MtPAR MYB transcription factor acts as an on switch for proanthocyanidin biosynthesis in truncatula

Jerome Verdiera,1, Jian Zhaoa,1, Ivone Torres-Jereza, Shujun Gea,b, Chenggang Liua, Xianzhi Hea, Kirankumar S. Mysorea, Richard A. Dixona,2, and Michael K. Udvardia

aPlant Biology Division, Samuel Roberts Noble Foundation, Ardmore, OK 73401; and bCollege of Agronomy, Agricultural University of Hebei, Baoding 071001, China

Contributed by Richard A. Dixon, December 20, 2011 (sent for review September 19, 2011) MtPAR (Medicago truncatula proanthocyanidin regulator) is an MYB biosynthesis in M. truncatula and other ; only a WD40- family transcription factor that functions as a key regulator of proan- repeat TF has been implicated as a positive regulator of PA bio- thocyanidin (PA) biosynthesis in the model Medicago trun- synthesis in M. truncatula seeds (15). catula. MtPAR expression is confined to the seed coat, the site of PA From a -wide study of TF and other genes that are accumulation. Loss-of-function par mutants contained substantially activated during seed development in M. truncatula, >30 seed- less PA in the seed coat than the wild type, whereas levels of antho- induced TFs were targeted for functional characterization by us- cyanin and other specialized metabolites were normal in the mu- ing reverse . One of these encodes an MYB family TF that tants. In contrast, massive accumulation of PAs occurred when we found regulates PA biosynthesis in M. truncatula seeds. Ec- MtPAR was expressed ectopically in transformed hairy roots of Med- topic expression of the gene in transformed hairy roots led to PA icago. Transcriptome analysis of par mutants and MtPAR-expressing biosynthesis and accumulation. Therefore, this MYB TF has the hairy roots, coupled with yeast one-hybrid analysis, revealed that potential to boost tannin levels for forage improvement. MtPAR positively regulates genes encoding enzymes of the flavo- noid–PA pathway via a probable activation of WD40-1. Expression of Results MtPAR in the forage legume alfalfa (Medicago sativa) resulted in MtPAR Encodes an MYB TF with Seed Coat-Specific Expression. We detectable levels of PA in shoots, highlighting the potential of this used the M. truncatula Gene Expression Atlas (MtGEA) (16) to gene for biotechnological strategies to increase PAs in forage le- select seed-induced TF genes for genetic characterization. One gumes for reduction of pasture bloat in ruminant animals. of these, MtPAR, encodes a putative MYB TF of the R2R3 class, based on the presence of highly conserved R2 and R3 MYB DNA condensed tannin | forage quality | metabolic engineering binding domains at the N-terminal end of the protein (17). MtPAR (Affymetrix probe set Mtr.50541.1.S1_at) is expressed in a seed- roanthocyanidins (PAs; also called condensed tannins) are specific manner (Fig. S1A; or see MtGEA, http://mtgea.noble.org/ Poligomers of flavan-3-ol units and are prominent compounds v2/), with maximal expression at 24 d after pollination (DAP; Fig. in seed coats, , , flowers, and bark of many 1). Quantitative RT-PCR (qRT-PCR) indicated that the gene is species (1). PAs and their monomeric building blocks catechin specifically expressed in the seed coat but not in the embryo or and epicatechin are antioxidants with beneficial effects on hu- endosperm of mature seeds (Fig. 1). Distance analysis revealed no man health, including cardioprotective (2), anticancer (3), and close relationship between MtPAR and MYB TFs involved in the anti-inflammatory activities (4). PAs in forage bind to regulation of anthocyanin (e.g., LAP proteins from M. truncatula proteins and slow their fermentation in the rumen, reducing or ANTHOCYANIN1 from Solanum lycopersicum) (18, 19) or PA microbial production of methane and thereby protecting the biosynthesis (e.g., TRANSPARENT TESTA2 from A. thaliana or animal from potentially lethal pasture bloat (1). Moderate levels MYBPA1 and MYBPA2 from Vitis vinifera) (refs. 9 and 20; Fig. of PAs in forages also improve nitrogen nutrition, reduce urinary S1B). The closest homolog of MtPAR is an MYB protein nitrogen excretion, and help counter intestinal parasites (5). (GmMYB115; GenBank accession no. Q0PJG9) from Glycine Unfortunately, many legumes, including the world’s most im- max (soybean) of unknown function. portant forage legume, alfalfa (Medicago sativa), have insufficient PA levels to prohibit bloat in grazing animals. Therefore, in- par Mutants Are Defective in Seed Coat PA Accumulation. We isolated creasing the PA content of shoots is an important target of bio- four independent mutants with retrotransposon insertions in the technology in alfalfa and other forage legumes. MtPAR gene through a PCR screen of DNA from a Tnt1-insertion PA biosynthesis and its regulation are quite well understood in mutant population (21). Tnt1 insertions were found in the second the model nonlegume species (6). Much of the exon of MtPAR in mutant line NF4419 and in the third exon knowledge of PA biosynthesis in Arabidopsis has come from the in lines NF2466, NF1358, and NF3308 (Fig. 2A). Homozygous analysis of transparent testa (tt) mutants, which exhibit reduced seed pigmentation (6, 7). Twenty TT genes have been characterized and encode enzymes or transporters involved in PA biosynthesis and Author contributions: J.V., J.Z., R.A.D., and M.K.U. designed research; J.V., J.Z., I.T.-J., S.G., storage or proteins that regulate PA production (6, 8). The latter C.L., X.H., and K.S.M. performed research; J.V., J.Z., R.A.D., and M.K.U. analyzed data; and include a set of transcription factors (TFs): TT2, an MYB family J.V., J.Z., R.A.D., and M.K.U. wrote the paper. TF; TT8, a bHLH TF; and TTG1, a WD40 protein, which together The authors declare no conflict of interest. form a ternary complex that regulates transcription of anthocya- Freely available online through the PNAS open access option. nidin reductase (ANR; the enzyme producing the epicatechin Data deposition: The MtPAR sequence reported in this paper has been deposited in the building block of PAs) (9–11). GenBank database (accession no. HQ337434). Recently, key genes/enzymes and a precursor transporter 1J.V. and J.Z. contributed equally to this work. (MATE1) involved in PA biosynthesis have been isolated and 2To whom correspondence should be addressed. E-mail: [email protected]. – characterized from the model legume Medicago truncatula (12 This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. 14). However, relatively little is known about the regulation of PA 1073/pnas.1120916109/-/DCSupplemental.

1766–1771 | PNAS | January 31, 2012 | vol. 109 | no. 5 www.pnas.org/cgi/doi/10.1073/pnas.1120916109 Downloaded by guest on September 27, 2021 MtPAR expression (Fig. S2C). Both soluble and insoluble PA levels in par seeds were significantly reduced compared with their segregant controls. Soluble PA concentration was ∼50% lower and insoluble PA concentration was up to 80% lower in the mutants than in the sibling WTs (Fig. 2D). High-performance liquid chromatography (HPLC) followed by normal-phase HPLC with postcolumn derivatization with DMACA reagent indicated that mutant and WT seeds exhibited a similar size distribution of PAs (Fig. S3). Spectrophotometric analysis of anthocyanin concentra- Fig. 1. Gene expression of MtPAR. Expression profile of MtPAR through tion revealed no significant difference between par mutant and WT seed development (10–36 DAP) according to the MtGEA (average of bi- seeds (Fig. S2D). Thus, MtPAR regulates PA but not anthocyanin ological triplicates with SD) and in seed tissues (SC, seed coat; E/Eo, embryo biosynthesis in seeds. and endosperm) according to qRT-PCR data (average of technical triplicates with respective SD). Ectopic Expression of MtPAR Induces PA Biosynthesis. To sub- stantiate a role for MtPAR in PA biosynthesis, we transformed M. insertion mutants of all four lines showed a reduction in pigmen- truncatula hairy roots with the MtPAR cDNA coupled to the con- tation of mature seed compared with the wild-type (WT) control stitutively active CaMV-35S promoter (22). Agrobacterium rhizo- (Fig. 2B and Fig. S2A). MtPAR transcript levels in developing genes (strain ARqua 1; ref. 23) was used to transfer the p35S:: fl < C MtPAR construct into M. truncatula together with a green uores- seed of the four mutants were 5% of the WT level (Fig. S1 ). fi 4-Dimethylaminocinnamaldehyde (DMACA) staining sug- cent protein (GFP) gene that enabled identi cation of transformed gested a decrease in the PA content of mature mutant seeds hairy roots (Fig. 3A). Ectopic expression of MtPAR in Medicago compared with WT (Fig. 2C). DMACA staining of developing hairy roots was checked by qRT-PCR (Fig. 3B). Initial observations seed revealed a similar, gradual accumulation of PA from 10 to 16 of unstained hairy roots revealed a decrease in red pigmentation in DAP in both mutant and WT. Differences between mutant and transgenic roots containing the p35S::MtPAR construct compared with control-transformed roots containing a p35S::GUS (β-glucu- WT first became apparent at ∼20–24 DAP (Fig. S2B), coinciding ronidase) construct (ref. 24; Fig. 3A). No differences in root growth with maximal MtPAR expression in the WT. DMACA staining was or other morphological features were observed between roots confined largely to seed coats, mirroring the tissue specificity of containing p35S::MtPAR and controls. Subsequent staining of hairy roots with DMACA revealed a dramatic difference between p35S::MtPAR-andp35S::GUS-transformed roots; the former stained dark blue-green with DMACA, indicating the presence of PAs, whereas the controls did not (Fig. 3A). Soluble PA levels were low in hairy roots of control trans- formed plants (p35S::GUS) but were up to 100-fold higher in some p35S::MtPAR lines (Fig. 3B). Moreover, a positive corre- lation (r2 = 0.82) was observed between soluble PA concentra- tion and MtPAR transcript levels in the different transgenic lines (Fig. 3B). In contrast, insoluble PA levels were very low in

A GFP marker Light image DMACA staining PLANT BIOLOGY p35S::PAR p35S::GUS

Anthocyanin content Soluble PA content B 40 12 35 10 30 R² = 0.8254 25 8 20 6 15 4 10 5 2 Fig. 2. Analysis of par mutants in Medicago.(A) MtPAR gene model with (μmol FW) Epi. equiv./g of Soluble PA concentration Soluble PA (μg g FW) of Cy3G equiv./ 0 0 position of different Tnt1 insertions and the names of the corresponding concentration Anthocyanin 0 0.05 0.1 0.15 0.2 0.25 independent mutant lines. Introns are represented by a line, and exons are MtPAR expression relative to ACTIN represented by a rectangle. (B) Effect of mutation on mature seed pig- mentation for the NF4419 mutant line. A similar phenotype was observed in Fig. 3. Ectopic expression of MtPAR in Medicago hairy roots. (A) Phenotype other mutant lines (Fig. S2A). (C) DMACA staining of mature seeds from of MtPAR ectopic expression transformants in hairy roots. (Left) GFP de- NF2466 mutant line (similar phenotype in other mutant lines; Fig. S2A). (D) tection as a transformation marker. (Center and Right) Unstained and Levels of extractable PAs (soluble and insoluble) with respect to their null DMACA-stained hairy roots. (B) Correlations between relative expression of segregant controls. Values represent averages and SD from three biological MtPAR with soluble PA and anthocyanin concentrations in 14 different hairy replicates. root transformants. Cy3G, cyanidin 3-O-glucoside; Epi, (-)-epicatechin.

Verdier et al. PNAS | January 31, 2012 | vol. 109 | no. 5 | 1767 Downloaded by guest on September 27, 2021 control hairy roots and were not increased as a result of ex- To identify genes that might be regulated directly by MtPAR, pression of MtPAR (Fig. S2E). Anthocyanin concentration was we compared genes that were repressed in the par mutants with high in hairy roots of control plants but decreased with in- those induced in p35S::MtPAR lines relative to their appropriate creasing soluble PA levels in p35S::MtPAR lines (Fig. 3B). controls. Twelve genes satisfied both criteria, and 8 of these encode enzymes involved in PA and anthocyanin biosynthesis MtPAR Regulates Expression of PA Biosynthesis Genes. To determine (Table 1). Among these were genes encoding ANS (Mtr.14017.1. the mechanism by which MtPAR triggers PA biosynthesis, we S1_at) and ANR (Mtr.44985.1.S1_at), which carry out the last carried out transcriptome analysis of mutant and WT sibling two steps in the synthesis of epicatechin, the building block for seeds and of p35S::MtPAR and p35S::GUS-transformed roots, PAs in Medicago (Fig. 4A). An integrated analysis of flavonoid using Affymetrix Medicago GeneChips. Comparisons of tran- pathway expression including all gene family members was car- script levels in seed at 20 DAP identified 49 genes that were ried out by using transcriptomic data from mutant seeds and differentially expressed (transcript ratio of <0.5 or >2; P < 0.05) ectopic expression in hairy roots (Fig. 4B). This analysis con- fi between par mutants (lines NF2466, NF3308, and NF4419) and rmed the regulatory effect of MtPAR on the expression of two central enzymes in the flavonoid pathway (i.e., CHS and F3H) their WT siblings. Of these, 38 genes exhibited lower and 11 fi genes exhibited higher transcript levels in the mutants (Table and one enzyme, ANR, speci cally regulating the switch between S1). According to GeneBins ontology (25), 14 of the genes that PA and anthocyanin synthesis. To assess the impact of the par mutation on the flavonoid were repressed in the mutants encode enzymes involved in fla- metabolic pathway, we performed metabolite profiling using ultra- vonoid biosynthesis (Table S1). Some of these genes/enzymes are performance liquid chromatography (UPLC) coupled to electro- required for both PA and anthocyanin synthesis (e.g., chalcone fl fl ′ spray ionization quadrupole time-of- ight mass spectrometry synthase, CHS; avonoid 3 -hydroxylase, F3H; and anthocyani- (MS). Out of 74 specialized metabolites identified in mature seeds, din synthase, ANS), whereas others act downstream of ANS and 19 were altered significantly in amount in par mutants compared fi are speci c to PA biosynthesis (e.g., ANR, glucosyltransferase with WT controls (Table S3). These metabolites belonged mainly UGT72L1). Genes that were more highly expressed in the to four groups of compounds: coumaric acid and coumarin, tri- mutants were mostly of unknown function (Table S1). terpene saponins, epicatechins, and flavonoid glycosides (two, In M. truncatula hairy roots transformed with p35S::MtPAR, 171 eight, two, and seven metabolites respectively). Although the genes were significantly altered in expression compared with amount of individual saponins changed in the par mutants, the p35S::GUS-transformed controls (transcript ratio of <0.5 or >2; total amount of saponins was not significantly different between P < 0.05; Table S2). Of these, 115 exhibited higher transcript levels mutants and their WT siblings. The same was true of coumaric acid in p35S::MtPAR roots. Eleven of the 115 genes coded for putative and coumarin. In contrast, total epicatechin content was sub- enzymes of flavonoid biosynthesis (e.g., CHS, F3H, and ANS). stantially lower (by 45.9%) and flavonoid glycoside content was

Table 1. The 11 common genes that are down-regulated in loss-of-function mutant lines and up-regulated in ectopic expression transformant lines par mutant lines 35S::MtPAR hairy roots

Representative Ratio par/ Ratio PAR/ Probe sets Target description public ID WT PQGUS PQ

Mtr.20567.1.S1_at Type III polyketide synthase; 1115.m00010 0.007 2.6E-02 0.0E+00 4.701 1.9E-02 0.0E+00 naringenin-CHS Mtr.36333.1.S1_at Similar to UP|Q84JJ4 F3H BE248436 0.027 8.9E-03 0.0E+00 2.296 8.9E-02 9.2E-13 Mtr.6517.1.S1_at Similar to UP|Q84J65 gray BQ147749 0.049 5.7E-03 0.0E+00 2.637 1.5E-03 0.0E+00 pubescence F3H Mtr.14017.1.S1_at Weakly similar to UP|LDOX_ARATH TC99980 0.063 3.0E-02 0.0E+00 3.857 6.3E-04 0.0E+00 (Q96323) ANS Mtr.39897.1.S1_at Similar to UP|P93697 CPRD12 TC105988 0.080 1.7E-02 1.0E-231 2.063 4.5E-02 3.6E-22 protein Mtr.14428.1.S1_x_at Naringenin-CHS; Type III 1115.m00011 0.119 1.0E-01 0.0E+00 2.235 9.3E-03 2.1E-70 polyketide synthase Mtr.44985.1.S1_at ANR TC98546 0.166 2.8E-02 4.2E-28 2.334 3.7E-01 2.5E-54 Mtr.50541.1.S1_at MtPAR gene, Myb, DNA-binding; 1054.m00009 0.197 2.9E-02 9.3E-16 104.692 4.6E-04 0.0E+00 homeodomain-like Mtr.28714.1.S1_at Homolog to PRF|1609233A| BI311259 0.215 2.8E-02 4.2E-73 2.333 9.6E-02 5.3E-181 226868|1609233A CHS 3 Mtr.10917.1.S1_at Similar to UP|C773_SOYBN TC108343 0.367 1.2E-02 3.0E-109 4.242 1.1E-01 0.0E+00 (O48928) cytochrome P450 77A3 Mtr.26465.1.S1_s_at Similar to UP|PEAM_SPIOL 1520.m00027 0.421 3.4E-04 0.0E+00 2.180 1.1E-01 4.8E-04 (Q9M571) phosphoethanolamine N-methyltransferase Mtr.37221.1.S1_at Homolog to UP|Q43437 TC100154 0.477 3.8E-02 5.7E-07 3.033 3.7E-01 6.1E-150 photosystem II type I chlorophyll a/b-binding protein precursor

Affymetrix ID, putative annotation, TC, expression ratios between par mutant vs. WT or overexpressing lines vs. control with their respective P and Q values indicated. Significant P values are indicated in bold.

1768 | www.pnas.org/cgi/doi/10.1073/pnas.1120916109 Verdier et al. Downloaded by guest on September 27, 2021 higher (by 23.2%) in par mutants than in WT controls (Fig. 4C). expression in par mutants by qRT-PCR. Significantly, WD40-1 The reduced amount of epicatechin in the par mutants mirrored transcript levels were between 15 and 50 times lower in the dif- the reduced level of soluble PAs (Fig. 2D). ferent par mutant seeds than in WT sibling seed controls at 20 DAP Together, the results of genetic, transcriptomic and metab- (Fig. 5A). In contrast, PAR transcript levels were unaffected by olomic analyses indicate that MtPAR functions as a positive mutations in the WD40-1 gene in 16 DAP seeds (15). Furthermore, regulator of PA biosynthesis in M. truncatula. MtPAR overexpression induced MtWD40-1 expression in Medicago hairy roots (Fig. 5B). Direct interactions between PAR and WD40- MtPAR Acts Upstream of WD40-1. Previously, a WD40 repeat pro- 1 and ANR promoters were investigated by using yeast one-hybrid tein, orthologous to Arabidopsis TTG1, was identified in M. trun- assay (Fig. 5C). MtPAR activated transcription from the WD40-1 catula and called MtWD40-1 (15). M. truncatula wd40-1 mutants promoter, but not the ANR promoter (Fig. 5C). display a drastic decrease of soluble and insoluble PA (15). How- ever, overexpression of MtWD40-1 in M. truncatula hairy roots Ectopic Expression of MtPAR Induces PA Accumulation in Alfalfa results in an increase of anthocyanin concentration without af- Leaves. To investigate whether MtPAR can be used to engineer fecting PA concentration. We compared published transcriptome PA production in alfalfa leaves, we generated >300 transgenic data of wd40-1 mutants (15) to those of par mutants. Out of the 38 alfalfa lines overexpressing MtPAR. Driven by the constitutive 35S genes that were down-regulated in 20 DAP seeds of par mutants promoter, MtPAR expression levels in leaves and stems were compared with WT controls, 16 were also down-regulated in wd40- significantly higher in almost half of the tested transgenic lines 1 mutant seeds collected at 16 DAP (Table S4). Moreover, 14 out than in the controls. Analysis of PAs from extracts of leaves and of the 16 are related to flavonoid biosynthesis according to Gen- stems with the DMACA-based assay indicated that low levels of ebins ontology (see complete list and annotations in Table S4). To PAs accumulated in these transgenic lines. PA production was – test whether MtPAR and MtWD40-1 act via a common regulatory positively correlated with MtPAR expression level (Fig. S4 A D). fi pathway to induce target genes, we measured WD40-1 gene The presence of PAs was further con rmed by using tandem MS. MS ion peak signals for epicatechin (M-H+, 289) and epicatechin hexoside (M-H+, 450.8), as well as epicatechin dimer (M-H+, 577), were present in extracts from transgenic alfalfa leaves ec- topically expressing MtPAR, whereas no signals were detected in extracts from WT alfalfa (Fig. S4E). Discussion MtPAR Is a Positive Regulator of PA Biosynthesis. We have charac- terized MtPAR, a member of the MYB R2R3 TF family, in M. truncatula. Its closest known homolog is an MYB protein of soybean that remains uncharacterized. Our data indicate that MtPAR plays a specific role in the regulation of PA biosynthesis in Medicago seeds. First, MtPAR gene expression is confined to the seed coat, the site of PA accumulation in developing seed. Sec- ond, loss-of-function (Tnt1 insertion) par mutants accumulate PLANT BIOLOGY

Fig. 4. Role of MtPAR in regulation of the flavonoid pathway. (A) Sche- matic representation of the flavonoid biosynthetic pathway leading to PAs and anthocyanins. (*) represents enzymes for which transcript levels are significantly affected in both mutant lines and overexpressing trans- formants. PAL, L-phenylalanine ammonia-lyase; C4H, cinnamate 4-hydroxy- lase; 4CL, 4-coumarate CoA ligase; CHI, chalcone isomerase; FLS, flavonol synthase; DFR, dihydroflavonol 4-reductase; LAR, leucoanthocyanidin re- ductase; OMT, O-methyltransferase; UFGT, UDP flavonoid glucosyl trans- ferase; RT, rhamnosyl transferase. (B) Cumulative expression values of different probe sets encoding putative genes involved in flavonoid bio- synthesis in par mutant seeds at 20 DAP and in hairy roots overexpressing MtPAR. Averages of the three biological replicates are indicated with their Fig. 5. PAR activates WD40-1 expression. (A and B) Transcript levels of respective SD. PAL gene expression is constituted by cumulative expression MtWD40-1 in seeds (20 DAP) from different par lines (A) and in hairy root of 8 different probe sets encoding for putative PAL enzyme; 4CL by 17 probe transformants (B). Relative expression was calculated from qRT-PCR data sets; CHS, 31 probe sets; CHI, 10 probe sets; F3H, 12 probe sets; FLS, 3 probe with respect to the geometric average of transcript levels of two house- sets; DFR, 4 probe sets; LAR, 1 probe set; ANS, 3 probe sets; ANR, 2 probe keeping genes, MSC27 and PDF2.(C) Yeast one-hybrid analysis. (Top) The sets; and GT (UGT72L1), 6 probe sets. Different probe set IDs for each gene reporter-expressing cassette used in yeast one-hybrid assays. The promoters are indicated in SI Materials and Methods. Expression values for WT were of MtWD40-1 (27 to −725) and MtANR (−1to−643) were cloned in front of normalized against respective control lines and adjusted to 1. (C) Statistically the reporter gene AUR1-C, an antibiotic resistance gene that confers Aur- significant changes in flavonoid content in mature par seeds expressed rel- eobasidin A (AbA) resistance in yeast. (Middle) No basal expression of WD40- ative to the concentration in segregant WT lines. All other flavonoid content 1 promoter and ANR promoter was detected in yeast. (Bottom) Yeast changes are indicated in Table S3. Averages of the three biological replicates growth assays after the Y1H reporter strains were transformed with plas- are indicated with their respective SD. mids carrying cassettes constitutively expressing PAR effector.

Verdier et al. PNAS | January 31, 2012 | vol. 109 | no. 5 | 1769 Downloaded by guest on September 27, 2021 substantially less PA in the seed coat than do WT controls. Third, MtPAR expression and would, therefore, fail to induce PA anthocyanin levels remain normal in par mutant seed despite the biosynthesis. existence of a common pathway that generates precursors for PA and anthocyanin biosynthesis. Fourth, no aberrant phenotype MtPAR as a Tool for Bioengineering PA Content in Legumes. Bio- apart from pale seed color is evident in any organ of par mutants. technological solutions are being sought to enhance PA levels in Fifth, ectopic expression of MtPAR in roots leads to production of legume shoots to reduce pasture bloat. TFs are attractive in this soluble PAs in an organ that normally does not accumulate PAs. respect because they have the potential to activate whole meta- Sixth, genes that are repressed in the seed of par mutants and bolic pathways. For example, ectopic expression of PAP1 led to induced in MtPAR overexpressing roots are largely involved in anthocyanin production in tobacco and Arabidopsis (31, 32), as flavonoid and PA biosynthesis. These putative target genes of did LAP1 in alfalfa (18). However, similar approaches to in- MtPAR protein activity include three CHS genes, two F3H genes, crease PA levels in legume foliage have not yet succeeded. and one ANS gene, which are required for both PA and antho- MtPAR offers promise for bioengineering of PA content in cyanin production, as well as the ANR gene, which is required for crop and pasture species. Indeed, ectopic overexpression of PA production alone. MtPAR led to massive accumulation of PAs in M. truncatula It is intriguing that anthocyanin levels were unaffected, whereas hairy roots, as a result of increases in the expression of multiple PA levels were substantially reduced in par mutant seed compared genes involved in PA/anthocyanin biosynthesis, including CHS, with the WT, given that many of the genes required for both PA F3H, ANS, and ANR. This key result shows that MtPAR ex- and anthocyanin biosynthesis were repressed in the mutant. pression is sufficient to induce PA accumulation in organs other Several hypotheses could explain these results. First, if MtPAR is than the seed where it functions normally, although the TF not expressed in cells accumulating anthocyanins, the decrease of fl activity of MtPAR may require additional proteins, such as PA could be counterbalanced by the increase of avonoid glyco- MtWD40-1. Stable, constitutive overexpression of MtPAR in sides and not anthocyanins. Another explanation could involve alfalfa resulted in accumulation of PA in shoots, albeit to a level metabolic channeling. If ANS and ANR are physically coupled, substantially below that needed to provide effective bloat pro- the product of ANS activity, 3-OH-anthocyanidin, would be tection (∼20 mg/g dry weight). Nonetheless, this finding repre- converted preferentially to epicatechin (and ultimately PA) by sents an advance in PA pathway engineering in legumes that ANR, rather than being glycosylated for anthocyanin production; theoretical decoupling of ANS and ANR could result in a relative may lead to improvement of alfalfa to prevent pasture bloat in increase in anthocyanin over epicatechin. the future. Materials and Methods MtPAR Activates MtWD40 in the Regulatory Network Controlling PA fi fi Biosynthesis. fl Gene Identi cation and Distance Analysis. The MtPAR gene was identi ed by Transcriptional regulation of avonoid biosynthesis is fi fi poorly understood in legumes. In the nonlegume Arabidopsis,six its seed-speci c expression pro le (probe set ID Mtr.50541.1.S1_at) by using the MtGEA Web server (http://mtgea.noble.org/v2/; refs. 16 and 33). Align- loci are known to have regulatory functions in PA biosynthesis, ment of the deduced amino acid sequences of MtPAR and other proteins of TT1, TT2, TT8, TT16, TTG1,andTTG2 (6). TT1 and TT16 encode fi the MYB R2R3 family was carried out by using ClustalW in the Geneious a zinc nger and a MADS box protein, respectively, and are es- software suite (Biomatters). The tree based on distance was built by using sential for seed pigmentation (26, 27). TTG2 encodes a WRKY a neighbor-joining algorithm with 100 bootstrap replicates. The R2R3 do- TF, which acts downstream of TTG1 (28). TT2, TT8, and TTG1 main of each MYB factor was identified by using the PFAM protein family encode a MYB (9), a bHLH (29), and a WD40 protein (30), re- database (34). spectively, which interact to form a ternary TF complex. Mutation in any one of these TFs affects PA concentrations in seeds via Insertional Mutant Screening. Generation of the M. truncatula Tnt1 in- down-regulation of flavonoid biosynthetic genes (8). In the Medi- sertional mutant population and growth of R1 seeds were done as described cago par mutants, we also observed down-regulation of key genes (21). Reverse genetic screening for Tnt1 retrotransposon insertions in MtPAR of the flavonoid pathway. was performed by using a nested PCR approach (35). PCR products from fi fi Little is known about regulation of PA biosynthetic genes in target mutant lines were puri ed with QIAquick PCR puri cation kit (Qia- fi Medicago. A single WD40-repeat TF, MtWD40-1, orthologous to gen) and sequenced by using Tnt1 primers to con rm insertions in MtPAR. Arabidopsis AtTTG1, was identified as a positive regulator of PA Analysis of PAs, Anthocyanins, and Flavonoids. DMACA staining was used to biosynthesis in M. truncatula seeds (15). We compared the action evaluate qualitative changes in the PA contents of mature seeds. Seeds were of MtWD40-1 and MtPAR in M. truncatula, and it appears that stained overnight and destained in ethanol for observation. Extraction and both genes may belong to the same regulatory network. Both analysis of flavonoids from seeds and hairy roots of M. truncatula by UV mutants exhibit a substantial decrease of PA levels in seed (Fig. spectroscopy, DMACA staining, and reverse- or normal- phase HPLC coupled 2D; ref. 15). Furthermore, transcriptomic analysis revealed that to postcolumn DMACA derivatization, UV diode array detection, or MS are a common set of genes was down-regulated in mutants defective described in SI Materials and Methods. in these genes. Gene expression analysis also revealed a decrease of MtWD40-1 gene expression in par mutant lines, suggesting that Ectopic Expression of PAR in Medicago Hairy Roots and Alfalfa Plants. Cloning MtPAR regulates MtWD40-1 expression. The converse was not of MtPAR and expression in transgenic hairy roots of M. truncatula and plants the case, as MtPAR expression was not affected in wd40-1 mutants of alfalfa were performed as described in SI Materials and Methods. (15). A yeast one-hybrid assay revealed a direct interaction be- tween PAR and the WD40-1-promoter in yeast, but no direct RNA Extraction, Microarray, and qRT-PCR Analysis. Total RNA was isolated from fi interaction with the ANR promoter, which implies that, in planta, developing seeds (20 DAP) by using a modi ed cetyltrimethylammonium fl – bromide method (36) and from hairy roots and alfalfa shoots by using TRIzol PAR regulates genes encoding enzymes of the avonoid PA ’ pathway via its direct role on WD40-1 expression. This finding may reagent, according to the manufacturer s instructions (Invitrogen). Ten micrograms of total RNA from each sample was DNase treated (Turbo DNase; explain why ectopic expression of MtPAR, but not of MtWD40-1, Ambion) and purified (RNeasy MinElute Cleanup kit; Qiagen), according to resulted in PA biosynthesis in roots (Fig. 3; ref. 15). If a complex ’ fl manufacturer s instructions. Microarray and qRT-PCR analyses were as de- of TFs is required to induce avonoid biosynthesis genes in scribed in SI Materials and Methods. PCR primers are given in Table S5. Medicago, as is the case in Arabidopsis (11), then ectopic expres- sion of MtPAR, and consequent induction of MtWD40-1, may Yeast One-Hybrid Assay. Yeast one-hybrid assay was performed by using the have provided the requisite TFs for PA biosynthesis in roots. In Matchmaker Gold Yeast One-Hybrid System (Clontech), as described in SI contrast, ectopic expression of MtWD40-1 would not induce Materials and Methods.

1770 | www.pnas.org/cgi/doi/10.1073/pnas.1120916109 Verdier et al. Downloaded by guest on September 27, 2021 ACKNOWLEDGMENTS. We thank Yuhong Tang and Stacy Allen for assis- Department of Agriculture (USDA)/National Institute of Food and Agricul- tance with Affymetrix data analysis; David Huhman and Lloyd Sumner for ture Grant 2010-65115-20373; USDA/Cooperative State Research, Education, UPLC-MS analysis; Pascal Ratet and Million Tadege for the construction of and Extension Service–National Research Initiative Plant Genome Program the Tnt1 mutant population; and Jiangqi Wen and Xiaofei Cheng for the Project Grant 2006-35300-17143; Forage Genetics International; and the screening of this population for par mutants. This work was supported by US Samuel Roberts Noble Foundation.

1. Dixon RA, Xie DY, Sharma SB (2005) Proanthocyanidins—a final frontier in flavonoid 19. Mathews H, et al. (2003) Activation tagging in tomato identifies a transcriptional research? New Phytol 165:9–28. regulator of anthocyanin biosynthesis, modification, and transport. Plant Cell 15: 2. Serafini M, et al. (2003) Plasma antioxidants from chocolate. Nature 424:1013. 1689–1703. 3. Ahmad N, Gupta S, Mukhtar H (2000) Green tea polyphenol epigallocatechin-3-gal- 20. Terrier N, et al. (2009) Ectopic expression of VvMybPA2 promotes proanthocyanidin late differentially modulates nuclear factor kappaB in cancer cells versus normal cells. biosynthesis in grapevine and suggests additional targets in the pathway. Plant – Arch Biochem Biophys 376:338–346. Physiol 149:1028 1041. 4. Noreen Y, Serrano G, Perera P, Bohlin L (1998) Flavan-3-ols isolated from some me- 21. Tadege M, et al. (2008) Large-scale insertional mutagenesis using the Tnt1 retro- – dicinal plants inhibiting COX-1 and COX-2 catalysed prostaglandin biosynthesis. transposon in the model legume Medicago truncatula. Plant J 54:335 347. 22. Odell JT, Nagy F, Chua N-H (1985) Identification of DNA sequences required for ac- Planta Med 64:520–524. tivity of the cauliflower mosaic virus 35S promoter. Nature 313:810–812. 5. Aerts RJ, Barry TN, McNabb WC (1999) Polyphenols and agriculture: Beneficial effects 23. Quandt H, Puhler A, Broer I (1993) Transgenic root nodules of Vicia hirsuta: A fast and of proanthocyanidins in forages. Agric Ecosyst Environ 75:1–12. efficient system for the study of gene expression in indeterminate-type nodules. Mol 6. Lepiniec L, et al. (2006) Genetics and biochemistry of seed flavonoids. Annu Rev Plant Plant Microbe Interact 6:699–706. Biol 57:405–430. 24. Jefferson RA, Kavanagh TA, Bevan MW (1987) GUS fusions: β-glucuronidase as 7. Abrahams S, Tanner GJ, Larkin PJ, Ashton AR (2002) Identification and biochemical a sensitive and versatile gene fusion marker in higher plants. EMBO J 6:3901–3907. characterization of mutants in the proanthocyanidin pathway in Arabidopsis. Plant 25. Goffard N, Weiller G (2007) GeneBins: A database for classifying gene expression – Physiol 130:561 576. data, with application to plant genome arrays. BMC Bioinformatics 8:87. 8. Debeaujon I, et al. (2007) Seed coat development and dormancy. Seed Development, 26. Nesi N, et al. (2002) The TRANSPARENT TESTA16 locus encodes the ARABIDOPSIS Dormancy and Germination, eds Bradford K, Nonogaki H (Blackwell, Oxford), pp BSISTER MADS domain protein and is required for proper development and pig- 25–49. mentation of the seed coat. Plant Cell 14:2463–2479. 9. Nesi N, Jond C, Debeaujon I, Caboche M, Lepiniec L (2001) The Arabidopsis TT2 gene 27. Sagasser M, Lu G-H, Hahlbrock K, Weisshaar B (2002) A. thaliana TRANSPARENT encodes an R2R3 MYB domain protein that acts as a key determinant for proantho- TESTA 1 is involved in seed coat development and defines the WIP subfamily of plant cyanidin accumulation in developing seed. Plant Cell 13:2099–2114. zinc finger proteins. Genes Dev 16:138–149. 10. Debeaujon I, et al. (2003) Proanthocyanidin-accumulating cells in Arabidopsis testa: 28. Johnson CS, Kolevski B, Smyth DR (2002) TRANSPARENT TESTA GLABRA2, a trichome regulation of differentiation and role in seed development. Plant Cell 15:2514–2531. and seed coat development gene of Arabidopsis, encodes a WRKY transcription 11. Baudry A, et al. (2004) TT2, TT8, and TTG1 synergistically specify the expression of factor. Plant Cell 14:1359–1375. BANYULS and proanthocyanidin biosynthesis in Arabidopsis thaliana. Plant J 39: 29. Nesi N, et al. (2000) The TT8 gene encodes a basic helix-loop-helix domain protein 366–380. required for expression of DFR and BAN genes in Arabidopsis siliques. Plant Cell 12: – 12. Xie DY, Sharma SB, Dixon RA (2004) Anthocyanidin reductases from Medicago 1863 1878. truncatula and Arabidopsis thaliana. Arch Biochem Biophys 422:91–102. 30. Walker AR, et al. (1999) The TRANSPARENT TESTA GLABRA1 locus, which regulates 13. Pang Y, Peel GJ, Wright E, Wang Z, Dixon RA (2007) Early steps in proanthocyanidin trichome differentiation and anthocyanin biosynthesis in Arabidopsis, encodes a WD40 repeat protein. Plant Cell 11:1337–1350. biosynthesis in the model legume Medicago truncatula. Plant Physiol 145:601–615. 31. Borevitz JO, Xia Y, Blount J, Dixon RA, Lamb C (2000) Activation tagging identifies 14. Zhao J, Dixon RA (2009) MATE transporters facilitate vacuolar uptake of epicatechin a conserved MYB regulator of phenylpropanoid biosynthesis. Plant Cell 12: 3′-O-glucoside for proanthocyanidin biosynthesis in Medicago truncatula and Arabi- 2383–2394. dopsis. Plant Cell 21:2323–2340. 32. Xie D-Y, Sharma SB, Wright E, Wang ZY, Dixon RA (2006) Metabolic engineering of 15. Pang Y, et al. (2009) A WD40 repeat protein from Medicago truncatula is necessary proanthocyanidins through co-expression of anthocyanidin reductase and the PAP1 for tissue-specific anthocyanin and proanthocyanidin biosynthesis but not for tri- MYB transcription factor. Plant J 45:895–907. – chome development. Plant Physiol 151:1114 1129. 33. He J, et al. (2009) The Medicago truncatula gene expression atlas web server. BMC 16. Benedito VA, et al. (2008) A gene expression atlas of the model legume Medicago Bioinformatics 10:441. – truncatula. Plant J 55:504 513. 34. Bateman A, et al. (2004) The Pfam protein families database. Nucleic Acids Res 32 17. Dubos C, et al. (2010) MYB transcription factors in Arabidopsis. Trends Plant Sci 15: (Database issue):D138–D141. 573–581. 35. Cheng X, Wen J, Tadege M, Ratet P, Mysore KS (2011) Reverse genetics in medicago 18. Peel GJ, Pang Y, Modolo LV, Dixon RA (2009) The LAP1 MYB transcription factor truncatula using Tnt1 insertion mutants. Methods Mol Biol 678:179–190. orchestrates anthocyanidin biosynthesis and glycosylation in Medicago. Plant J 59: 36. Verdier J, et al. (2008) Gene expression profiling of M. truncatula transcription factors 136–149. identifies putative regulators of grain legume seed filling. Plant Mol Biol 67:567–580. PLANT BIOLOGY

Verdier et al. PNAS | January 31, 2012 | vol. 109 | no. 5 | 1771 Downloaded by guest on September 27, 2021