Abstract

Introduction

Environmental stresses cause significant loss of plant productivity. Abiotic stresses, e.g. light, UV, temperature, drought, soil salinity, air pollution, and mechanical damage, directly affect crop production (Vickers et al., 2009). For example, the US drought of 2012, which was the most severe in at least 25 years, had a serious impact on production and yields of corn and soybean, leading to economic damage (Gilbert, 2012). This observation strongly suggests that a better understanding of the stress response is critically important for agricultural and economic performance to improve the tolerance of crops against such environmental stresses using phytochemical genomics and crop breeding (Varshney et al., 2012).

Flavonoids are a representative group of secondary metabolites that have been recently called ‘specialized metabolites’ because plants synthesize species-specific metabolites (Harborne and Williams, 2000). A series of flavonoid subgroups such as flavonols, (iso)flavones, flavanones, flavan-3-ols, proanthocyanidins, and anthocyanins has been recognized in over 500 food items by the US Department of Agriculture as dietary flavonoids with consistent evidence of beneficial effects in humans. Such metabolites are expected to be useful for crop breeding because they play a role in allelopathy, which has positive or negative effects on the growth of target organisms such as fungi, insects, and plants (Harborne and Williams, 2000; Bais et al., 2003; Field et al., 2006).

Biotic and abiotic environmental stresses induce the accumulation of flavonoids in plants (Dixon and Paiva, 1995). Under the influence of abiotic stresses such as phosphate, nitrogen, light, temperature, UV, and drought, flavonoids accumulate in the model plant Arabidopsis thaliana (Arabidopsis) (Rowan et al., 2009; Stracke et al., 2010a; Catala et al., 2011; Koops et al., 2011; Kusano et al., 2011; Lei et al., 2011; Tohge et al., 2013). A common factor with abiotic stress is the accumulation of reactive oxygen species (ROS) such as hydroxyl radicals ·OH, superoxide radicals , and hydrogen peroxide H₂O₂. Therefore, it can be questioned whether flavonoids effectively scavenge ROS. Arabidopsis can be utilized to address this question as its flavonoid biosynthetic pathway has been elucidated at the metabolite and gene levels. Fifty-four flavonoids accumulate in an organ-specific manner in Arabidopsis (Saito et al., 2013), and 35 genes encoding biosynthetic enzymes or transcription factors are responsible for aglycone formation and subsequent tailoring modifications to produce these flavonols and anthocyanins (Yonekura-Sakakibara et al., 2008). The flavonol regulators MYB12/PFG1 (PRODUCTION OF FLAVONOL GLYCOSIDES1), MYB11/PFG2, and MYB111/PFG3 regulate expression of the four early biosynthetic genes chalcone synthase (CHS), chalcone isomerase (CHI), flavanone 3-hydroxylase (F3H), flavonoid 3′-hydroxylase (F3′H), and flavonol synthase (FLS) (Stracke et al., 2007, 2010b), whereas the anthocyanin regulators MYB75/PAP1 (PRODUCTION OF ANTHOCYANIN PIGMENT1), MYB90/PAP2, MYB113, and MYB114 control the expression of the late biosynthetic genes dihydroflavonol reductase (DFR) and leucoanthocyanidin dioxygenase/anthocyanidin synthase (LDOX/ANS) (Borevitz et al., 2000; Gonzalez et al., 2008; Dubos et al., 2010).

Manipulating the expression of flavonoid biosynthetic genes alters the accumulation of flavonoids in Arabidopsis and other plants (Shirley et al., 1995; Borevitz et al., 2000; Butelli et al., 2008, 2012). Subsequent omics profiling using these genetically engineered plants has led to the identification of new functions of biosynthetic genes and the clarification of biosynthetic pathways (Fukushima et al., 2009; Saito and Matsuda, 2010). Although a few reports have discussed the interaction between flavonoid accumulation and tolerance to abiotic stresses such as UV-B irradiation (Stracke et al., 2010a; Kusano et al., 2011), flavonoid function remains to be determined experimentally. In the present study, we report flavonoids as in vitro ROS scavengers that lead to enhanced oxidative and drought tolerance in vivo using metabolome and transcriptome profiling in the wild type, MYB12/PAP1 overexpressors, transparent testa4 (tt4) as a flavonoid-deficient mutant, and flavonoid-deficient MYB12 or PAP1 overexpressors.

Results

Visible phenotype is conferred by co-expression of the flavonoid-specific transcription factors MYB12 and PAP1

The WOX1 (double overexpressor1 of both MYB12 and PAP1) lines were prepared by crossing MYB12OX (MYB12 overexpressor) (Mehrtens et al., 2005) and pap1-D (production of anthocyanin pigment1-Dominant) (Borevitz et al., 2000) to address the function of flavonoids during environmental stress (Figure​(Figure1a).1a). Both MYB12 and PAP1 were concurrently overexpressed in WOX1 lines (Figure​(Figure1b).1b). MYB12/PAP1 overexpression did not result in significant differences in plant size in the MYB12OX, pap1-D, or WOX1 lines compared with that in the wild type Columbia-0 (Col-0) under our growth conditions (Figure S1 in Supporting Information). During the seedling stage, the pap1-D and WOX1 lines exhibited purple pigmentation due to the accumulation of anthocyanin in petioles, hypocotyls, and roots. In contrast, the MYB12OX line had apparently the same phenotype as that of the wild type (Figure​(Figure1a).1a). Purple pigmentation was observed in the adult stage, particularly in the veins, petioles, and stems of the pap1-D and WOX1 lines (Figure S1).

Open in a separate window

Figure 1

Phenotypes of the wild type, MYB12OX, pap1-D, and WOX1 lines.(a) Phenotypes of seeds and aerial parts. The red arrows indicates coloration from anthocyanin accumulation in 3-week-old plants.(b) Steady-state MYB12 and PAP1 mRNA levels by real-time PCR. The aerial parts of 3-week-old plants grown on growth medium were used in this analysis. Steady-state MYB12 and PAP1 mRNA levels in MYB12OX and pap1-D were both considered as 100%. Standard deviations (error bars) were calculated from the results of three biological replicates.

The quantity and diversity of flavonoids is increased by overexpressing MYB12/PAP1

The dominant flavonoids in MYB12OX and pap1-D are flavonol glycosides and anthocyanins (Mehrtens et al., 2005; Tohge et al., 2005a; Stracke et al., 2010b). Flavonoid-targeted profiling was performed by liquid chromatography photodiode array mass spectrometry (LC PDA MS) to determine the differences in the quantity and profiles of flavonols and anthocyanins between the wild type and MYB overexpressors. The LC PDA MS procedure provided UV spectra for deciphering the peaks of flavonol glycosides and anthocyanins in addition to MS spectra (Yonekura-Sakakibara et al., 2008). The level of total anthocyanins in all parts increased significantly in the pap1-D and WOX1 lines (approximately a 20-fold increase), whereas that of total flavonols increased slightly in the MYB12OX and WOX1 lines compared with that in the wild type (approximately a two-fold increase) (Figure​(Figure2a2a and Figure S2). The WOX1 lines accumulated enhanced levels of the glycosides of kaempferol, quercetin, and cyanidin compared with each single overexpressed line (Figure S3).

Open in a separate window

Figure 2

Flavonoid contents of the wild type and MYB overexpressors grown under general conditions.(a) Total flavonols and anthocyanins in each part of 3-week-old plants. Differences between the wild-type and other plants analyzed using Student's t-test were statistically significant (*P < 0.05). Differences and standard deviations (error bars) were calculated from the results of three biological replicates. PH, petiole and hypocotyl; N.D., not detected.(b) Representative chromatogram of the wild type and WOX1-1 at 520 nm. The black line chromatogram represents the wild type, and the purple line chromatogram represents WOX1-1. See Tables S1 and S2 for details of the detected peaks. Such anthocyanins were also detected in pap1-D. L, leaf; PH, petiole and hypocotyl; R, root.(c) Structure of A11.

Twenty-three anthocyanins (A18–A40), probably including isomers of the major anthocyanins (Rowan et al., 2009), were assigned as minor anthocyanins in the pap1-D and WOX1 lines during profiling (Figure​(Figure2b).2b). In addition to the major anthocyanin A11 (Figure​(Figure2c)2c) and its intermediates (A1–A10) (Table S1) (Nakabayashi et al., 2009), most of the minor anthocyanins accumulated in petioles and hypocotyls (PH). Of those, the five minor anthocyanins (A21, A22, A28, A30, and A39) with caffeoyl or feruloyl moieties (Table S2) were assigned on the basis of UV (approximately λ_max 520 nm) and MS spectra (difference and fragment ion). This metabolic plasticity revealed the dynamic nature of specialized metabolism in plants, thus suggesting an adaptive response to biosynthetic diversity.

Flavonol overaccumulation in the MYB12OX and WOX1 lines was not prominent (Figure​(Figure2a).2a). Thus, the effect of MYB12 overexpression on flavonol accumulation was examined under light-shielding conditions using black paper (Figure S4). The effects of MYB12 overexpression were more clearly observed by comparing the wild type and MYB overexpressors grown under the light-shielding conditions because flavonols and anthocyanins did not accumulate in the roots of Arabidopsis wild-type plants grown under similar light-shielding and soil conditions (Hemm et al., 2004). Growth under the light-shielding conditions revealed that flavonols accumulated significantly in all parts of the MYB12OX and WOX1 lines compared with the wild type (greater than seven-fold increase) (Figure​(Figure3a).3a). Accumulation of major flavonols (f1, f2, f3, f5, f6, and f8) increased significantly in the hypocotyls and transient zone (HTz) and roots of the MYB12OX and WOX1 lines (Figure​(Figure3b,c).3b,c). This result indicates that MYB12 overexpression increases flavonol biosynthesis systemically in the plants.

Open in a separate window

Figure 3

Flavonoid contents of the wild type and MYB overexpressors grown under the light-shielding condition.(a) Total flavonols and anthocyanins in each part of 2-week-old plants. Differences between the wild type and other plants analyzed using Student's t-test were statistically significant (*P < 0.05). Differences and standard deviations (error bars) were calculated from the results of three biological replicates. HTz, hypocotyl and transient zone; N.D., not detected.(b) Representative chromatogram of the wild type and WOX1-1 at 340 nm. Details of the detected peaks are presented in Table S1.(c) Structure of flavonols.

Stress-related gene expression is unaffected by increased flavonoid biosynthesis

We further prepared a tt4 mutant line lacking CHS (Shirley et al., 1995) and the flavonoid-deficient MYB overexpressors MYB12OX/tt4 and pap1-D/tt4 that were obtained by crossing MYB12OX or pap1-D with tt4, respectively (Figure​(Figure4).4). This was done to explore the effect of MYB overexpression on the accumulation of a wide range of metabolites including flavonoids by untargeted metabolome profiling of whole plants using LC quadrupole time-of-flight MS (LC QTOF MS) (Matsuda et al., 2010). Principal component analysis (PCA) of the processed data suggested that accumulation of flavonols and anthocyanins was the main contributor to the separation in the scatterplot (Figure​(Figure5a).5a). The results of hierarchical cluster analysis (HCA) revealed that MYB12 and PAP1 overexpression primarily induced overaccumulation of flavonols and anthocyanins (Figure​(Figure5b).5b). Flavonols and anthocyanins were not detected in tt4, MYB12OX/tt4, or pap1-D/tt4.

Open in a separate window

Figure 4

Phenotype of wild-type, tt4, the flavonoid-deficient MYB overexpressors, and the MYB overexpressors.Upper: Phenotype of seeds and 3-week-old plants.Lower: Relative MYB12 and PAP1 expression in the aerial parts of 3-week-old plants grown on GM plates. Steady-state MYB12 and PAP1 mRNA levels in MYB12OX and pap1-D were considered 100%, respectively. Standard deviations (error bars) were calculated from the results of 3 biological replicates.

Open in a separate window

Figure 5

Metabolome profiling using liquid chromatography quadrupole time-of-flight mass spectrometry.(a) Metabolo-type of the wild-type, tt4, flavonoid-deficient MYB overexpressors, and MYB overexpressors. Whole parts of 3-week-old plants grown on growth medium plates in a growth chamber were analyzed (three biological replicates). Processed data including 380 metabolite peaks after alignment, deisotoping, noise cutting, and normalization were used for principal component analysis (PCA) using SIMCA-P 11.5 software. The plot of principal component 1 (PC1, 35.1%) vs. principal component 2 (PC2, 17.8%) is presented. Gray and brown dotted circles represent the flavonoid-deficient and flavonoid-overaccumulating plants, respectively. Col-0, black squares; tt4, orange circles; MYB12OX/tt4, green triangles; pap1-D/tt4, purple squares; MYB12OX, red circles; pap1-D, blue diamonds; WOX1-1, pink crosses; and WOX1-2, yellow triangles. (b) Hierarchical cluster analysis using the processed metabolome data. Pearson's correlation analysis was used. Relative log2-transformed values of signal intensities were calculated for the wild-type. The single arrow indicates the cluster of flavonols, and the double arrow indicates the cluster of anthocyanins. Lower: Enlarged view of the flavonol and anthocyanin clusters. Bold indicates assigned monoisotopic ion, and asterisk indicates flavonoid-derived ion.

Transcriptome profiling using DNA microarrays of the aerial parts of the wild type and other plants was performed to examine whether MYB overexpression or flavonoid overaccumulation were responsible for the expression of stress-related genes involved in both the biotic and abiotic stress response (Figure​(Figure6,6, upper part). Profiling revealed that the expression of 907 genes was significantly induced in the mutants compared with the wild type (P < 0.05; log₂ fold change ≥1.0) (Dataset S1). Of these genes, 245 were categorized into the Gene Ontology term ‘response to biotic or abiotic stimulus’ in the Gene Ontology annotations of TAIR10. However, of those, only six genes (A5GlcMalT, F3H, A5GlcT, TT19, DFR, and A3G2″XylT) were commonly induced in the PAP1-overexpressing pap1-D/tt4, pap1-D, and WOX1 lines, whereas no stress-related genes were induced in the anthocyanin-overaccumulating pap1-D and WOX1 lines (Figure​(Figure6,6, lower part). No stress-related genes were commonly induced in the MYB12-overexpressing MYB12OX/tt4, MYB12OX, and WOX1 lines or in the flavonol-overaccumulating MYB12OX and WOX1 lines. In addition, untargeted profiling using the 907 genes was performed to investigate the expression of genes annotated as ‘unknown’. The expression of six additional genes was identified as commonly altered in the PAP1-overexpressing lines (Dataset S1). However, those genes, including three flavonoid biosynthetic genes, appeared to be unrelated to the improvement of stress tolerance. These results suggest that enhanced MYB expression and flavonoid accumulation have no effect on the expression of stress-related genes.

Open in a separate window

Figure 6

Transcriptome profiling using a DNA microarray.The processed transcriptome data including 245 stress-related genes were used for hierarchical cluster analysis. Pearson's correlation analysis was used. Relative log₂-transformed signal intensity values were calculated for the wild type. The triple arrow indicates the cluster of flavonoid-related biosynthetic genes. The lower part shows an enlarged view of the flavonoid gene cluster. F3H, flavanone 3-hydroxylase (At3 g51240); A5GlcT, anthocyanin 5-O-glucosyltransferase (At4 g14090); DFR, dihydroflavonol reductase (At5 g42800); A5GlcMalT, anthocyanin malonyltransferase (At3 g29590); GSTF12, glutathione S-transferase12 (At5 g17220); A3G2″XylT, anthocyanin 3-O-glucoside: 2″-O-xylosyltransferase (At5 g54060).

Enhanced tolerance of abiotic stress by overaccumulation of flavonols and anthocyanins

The radical scavenging activity of flavonoids found in Arabidopsis was assayed using 2,2-diphenyl-1-picrylhydrazyl (DPPH) to investigate the contribution of flavonoid accumulation to stress tolerance. The DPPH assay is an easy method for evaluating the activity of flavonoids on the basis of the established reaction mechanism of DPPH and flavonoids (Seyoum et al., 2006). Quercetin 3-O-glucoside (Q3G) and cyanidin 3-O-glucoside (C3G) exhibited approximately 10-fold higher activity than kaempferol 3-O-glucoside (Figure​(Figure7a),7a), suggesting that the highly modified flavonoid glycosides detected in this study have radical scavenging activities. Thus, the radical scavenging activity of the crude extract from each plant was examined. Extracts from the pap1-D and WOX1 lines exhibited greater than two-fold higher activity than those from the wild type (Figure​(Figure7b),7b), providing compelling evidence that the highly modified flavonoids in Arabidopsis are responsible for radical scavenging activity in vitro. All plants were exposed to oxidative stress using methyl viologen, which generates and accumulates ROS in plants (Fujita et al., 2012). As a result, the pap1-D and WOX1 lines exhibited a tolerance to oxidative stress three times greater than the wild-type (Figure​(Figure77c).

Open in a separate window

Figure 7

Enhancement of oxidative tolerance by flavonoid overaccumulation.(a) Radical scavenging activity of kaempferol 3-O-glucoside (K3G), quercetin 3-O-glucoside (Q3G), isorhamnetin 3-O-glucoside (I3G), and cyanidin 3-O-glucoside (C3G) against 2,2-diphenyl-1-picrylhydrazyl (DPPH). Trolox, which is a vitamin E analog, was the positive control. Differences and standard deviations (error bars) were calculated from the results of three replicates.(b) Relative radical scavenging activity of the acidic methanol–water extracts from the wild type, tt4, flavonoid-deficient MYB overexpressors, and MYB overexpressors against DPPH. Relative activity values for the wild type were calculated for this analysis. Differences between the wild type and other plants analyzed using Student's t-test were statistically significant (*P < 0.05). Differences and standard deviations (error bars) were calculated from the results of three biological replicates.(c) Relative chlorophyll levels in the wild type, tt4, flavonoid-deficient MYB overexpressors, and MYB overexpressors after oxidative stress using methyl viologen. Relative values of the wild-type chlorophyll level were calculated. Differences between the wild type and other plants analyzed using Student's t-test were statistically significant (*P < 0.05). Differences and standard deviations (error bars) were calculated from the results of three biological replicates.

The wild-type and mutants were exposed to drought stress, which also increases the production and accumulation of ROS (Rivero et al., 2007; Hernandez et al., 2009). The number of PAP1-overexpressing lines that survived was more than two times higher than that of the wild type, which survived stochastically (Figure​(Figure8a8a and Figure S6). Untargeted profiling using LC QTOF MS was performed using samples harvested after rewatering to check the contribution of anthocyanin accumulation to drought tolerance (Figure​(Figure8b).8b). Accumulation of the major anthocyanin A11 most prominently contributed to the separation of the pap1-D and WOX1 lines on PC1 in PCA (Figure​(Figure88c).

Open in a separate window

Figure 8

Enhancement of drought tolerance by flavonoid overaccumulation.(a) Phenotype of the wild type, tt4, flavonoid-deficient MYB overexpressors, and MYB overexpressors under drought stress. Top: three plants were grown on soil in one pot. Ten pots were prepared in each experiment. Middle: a difference in wilting was observed among the MYB overexpressors and other lines after 14 days; Bottom: all plants were rewatered after 16 days. Images were obtained, after repositioning the pots for photography, 3 days after rewatering. The total number of surviving plants is shown in the experiment with 30 plants. Differences in average numbers of surviving plants in the pots between the wild type and other plants analyzed using Student's t-test were statistically significant (*P < 0.05). The reproducible results were obtained from two independent experiments.(b) Metabolo-type of the samples after rewatering in the drought stress experiment. The freeze-dried samples were analyzed by untargeted profiling using liquid chromatography quadrupole time-of-flight mass spectroscopy. Processed data, including 2084 metabolite peaks after alignment, de-isotoping, noise cutting, and normalization were used for PCA using simca-P 11.5 software. The plot of PC1 (39.0%) versus PC2 (17.5%) is presented. Col-0, black squares; tt4, orange circles; MYB12OX/tt4, green triangles; pap1-D/tt4, purple squares; MYB12OX, red circles; pap1-D, blue diamonds; WOX1-1, pink crosses; and WOX1-2, yellow triangles.(c) Relative intensity of A11 after drought stress and rewatering. Differences of the intensities between the wild type and other plants were analyzed using Student's t-test (*P < 0.05). Differences and standard deviations (error bars) were calculated from the results of three biological replicates.(d) H₂O₂ generation in the wild type, tt4, flavonoid-deficient MYB overexpressors, and MYB overexpressors after drought stress. H₂O₂ production was detected by 3,3′-diaminobenzidine polymerization. The representative results are shown, and the reproducible results were obtained from three biological replicates.

To investigate the flavonoids in Arabidopsis that mitigate ROS in vivo, 3,3′-diaminobenzidine (DAB) staining, which is a marker of H₂O₂ accumulation in the presence of peroxidase (Thordal-Christensen et al., 1997), was performed using the detached leaves of 3-week-old plants incubated under drought conditions. The DAB staining patterns showed that greater anthocyanin accumulation mitigated H₂O₂ accumulation (Figure​(Figure88d).

The relative water content in whole plants and detached leaves of the MYB12OX, pap1-D, and WOX1 lines showed a slightly greater decreasing trend in the water loss experiments than that of the wild type (Figure​(Figure9a,b),9a,b), suggesting a positive correlation between flavonoid accumulation and the prevention of water loss.

Open in a separate window

Figure 9

Water loss in the wild type, MYB12OX, pap1-D, WOX1-1, and WOX1-2.(a) Water loss in plants grown on soil during drought stress. Five plants were grown in a pot. The weight of the pot was calculated for water loss. Differences and standard deviations (error bars) were calculated from the results of six biological replicates. Differences between the wild type and other plants were analyzed using Student's t-test. The star indicates statistical significance (P < 0.05) between the wild type and MYB12OX, pap1-D, and WOX.(b) Water loss in detached leaves (10 leaves) of 3-week-old plants. Differences and standard deviations (error bars) were calculated from the results of three biological replicates. Differences between the wild type and other plants were analyzed using Student's t-test. There was no statistically significant difference between the wild type and other plants.

Discussion

There is growing interest in the functions of specialized metabolites in plants because such functions may be beneficial for agriculture (Varshney et al., 2012). Metabolomics and transcriptomics are powerful tools for understanding cellular responses under a given condition and for identifying the functions of genes and metabolites. Many responses to environmental stresses are common among plant species, including transcriptional activation of stress-related genes for acquisition of stress tolerance, accumulation of compatible solutes to retain water in cells and of antioxidants to protect cells against environmental stresses. However, no direct experimental evidence links flavonoids and stress tolerance. In general, wild-type, flavonoid-deficient mutant(s) and flavonoid-specific MYB overexpressor(s) are useful for discussing stress tolerance due to flavonoid accumulation. However, the bottleneck in the experimental design is to precisely evaluate the effect of flavonoid accumulation or MYB overexpression in a given result. Therefore, a series of Arabidopsis lines overaccumulating and lacking flavonoids under MYB overexpression was required to identify flavonoid function. In this study, the effects of both MYB overexpression and flavonoid overaccumulation on oxidative and drought tolerance were distinguished by comparing the wild-type, tt4, the flavonoid-deficient MYB overexpressors MYB12OX/tt4 and pap1-D/tt4, the single MYB overexpressors MYB12OX and pap1-D, and the double MYB overexpressors WOX1-1 and WOX1-2. We showed that flavonoid accumulation, which results in radical scavenging activity in vitro, leads to the enhancement of oxidative and drought tolerance in vivo. We also showed that flavonoid accumulation tends to prevent water loss.

Data from metabolome, transcriptome, and plant hormone analyses under the unstressed growth condition suggested that overexpressing MYB and flavonoid overaccumulation had no effect on the ABA level (Figure S5), which is a representative plant hormone involved in signaling in abiotic stress responses and in the expression of ABA biosynthetic genes and stress-related genes excluding the six flavonoid biosynthetic genes (Figure​(Figure6).6). The MYB-overexpressing plants did not show enhanced drought tolerance through ABA-induced abiotic stress response mechanisms under the unstressed condition. EDL3, encoding an F-box protein, functions positively during ABA signaling and negatively during anthocyanin biosynthesis (Koops et al., 2011). EDL3 expression did not change significantly in the transcriptome and real-time PCR analyses, indicating that the stress signaling pathway including ABA and EDL3 for both the flavonoid biosynthetic pathway and abiotic stress response was not affected by the induction of flavonoid biosynthesis. Increased levels of the plant hormones ABA and cytokinins play a key role in drought tolerance in tobacco (Nicotiana tabacum) (Rivero et al., 2007), suggesting that the slight increase in cytokinin accumulation in the MYB overexpressors was unrelated to enhanced drought tolerance (Figure S5). The relationship between flavonoid and cytokinin biosynthesis remains to be clarified and further research is necessary.

Our results suggest that anthocyanin overaccumulation is directly associated with oxidative and drought tolerance, given that the anthocyanin-overaccumulating plants showed such tolerance whereas the flavonoid-deficient overexpressing plants exhibited apparently the same phenotype as the wild type and tt4 (Figures​(Figures7c7c and ​and8a).8a). Cyclic single electron reduction/oxidation of the methyl viologen reagent is a critical event (redox cycling). Redox cycling is linked to two important events related to the development of toxicity: depletion of NADPH and generation of ROS (Winterbourn, 1981; Bus and Gibson, 1984). Mitigation of ROS accumulation by anthocyanin overaccumulation resulted in oxidative tolerance in the pap1-D and WOX1 lines (Figure​(Figure7c).7c). Drought stress induces the accumulation of ROS in plants (Vickers et al., 2009). Staining with DAB, which is a marker of H₂O₂ accumulation in vivo (Thordal-Christensen et al., 1997), clearly showed that anthocyanin overaccumulation reduced the accumulation of H₂O₂ in vivo (Figure​(Figure8d).8d). A common factor between oxidative and drought stress is the accumulation of ROS, which are mainly produced in chloroplasts and peroxisomes (Kim et al., 2008). Reactive oxygen species play an important role as signaling molecules that initiate stress responses in plants. However, overaccumulation of ROS causes a chain reaction of ROS production. Cytosolic H₂O₂ enters the vacuoles and subsequently initiates a chain reaction resulting in excess ROS accumulation (Bolouri-Moghaddam et al., 2010; Agati et al., 2012). The increased generation of ROS in the cellular components is attributed to a disturbance in cellular homeostasis such as programmed cell death (Vickers et al., 2009) or death by cytotoxicity (Kim et al., 2008). Our results suggest that anthocyanins, which generally accumulate in the vacuoles (Hernandez et al., 2009), prevent oxidative damage, making it possible for the anthocyanin-overaccumulating plants to exhibit oxidative and drought tolerance. The function of anthocyanins against ROS can be inferred from their strong antioxidative activity in vitro. Cyanidin-type anthocyanins and quercetin-type flavonols were responsible for the strong radical scavenging activity in the in vitro DPPH assay (Figure​(Figure7)7) (Seyoum et al., 2006). The presence of hydroxyl group(s) on the aglycone moiety is important for scavenging DPPH, whereas the sugar and acyl moieties do not appear to be essential (Tohge et al., 2005b; Seyoum et al., 2006). The mechanism of mitigation of ROS by such anthocyanins remains to be clarified. Therefore for the next step it will be important to reveal what type of ROS can be mitigated by such anthocyanins. Partial evidence for flavonoids as antioxidants in vivo is accumulating gradually. Flavonoid deficiency increases H₂O₂ in Arabidopsis (Zhang et al., 2012). Anthocyanins are suitably located to account for the enhanced rates of H₂O₂ production (Chalker-Scott, 2002). In addition to anthocyanins, other specialized metabolites such as hydroxycinnamate esters, which are believed to be radical scavengers (Ferreres et al., 2011), also accumulate in vacuoles. Overaccumulation of hydroxycinnamate ester confers UV-B tolerance in Arabidopsis (Jin et al., 2000). These results suggest that specialized metabolites play a role in scavenging ROS in vacuoles to enhance stress tolerance.

It has been suggested that anthocyanins as antioxidants are likely to be osmoregulators that maintain water homeostasis (Do and Cormier, 1990; Chalker-Scott, 2002; Hughes et al., 2010, 2013; Hughes, 2011; Sperdouli and Moustakas, 2012), together with compatible solutes such as sugars (e.g. mannitol, sorbitol, and trehalose) and amino acids (e.g. proline), which help to adjust homeostasis and scavenge ROS in various plant species (Chen and Murata, 2002). We revealed that flavonoid overaccumulation prevented water loss from whole plants and detached leaves (Figures​(Figures8a8a and ​and9).9). Our data suggest that anthocyanins in Arabidopsis may play a role in water homeostasis by scavenging ROS into vacuoles under drought stress. Stomatal closure reduces plant water loss through transpiration during drought stress (Hirayama and Shinozaki, 2010). A previous study in tobacco with an altered level of the antioxidant ascorbic acid reported a change in stomatal closure, leading to the prevention of wilting and water loss. In comparative analysis with the wild type, plants overexpressing or suppressing dehydroascorbate reductase (DHAR) revealed that increasing the redox state of ascorbic acid resulted in an increase in the transpiration rate and stomatal conductance under normal growth conditions. Furthermore, guard cells with a higher redox state of ascorbic acid were less responsive to H₂O₂ or ABA signaling, and the plants exhibited increased water loss after drought conditions were initiated (Chen and Gallie, 2004). Taken together with previous results and our data, there may be no direct association between anthocyanins, which do not accumulate in substantial amounts, guard cells (Gallagher and Smith, 2000), and stomatal closure.

Flavonoids, which have various useful in vitro activities (Harborne and Williams, 2000), are useful in crop production. Increased flavonoid accumulation was achieved by crossing MYB12OX and pap1-D. As expected, the major difference between the other lines was flavonoid accumulation (Figure​(Figure5a).5a). In fact, the WOX1 lines accumulated higher levels of flavonols and anthocyanins than each single overexpressing line. The profile of flavonols and anthocyanins in the WOX1 lines was identical to that in the MYB12OX or pap1-D lines. The enhanced flavonoid overaccumulation in the WOX lines was probably initiated independently by MYB12 and PAP1. The Arabidopsis triple mutant myb11myb12myb111 suggests the independent function of MYB12 and PAP1 (Stracke et al., 2010b). Ectopic overexpression of single MYB transcription factors resulted in the induction of flavonoid biosynthesis without significantly inhibiting plant growth (Figure S1). The combined overexpression of MYB12 and PAP1 also resulted in no significant difference in the WOX1 lines. Furthermore, the flavonoid-deficient MYB12OX/tt4 and pap1-D/tt4 lines were generated by crossing tt4 and MYB12OX or pap1-D, respectively, with no or minimal growth penalty. A recent report revealed that MYB12OX is UV-B tolerant (Stracke et al., 2010a), and MYB12-overexpressing tobacco is tolerant against insects such as Spodoptera litura and Helicoverpa armigera because of flavonoid overaccumulation (Misra et al., 2010). Purple tomato fruit, which ectopically overexpresses two transcription factors Delila and Rosea1 from snapdragons, has increased shelf-life and reduced susceptibility to Botrytis cinerea, an important post-harvest pathogen (Zhang et al., 2013). These reports suggest that flavonoid overaccumulation can be a key to concurrently enhancing tolerance to multiple stressors. Our results also suggest a useful indication of how to generate promising novel crop varieties with enhanced stress tolerance. Cross-fertilization using varieties containing high amounts of flavonoids will provide new insights into tolerance to biotic and abiotic stresses, potentially leading to increased crop production. Flavonoids are hypothetically produced as antioxidants in the late response against UV-B stress (Kusano et al., 2011). Insights into biotic and abiotic stress tolerance due to flavonoid accumulation in crops will be elicited during long-term stress exposure.

Methods

Acknowledgments

References