Soybean cyclophilin GmCYP1 interacts with an isoflavonoid regulator GmMYB176

Cyclophilins (CYPs) belong to the immunophilin superfamily with peptidyl-prolyl cis-trans isomerase (PPIase) activity. They catalyze the interconversion of the cis- and trans-rotamers of the peptidyl-prolyl amide bond of peptides. A yeast-two-hybrid screening using the isoflavonoid regulator GmMYB176 as bait identified GmCYP1 as one of the interacting proteins in soybean embryos. GmCYP1 localizes both in the nucleus and cytoplasm, and interacts in planta with GmMYB176, in the nucleus, and with SGF14l (a soybean 14-3-3 protein) in the nucleus and the cytoplasm. GmCYP1 contains a single cyclophilin-like domain and displays a high sequence identity with other plant CYPs that are known to have stress-specific function. Tissue-specific expression of GmCYP1 revealed higher expression in developing seeds compared to other vegetative tissues, suggesting their seed-specific role. Furthermore, GmCYP1 transcript level was reduced in response to stress. Since isoflavonoids are involved in plant stress resistance against biotic and abiotic factors, the interaction of GmCYP1 with the isoflavonoid regulators GmMYB176 and 14-3-3 protein suggests its role in defense in soybean.

stress and pathogen infection 26 . Ectopic expression of Thellungiella halophile CYP, ThCYP1, in fission yeast and tobacco cells increased salt tolerance 27 . Transgenic Arabidopsis plants overexpressing pigeon pea CYP (CcCYP1) showed enhanced PPIase activity under stressed conditions, which correlated with their increased tolerance against drought, salinity and high temperature 28 . Similarly, overexpression of cotton CYP (GhCYP1) in tobacco plants conferred tolerance against salt stress and fire-blight disease 29 . Together, these findings clearly demonstrate a role for plant CYPs in stress tolerance.
Soybean (Glycine max) is a grain legume belonging to the family Fabaceae. Soybean seeds provide a major supply of oil, protein and beneficial plant natural compounds such as isoflavonoids and saponins. The soybean genome contains 88,647 predicted transcripts and 56,044 protein coding loci located on 20 different chromosomes 30 . Previously, we performed a genome-wide analysis of soybean CYPs and identified 62 CYP genes 15 . Among these, GmCYP1 has been shown to act as a "helper" to Phytophthora sojae RXLR effector Avr3b by activating its hydrolase activity in plant cells 31 . The protein-protein interaction between GmCYP1 and Avr3b was shown to be isoform-specific since GmCYP1 paralogs failed to interact with Avr3b. Here we present a molecular characterization of GmCYP1 covering its sequence analysis, phylogeny, temporal and spatial expression, subcellular localization, and provide the evidence for its possible role in isoflavonoid biosynthesis and stress response in soybean.

Results and Discussion
Isolation, sequence analysis and phylogeny of GmCYP1. GmCYP1 was identified in our Y2H screening as a protein that demonstrated protein-protein interaction with the isoflavonoid regulator GmMYB176. The Y2H assay was performed to identify GmMYB176-interacting proteins using GmMYB176 as the bait protein and proteins from soybean embryos (50-60 days after pollination) as prey. Of the several hundred yeast colonies screened, 6.5% contained a sequence corresponding to GmCYP1 (accession #AF456323, locus Glyma.11G098700). GmCYP1 is predicted to contain only one exon (519 bp), and is located on the long arm (q arm) of chromosome 11, approximately 16 Mb from the centromere. It encodes a single domain protein of 172 amino acid residues with a calculated molecular mass of 18.22 kDa and a pI of 8.69. The cyclophilin-like domain in GmCYP1 is predicted in between the amino acid residues 7 and 169.
In order to find sequences closely related to GmCYP1, a protein-protein BLAST (BLASTP) was performed using GmCYP1 as a query against the NCBI non-redundant protein database. A list of 12 high-scoring and previously characterized CYPs is shown in Table S1. Alignment of the deduced sequence of GmCYP1 with previously characterised CYPs from several different plant species, human, yeast, and two Arabidopsis multi-domain CYPs (AtCYP40 and AtCYP63) revealed two general features (Fig. 1). First, three amino acid residues that critically affect PPIase activity (R 55 , F 60 and H 126 ) 32 are conserved in all CYPs aligned. Second, the tryptophan residue (W 121 ) implicated in substrate cyclosporinA binding 32,33 is present in all of the CYPs studied except in the multi domain CYPs.
A phylogenetic analysis of GmCYP1 and other functionally characterized plant CYPs clustered GmCYP1 close to GhCYP1, CcCYP1 and PvCYP from cotton, pigeon pea and common bean, respectively. Both GhCYP1 and CcCYP1 are known to have stress-specific function (Fig. 2). Overexpression of GhCYP1 in tobacco conferred increased tolerance to biotic and abiotic stress 29 . Similarly, Arabidopsis plants overexpressing CcCYP1 showed higher PPIase activity during stress and increased tolerance against multiple abiotic stresses as compared to control 28 . Differential accumulation of PvCYP transcripts in response to various external stimuli suggested that it may possess a stress-related function 24 . The high amino acid sequence identity of GmCYP1 with GhCYP1 (96%), CcCYP1 (81%) and PvCYP (96%), suggests similar possible functions of GmCYP1 in stress response. GmCYP1 localizes in the nucleus and the cytoplasm. To further study the subcellular localization of GmCYP1, a translation fusion of GmCYP1 with YFP was created under the control of 35S promoter, and transiently expressed in N. benthamiana leaves. Although there was no predicted nuclear localization sequence in GmCYP1, confocal imaging of the GmCYP1-YFP-infiltrated tobacco leaves showed both nuclear and cytoplasmic localization (Fig. 3a). Nuclear localization of GmCYP1 was confirmed by co-expression of GmCYP1-YFP and NLS-CFP.
Molecules of size smaller than 20-40 kDa, such as ions, water, and small proteins, can pass through the nuclear pore complex by diffusion 34 , whereas movement of larger molecules (70 kDa or higher) entails an active transport system 35 , mediated by transport receptors and signal peptides 36 . The size of GmCYP1-YFP (45.22 kDa) is not considerably larger than the size of molecules that can pass through the nuclear pore complex by diffusion. It is also possible that heterologous protein GmCYP1-YFP, expressed in N. benthamiana is cleaved by the endogenous host proteases and only the cleaved YFP fragments are localized to the nucleus. The confirmation of the YFP signal in the nucleus arising from the intact GmCYP1-YFP and not from the cleaved product of a fusion protein was performed by Western blot analysis (Fig. 3b). Therefore, it is not clear whether GmCYP1-YFP localization in the nucleus was due to passive diffusion or to active transport. Regardless, its nuclear localization indicates a possible role in the regulation of gene expression.
GmCYP1 interacts with GmMYB176 in planta. GmCYP1 was identified as one of the interacting proteins of GmMYB176 in the Y2H assay. Protein chaperones often bind with misfolded bait proteins when a protein is overexpressed or heterologously expressed. To confirm that GmCYP1 is a true GmMYB176-interacting protein, we performed a targeted Y2H assay using GmCYP1 and GmMYB176. However, our targeted Y2H assay failed to verify the interaction between GmCYP1 with GmMYB176. This result led us to hypothesize that there may be an indirect interaction between GmCYP1 and GmMYB176 via involvement of protein (s) that may not be conserved between the species. Therefore, a BiFC assay 37 was carried to further investigate in planta interaction between GmCYP1 and GmMYB176. Translational fusions of GmMYB176 and GmCYP1 were created with N-terminal (YN) or C-terminal (YC) halves of YFP 38 and co-expressed in N. benthamiana leaf epidermal cells in the following combinations: (A) GmCYP1-YN and GmMYB176-YC, (B) GmMYB176-YN and GmCYP1-YC. The negative controls used for the experiment were co-expression of GmCYP1-YN and -YC only or GmCYP1-YC and -YN only. As shown in Fig. 4a, GmCYP1 interacts with GmMYB176 in planta, and the interaction between GmCYP1 and GmMYB176 was strong in the nucleus. Similar results were obtained for the reciprocal combinations. No YFP signal was detected during co-expression of control constructs -YC or -YN with GmCYP1-YN or GmCYP1-YC, respectively, confirming that YFP signal was due to the protein-protein interaction in planta. Furthermore, we measured the strength of interaction by using FRET approach. It is a powerful tool for non-invasive monitoring of protein-protein interactions that involves transfer of energy between two closely positioned fluorophores 39 , and has been widely used to determine the interactions between proteins in living plant cells [40][41][42] . Translational fusions of GmMYB176 with YFP and GmCYP1 with CFP were co-expressed in N. benthamiana leaf epidermal cells and analysed for FRET efficiency between the reporters in vivo. The results revealed a FRET efficiency of 23.8%, indicating a close proximity of GmMYB176 and GmCYP1 and their co-existence in a complex (Fig. 4b).
Empty vectors containing YFP only and CFP only were used as negative control which showed a FRET efficiency of 0.5%. A FRET value of 8-10% was considered as the background level in previous studies 43,44 .
When A. tumefaciens GV3101 strains carrying BiFC plasmids containing GmCYP1-YN or GmCYP1-YC were co-infiltrated into N. benthamiana leaves and visualized by confocal microscope, a strong yellow fluorescence was observed in the nucleus and relatively weaker fluorescence in the cytoplasm, suggesting that GmCYP1 forms a homodimer in both cellular compartments in planta (Fig. 4a). Even though recombinant hCYP-A has been reported to form monomers, dimers, and trimers when expressed in E. coli 45 , there is no published literature on homo-dimerization of plant CYPs. Search for predicted motifs in GmCYP1 identified a putative phosphorylation (pST binding) site within the GmCYP1 sequence (97-ENFVKKHTGPGILSM-112), where T 105 is potentially phosphorylated. The pST binding motifs are binding sites for 14-3-3 family of proteins. 14-3-3 protein functions as a dimer to bind with its client proteins where each monomer in the dimer is capable of interacting with a  to Table S1 for accession numbers) were aligned by ClustalW, and imported into BOXSHADE 3.21 for shading. Identical amino acids are shown in the dark box and similar amino acids are indicated by the grey box. Amino acid residues involved in PPIase activity (R 55 , F 60 and H 126 ) (Zydowsky et al. 32 ) and CsA binding (W 121 ) (Liu et al. 33 ; Zydowsky et al. 32 ) are highlighted with green and red, respectively. Secondary structure is shown below the alignment. The relative positions of amino acids indicated for PPIase activity and CsA binding sites, and the secondary structure features are based on hCYP-A (Kallen et al., 1991).
Scientific RepoRts | 7:39550 | DOI: 10.1038/srep39550 separate client protein. The dimeric nature of 14-3-3 proteins allows them to serve as scaffolds by bringing two regions of the same proteins into proximity or two different proteins together 46 . We have previously demonstrated that GmMYB176 interacts with 14-3-3 proteins, thereby affecting its subcellular localization 38 . The interactions between GmCYP1 and GmMYB176 in the Y2H screen and in BiFC assay where soybean and N. benthamiana 14-3-3 s possibly bring the two proteins together could be explained if GmCYP1 is a true client of 14-3-3 protein. Therefore, we performed a BiFC assay between SGF14l (a soybean 14-3-3) and GmCYP1. Indeed, GmCYP1 interacted with SGF14l in planta ( Fig. 4), suggesting that 14-3-3 proteins may act as a scaffold to facilitate binding of GmCYP1 and GmMYB176. Despite that the mechanism and consequence of GmCYP1 dimerization are not yet known, it is possible that binding of 14-3-3 with GmCYP1 could bring two GmCYP1 monomers together to produce fluorescence in the BiFC assay. Further, it is not clear whether the binding of GmCYP1 and GmMYB176 mediated by SGF14l is involved in the CHS8 gene regulation, and subsequent isoflavonoid biosynthesis in soybean, or to some other, as-yet unknown function.
GmCYP1 is expressed ubiquitously in soybean tissues. To study the temporal and spatial expression of GmCYP1 in soybean, a detailed transcript analysis using quantitative PCR was performed. Total RNA isolated from different tissues of soybean cultivar Harosoy63, at several different developing stages, was used in the analysis. As shown in Fig. 5a, GmCYP1 was expressed in all soybean tissues, albeit at various levels. Transcript accumulation was higher in embryos compared to that in other tissues. The level of GmCYP1 transcript increased in soybean embryos during the late developmental stages, showing highest levels (a 3-fold increase) in the embryos at 60 and 70 days after pollination compared to that in embryos at 30, 40 or 50 days after pollination (mid developmental stage) (Fig. 5a). No GmCYP1 transcripts were detected at early embryo developmental stages or in mature seeds 15 . Seed coat tissues accumulated 23-and 18-fold less GmCYP1 transcripts compared to the embryos at 60 and 70 days after pollination, respectively. The seed coat is rich in defensive and pathogen-related proteins 47 . Among all the soybean tissues and organs tested for the expression of GmCYP1, the level of expression was lowest in the seed coat.
For an in-depth analysis of GmCYP1 expression, we constructed a reporter vector containing the 1,148 bp GmCYP1 promoter fragment (upstream of the translation start site) to drive GUS gene expression. Agrobacterium strain containing the pGmCYP1pro-GUS was transformed into wild type Arabidopsis. Transgenic lines (8-10 independent T 2 progenies) carrying GmCYP1pro-GUS were selected for measuring GUS activity in different tissues during development. As observed in soybean tissues, the GmCYP1 promoter was active in most tissues of Arabidopsis (Fig. 5b). Strong GUS activity was observed in leaves and roots of young seedling while relatively less

vulgaris).
Scientific RepoRts | 7:39550 | DOI: 10.1038/srep39550 activity was found in the hypocotyl. This study revealed additional information on tissue-specific expression of GmCYP1. For example, in leaves, GUS staining was more pronounced in the vascular tissues. Similarly, GmCYP1 promoter was active in flower buds, stamen, stigma and silique walls but not in flower stem, style, petal or seeds. Failure to observe GUS activity in seeds could be due to weaker activity of the GmCYP1 promoter in the seed coat, and is supported by GmCYP1 expression in seed coat as shown in Fig. 5a. Similar results were observed for soybean chalcone synthase (CHS) gene promoters. Both CHS7 and CHS8 promoter driven GUS activities were absent in Arabidopsis seeds, despite the fact that CHS7 and CHS8 transcripts were present in the seed coat in soybean 48 . It is possible that the observed differences of gene expression in soybean and Arabidopsis may be caused by the presence or absence of the required regulatory factors in the specific tissue or developmental stage. Like GmCYP1, the Arabidopsis ortholog of GmCYP1, ROC1 (AGI:At4g38740), also exhibited higher transcript accumulation in seeds than in other tissues (Fig. 5c). The normalized mean expression data of ROC1 was compiled from AtGenExpress Visualization Tool (http://jsp.weigelworld.org/expviz/expviz.jsp). The expression of ROC1 increased gradually, and approx. 7-fold throughout seed development, from a relatively low value (0.55) at the mid globular stage (stage 3), to values of 3.2-3.75 during the later developmental stages (early curled cotyledon embryos, stage 8) to green cotyledon embryos (stage 10) (Fig. 5d). The similar expression pattern during embryo development of GmCYP1 and its Arabidopsis ortholog ROC1 suggests a conserved role for GmCYP1 and ROC1 in seed development.
To identify putative cis-elements that regulate the expression of GmCYP1 gene in soybean, we performed in silico motif analysis of 1148 bp upstream of translational start site using PlantPan2.0 database (http://plantpan2. itps.ncku.edu.tw/promoter.php). The transcription factors specific to soybean was selected during the analysis. Transcription factors that are known for their role in stress and hormonal pathways are shown in Fig. 6. The analysis identified several sequence motifs that are recognized by a number of key factors such as bZIP, MYB, WRKY, bHLH, AP2, NAC factors have been identified. Besides their role in normal plant growth and development, the stress-specific roles of these factors have been well documented in crop plant and model species 49-51 . GmCYP1 expression is reduced in response to stress. P. sojae effector Avr3b contains Nudix hydrolase activity in planta that is required for the virulence of the pathogen in soybean 52 . Recently, it has been shown   Error bars indicate the standard deviation of the mean. The stages of seed developments are: stage 3, mid globular to early heart embryos; stage 4, early to late heart embryos; stage 5, late heart to mid torpedo embryos; stage 6, mid to late torpedo embryos; stage 7, late torpedo to early walking-stick embryos; stage 8, walking-stick to early curled cotyledons embryos; stage 9, curled cotyledons to early green cotyledon embryos; and stage 10, green cotyledon embryos.
Scientific RepoRts | 7:39550 | DOI: 10.1038/srep39550 that GmCYP1 acts as a 'helper' by directly interacting with Avr3b, and modulating the hydrolase activity of it in soybean 31 . This GmCYP1-Avr3b interaction is required for the virulence and avirulence functions of Avr3b in soybean. Treatment of soybean hypocotyls with AgNO 3 is known to induce defense response and phytoalexin accumulation 53 . Here we measured the accumulation of GmCYP1 transcripts in response to stress by treating etiolated soybean hypocotyls with AgNO 3 and monitoring GmCYP1 expression 24, 48 or 72 h post-treatment. The results revealed that AgNO 3 treated soybean hypocotyls accumulate reduced level of GmCYP1 transcripts compared to the control hypocotyls at all the time points under the study (Fig. 7a). The difference in the level of GmCYP1 transcript accumulation between AgNO 3 treated and control hypocotyls was more pronounced at 24 h compared to 48 or 72 h post-treatment.
Isoflavonoid phytoalexins are host-produced antimicrobial compounds that are massively induced by pathogen attack or any other stress 54,55 . GmMYB176 regulates isoflavonoid biosynthesis by regulating CHS8 gene expression 56 . Since GmCYP1 interacts with GmMYB176, and its expression is down-regulated upon stress, whereas phytoalexin biosynthesis is induced upon stress, we measured the expression levels of isoflavonoid biosynthetic genes GmCHS8, isoflavone synthase (GmIFS2) and a isoflavonoid-specific prenyltransferase (GmPT) in response to AgNO 3 treatment. Our results revealed that expression of all these genes were induced upon AgNO 3 treatment albeit at different levels (Fig. 7b). GmCHS8 is the first enzyme in the flavonoid pathway, and as compared to control, its transcripts were accumulated at 6, 12 and 5.8 fold higher at 24, 48 and 72 hours, respectively after AgNO 3 treatment. GmIFS2 is a key legume-specific enzyme that introduces the isoflavonoid branch in the flavonoid pathway in legumes. Transcript levels of GmIFS2 were 11.6, 17 and 3.7 fold higher after 24, 48 and 72 hours, respectively in AgNO 3 treated samples compared to control. To confirm if downstream phytoalexin biosynthetic genes are induced upon stress, we measured the transcript levels of GmPT. Our results demonstrated that GmPT transcripts accumulated at 121, 64 and 3.5 fold greater than control at 24, 48 and 72 hours, respectively after AgNO 3 treatment (Fig. 7b). A significantly higher difference in the expression of isoflavonoid genes were observed only when there was a significant reduction of GmCYP1 gene expression (Fig. 7).
Several studies have shown that pathogen effectors interact with plant helper proteins for their activation and proper function 57,58 . Activated effector proteins bind with plant targets to suppress plant defenses and otherwise enable pathogen growth. For effectors encoded by Avr genes, activation can also result in detection by an immune receptor encoded by a resistance (R) gene. Thus, activation of Avr3b by GmCYP1 triggers immunity in soybean cultivars containing Rps3b, whereas activation of Avr3b in soybean cultivars lacking Rps3b enables pathogen growth 31 . Unlike many other plant CYPs that function in protecting plants during biotic and abiotic stress 29,59 , GmCYP1 acts as a susceptibility factor in soybean, at least in instances when Avr3b does not trigger immunity 31 . The expression of susceptibility proteins is generally induced during infection in susceptible plants 60 . However, our results show that GmCYP1 transcripts are reduced in stressed plants compared to controls. This finding, together with the results that show GmCYP1 interacts with GmMYB176, suggests a role for GmCYP1 as a negative regulator of plant defense and isoflavonoid biosynthesis in soybean. The high expression of GmCYP1 in the late stage developing embryos cannot be explained by this hypothesis because isoflavonoid biosynthesis is active in seed tissues, albeit this is seed isoflavonoid and not phytoalexin isoflavonoid biosynthesis.
Overall, this study presents a detailed analysis of GmCYP1. The presence of predicted cyclophilin domain, subcellular localization and its sequence homology with other identified CYPs from other organisms provides insights into its putative function. The interaction of GmCYP1 and GmMYB176 is particularly intriguing because of the conditional functionality of GmCYP1 in effector-triggered immunity or susceptibility to the pathogen P. sojae. It seems more than coincidence that the biosynthesis of isoflavonoid phytoalexins, being necessary for resistance to P. sojae, is also connected to GmCYP1. Since isoflavonoids are involved in plant stress resistance against biotic and abiotic factors [61][62][63] , the interaction of GmCYP1 with isoflavonoid regulators and its potential role as a suppressor of plant defense merits further investigation.  Error bars indicate SEM of two biological replicates and three technical replicates for each biological replicate. The data were normalized against the SUBI-3 gene for GmCYP1 and CON4 gene for isoflavonoid genes. Asterisk (*) indicate significant difference between the samples using Student's t-test.
using the primers listed in Table S2. The amplicons were cloned into a pGEM-T Easy vector (Promega, USA), and its sequence verified. SOYBEAN UBIQUITIN-3 (SUBI-3) or CON4 was used as a reference gene for data normalization and to calculate the relative mRNA levels. The data were analyzed using CFX manager (Bio-Rad, USA).
For in planta protein-protein interaction study, pDONR-Zeo-GmCYP1 was recombined separately with pEarlyGate201-YN and pEarlyGate202-YC to obtain pEG201-GmCYP1-YN and pEG202-GmCYP1-YC, respectively. The recombinant plasmids were transformed into E. coli DH5α, PCR screened, and then transformed into A. tumefacians GV3101. For FRET analysis, pDONR221-GmMYB176 and pDONR-Zeo-GmCYP1 were recombined into the pEG101 or pEG102 upstream of the YFP or CFP sequence under the control of the CaMV 35 S promoter. The recombinant plasmids were transformed into E. coli DH5α , PCR screened, and then transformed into A. tumefacians GV3101.
The promoter fragment of GmCYP1 (1148 bp) was amplified using the primers GmCYP1-P-F: 5′ -GGGGACAAGTTTGTACAAAAAAGCAGGCTCATCGTACTGCGATTTGAACGCAAGACTTC-3′ , and GmCYP1-P-R: 5′ -GGGGACCACTTTGTACAAGAAAGCTGGGTAGTTGCTGCAGAGAAGAGAAGCGTA AATGAC-3′ , cloned into pDONR-Zeo (Invitrogen, USA), as described previously, followed by the recombination in the destination vector pMDC162 to obtain pGmCYP1pro-GUS. The pGmCYP1pro-GUS was transformed into A. tumefaciens GV3101 by electroporation, and then into wild-type Arabidopsis Col-0 by floral dip method 66 . Subcellular localization and bimolecular fluorescent complementation assay. The subcellular localization of GmCYP1 was studied by infiltrating A. tumefaciens GV3101 carrying pEG101-GmCYP1 into N. benthamiana leaves, as described by Sparkes et al. 67 . For co-expression, equal volumes of two construct-bearing strains, suspended in Gamborg's solution, were mixed together and then infiltrated into N. benthamiana leaf epidermal cells. The protein expression was visualized by confocal microscopy using a Leica TCS SP2 inverted confocal microscope. An excitation wavelength of 514 nm was used for YFP imaging, and 525-545 nm emissions were collected. For visualization of CFP, an excitation wavelength of 458 nm was used, and emissions were collected between 465-495 nm.
Fluorescence resonance energy transfer (FRET) assay. Equal volumes of two construct-bearing strains containing YFP and CFP fusions (in Gamborg's solution), were mixed together and then infiltrated into N. benthamiana leaf epidermal cells. The protein expression was visualized by confocal microscopy using a Leica TCS SP2 inverted confocal microscope. An excitation wavelength of 458 nm and 514 nm were used for CFP and YFP imaging respectively. FRET acceptor bleaching, with CFP as donor and YFP as acceptor, was carried out by following Leica confocal application manual. The average of the FRET efficiency was calculated from multiple samples (n > 15).
Histochemical GUS assay. For histochemical GUS staining, T 2 transgenic Arabidopsis tissues were used.
Tissues were incubated in dark in a solution containing 100 mM sodium phosphate buffer pH 7.0, 10 mM EDTA, 0.05% Triton X-100, 1 mM potassium ferricyanide, 1 mM potassium ferrocyanide and 1 mM X-Gluc (5-Bromo-4-chloro-3-indolyl-β -D-glucuronide) for 16 h at 37 °C with gentle shaking. De-staining was carried out with 95% ethanol for 4 to 6 times. A LeicaM2 FLIII TM microscope with a QImaging Retiga 2000 R camera was used to take the photographs. For staining of silique and seeds, improved clearing method was used 68 .
Yeast two-hybrid assay. Yeast two-hybrid assay (Y2H) was performed using the Matchmaker ® Gold Two-Hybrid System (Clontech Laboratories, Inc., USA). Briefly, GmMYB176 was cloned into the vector pGBKT7 as the bait. The cDNA library as prey was generated by SMART ™ cDNA Synthesis technology (Clontech Laboratories, Inc., USA) from soybean embryos (50-60 days after pollination) and fused to GAL4 activation domain. Screening was performed by co-transformation of bait and prey using Mate & Plate TM library system (Clontech Laboratories, Inc., USA). After transformation, yeast cells were spread on SD/Ade/-His/-Leu/-Trp agar plates, and incubated at 30 °C for 5 days.