The Transcriptional Repressor MYB2 Regulates Both Spatial and Temporal Patterns of Proanthocyandin and Anthocyanin Pigmentation in Medicago truncatula

Identification of Mt MYB2 in M. truncatula

Recently, Mt MYB5 was identified as major regulator of the PA biosynthetic pathway in M. truncatula, and Medtr5g079670, a putative repressor, was found as one of the most highly induced genes in hairy roots generated from M. truncatula A17 plants overexpressing MYB5 (Figure 1A; Liu et al., 2014). Since Medtr5g079670 encodes a 213-amino acid R2R3 MYB transcription factor that shows highest sequence identity with the anthocyanin repressor MYBL2 among M. truncatula MYB proteins, and MYBL2 was also found to be a downstream target of Arabidopsis MYB5 (Dubos et al., 2008; Li et al., 2009), we named Medtr5g079670 as Mt MYB2. MYB2 transcript level was strongly reduced in myb5 mutant seeds (Figure 1A). MYB2 expression was also highly induced by overexpression of the PA regulator MYB14 in M. truncatula hairy roots and overexpression of the anthocyanin regulator Mt LAP1 in transgenic plants of alfalfa (Medicago sativa) (Figure 1A; Supplemental Figure 2). The transcript level of MYB2 was not reduced in myb14 mutant seeds (Figure 1A).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Identification of M. truncatula MYB2.

(A) Microarray analysis of Mt MYB2 (probe set ID Mtr.34401.1.S1_s_at) expression in M. truncatula hairy roots overexpressing Mt MYB14 and Mt MYB5 and in Mt myb5 and Mt myb14 mutant seeds. Microarray experiments were performed on three and two independent biological replicates for hairy roots and mutants, respectively. The relative expression was calculated with normalization to tubulin (probe set ID Mtr.40026.1.S1_at). Error bars indicate standard deviations.

(B) Phylogenetic tree of repressor MYB transcription factors based on the N-terminal protein sequences including R2R3 domain or only R3 domain for small MYB proteins. The phylogenetic tree was constructed using MEGA 6 (Tamura et al., 2013) with 1000 bootstrap replicates. Numbers indicate the percentage of consensus support.

(C) Tnt1 retrotransposon insertion positions in the Mt MYB2 gene. Arrows indicate the insertion positions in myb2-1 (NF10162) and myb2-2 (NF19285). Solid boxes indicate exons and thinner lines indicate introns. The positions of primer pairs used for the PCR (bottom panel) are shown by arrows.

In a phylogenetic analyses of repressor MYB proteins based on N-terminal protein sequences encompassing the R2R3 domain or R3 domain (for the R3-MYB group) or the entire protein sequences (Figure 1B; Supplemental Figures 3 and 4 and Supplemental Files 1 and 2), MYB2 was clustered with a group of proteins that have been identified as negative regulators of anthocyanin and/or PA biosynthesis in other plant species (Aharoni et al., 2001; Albert et al., 2014; Yoshida et al., 2015; Cavallini et al., 2015). One R2R3 MYB protein, MYB 530 (Medtr4g485530.1) from M. truncatula was also found in the same clade, which is separated from the MYB4 lignin/phenylpropanoid MYBs. Protein sequence alignment of the Fa MYB1-like proteins and MYBL2 indicates that the R2 domain and R3 domain containing the bHLH interaction residue ([DE]Lx2[RK]x3Lx6Lx3R; Zimmermann et al., 2004) are highly conserved and all of the proteins except Fa MYB1 additionally contain the C1 motif (LlsrGIDPxT/SHRxI/L; Shen et al., 2012) (Supplemental Figure 5). More importantly, a putative EAR-like repressive motif (LxLxL) is present near the C-terminal end of the proteins. Some of the Fa MYB1-like proteins, such as cotton (Gossypium hirsutum) Gh MYB6 and Vv MYBC2, additionally contain a TLLLFR repression domain that was originally identified in MYBL2. MYB2 does not have a TLLLFR domain but instead possesses a semiconserved EAR sequence at the C-terminal end in addition to a canonical EAR motif in the C2 domain (pdLNLD/ELxiG/S; Shen et al., 2012) (Supplemental Figure 5).

In the microarray data of the Medicago truncatula Gene Expression Atlas (http://mtgea.noble.org/v2/probeset.php?id=Mtr.34401.1.S1_s_at) obtained using the probe set Mtr.34401.1.S1_s_at as query, MYB2 transcripts are strongly expressed in flower and developing seeds, and this expression pattern is highly overlapping with the expression profile of MYB5 (Liu et al., 2014), suggesting that MYB2 might have roles as a negative regulator of MBW complexes in these tissues.

Mt MYB2 Is a Negative Regulator of Anthocyanin Biosynthesis

To better understand the functions of MYB2, we screened the Tnt1 retrotransposon mutagenized M. truncatula R108 population (Tadege et al., 2008). Two independent MYB2 Tnt1 insertion lines were isolated. Tnt1 insertions were located in the first and second exon of the MYB2 gene in lines NF10162 and NF19285, respectively (Figure 1C). RT-PCR could not detect the full-length MYB2 transcript in either line, indicating that they are null mutants (Figure 1C). No morphological phenotypic differences between mutant and wild-type plants were detected during the entire developmental period under our growth conditions, and both lines were used for further analysis.

At the seedling stage, myb2 mutants showed more anthocyanin accumulation in hypocotyl, junction of hypocotyl and root, and occasionally abaxial side of true leaves compared with the wild type when plants were grown on Murashige and Skoog medium (Figure 2A, top panel). Expanded anthocyanin accumulation was also observed in the flower of myb2 mutant plants. The flower banner of wild-type M. truncatula R108 had only a small spot of anthocyanin accumulation in the center, whereas the mutants exhibited a circular purple area around the fused keel petal (Figure 2A, bottom panel). The myb2 plants accumulated ∼2.5- and 1.7-fold more anthocyanin than the wild type in seedling and flower, respectively (Figure 2B). However, purple spots on the adaxial surface of leaf and stem epidermal cells did not show any change in number, spot size, or color intensity of anthocyanin pigmentation, indicating that MYB2 function might be required under specific environmental conditions in these tissues or have functional redundancy with other anthocyanin suppressors during vegetative growth.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

MYB2 Is a Negative Regulator of Anthocyanin Biosynthesis.

(A) Phenotypes of myb2-1 (NF10162) and myb2-2 (NF19285). Arrows indicate anthocyanin accumulation in the hypocotyl (7-d-old seedlings, top panel) and in the middle of banner petals (bottom panel). Magnified views of areas of anthocyanin accumulation are shown in the boxes.

(B) Anthocyanin levels in myb2 mutants. Error bars represent sd of three biological replicates. **P < 0.01 and *P < 0.05, Student’s t test.

(C) Phenotypes of M. truncatula (A17) hairy root cultures of vector control and two independent transgenic MYB2 overexpression lines. Purple color of control line indicates accumulation of anthocyanin.

(D) Anthocyanin levels in M. truncatula (A17) hairy root cultures. Data are means ± sd (n = 3). **P < 0.01, Student’s t test.

(E) qRT-PCR analysis of transcript levels in transgenic MYB2 overexpression lines in (C) compared with control. Data are means ± sd of three biological replicates with two technical replicates. *P < 0.01, Student’s t test.

(F) qRT-PCR analysis of transcript levels in myb2 seedlings (7 d old) compared with control R108. Data are means ± sd of three biological replicates with two technical replicates. **P < 0.01 and *P < 0.05, Student’s t test.

To demonstrate that MYB2 can inhibit anthocyanin accumulation in planta, we generated MYB2-overexpressing transgenic hairy roots in M. truncatula ecotype A17. The purple anthocyanin accumulation characteristic of wild-type roots was absent in the MYB2-overexpressing roots (Figure 2C). Some lines of MYB2-transformed hairy roots showed retarded growth (Figure 2C, MYB2-ox-2). The anthocyanin level was highly reduced in both overexpression lines, and this correlated with suppression of ANS and DFR1 transcript levels (Figures 2D and 2E). To gain insights into MYB2-regulated genes in M. truncatula, three independent biological replicates of MYB2-expressing and vector control hairy roots were subjected to transcriptome analysis using Affymetrix M. truncatula GeneChips. In total, 326 and 226 probe sets were identified to be induced and suppressed in MYB2-overexpressing hairy roots, respectively (transcript ratio of >2 or <0.5; P < 0.05) (Supplemental Data Set 1). According to GeneBins ontology (Goffard and Weiller, 2007), 20 probe sets of genes that were repressed in the MYB2 hairy roots encode enzymes involved in the biosynthesis of secondary metabolites (Table 1; Supplemental Figure 6). An acyltransferase (probe set ID Mtr.44986.1.S1_at) that is annotated as malonyl-CoA:isoflavone 7-O-glucoside malonyltransferase (Medicago genome 4.0 Medtr6g015830.1) and possibly involved in modification of anthocyanin was identified as the most repressed transcript (Table 1; 25-fold reduction), and transcripts encoding characterized enzymes required for flavonoid biosynthesis, including DFR1 and 2, F3′H, ANS (Medtr5g011250.1), CHS, and UGT78G1, were also affected, indicating that MYB2 plays a role as a general suppressor of anthocyanin biosynthesis (Table 1). The changes in expression of a selection of these genes was confirmed with qRT-PCR, and, in addition, strong reduction of TT8 transcript levels and mild reduction of MYB5 transcripts were found (Supplemental Figure 7). These latter genes were not represented in the microarray data.

Target gene expression was also analyzed in wild-type and mutant seedlings at 7 d after germination. CHS1, DFR1, ANS, and F3H were upregulated in myb2 seedlings (Figure 2F). DFR2 expression was not detected in either wild-type or myb2 mutant seedlings. TT8 and MYB5 expression were also increased in the myb2 mutant, as was expression of MYB530, which was identified as the closest homolog of MYB2 in the Medicago genome. This indicates that regulators of the anthocyanin pathway genes are also regulated by MYB2 (Figure 2F).

MYB2 Is a Negative Regulator of PA Biosynthesis

To determine if mutations in MYB2 have an effect on PA accumulation, we observed the seed color during the entire program of seed growth and maturation. There was no difference detected between the wild type and mutant alleles in the color of the seed coat from bright green to brown during seed development to desiccation (Figure 3A). However, staining with DMACA indicated more accumulation of PAs and/or their flavan-3-ol precursors in the myb2 mutant seed coat (Figure 3B). Wild-type seeds at 10 DAP (days after pollination) only accumulated purple/blue-staining PA around the hilum area. The staining was restricted to the same area at 12 DAP but then began to expand to the seed body until staining covered the entire seed coat at 20 DAP. At this time, the hilar groove had stronger staining than other areas, but staining then became evenly distributed and saturated by 30 DAP. Expansion of DMACA staining started earlier and was always stronger on the micropyle side compared with the chalazal side before 20 DAP in wild-type seeds (Figure 3B). In myb2 mutant seeds, DMACA staining was already detected in half of the seed body from the hilar side at 12 DAP (Figure 3B), and it covered the whole seed coat from 14 DAP although the staining was stronger in the hilar area (proximal side) as also seen in the wild-type seed coat (Figure 3B). The staining was stronger in both mutant alleles through seed development compared with the wild type, but we could not see a difference in mature seed (from 30 DAP) due to the saturation of DMACA staining (Figure 3B). Since it has been shown that the wd40-1 mutant from M. truncatula has a defect in PA biosynthesis (Pang et al., 2009), the seed staining phenotype of the wd40-1 mutant was analyzed in parallel. The dry seeds from wd40-1 plants showed a yellowish seed color (Figure 3A) and no purple/blue staining was detected with DMACA (Figure 3B); this confirms that the purple/blue staining in the wild type and myb2 mutant is indeed due to PAs.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

MYB2 Is a Negative Regulator of PA Biosynthesis.

(A) Phenotypes of wild type (R108), myb2-1 (NF10162), myb2-2 (NF19285), and wd40-1 (NF0977) mutants during seed development.

(B) DMACA-stained seeds of wild type (R108), myb2-1 (NF10162), myb2-2 (NF19285), and wd40-1 (NF0977) mutants. Bars = 2 mm.

(C) Soluble PA levels as quantified with DMACA reagent and expressed as epicatechin equivalents. Data are means ± sd (n = 6, three biological replicates). *P < 0.01, Student’s t test.

(D) Insoluble PA levels as quantified by the butanol-HCl method and expressed as procyanidin B1 equivalents. Data are means ± sd (n = 6, three biological replicates). *P < 0.01, Student’s t test.

(E) Normal-phase HPLC analysis, with postcolumn derivatization with DMACA, of soluble PA fractions from wild-type (top), myb2-1 (middle), and myb2-2 (bottom) seeds at 30 DAP. Retention time for the major peaks is noted.

(F) Phloroglucinolysis analysis of 30 DAP seeds of wild-type (top), myb2-1 (middle), and myb2-2 (bottom) plants. Retention time for phloroglucinol-epicatechin and epicatechin is noted.

(G) qRT-PCR analysis of transcript levels in myb2 seeds (12 DAP) compared with control R108. Data are means ± sd of three biological replicates with two technical replicates. **P < 0.01 and *P < 0.05, Student’s t test.

(H) Seed coat phenotypes of control (Col-0) and MYB2 overexpressing Arabidopsis plants.

(I) qRT-PCR analysis of transcript levels in MYB2-overexpressing Arabidopsis plants compared with control. Young siliques were used for analysis. Data are means ± sd of three biological replicates with two technical replicates. *P < 0.01, Student’s t test.

To quantify PA accumulation, we extracted soluble and insoluble (nonextractable) PAs separately. Soluble PA levels were significantly increased in myb2 mutant seeds at every stage of development, including 30 DAP and dry (40 DAP) seeds, with a fold-change increase of ∼1.5 in mutants compared with the wild type (Figure 3C). Soluble PA level was also increased in flowers and young pods at 4 to ∼6 DAP, although the absolute level of PA was low (<20 µg epicatechin equivalents g⁻¹ fresh weight) (Supplemental Figure 8). Measurement of insoluble PA based on butanol-HCl hydrolysis showed that myb2 mutants have more nonextractable PAs at 16, 20, and 30 DAP (Figure 3D). No significant difference was detected in seeds at 12 and 40 DAP.

To determine whether the composition of soluble PA was altered in mutant seeds, size heterogeneity of soluble PAs was determined by normal-phase HPLC coupled to postcolumn derivatization with DMACA reagent (Peel and Dixon, 2007). At 16 DAP, both wild-type and myb2 mutant seeds had low levels of free epicatechin monomer along with a range of oligomers (Supplemental Figure 9). In 30 DAP and dry seeds, free monomers were not detected and most of the soluble PA fraction was shifted to a higher degree of polymerization, leaving only a trace amount of smaller units in dry seeds (Figure 3E; Supplemental Figure 9). At all three developmental stages, the myb2 mutant produced a similar profile but higher amounts of PA compared with the wild type, indicating no major change in the size of soluble PAs in the myb2 mutant (Figure 3E; Supplemental Figure 9). Phloroglucinolysis analysis (Pang et al., 2007) was then used to determine the monomeric composition and mean degree of polymerization after purification of the soluble PAs on Sephadex LH20 resin. Although quantitatively more PA was observed from myb2 mutants by this method, the polymers in both the wild type and mutant released similar proportions of phloroglucinol-epicatechin extension units and epicatechin starter units, with similar mean degrees of polymerization (wild type, 15.95 ± 0.49; myb2, 17.05 ± 1.77 in dry seeds) (Figure 3F; Supplemental Figure 10). These data indicate that MYB2 influences the quantity but not the quality of PAs in M. truncatula seeds.

We next analyzed transcript levels of PA pathway genes in 12 DAP seeds dissected from pods of wild-type and mutant plants. qRT-PCR results showed that many of the PA biosynthetic pathway genes were upregulated in myb2 mutant seeds. EBGs including CHS1, F3H, and F3′H showed more than 2-fold increases and a smaller but significant induction of ANS was also detected. The transcript level of LDOX (Medtr3g072810.1), which is a homolog of ANS (Medtr5g011250.1) and was previously identified as a downstream target of MYB14 and MYB5 (Liu et al., 2014), was highly increased in the myb2 mutant. Among the LBGs, the homolog of Arabidopsis TT19 was induced more than 2-fold, and ANR, LAR, UGT72L1, and MATE1 expression were also upregulated. Interestingly, the most highly differentially expressed gene was MYB730 (Medtr2g088730.1), which was identified as encoding a small R3-MYB in an M. truncatula genome search and the close homolog of CPC. We also observed increases in MYB530 and TT8 transcript levels in myb2 seeds (Figure 3G).

To test if MYB2 can inhibit PA accumulation, transgenic Arabidopsis plants were generated using an MYB2 overexpression construct. Transgenic plants produced seeds with yellow seed coat color (Figure 3H; Supplemental Figure 11A), and qRT-PCR results showed that expression of PA pathway genes, including BAN (ANR), were highly reduced in young siliques (Figure 3I; Supplemental Figure 11B). MYB2 overexpression lines failed to accumulate anthocyanin in the base of the cotyledon and hypocotyl at the seedling stage. MYB2-overexpressing plants also had no anthocyanin at the bases of the primary stem and side branches, but showed normal trichome development on stems (Supplemental Figures 11C and 11D). Thus, ectopic expression of MYB2 suppresses both PA and anthocyanin biosynthesis in Arabidopsis.

Molecular Mechanism of MYB2 in Relation to MBW Complexes

We have previously shown that a MYB5-TT8-WD40-1 complex can activate the M. truncatula ANR promoter directly in Arabidopsis protoplasts in dual luciferase assays (Liu et al., 2014). To test whether MYB2 can suppress MBW complex activity by virtue of the EAR motif found in its C-terminal region (LNLDL, residues 180 to 184), MYB2 or MYB2_m1, in which the three Leu residues were changed to Ala (ANADA), was cotransfected into Arabidopsis protoplasts along with various combinations of MBW components. Whereas intact MYB2 could suppress ANR promoter activity down to as low as 4% of control activity when coexpressed with MYB5-TT8-WD40-1 constructs, ANR promoter activation was only partially suppressed when the mutated version of MYB2 was included (50% induction of the activity without MYB2) (Figures 4A and 4B). Since a semiconserved EAR motif sequence (MDIDL, residues 207 to 211) was also found at the C-terminal end of the MYB2 protein, we introduced a second mutation (MYB2_m2) by exchanging the Ile and Leu residues with Ala (MDADA) (Figure 4A). This replacement alone relieved MYB2 repression to a lesser extent compared with the mutation in MYB2_m1, but combination of the mutations in both potential suppressive domains completely abolished MYB2 repressive activity on the ANR promoter (Figure 4B). The same result was observed with different promoters (LAR_pro, CHS1_pro, LDOX_pro, and DFR1_pro) that are known to be direct targets of TT2/MYB5-TT8-TTG1 complexes in Arabidopsis (Figure 4C). MYB2_m1m2-AD (Gal4 DNA activation domain) fusion protein did not itself activate the ANR promoter, indicating that MYB2 protein itself may not bind directly to the promoter (Figure 4B).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Regulatory Mechanism of MYB2 on Target Promoters.

(A) Analysis of repression motifs of MYB2 protein. The conserved and semiconserved EAR motifs are underlined and the five amino acids changed to Ala are indicated by shadow in two modified versions of the MYB2 protein (m1 and m2).

(B) Activation and repression of the ANR promoter in transient expression assay using Arabidopsis protoplasts. Various combinations of M. truncatula effectors (MYB5, TT8, WD40-1, MYB2, MYB2_m1, MYB2_m2, MYB2_m1m2, and MYB2_m1m2-AD) were used to transfect Arabidopsis protoplasts along with the ANR promoter fused with luciferase reporter. Data are means ± sd (n = 3 biological replicates). Significant difference was determined by one-way ANOVA (LSD, P = 0.05) and indicated by letters.

(C) Activation and repression assays of the various target promoters. Data are means ± sd (n = 3 biological replicates). Significant difference for each promoter was determined by one-way ANOVA (LSD, P = 0.05) and indicated by letters.

(D) BiFC assay showing the physical interaction between MYB2 and TT8 or MYB5 (TT8-mediated). eYFP signals are localized in the nucleus and are shown in green and chloroplast autofluorescence is shown as magenta.

(E) Dosage-dependent suppression of MBW complex activities by MYB2. Serial dilutions of the MYB2 construct were cotransfected with different effector combinations (MYB14-TT8-WD40-1, MYB5-TT8-WD40-1, and MYB5-MYB14-TT8-WD40-1). Data are means ± sd (n = 3 biological replicates).

(F) Putative mechanism of MYB2 and MYB5/MYB14 MBW activator complexes to differentially regulate the promoter activity of a target gene. MYB2 interaction with MYB5/MYB14 MBW activator complexes is mediated by TT8. MYB2-TT8 dimer may bind to the complex or replace one of the activator MYBs with MYB2, leading to suppression of transcription in a dosage dependent manner.

To test physical interactions between MYB2 and MBW components, bimolecular fluorescence complementation (BiFC) assays were performed in Arabidopsis protoplasts. The complementary eYFP signal was only observed when MYB2 and TT8 constructs were cotransfected, whereas MYB2 did not show any interaction with MYB5, MYB14, or WD40-1 proteins (Figure 4D; Supplemental Figure 12). However, MYB2 also showed a TT8-dependent interaction with MYB5 or MYB14 (Figure 4D; Supplemental Figure 12), indicating that MYB2 can be incorporated in the MBW complex by making physical interactions with MYB5 or MYB14 that result in an inhibitory complex and that are mediated by TT8. This conclusion is further supported by the efficient suppression of promoter activity by MYB2 when TT8 was cotransfected (Figure 4B; Supplemental Figure 13). No dilution effect of MYB2_m1m2 on MYB5 activity in transactivation assays indicates that MYB2 functions as an active inhibitor rather than as a passive competitor, since MYB2_m1m2 protein interacts normally with TT8 (Figures 4B and 4D). However, the fact that MYB2_m1m2-AD fusion protein did not enhance ANR promoter activity when combined with MYB5 (Figure 4B), in contrast to the case of synergistic promoter activation upon cotransfection of MYB5 and MYB14 (Liu et al., 2014), suggests that a specific configuration of the activator complex is required upon binding to the target promoter to enhance transcription.

To define the inhibitory mechanism of MYB2 on MYB5 or MYB14 MBW complexes, dosage-dependent suppression of ANR promoter activity was tested. When serial dilutions of MYB2 construct were cotransfected along with different MBW complexes, the MBW MYB14-TT8-WD40-1 complex showed the most sensitive response to low concentrations of MYB2 (more than 50% reduction at the ratio of MYB14 to MYB2 of 1:0.03125) compared with MYB5-TT8-WD40-1 and MYB5-MYB14-TT8-WD40-1 combinations, which had similar 20% reductions (Figure 4E). Since the fold induction of promoter activity with MYB5-MYB14-TT8-WD40-1 combinations compared with no effector was the highest (747.7-fold; Supplemental Figure 14), these data indicate that coregulation of target promoters by MYB5 and MYB14 has an additive effect in terms of transcriptional activation and resistance to repressor(s), which is mainly attributed to the presence of MYB5 (Figure 4F).

Spatio-Temporal Regulation of PA Accumulation by MYB2 in M. truncatula Seeds

To determine the spatio-temporal patterns of PA accumulation in myb2 mutant seeds, detailed histochemical analysis was performed at different stages of seed development. Longitudinal sections from wild-type and myb2-1 (NF10162) seeds were stained with toluidine blue O (TBO), which was reported to show specific greenish-blue staining indicating deposition of PA or PA precursors in Arabidopsis (Debeaujon et al., 2003). The green staining was localized to the seed coat, but it was mainly detected in the epidermal cell layer on the hilar side of the seed coat (Supplemental Figure 15), while it was observed in the endothelium of the Arabidopsis seed coat (Debeaujon et al., 2003). In the hilar groove and the proximal region close to the micropyle and embryo, the vacuoles at the bottom of the epidermal cells were filled with green staining, and smaller vacuoles or vesicle structures that had weak TBO staining were detected in the upper side of the cells in both R108 and myb2 mutant plants (Supplemental Figures 15B, 15B-III, 15C, 15E, 15E-III, and 15F). In the chalaza, enlarged multiple vacuoles with green staining were found in multiple or closely packed macrosclereids (Supplemental Figures 15B, 15B-II, 15C, 15E, 15E-II, and 15F). However, there was no indication of PA accumulation or PA-vacuole formation in the epidermal layer of the distal side of the seed coat (Supplemental Figures 15B, 15B-IV, 15E, and 15E-IV). We also could not observe PA staining in hypodermal cells (osteosclereid precursors) and inner parenchyma cells throughout the entire seed coat, except the mycropylar-chalazal extension region where the seed coat is more enlarged. In this region, the staining was expanded to a group of round parenchyma cells close to vascular bundles, and small prevacuolar structures containing PA were found inside or outside of the central vacuole (Supplemental Figures 15B, 15B-I, 15C, 15E, 15E-I, and 15F).

In the wd40-1 mutant, the staining in the epidermal layer completely disappeared, confirming that the TBO staining in the same cell files of both wild-type and myb2 seeds was indeed PA deposition (Supplemental Figures 15G, 15H, 15H-I to 15H-IV, and 15I). Green TBO staining was only detected in the xylem tissue of chalazal vessels in wd40-1 seed (Supplemental Figure 16).

To compare the proximal-distal seed axis at the same longitudinal position, transverse sections of 10 DAP wild-type and myb2-1 (NF10162) mutant seeds were made. In the section across the developing cotyledon, PA in the large vacuole was more densely stained on the hilar side of myb2 seeds, although the vacuolar size was similar to that of the wild type (Supplemental Figures 17B, 17B-I, 17D, and 17D-I). In the middle section dissecting the hilum, only the myb2 mutant had faint green staining in the small vacuolar structures of epidermal cells at the boundary of the proximal and distal region (Supplemental Figures 18B, 18B-III, 18D, and 18D-III). There was no sign of PA accumulation on the distal side of the seed body in either wild-type or myb2 mutant seeds (Supplemental Figures 17B, 17B-III, 17D, 17D-III, 18B, 18B-IV, 18D, and 18D-IV). In the wd40-1 mutant, the generation of an empty vacuole of normal size in the epidermal layer was more obvious in the transverse sections (Supplemental Figures 17E, 17F, 17F-I to 17F-III, 18E, 18F, and 18F-I to 18F-IV).

To examine differences in spatial patterns of PA accumulation at a more mature stage, transverse sections of wild-type R108, myb2-1 (NF10162), and wd40-1 were compared at 12 DAP (bent cotyledon stage). On the proximal side of the upper region across the hypocotyl of the developing embryo, the vacuoles containing PA in the epidermis were enlarged and occupied almost half of the cytoplasm, with bright green TBO staining compared with 10 DAP seeds (Figures 5B, 5B-I, 5D, and 5D-I). In parenchyma cells of the middle region, the small vacuoles remained as aggregated but still separate granular structures (Figures 5H, 5H-I, 5J, and 5J-I). The PA staining sometimes appeared to be diffused in the entire central vacuole but with stronger staining near the tonoplast (Figures 5H, 5H-I, 5J, and 5J-I). The most distinct difference in spatial distribution of PAs between the wild type and myb2 mutant at this stage was detected in the boundary between the proximal and distal axis. The epidermal cells of myb2 had vacuoles positioned at both ends of the cells that stained densely with TBO (Figures 5D, 5D-III, 5J, and 5J-III). In the wild-type seeds, only some of the vacuoles were weakly stained with green color in the upper section and no staining was found in the middle area (Figures 5B, 5B-III, 5H, and 5H-III). Although the sizes of the cell files were smaller and the cytoplasmic staining was denser, small vacuoles with weak TBO staining were seen in the myb2 mutant even on the distal side of the seed (Figures 5D, 5D-IV, 5J, and 5J-IV). These data indicate that PA accumulation in the vacuole or vesicle is proceeding in the seed body of the myb2 mutant even though the differentiation of the cell file based on cell elongation, vacuole formation, and cuticle accumulation has progressed to the same phase as in the wild type. In the wd40-1 mutant, the cell file developed as normal but the vacuole was empty and there was no staining detected with TBO (Figures 5E, 5F, 5F-I to 5F-IV, 5K, 5L, and 5L-I to 5L-IV). The development of cuticle was not affected in myb2 or wd40-1 seeds compared with the wild type at both 10 and 12 DAP (Figure 5; Supplemental Figures 15, 17, and 18). We could not obtain TBO staining data for older seeds due to the limitation of permeability of fixative and resin.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

Regulation of Progressive PA Deposition in the M. truncatula Seed Coat by MYB2.

(A), (C), and (E) DMACA-stained seeds of wild-type, myb2-1 (NF10162), and wd40-1 plants at 12 DAP.

(G), (I), and (K) DMACA-stained seeds of wild-type, myb2-1 (NF10162), and wd40-1 plants at 12 DAP.

(B) and (H) Transverse section of R108 seeds at 12 DAP stained with TBO.

(D) and (J) Transverse section of myb2-1 (NF10162) seeds at 12 DAP stained with TBO.

(F) and (L) Transverse section of wd40-1 seeds at 12 DAP stained with TBO.

At 12 DAP (bent cotyledon stage of embryo), progressive differentiation of the cell files from proximal (p) to distal (d) axis was observed. A thick cuticle layer on the outer surface of the hilar side indicated that the proximal side was more differentiated compared with the distal side, where small cells with dense cytoplasm were observed. The magnified views of the regions boxed in (B), (D), (F), (H), (J), and (L) are denoted by I, II, III, and IV. Double-headed arrow indicates the proximal-distal axis. Dashed lines indicate the position of the cross section. White arrowhead indicates the nucleus containing nucleolus aligned in the middle of the macrosclereid cells. Black arrowheads denote the PA-positive central vacuole with granule structures of small vacuoles. Arrows indicate the vacuole with accumulation of PAs or PA precursors. Magnified views of PA-containing vacuole are shown in the boxes in (D-IV) and (J-IV). CV, chalazal vessel; EN, endosperm; EP, epidermis; H, hilum; HG, hilar groove; HP, hypocotyl of embryo; HY, hypodermis; PR, parenchyma. Bars = 1 mm in (A), (C), (E), (G), (I), and (K), 0.25 mm in (B), (D), (F), (H), (J), and (L), 10 µm in (B-I) to (B-IV), (D-I) to (D-IV), (F-I) to (F-IV), (H-II) to (H-IV), (J-II) to (J-IV), and (L-II) to (L-IV), and 10 µm in (H-I), (J-I), and (L-I).

To confirm that the blue-green TBO staining in the epidermal layer represents PAs and/or PA precursors, 12 DAP seeds were sectioned after DMACA staining (Figure 6). In the longitudinal sections, strong purple staining with DMACA was colocalized with TBO staining in macrosclereids and some parenchyma cells around the hilum area in both R108 and myb2-1 (NF10162) mutant seeds. The coloration was more expanded in the epidermal layer of myb2 seeds (Figures 6A and 6B). The same expansion was also observed in transverse sections across the upper side of the seeds (Figures 6E and 6F). DMACA staining was dispersed in the entire epidermal layer rather than being restricted to the vacuole, but it appeared most intense near the tonoplast, and strong purple aggregations attached to the vacuolar membrane were observed in DMACA-positive cells (Figures 6E, 6E-I, 6F, and 6F-I). Weak purple staining was continuous to the seed body only in myb2 mutants (Figures 6E, 6E-II, 6F, and 6F-II). There was a complete absence of DMACA staining in wd40-1 seeds (Figures 6C, 6G, 6G-I, and 6G-II). In mutant myb14-1 seeds, DMACA staining disappeared in the epidermal cells in the hilum groove and in parenchyma cells. The weak staining in other areas compared with the wild type is possibly due to the redundancy of Mt MYB14 with the other MYB transcription factors PAR or Mt MYB5 (Figures 6D, 6H, 6H-I, and 6H-II).

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.

DMACA Staining of PA Accumulation in M. truncatula Seeds.

(A) Longitudinal section of R108 seeds at 12 DAP.

(B) Longitudinal section of myb2-1 (NF10162) seeds at 12 DAP.

(C) Longitudinal section of wd40-1 seeds at 12 DAP.

(D) Longitudinal section of myb14-1 seeds at 12 DAP.

(E) Transverse section of R108 seeds at 12 DAP.

(F) Transverse section of myb2-1 (NF10162) seeds at 12 DAP.

(G) Transverse section of wd40-1 seeds at 12 DAP.

(H) Transverse section of myb14-1 seeds at 12 DAP.

Wild-type (R108), myb2-1, wd40-1, and myb14-1 seeds at 12 DAP were stained with 0.5% DMACA in absolute ethanol containing 0.8% w/v HCl and used for sectioning. The magnified views of the regions boxed in (E) to (H) are denoted by I and II. Dashed lines indicate the position of the cross section. The magnified view of the epidermal layer is shown in the boxes in (E-I), (E-II), (F-I), (F-II), (G-I), (G-II), (H-I), and (H-II). Arrowheads indicate the end points of DMACA-stained cells in the epidermal layer and arrows denote intense PA coloration with DMACA near the tonoplast. Bars = 0.25 mm in (A) to (D), 0.25 mm in (E) to (H), and 50 µm in (E-I), (E-II), (F-I), (F-II), (G-I), (G-II), (H-I), and (H-II).

MYB2 Represses ANR in the M. truncatula Seed Coat

Given that ANR is a key enzyme specific for PA biosynthesis in the seed coat and is an immediate target of the MBW complex and MYB2, we compared the ANR expression pattern in wild-type and myb2 mutant seeds to determine where MYB2 protein suppresses ANR expression in vivo. In young wild-type seeds, an ANR_pro:GUS construct was strongly expressed in the hilar area, although weak GUS activity was detected in the entire seed coat (Figure 7A). In the hilar area, strong ANR expression was diffused into parenchyma cells, which normally accumulate PA (Figures 7A, 7E, and 7I). In the seed body beside the hilar region, weak GUS activity was only detected in the epidermal layer of the seed coat in developing seeds (Figures 7E and 7F). In the myb2 mutant background, strong ANR expression was detected in the outside of the hilar area, and GUS activity was more evenly distributed in the entire seed coat (Figure 7B). However, the expression still appeared to be strongest in the outermost cell layer (Figures 7J and 7K). These results are consistent with the expanded regions of PA accumulation in the myb2 mutant seeds.

• Download figure
• Open in new tab
• Download powerpoint

Figure 7.

Regulation of ANR Expression by MYB2.

(A) Representative expression patterns of ANR_pro:GUS construct in two independent transgenic lines of wild-type R108 background.

(B) Representative expression patterns of ANR_pro:GUS construct in two independent transgenic lines of myb2 mutant background.

(C) and (D) Representative expression patterns of MYB2_pro:GUS construct in two independent transgenic lines of M. truncatula R108.

(E) and (F) Longitudinal section of ANR_pro:GUS R108 seed.

(G) and (H) Longitudinal section of MYB2_pro:GUS R108 seed.

(I) Transverse section of ANR_pro:GUS R108 seed.

(J) and (K) Transverse section of ANR_pro:GUS myb2 seed.

(M) to (O) Transverse section of MYB2_pro:GUS R108 seed.

(L) and (P) Representative expression patterns of MYB2_pro:GUS construct in seedling and flower.

Arrows indicate hypocotyl and the center of banner petal, which have ectopic anthocyanin accumulation in myb2 mutant. Arrowheads indicate hilum. (F), (H), (K), and (N) and (O) are the magnified views of the regions boxed in (E), (G), (J), and (M), respectively. Bars = 1 mm in (L) and (P), 0.5 mm in (G), 0.2 mm in (E) and (M), 0.1 mm in (I) and (J), 50 µm in (H) and (K), 25 µm in (N) and (O), and 10 µm in (F).

Additionally, we examined the activity of a MYB2_pro:GUS reporter gene construct in M. truncatula seed and found that MYB2 expression was excluded from the hilar region (Figures 7C, 7D, 7G, and 7M), conversely to PA accumulation and ANR expression. MYB2 expression was strong in the seed body but became faint as seeds matured (Figures 7C and 7D), which is consistent with the microarray data. MYB2 promoter activity was strongest in the epidermal layer circumventing the seed body and was more diffused into inner cell files, including the endosperm in the distal region of the seed body (Figures 7G, 7H, and 7M to 7O). Faint GUS activity was also detected in the parenchyma cell files close to chalazal vessel branches (Figure 7M). Overall, these expression analyses indicate that MYB2 represses PA accumulation in the seed body by suppressing ANR expression at the early stage of seed development.

We also checked MYB2_pro:GUS activity in other tissues, including flower and seedling, which exhibit ectopic anthocyanin accumulation in the myb2 mutant. MYB2 expression was detected in the hypocotyl and in the junction between the hypocotyl and root of 7-d-old seedlings (Figure 7L). MYB2 was also expressed in the middle of the banner of the flower, making a semicircle with stronger GUS staining at the periphery (Figure 7P). In the strongest expression line analyzed, GUS activity in the flower was more diffused into the center of the banner and in the base of the wing and keel petals (Supplemental Figure 19A). GUS activity was detected in the cotyledon and root, where strongest staining was observed in the vasculature and lateral root tips (Figure 7L; Supplemental Figure 19B). In the shoot apex, promoter activity was specifically localized in the trichomes of the inflorescence meristem, young flowers, and leaves, with more diffused expression at wounding sites (Supplemental Figures 19C and 19D). Relatively strong MYB2 expression was detected in the developing pod and its trichomes (Supplemental Figures 19E and 19F). We also observed GUS staining in the tip of the style, anther, and ovary (Supplemental Figure 19A).