line of evidence in supporting the roles of GmMYBJ3 in isoflavonoid biosynthesis. Because levels of total seed isoflavonoids were increased only by 1.5-fold in the soybean transgenic plants, which is only about half of those in the soybean seeds of the maize chimeric CRC gene transgenic plants [24], the function of GmMYBJ3 is likely weaker than the maize CRC in regulating the phenylpropanoid pathway. Furthermore, C1 is a R2R3 MYB TF involved in anthocyanin accumulation in maize and must interact with R (bHLH TF) to be effective [43], whereas AtMYB11, AtMYB12 and AtMYB111 all can regulate flavonol synthesis by themselves [27,44]. Further study is required to determine whether GmMYBJ3 needs to interact with other protein factors to enhance its function.
Previous study showed that overexpression of LjMYB14 up-regulated expression of PAL, C4H, 4CL, IFS and IFR genes in Lotus [34], and silencing GmMYB176 down-regulated the expression level of CHS8 and reduced the isoflavonoid content in soybean hairy roots [41]. We speculate that GmMYBJ3 may affect soybean isoflavonoid biosynthesis through regulation of the structural genes for enzymes in the biosynthesis pathway. Of the twelve members in the  genes, PAL1, CHS8, IFS2, CHI1A and F3H, involved in the isoflavonoids synthesis in the seeds of positive T 2 transgenic lines. WT, wild type; OE, independent GmMYBJ3 overexpression lines. The expression data of transgenic and wild-type lines represent the mean ± SD of three independent plants, with each plant having three technical replicates. * and ** represent significant differences at P 0.05 and 0.01, respectively, between the transgenic lines and WT.
https://doi.org/10.1371/journal.pone.0179990.g008 CHI family, CHI1A, PAL1 and IFS2 are involved in seed isoflavonoid biosynthesis [35], and CHS8 plays a crucial role in isoflavonoid biosynthesis [14,15]. Our analysis of the PAL1, IFS2, CHS8 and CHI1A expression in the seeds of wild-type and transgenic plants indicates that GmMYBJ3 does not have effect on PAL1 and IFS2, but does have effect on CHS8 and CHI1A. This result suggests that GmMYBJ3 is able to activate the expression of CHS8. These results all indicate that GmMYBJ3 participates in isoflavonoid biosynthesis by up-regulating the expression levels of CHS8 and CHI1A in soybean seeds.
In conclusion, we have isolated a novel R2R3 MYB TF, GmMYBJ3, from soybean. GmMYBJ3 encodes a nucleus-localized protein and has a transcription activation domain. GmMYBJ3 activates the expression of CHS8, thus enabling to enhance the total and individual isoflavone contents, especially glycitin and genistin, in soybean seeds by up-regulating the genes encoding the enzymes involved in the isoflavonoid biosynthesis pathway.

Plants materials
Soybean (Glycine max L.) cv. Williams 82 was used for expression analysis of GmMYBJ3 in various soybean tissues. The cultivar was planted in the Crop Experimental Station of Jilin University near Changchun, China. Roots, stems and leaves were sampled between the ternate leaf stage and before flowering; flowers were sampled at the flowering stage; embryos were tagged on the first day of pollination and then collected on the 20th, 30th, 40th and 50th day after pollination and pod wall was separated from the 50th day embryo. All of the tissues were collected randomly from three independent plants, quickly frozen in liquid nitrogen and then stored at −80˚C.
Cultivar Jilin 35 was used as the host plant for genetic transformation and overexpression analysis of GmMYBJ3 in soybean. The T 2 transgenic plants generated and wild-type Jilin 35 were grown in the transgenic controlled field, Jilin University. The tissues for expression analysis of the genes were collected from the transgenic plants as described above.

Isolation and sequence analysis of the GmMYBJ3 gene
Isoflavonoid content QTL analysis allowed identification of a MYB TF (Glyma.06g193600) in the QTL region. Therefore, it was used an inquiry to search the soybean genome data (http:// soybase.org/). A cDNA clone (GenBank accession number: KU664645) that is highly homologous to plant MYB TFs was identified and this cDNA clone was designated as GmMYBJ3. To reproduce the cDNA of the gene, total RNA was extracted from 50 d embryos of Williams 82 and reverse transcription polymerase chain reaction (RT-PCR) was performed using the primer pair, 5 0 -GAGCCCAAAGGGATCAAA -3 0 (forward) and 5 0 -CGACCTCAAGTCCGCTAC -3 0 (reverse) designed based on the cDNA sequence. The RT-PCR reaction contained 18.2 μL H 2 O, 1 μL of the first strand of cDNA, 1 μL (10 μM) of each primer, 2.5 μL of 10x PCR reaction buffer, 2 μL of dNTP mix and 0.3 μL of ExTaq DNA polymerase (Takara, China). The PCR reaction program was 95˚C for 4 min, followed by 30 cycles (94˚C for 40 s, 58˚C for 40 s and 72˚C for 80 s) and 72˚C for 10 min at the final extension. The amplified fragments were cloned into the pMD-18T vector (Takara, China) and sequenced for confirmation. Phylogenetic analysis of the GmMYBJ3 gene with foreign genes was performed using MEGA 7 with the Neighbor-Joining (NJ) algorithm [45].

Subcellular localization assay
The coding region of GmMYBJ3 lacking a stop codon was ligated to the N-terminus of GFP to construct the pBI121-GmMYBJ3-GFP construct under the control of the cauliflower mosaic virus 35S (CaMV 35S) promoter in the pBI121 vector. This construct was then introduced into the Agrobacterium tumefaciens Strain EHA105 and used for subcellular localization analysis with transient transformation in onion epidermal cells. The Agrobacterium culture was prepared as described by Yang et al. [46]. The onion cells were dipped in prepared Agrobacterium solution for 30 min and then plated on Murashige and Skoog (MS) medium and incubated at 25˚C in the dark for 16-48 h. Hoechst No.33342 was used to stain the nuclei of the onion cells and the transformed onion cells were observed using a confocal microscope (Olympus, Japan).

Transactivation analysis in the yeast GAL4 system
The GmMYBJ3 coding sequence was cloned into the GAL4 DNA-BD binding domain in the pGBKT7 vector. The transactivation assay (PT3024-1) was performed as described by its manufacturer (Clontech, USA). The pGBKT7-GmMYBJ3 plasmid was transformed into the yeast strain AH109 by the lithium acetate-mediated method [47], and the transformants were screened on SD/-Trp at 28˚C for 2 d. Then, the transformants from SD/-Trp were streaked onto the solid medium SD/-Trp/-His/-Ade plus 3-AT. The plates were incubated for 3 d until they were used for a β-galactosidase assay. For the β-galactosidase assay, the transformant cells were imprinted onto Whatman filter paper and lysed by freezing with liquid nitrogen for 10 s and then thawing at room temperature. This process was repeated 3 times to completely lyse the cells. The filter was incubated in 2.5 ml of Z buffer containing 0.8 mg of 5-bromo-4-chloro-3-indolyl β-D-galactopyranoside supplemented with 21.51 g L -1 Na 2 HPO 4 Á12 H 2 O, 6.22 g L -1 NaH 2 PO 4 ÁH 2 O, 0.75 g L -1 KCl, and 2.46 g L -1 MgSO 4 Á7H 2 O at 30˚C for 30 min to 8 h. The color reaction was monitored.

Transient expression
To construct the reporter plasmid, the 1577-bp CHS8 promoter (F: 5'-AGCTGAGCAAGTAT ACCAACC-3'; R: 5'-GAGGTTGAAATGAAGGTGTGC-3') was amplified by PCR and cloned into the pCAMBIA1301 plasmid to replace the CaMV 35S promoter. To construct the effector plasmid, the full coding region of GmMYBJ3 was cloned into the pCB35SR1R2-GFP plasmid that was constructed as below and under the control of the CaMV 35S promoter. The reporter and the effector plasmids were introduced into the Agrobacterium EHA105 strain for further analysis. Agrobacterium-mediated transient expression was performed on the leaves of 4-week-old tobacco seedlings (Nicotiana tabacum cv. NC89) [46]. GUS activity analysis was performed after approximately 48 h [48]. All transfection experiments and GUS activity analysis were performed at least three times.

Quantitative real-time RT-PCR analysis
Total RNA was extracted from the soybean tissues sampled above using an RNAprep pure Plant Kit (Tiangen, China) and cDNA was synthesized using MMLV reverse transcriptase (Takara, China). Quantitative real-time RT-PCR analysis was performed with Applied Biosystems 7500 (Applied Biosystems, USA) using SYBR Premix Ex Taq (Takara, China). Each reaction contained 10 μL of SYBR Green I, 2 μL of cDNA samples, 0.4 μL of ROX Reference Dye II and 0.4 μL of 10 μM gene-specific primers for a final volume of 20 μL. The reaction was performed at 95˚C for 30 s, followed by 40 cycles of 95˚C for 5 s, 60˚C for 34 s and 72˚C for 30 s. Three biological replicates, with three technical replicates per biological replicate, were applied for the experiment. The soybean β-tubulin gene (GenBank accession No. GMU12286) was used as an internal control. The gene-specific primers used for real-time quantitative RT-PCR and their accession numbers are listed in Table 1. The data were analyzed using the comparative C T method based on C T values [49].

Isoflavonoid analysis by high-performance liquid chromatography (HPLC)
Approximately 0.5 g of dry soybean seeds was ground, mixed with 10 mL of 80% methanol in distilled water followed by sonication for 20 min, and subsequently incubated at 80˚C for 14 hours. The mixture was filtered through a 0.45-μm filter and transferred to 5-mL HPLC vials. Aliquots of this filtrate (20 μL) were utilized for HPLC analysis.

Plasmid construction and plant transformation
The GATEWAY cloning technology was used to construct the plant expression vector [50]. First, the GmMYBJ3 coding region was cloned into the entry vector pDONR221 (Invitrogen, USA) via a BP recombinant reaction. Second, the fragment in the entry vector was transferred to the destination vector, pCB35SR1R2-GFP, based on the LR recombinant reaction. The pCB35SR1R2-GFP-GmMYBJ3 construct was heat shocked into A. tumefaciens EHA105 and then transformed into embryonic tips of soybean cv. Jilin 35. The transformed soybean embryonic tips were grown at 25˚C/22˚C with a light/dark cycle of 16 h light and 8 h darkness and a relative humidity of 60%. Transgenic T 0 and T 1 plants were cultured in a phytotron at 28˚C/ 22˚C with a light/dark cycle of 16 h light and 8 h darkness and a relative humidity of 60%.

Southern blot analysis
Southern blot analysis was performed to detect transgenic plants. DNA was extracted from the young leaves of T 2 transgenic and wild-type plants (Jilin 35) using the CTAB method [51]. Approximately 10 μg of genomic DNA was digested with Hind III (Takara, China). The digests were separated on a 0.8% agarose gel and blotted onto a Hybond-N+ nylon membrane (Roche Applied Science, Germany). The 501-bp PCR fragments (amplified with primer 5-ACCCACG TCATGCCAGTT-3 and 5'-CTAGGGGGATCTACCATG-3') containing the bar gene (phosphinothricin acetyltransferase gene) coding region was labeled with Digoxigenin (DIG)-high prime and used as a probe for hybridization. Probe labeling, prehybridization, hybridization, membrane washing and signal detection were performed according to the manufacturer's instruction (Roche Applied Science).

Statistical analysis
Statistical analysis was completed using the SPSS 21.0 program. Differences were defined at the two-tailed significance level of P 0.05.
Supporting information S1 Table. The correlation between the GmMYBJ3 expression level and total isoflavonoid content. Ã represent significant (P 0.05) correlation at the level of 2-tailed. (DOCX)