https://orcid.org/0000-0003-3280-5076
ABSTRACT
Carthamin, a dimeric quinochalcone that is sparingly soluble in water, is obtained from the yellow-orange corolla of fully blooming safflower (Carthamus tinctorius L.) florets. Carthamin is a natural red colorant, which has been used worldwide for more than 4500 years and is the major component of Japanese ‘beni’ used for dyeing textiles, in cosmetics and as a food colorant. The biosynthetic pathway of carthamin has long remained uncertain. Previously, carthamin was proposed to be derived from precarthamin (PC), a water-soluble quinochalcone, via a single enzymatic process. In this study, we identified the genes coding for the enzyme responsible for the formation of carthamin from PC, termed ‘carthamin synthase’ (CarS), using enzyme purification and transcriptome analysis. The CarS proteins were purified from the cream-colored corolla of safflower and identified as peroxidase homologs (CtPOD1, CtPOD2 and CtPOD3). The purified enzyme catalyzed the oxidative decarboxylation of PC to produce carthamin using O2, instead of H2O2, as an electron acceptor. In addition, CarS catalyzed the decomposition of carthamin. However, this enzymatic decomposition of carthamin could be circumvented by adsorption of the pigment to cellulose. These CtPOD isozymes were not only expressed in the corolla of the carthamin-producing orange safflower cultivars but were also abundantly expressed in tissues and organs that did not produce carthamin and PC. One CtPOD isozyme, CtPOD2, was localized in the extracellular space. Based on the results obtained, a model for the stable red pigmentation of safflower florets during flower senescence and the traditional ‘beni’ manufacturing process is proposed.
Introduction
Carthamin (Fig. 1, structure 1), a historic red colorant that is sparingly soluble in water, is produced from yellow-orange corollas of fully blooming florets of safflower (Carthamus tinctorius L., Benibana; Fig. 2), an asteraceous plant (Yue et al. 2013). The use of carthamin as a red colorant dates back to ancient Egypt (4500 years ago). In Japan, the carthamin-based red pigment has been called ‘beni’ (meaning ‘red’) and preferentially used for more than 1400 years for dyeing textiles (such as Kimonos), in cosmetics (such as lipsticks), as a food colorant (for Japanese sweets) and a herbal medicine (Kosoto 2007).
Chemical structures of safflower quinochalcones. 1, carthamin; 2, PC; 3a, SYB; 3b, ASYB; 4, HSYA. A cyclohexanonedienol-type A-ring structure is labeled with ‘A’ in 1.
Safflower cultivars and stages of capitulum development. (A) Capitulum of the orange cultivar ‘Orange Maruba-shu’. (B) Capitulum of the cream cultivar ‘Shirobana Maruba-shu Cream’. (C) Developmental stages of safflower capitulum. For further details, see the ‘Materials and Methods’ section.
Purification of CarS from the corolla of safflower cultivars with different floret colors. (A) SDS-PAGE of the purified CarS from the corolla of the orange cultivar. Red arrowheads indicate the protein bands of the purified enzyme. (B) SDS-PAGE of the purified CarS from the corolla of the cream cultivar. Red arrowheads indicate the protein bands of the purified enzyme. (C) Elution profiles of CarS activity (upper graph) and peroxidase activity (lower graph) during Superdex 200 10/300 GL gel filtration chromatography (step 7; see the ‘Materials and Methods’ section).
The elucidation of the chemical structure of carthamin has a long history since the 1910s (Kametaka and Perkin 1910). A monomeric O-glucosylquinochalcone structure was first proposed for carthamin by Kuroda (1930), in which ‘quinochalcone’ refers to a chalcone with a cyclohexanonedienol-type A-ring (Fig. 1). A half-century later, a planar structure for carthamin, which represents a dimeric C-glucosylquinochalcone structure, was proposed by two independent research groups (Obara and Onodera 1979, Obara et al. 1980, Takahashi et al. 1982). The stereo structure of carthamin was proposed in 1996 (Sato et al. 1996) and confirmed via the total chemical synthesis in 2019 (Azami et al. 2019). Yellow C-glycosylquinochalcones [i.e. hydroxysafflor yellow A (HSYA); for structure see Fig. 1, 4), safflor yellow B (SYB, 3a) or anhydrosafflor yellow B (ASYB, 3b) and precarthamin (PC, 2)] also accumulate in safflower florets and might be precursors of carthamin (Kazuma et al. 2000). Among these compounds, PC has been proposed to be the direct precursor of carthamin (Kumazawa et al. 1994, 1995, Kazuma et al. 1995). To date, these C-glycosylquinochalcones have exclusively been identified in safflower (Yue et al. 2013).
Nucleotide and deduced amino acid sequences of CtPOD1 from safflower. The underlined amino acid sequences are those determined with the enzyme purified from the cream cultivar by LC/MS/MS sequence analysis. The sequences shaded with orange are those determined with the enzyme purified from the orange cultivar.
Mechanistic aspects of carthamin formation from PC remain to be clarified. In planta, the formation of carthamin in the yellow-orange safflower florets appears to be a senescence-related process; the corolla of carthamin-producing safflower florets is yellow to orange in color until full blooming (stages 2 and 3; see Fig. 2) and subsequently changes to reddish color during floret wilting (stages 4 and 5) (Fig. 2). The red pigment formed in the corolla during floret senescence is identical to that of ‘beni’ (Kazuma et al. 2017). However, for traditional ‘beni’ production, yellow-orange corollas of fully blooming florets, but not wilted red corollas, are used. Important steps in the traditional ‘beni’ manufacturing process include disruption of corolla tissues and their prolonged exposure to air. It has been proposed that the red pigmentation of the safflower corolla is an enzymatic process catalyzed by polyphenol oxidase or peroxidase (Shimokoriyama and Hattori 1955). Saito and co-workers biochemically characterized the enzymatic formation of carthamin (Saito et al. 1983a, 1983b, 1985, 1986, 1993) before the structure of PC had been established. These authors observed that a fungal glucose oxidase and other heterologous oxidases (Saito 1992, 1993, Saito et al. 1998) oxidize their hydrogen donor substrates to produce H2O2, which in turn reacts with PC to produce carthamin; thus, they proposed that the red pigmentation of the safflower corolla is mediated by oxidases. In vitro assays revealed that horseradish peroxidase (HRP), a heterologous plant enzyme, is capable of catalyzing the formation of carthamin from PC in the presence (Kumazawa et al. 1995) and absence (Abe et al. 2020) of H2O2. However, native peroxidase activity in the crude extract from safflower corollas decomposes carthamin in vitro (Kanehira and Saito 1990). This observation appeared to contradict the possible involvement of peroxidase activity in carthamin synthesis in planta. In 2000, Cho and co-workers purified to apparent homogeneity an enzyme [PC decarboxylase, herein termed ‘carthamin synthase’ (CarS)], which catalyzed the conversion of PC to carthamin, from the yellow corolla of safflower. The purified enzyme was capable of catalyzing the formation of carthamin from PC in the absence of H2O2 (Cho and Hahn 2000, Cho et al. 2000). However, the complementary DNA (cDNA) coding for this enzyme has not been isolated; hence, the identity of this enzyme remains to be clarified.
In this study, we purified CarS from safflower cultivars with orange and cream corollas to near homogeneity. On the basis of partial amino acid sequences determined from the purified CarS preparations, we cloned the CarS cDNAs and determined that they encode homologs of peroxidase (CtPOD1, CtPOD2 and CtPOD3). Transcription and subcellular localization analyses of these CtPOD isozymes were performed to examine physiological aspects of the peroxidase-mediated red pigmentation of safflower.
Results
Quinochalcones and CarS activities in safflower corollas of different colors
First, we analyzed quinochalcones [HSYA, SYB (as ASYB) and PC] and CarS activities in orange and cream corollas of two safflower cultivars (‘Orange Maruba-shu’ and ‘Shirobana Maruba-shu Cream’, respectively) to detect corolla color-dependent differences. During stages 3–5, the corollas of the orange cultivar changed to red (Fig. 2), whereas those of the cream cultivar remained cream in color (i.e. no carthamin is produced). Of note, carthamin content could not be determined with the present analytical systems because of the difficulty of its extraction from the corolla.
The corolla of the orange cultivar mainly accumulated HSYA and SYB together with a small amount of PC, and HSYA and SYB were maximally accumulated at stage 4 (Supplementary Fig. S1A). In striking contrast, in the cream corolla, HSYA, SYB and PC were absent (Supplementary Fig. S1B). Light microscopic observations of cellulase-treated corollas (stage 3) of the orange cultivar showed that yellow pigments, which may include HSYA and SYB, exclusively accumulated in vacuoles of corolla cells, whereas red pigments (i.e. carthamin) were indicated to adsorb to insoluble extracellular matrices (Supplementary Fig. S2).
Appreciable amounts of H2O2-independent CarS activity were detected in crude extracts from the corollas of orange and cream cultivars, as assayed by assay method II. Given that quinochalcones, which hampered quantitative CarS assays, were virtually absent in the crude extracts from the cream cultivar, we were able to quantitatively assay CarS activity. The specific activity of CarS in the crude extracts from the cream corolla was 13.4 nmol min-1 (mg protein)-1 (i.e. 223 pkat (mg protein)-1). This CarS activity was substantially inhibited by 500 μM or higher concentrations of potassium cyanide (a peroxidase inhibitor; Dunford and Stillman 1976) in the presence or absence of H2O2 (Supplementary Fig. S3A, B, respectively). However, CarS activity was not inhibited by 1 mM N-phenylthiourea, a specific inhibitor of polyphenol oxidase (Nakayama et al. 2001) (Supplementary Fig. S3C).
Purification of CarS and identification of CarS cDNAs
We purified CarS from the crude extracts of the corolla (500 g) of the orange cultivar by a combination of Q Sepharose FF, SP Sepharose FF, Foresight CHT Type I, Superdex 200 10/30 GL, Bio-Scale CHT2-I and Mono S 5/50 column chromatographies (see the ‘Materials and Methods’ section for details). Two major CarS activity peaks were distinguished during Foresight CHT Type I hydroxyapatite column chromatography (see below), and we purified the protein responsible for the second-eluted activity peak. The purified enzyme yielded dual protein bands (indicated by red arrowheads) with similar molecular masses of ∼36 kDa by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (Fig. 3A). The purified enzyme was subjected to trypsin digestion, followed by liquid chromatography–tandem mass spectrometry (LC/MS/MS) analysis of the resulting peptides. The results showed that the protein bands were indistinguishable, based on the amino acid sequences determined with tryptic peptides, which showed high similarities to those of plant peroxidases (Fig. 4).
We then performed a transcriptome analysis of a cDNA library generated from orange safflower corollas (stage 3), which provided 54,102 unigene sequences. Amino acid sequences of peptides derived from the purified enzyme all coincided with those deduced from the nucleotide sequence of a single unigene, termed here CtPOD1, which encoded a homolog of a peroxidase (Fig. 4).
We also completed the purification of CarS from the corolla (130 g) of the cream cultivar following the purification steps summarized in Table 1 (see also the ‘Materials and Methods’ section for details). Again, two major CarS activity peaks (termed here peaks I and II; see Supplementary Fig. S4) were distinguished during the Bio-Scale CHT2-1 hydroxyapatite column chromatography. The protein responsible for activity peak II was purified to near homogeneity (Fig. 3B). SDS-PAGE of this enzyme preparation showed two major protein bands with approximate molecular masses of 36 kDa (see Fig. 3B, bands indicated by red arrowheads). The chromatographic behaviors of the protein bands were strongly correlated with those of CarS activity (Fig. 3C, upper graph) and a peroxidase activity as assayed by assay method III using 2,2′-azino-bis(3-ethylbenzthiazoline-6-sulfonic acid) (ABTS) as a hydrogen donor (Fig. 3C, lower graph). The purified enzyme preparation showed specific CarS activity of 264.4 nkat (mg protein)-1 (assay method II) and specific peroxidase activity of 94.24 μkat (mg protein)-1 protein (assay method III). Each of these protein bands (Fig. 3B) was excised from the gels and subjected to trypsin digestion. The amino acid sequences of the resulting peptides were determined by means of LC/MS/MS. The amino acid sequences for both bands coincided with those deduced from the nucleotide sequence of CtPOD1 (Fig. 4).
Purification of CarS (CtPOD1) from the crude extracts of the corollas of the cream cultivar
| Total protein | Total activity | Specific activity | Purification | Recovery | |
|---|---|---|---|---|---|
| Step | (mg) | (nkat) | (nkat/mg protein) | (-fold) | (%) |
| 1. Crude extract | 607 | 136 | 0.224 | 1 | 100 |
| 2. DEAE Sephadex | 208 | 116 | 0.558 | 2.5 | 84.9 |
| 3. SP Sepharose FF | 143 | 74.9 | 0.525 | 2.3 | 55 |
| 4. Ammonium sulfate fractionation | 56.5 | 13.8 | 0.246 | 1.1 | 10.2 |
| 5. CM Sepharose FF | 1.8 | 5.3 | 2.93 | 13 | 3.9 |
| 6. Bio-Scale CHT2-1 | 0.31 | 2.76 | 8.92 | 39.8 | 2 |
| 7. Superdex 200 | 0.00559 | 1.59 | 284 | 1,267 | 1.2 |
| Total protein | Total activity | Specific activity | Purification | Recovery | |
|---|---|---|---|---|---|
| Step | (mg) | (nkat) | (nkat/mg protein) | (-fold) | (%) |
| 1. Crude extract | 607 | 136 | 0.224 | 1 | 100 |
| 2. DEAE Sephadex | 208 | 116 | 0.558 | 2.5 | 84.9 |
| 3. SP Sepharose FF | 143 | 74.9 | 0.525 | 2.3 | 55 |
| 4. Ammonium sulfate fractionation | 56.5 | 13.8 | 0.246 | 1.1 | 10.2 |
| 5. CM Sepharose FF | 1.8 | 5.3 | 2.93 | 13 | 3.9 |
| 6. Bio-Scale CHT2-1 | 0.31 | 2.76 | 8.92 | 39.8 | 2 |
| 7. Superdex 200 | 0.00559 | 1.59 | 284 | 1,267 | 1.2 |
Purification of CarS (CtPOD1) from the crude extracts of the corollas of the cream cultivar
| Total protein | Total activity | Specific activity | Purification | Recovery | |
|---|---|---|---|---|---|
| Step | (mg) | (nkat) | (nkat/mg protein) | (-fold) | (%) |
| 1. Crude extract | 607 | 136 | 0.224 | 1 | 100 |
| 2. DEAE Sephadex | 208 | 116 | 0.558 | 2.5 | 84.9 |
| 3. SP Sepharose FF | 143 | 74.9 | 0.525 | 2.3 | 55 |
| 4. Ammonium sulfate fractionation | 56.5 | 13.8 | 0.246 | 1.1 | 10.2 |
| 5. CM Sepharose FF | 1.8 | 5.3 | 2.93 | 13 | 3.9 |
| 6. Bio-Scale CHT2-1 | 0.31 | 2.76 | 8.92 | 39.8 | 2 |
| 7. Superdex 200 | 0.00559 | 1.59 | 284 | 1,267 | 1.2 |
| Total protein | Total activity | Specific activity | Purification | Recovery | |
|---|---|---|---|---|---|
| Step | (mg) | (nkat) | (nkat/mg protein) | (-fold) | (%) |
| 1. Crude extract | 607 | 136 | 0.224 | 1 | 100 |
| 2. DEAE Sephadex | 208 | 116 | 0.558 | 2.5 | 84.9 |
| 3. SP Sepharose FF | 143 | 74.9 | 0.525 | 2.3 | 55 |
| 4. Ammonium sulfate fractionation | 56.5 | 13.8 | 0.246 | 1.1 | 10.2 |
| 5. CM Sepharose FF | 1.8 | 5.3 | 2.93 | 13 | 3.9 |
| 6. Bio-Scale CHT2-1 | 0.31 | 2.76 | 8.92 | 39.8 | 2 |
| 7. Superdex 200 | 0.00559 | 1.59 | 284 | 1,267 | 1.2 |
In addition, we purified the protein responsible for activity peak I (Supplementary Fig. S5, indicated by a red arrowhead), which was also subjected to trypsin digestion followed by LC/MS/MS analysis. The amino acid sequence determined from the protein band coincided with those deduced from the nucleotide sequences of other peroxidase paralogs (i.e. CtPOD2 and CtPOD3; for their nucleotide sequences and deduced amino acid sequences, see Supplementary Fig. S6A, B, respectively).
Enzymatic properties of CarS (CtPOD1) purified from safflower corolla
The CarS (CtPOD1) protein purified from the corolla of the cream cultivar (Fig. 3B) efficiently catalyzed the formation of carthamin from PC (Fig. 5A). Catalytic properties of the purified enzyme were examined using standard assay conditions (assay method II, see the ‘Materials and Methods’ section for details). The purified enzyme showed maximum CarS activity at approximately pH 5.0 (at 30°C) and 40°C (at pH 5.0) (Supplementary Fig. S7A, B, respectively). The enzyme displayed CarS activity in the absence of H2O2, and the maximal CarS activity was observed in the presence of 0.5 mM H2O2 (relative activity, 570% of the activity in the absence of H2O2) (Supplementary Fig. S8A). For comparison, HRP also displayed CarS activity in the absence of H2O2, with the maximal CarS activity detected in the presence of 0.25–0.5 mM H2O2 (relative activity, 370% of the activity in the absence of H2O2) (Supplementary Fig. S8B). When CtPOD1-catalyzed CarS reactions were carried out under N2-substituted conditions (see the ‘Materials and Methods’ section), CarS activity of CtPOD1 was substantially lower than that observed under aerobic conditions, irrespective of the presence or absence of 500 μM H2O2 (Fig. 6). This was also the case for CarS activity of HRP (Supplementary Fig. S9A–D). However, HRP showed full peroxidase activity under the N2-substituted conditions when ABTS and 100 μM H2O2 were used as an electron donor and acceptor, respectively (Supplementary Fig. S9E).
Formation of carthamin from PC catalyzed by CarS (CtPOD1). (A) The assay mixture for the CarS-catalyzed reaction (assay method II) was analyzed by HPLC (system B). Upper panel, before addition of the purified enzyme; middle panel, 3 min after enzyme addition; lower panel, 10 min after enzyme addition. The inset in each panel shows the appearance of the reaction mixture, showing time-dependent color changes. (B) Peroxidase-catalyzed decolorization of carthamin is circumvented by adsorption of nascent pigment to cellulose. Five nanomoles of PC was incubated with 16 pkat of the purified CarS (CtPOD1) at 30°C for 30 min in 100 μl of 100 mM sodium citrate buffer, pH 5.0, in the absence (1) or presence (2) of a dialysis tubing cellulose membrane (3 mm × 3 mm, six sheets per reaction mixture). Time-dependent color changes of the reaction mixtures were recorded (0, 5 and 30 min). In addition, the reaction mixtures were allowed to stand for 34 d at room temperature.
CarS-catalyzed reaction under aerobic and N2-substituted conditions. HPLC chromatograms of CtPOD1-catalyzed CarS reactions (A) in the absence of H2O2 under aerobic conditions, (B) in the presence of 500 μM H2O2 under aerobic conditions, (C) in the absence of H2O2 under N2-substituted conditions and (D) in the presence of 500 μM H2O2 under N2-substituted conditions are shown together with HPLC chromatograms of CarS reaction mixture at t = 0 (panel E), in which 500 μl methanol and 100 μl of 100 mM sodium citrate, pH 5.0, were added to the vial simultaneously. The insets in panels (A) through (D) show the appearance of the reaction mixtures in the vials. For experimental details, see the ‘Materials and Methods’ section.
When the effect of substrate concentration [PC] on the initial velocity of the purified CarS (CtPOD1) was examined at pH 5.0 and 30°C in the absence of H2O2, a linear relationship was obtained in the range of [PC] examined (up to 100 μM, see Supplementary Fig. S10). This finding suggested that the Km value of CarS (CtPOD1) for PC should be significantly greater than 100 μM, because the enzyme-catalyzed reactions proceeded with first-order kinetics when substrate concentrations were much lower than Km. Thus, the kcat/Km value was determined to be 0.108 s−1 μM−1 from the slope of the initial velocity vs [PC] plots. The kcat value of CarS (CtPOD1) was 10.4 s−1 at [PC] = 100 μM, corresponding to specific CarS activity of 289 nkat (mg protein)−1.
It must be mentioned that prolonged incubation of the CarS assay mixture caused the diminution of the red coloration (Fig. 5A), consistent with previous observations that carthamin was decomposed by a peroxidase-catalyzed process (Kanehira and Saito 1990). However, when the Car assays were carried out in the presence of cellulose sheets, carthamin produced was adsorbed to the cellulose sheet and the red coloration of the cellulose sheet remained unchanged during incubation and even after the reaction mixture was allowed to stand for 34 d at room temperature (Fig. 5B). This result suggested that carthamin adsorbed to cellulose did not undergo CtPOD1-catalyzed decolorization.
Expression analysis
Expression of CtPOD1, CtPOD2 and CtPOD3 in the safflower plant was examined by the quantitative real-time reverse transcription polymerase chain reaction (RT-PCR) analysis (Fig. 7). CtPOD1 was expressed in the buds (stage 1), corollas (stage 3), leaves and involucre of the orange cultivar, of which the buds, leaves and involucre do not accumulate carthamin (Fig. 7A). CtPOD1 was also abundantly expressed in the corollas (stage 3) of the cream cultivar (Fig. 7A) where PC and other yellow quinochalcones were absent (Supplementary Fig. S1). CtPOD2 was expressed in the orange and cream corollas, whereas no transcripts or only a low level of its transcripts were detected in the buds, leaves and involucre of the orange cultivar (Fig. 7B), which suggests that the expression of CtPOD2 was corolla-specific. CtPOD3 transcripts were the most abundantly detected in the involucre among the tissues and organs examined (Fig. 7C). These results prompted us to further examine whether CarS and peroxidase activities also were detectable in the organs other than corollas using the orange cultivar. The results showed that appreciable levels of activities of both enzymes were found in the crude extracts of buds, leaves and the involucre, in addition to corollas (Fig. 7D).
Relative transcript levels of CtPOD1–3 and activities of CarS and peroxidase in safflower organs. (A) CtPOD1 transcript levels, (B) CtPOD2 transcript levels, (C) CtPOD3 transcript levels and (D) specific CarS activity (upper panel) and specific peroxidase activity (lower panel). Bud, corolla at stage 1 of the orange cultivar; Corolla, corolla at stage 3 of the orange cultivar; Involucre, involucre at stage 3 of the orange cultivar; Leaf, leaf of the orange cultivar; Cream corolla, corolla at stage 3 of the cream cultivar. The expression levels are relative to that of the safflower actin gene (CtACT). Data are presented as the average of independent determinations of three technical replicates (± standard deviation).
Subcellular localization analysis
Signal sequences for secretion were identified in the deduced amino acid sequences of CtPOD2 and CtPOD3, whereas no transit peptide was unambiguously identified in the deduced amino acid sequence of CtPOD1, as analyzed with SignalP ver. 5.0 (Supplementary Fig. S11). To clarify the subcellular localization of these CtPOD isozymes, CtPOD1 and CtPOD2 proteins were C-terminally fused to the mVenus protein (a spectral variant of the green fluorescent protein) and were transiently expressed with non-tagged mTurquoise2 (another spectral variant of the green fluorescent protein) in Nicotiana benthamiana leaves, followed by confocal laser fluorescence microscopic observation. mVenus fluorescence signals from the expressed CtPOD2-mVenus protein were observed in the extracellular space of epidermal cells of the transformed N. benthamiana leaves (Fig. 8B). The CtPOD1-mVenus protein was not expressed in N. benthamiana leaves (Fig. 8A).
Subcellular localization analysis of CtPOD1 and CtPOD2. The CtPOD-mVenus and mTurquoise2 proteins expressed in N. benthamiana leaves were identified by yellow and light-blue fluorescence signals, respectively, via confocal laser scanning microscopy. (A) CtPOD1-mVenus. (B) CtPOD2-mVenus. From left to right, fluorescence from mVenus, fluorescence from mTurquoise2, merged fluorescence signals from mVenus and mTurquoise2 and brightfield image. In (B), magnified images are also shown below the fluorescence images. Note that yellow fluorescence from CtPOD2-mVenus is localized between light-blue fluorescence signals of mTurquoise2, which is in the cytoplasm of cells ‘a’ and ‘b’ identified in the brightfield image. Scale bar = 30 μm.
Discussion
In this study, we identified the genes coding for CarS of safflower, CtPOD1, CtPOD2 and CtPOD3, which catalyze the conversion of PC to carthamin, the final enzymatic step of carthamin synthesis in this plant. These CtPOD isozymes were shown to be homologs of peroxidase. The deduced amino acid sequences of CtPOD1, CtPOD2 and CtPOD3 were 50%, 38% and 41%, respectively, identical to that of HRP and shared conserved sequences for the peroxidase active site and heme-binding sites (see Supplementary Fig. S11). The present findings provide a further example of the involvement of a peroxidase in flavonoid biosynthesis, in addition to peroxidase involvement in aurone biosynthesis in Medicago truncatula (Farag et al. 2009).
The molecular mass of CtPOD1 predicted from the deduced amino acid sequence was 40.6 kDa, which was slightly larger than the estimated molecular masses (approximately 36 kDa) of the purified enzyme, thus showing micro-heterogeneity (Fig. 3B). The enzyme might have undergone post-translational processing, although no transit peptide was unambiguously identified in the deduced amino acid sequence of CtPOD1 (see the ‘Results’ section). During LC/MS/MS protein sequence analyses, we did not identify the N-terminal tryptic peptide predicted from the CtPOD1 cDNA sequence, which might be removed in planta after post-translational proteolytic cleavage. It is possible that proteolytic cleavage might take place at different sites of the N-terminal region of the precursor protein and give rise to the observed micro-heterogeneity of the purified CtPOD1 bands. It must be mentioned that when the purified CtPOD1 was treated with endoglycosidase H, the molecular masses of the CtPOD1 bands in SDS-PAGE gels were unchanged (Supplementary Fig. S12), which suggests that the observed micro-heterogeneity of the CtPOD1 bands was unrelated to N-glycosylation of proteins.
Unfortunately, we were unable to express the CtPOD isozymes as soluble, catalytically active recombinant proteins using heterologous expressions systems, such as those involving Escherichia coli, Saccharomyces cerevisiae and Pichia pastoris cells. Thus, we extensively characterized the catalytic properties of CarS (CtPOD1) purified from the cream corolla of safflower using the fraction no. 31 from activity peak II (Fig. 3B, C), because this fraction contained sufficient CarS activity for enzymatic characterization.
Consistent with the predicted amino acid sequence, the purified CarS (CtPOD1) displayed peroxidase activity, catalyzing the oxidation of a general peroxidase substrate (ABTS) by H2O2 (Fig. 3C). The specific peroxidase activity, 94.24 μkat (mg protein)-1 protein, was 356 times higher than its specific CarS activity (264.4 nkat (mg protein)-1protein). However, this CarS activity, corresponding to a kcat value of 10 s−1, was sufficiently high for the activity of an enzyme involved in plant specialized metabolism, which generally ranges from 10−2 s−1 to 102 s−1 with a median value of 2.5 s−1 (Bar-Even et al. 2011). The results of transcriptome analysis showed that 134 peroxidase paralogs were expressed in the orange corolla of safflower. This observation, together with the fact that HRP was capable of catalyzing the formation of carthamin from PC (see the ‘Results’ section), implied that other peroxidase paralogs expressed in safflower corollas might also be involved in carthamin synthesis. Consistent with this suggestion, two CarS activity peaks were distinguished during enzyme purification from the corollas of different cultivars. Of these peaks, peak I arose from two other peroxidase paralog(s) (CtPOD2 and CtPOD3), as identified by LC/MS/MS. In this connection, among the 134 peroxidase paralogs expressed in the orange corollas, two additional paralogs (termed here CtPOD4 and CtPOD5) showed higher transcripts per million (TPM) values than those of CtPOD1, CtPOD2 and CtPOD3, and the expression of CtPOD5 appeared to be corolla-specific, as analyzed by quantitative real-time RT-PCR analysis (Supplementary Fig. S13). However, the CtPOD4 and CtPOD5 proteins were not identified during the present trials of enzyme purification using different cultivars. Thus, only a limited number of peroxidase paralogs (CtPOD1, CtPOD2 and CtPOD3) were likely to be mainly involved in the red pigmentation of the safflower corolla among the many other expressed peroxidase paralogs. It must also be mentioned that glucose oxidase was previously proposed to mediate the formation of carthamin from PC in safflower (Saito 1993). However, involvement of this enzyme in the red pigmentation of the corolla of the orange cultivar used in this study is highly unlikely because no detectable amount (based on TPM values) of the transcript that was annotated as glucose oxidase was expressed therein.
Mechanistic aspects of the H2O2-independent formation of carthamin from PC catalyzed by CtPOD1 and HRP should be addressed in future studies.
Peroxidase activity in the crude extract of safflower corollas was previously shown to catalyze the decomposition of carthamin in vitro (Kanehira and Saito 1990). Indeed, when the assay mixture (method II) was incubated for a prolonged period, carthamin formed during 3-min incubation was converted to a pale-orange-colored substance (Fig. 5A). However, in the presence of cellulose in the reaction mixture, nascent red pigment was immediately adsorbed to it and remained red even after prolonged incubation, even after 1 month (Fig. 5B). Thus, the peroxidase-catalyzed decolorization of carthamin could be circumvented by adsorption to cellulose, the major component of the corolla cell walls. It is likely that, during the course of red pigmentation of safflower florets, CtPOD1 and related peroxidases act on PC, which is water-soluble, to produce carthamin, which is sparingly water-soluble and is immediately adsorbed to the cell wall of corolla cells to avoid enzymatic degradation and attain stable red pigmentation.
Transcription analysis revealed that the messenger RNA expression patterns of CtPOD1–3 in the safflower plant were not specifically associated with the red pigmentation of safflower florets (Fig. 7). For example, in addition to the corolla (stage 3) of the orange cultivar, which showed red pigmentation during stages 4 and 5 (see Fig. 2), CtPOD1 was abundantly expressed in the buds (stage 1), leaves and involucres of the same cultivar, none of which showed red pigmentation. Moreover, CtPOD1 was abundantly expressed in the corolla (stage 3) of the cream cultivar in which PC and other yellow quinochalcones were absent (Supplementary Fig. S1). Expression of CtPOD2 appeared to be corolla-specific, whereas CtPOD3 transcripts were most abundantly detected in the involucre among the tissues and organs examined (Fig. 7B, C). These results were consistent with the fact that strong CarS and peroxidase activities were present in the crude extracts of the PC-lacking tissues and organs (Fig. 7D) as well as the cream corolla. Thus, differences in the capability for red pigmentation of florets among the cultivars used in this study were not associated with expression levels of CtPOD1–3 but were most likely related to the availability of the precursor PC in these cultivars.
Signal sequences for secretion were unambiguously identified in the deduced amino acid sequences of CtPOD2 and CtPOD3 (Supplementary Fig. S11), and CtPOD2 was consistently shown to be localized in the extracellular space via subcellular localization analysis using a CtPOD2-mVenus fusion protein (Fig. 8). Unfortunately, we could not predict nor identify the subcellular localization of CtPOD1 (see the ‘Results’ section), which should be further addressed in future studies.
All of these observations suggest the following possible scenario for the red pigmentation of safflower florets. The CarS genes—CtPOD1, CtPOD2 and CtPOD3—are expressed in the corolla and other tissues of the safflower plant, irrespective of flower color (Fig. 7). In carthamin-producing cultivars, PC likely accumulates in a cellular compartment of the corolla cell that is physically separated from the CarS-containing cellular compartment. During the course of floret senescence, the structural integrity of the corolla cells is likely compromised, resulting in the incomplete cellular compartmentalization, which could allow CarS to colocalize with PC to produce carthamin. The traditional ‘beni’ manufacturing process includes corolla disruption, which would allow efficient mixing of PC with CarS to produce carthamin. The carthamin thus produced is immediately adsorbed to the corolla cell wall and other insoluble extracellular matrices (see Supplementary Fig. S2A), and thereby, stabilization of red pigmentation is attained.
Materials and Methods
Plant materials
Orange and cream cultivars of safflower (‘Orange Maruba-shu’ and ‘Shirobana Maruba-shu Cream’, respectively; Dainichi Shokai, Koga, Ibaraki, Japan) were grown in agricultural fields at Sendai, Miyagi, and at Chitose, Hokkaido, Japan. The developmental stages of safflower florets were defined as follows (see also Fig. 2C): stage 1, closed buds (<25 mm in diameter); stage 2, small florets emergent in the center of the involucre; stage 3, all florets fully developed and at anthesis; for orange cultivars, florets show little red pigmentation; stage 4, florets begin to wilt; for orange cultivars, florets begin to show red pigmentation; and stage 5, florets strongly wilted and the ovules are enlarged; for orange cultivars, florets show red pigmentation.
Chemicals
LIOFRESH Red CR (a carthamin-adsorbed cellulose powder) was obtained from TOYOCHEM, Tokyo, Japan. HRP and the dialysis tubing cellulose membrane that was used during enzyme assays (see Fig. 5B) were obtained from FUJIFILM Wako Pure Chemical, Osaka, Japan. Cellulase Onozuka R-10 and Macerozyme R-10 were the products of Yakult, Tokyo, Japan. ABTS and N-phenylthiourea were purchased from Nacalai Tesque, Kyoto, Japan, and Sigma-Aldrich, St Louis, MO, USA, respectively. All other chemicals were of analytical grade.
Isolation of PC, HSYA and SYB from safflower corollas
An appropriate amount of orange safflower corollas (typically 50 g fresh weight) was incubated overnight with 150 ml of 8:2 (v/v) mixture of methanol and water. The mixture was then filtered through a nylon net (100 µm) to obtain washed corollas. The washed corollas were pulverized in 50 ml acetone containing 0.1% (v/v) trifluoroacetic acid (TFA) in a mortar. The homogenate was centrifuged at 4°C and 15,000×g for 10 min. The supernatant was concentrated to a small volume by evaporation, followed by filtration through a 0.2-µm polytetrafluoroethylene (PTFE) filter. The filtrate was diluted with four volumes of water. A 4-ml aliquot of the mixture was subjected to preparative high-performance liquid chromatography (HPLC) performed using a Clarity Chromatography Station (LCScience, Nara, Japan): column, J’sphere ODS-M20 column (20 mm × 250 mm; YMC, Kyoto, Japan), flow rate, 6.0 ml min-1; solvent A, 0.1% TFA (v/v) in water; and solvent B, 9:1 mixture of acetonitrile and 0.1% TFA (v/v) in water. The column was previously equilibrated with 10% B. After injection into the column, the column was developed with a linear gradient from 10% B to 50% B in 70 min, followed by a linear gradient from 50% B to 100% B in 1 min. The column was then washed with 100% B for 5 min, followed by a linear gradient from 100% B to 10% B in 1 min, and subsequently washed with 10% B for 24 min. The chromatograms were obtained with detection at 400 nm using a variable wavelength UV detector (model SPD-10A, Shimadzu, Kyoto, Japan). The pigment peaks were collected and evaporated to dryness. The m/z values and absorption spectra of the purified pigments were consistent with those previously reported (see Supplementary Table S2), as analyzed by LC/MS (system A, see below). Weighed quantities of the purified pigments were dissolved in appropriate solvents [methanol for HSYA and SYB, and 80% (v/v) methanol for PC] to generate stock solutions of known concentrations. The molar extinction coefficient, Ε406 = 45,708 cm-1 M-1, of PC (in methanol) was reported elsewhere (Kazuma et al. 2000) and was used for quantification of PC in this study.
Preparation of carthamin
Carthamin was purified from a commercially available natural colorant, LIOFRESH Red CR (TOYOCHEM, Tokyo, Japan) as follows. An appropriate amount (0.5–1.0 g) of LIOFRESH Red CR was suspended in 10 ml of water. The suspension was vigorously vortexed and centrifuged at 10,000×g for 10 min. The supernatant was removed. To the precipitate, 2 ml dimethylformamide was added, vortexed and centrifuged at 10,000×g for 10 min. The supernatant was filtered through a 0.2-µm PTFE filter. The filtrate was diluted with four volumes of water, and a 4-ml aliquot was subjected to preparative HPLC as described in the preceding section. The chromatograms were obtained with detection at 520 nm using a variable wavelength UV detector (model SPD-10A, Shimadzu). The m/z value and absorption spectra of the purified pigment were consistent with those previously reported for carthamin (see Supplementary Table S2), as analyzed by LC/MS (system A, see below). Weighed quantities of the purified carthamin were dissolved in 80% (v/v) methanol to produce stock solutions of known concentrations. The molar extinction coefficient, Ε515 = 48,978 cm−1 M−1, of carthamin (in methanol) was reported elsewhere (Kazuma et al. 2000) and was used for the quantification of carthamin in this study.
Quinochalcone analysis
Weighed quantities of orange safflower corollas (typically 1–2 g fresh weight) were pulverized in liquid nitrogen in a mortar. Five milliliters, per gram of pulverized material, of an 8:2 (v/v) mixture of methanol and water was added and vigorously vortexed. The mixture was centrifuged at 4°C and 15,000×g for 10 min to remove insoluble materials. For the qualitative analysis, the supernatant was mixed with two volumes of a 1:1 mixture of methanol and chloroform and vigorously vortexed for 3 min, followed by centrifugation at 15,000×g for 10 min. The upper layer was recovered, and a 5-µl aliquot was subjected to LC/MS analysis (system A; see below). For the quantitative analysis, the supernatant was appropriately diluted with water and subjected to HPLC analysis (system B; see below).
System A—LC system, LC solution (Shimadzu); column, CAPCELL CORE C18 (2.1 mm × 100 mm; Shiseido, Tokyo, Japan); flow rate, 0.2 ml/min; solvent A, 0.05% (v/v) formic acid in water; and solvent B, 0.05% (v/v) formic acid in acetonitrile. The column was equilibrated with 8% B. After injection onto the column, the column was developed with 8% B for 5 min and then with a linear gradient from 8% B to 40% B in 45 min, followed by a linear gradient from 40% B to 90% B in 10 min. The column was then washed with 90% B for 10 min, followed by a linear gradient from 90% B to 8% B in 1 min and subsequently washed with 8% B for 9 min. The chromatograms were obtained with detection at 190–900 nm using a photodiode array system (model SPD-M20A, Shimadzu) and LCMS-2020 (Shimadzu) (scanned m/z range: negative, 200–1100; positive 200–1100). The m/z values, λmax values and typical retention times of some safflower quinochalcones under these analytical conditions are provided in Supplementary Table S2.
System B—LC system, LC solution; column, Mightysil RP-18GP II (4.6 mm × 150 mm; particle size, 5 µm; Kanto Kagaku, Tokyo, Japan), flow rate, 0.7 ml min-1; solvent A, 0.1% TFA (v/v) in water; and solvent B, a 9:1 mixture of acetonitrile and 0.1% TFA (v/v) in water. The column was equilibrated with 10% B. After injection onto the column, the column was isocratically developed with 10% B for 5 min and then with a linear gradient from 10% B to 40% B in 35 min, followed by a linear gradient from 40% B to 100% B in 5 min. The column was then washed with 100% B for 4 min, followed by a linear gradient from 100% B to 10% B in 1 min and subsequently washed with 10% B for 10 min. The chromatograms were obtained with detection at 190–900 nm using a photodiode array system (model SPD-M10A, Shimadzu). The standard curves of HSYA, SYB and PC were generated by analyzing the standards after appropriate dilutions of the stock solutions (see above).
Enzyme assays
Method I (CarS assay)—The assay mixture consisted of 1 μmol sodium citrate buffer, pH 5.0, 0.1 nmol PC and enzyme in a final volume of 15 μl. The reaction was started by the addition of enzyme, and the mixture was allowed to stand at room temperature for 5 min, during which color changes (yellow to pink; see Fig. 5) were visually observed.
Method II (CarS assay)—The assay mixture consisted of 1 μmol sodium citrate buffer, pH 5.0, 0.05 nmol PC and enzyme in a final volume of 10 μl. The reaction was started by the addition of enzyme, and the mixture was allowed to stand at 30°C for 1–10 min. The reaction was stopped by the addition of 190 μl methanol, followed by centrifugation at 15,000×g and 4°C for 10 min. The absorbance of the supernatant (50 μl) was measured at 515 nm using a UV-VIS spectrophotometer (model U-2910, Hitachi High-Tech Science, Tokyo, Japan) to quantify carthamin.
Method III (peroxidase assay)—The assay mixture consisted of 5 μmol sodium citrate buffer, pH 5.0, 50 nmol ABTS, 5 nmol H2O2 and enzyme in a final volume of 50 μl. The reaction was started by the addition of enzyme, and the initial velocity was calculated by measuring the absorbance at 420 nm using a UV-VIS spectrophotometer (model U-2910). The extinction coefficient of the oxidized form of ABTS, Ε420 = 3.6 × 104 cm−1 M−1 (Shin and Lee 2000), was used for unit calculations.
Purification of CarS
All operations were performed at 0–4°C.
Purification from orange corollas.
Step 1: Preparation of crude extract—Orange corollas (stage 2, 500 g; the involucre was removed) were suspended in 1,000 ml of 10 mM Tris-HCl, pH 7.5 (termed buffer A), containing 150 mM NaCl, and disrupted in a Waring blender. The homogenate was filtered through a cheesecloth and the filtrate was centrifuged at 5,600×g for 20 min, followed by centrifugation at 7,700×g for 20 min. The supernatant was extensively dialyzed against buffer A. After dialysis, the enzyme solution was centrifuged at 7,700×g for 20 min, followed by passage through filters (Advantec No. 1; Advantec Toyo, Tokyo, Japan).
Step 2: Q Sepharose FF—The filtrate was applied to a column of Q Sepharose FF (Cytiva, Tokyo, Japan; column volume, 300 ml) equilibrated with buffer A. The column was washed with buffer A. The CarS activity was detected in the flow-through and column-wash fractions, as assayed by method I. The active fractions were combined and concentrated by ultrafiltration using an Amicon Ultra-15 centrifugal filter unit (Ultracel-10 kDa; Millipore, USA), followed by passage through Advantec No. 1 filters.
Step 3: SP Sepharose FF—The filtrate was applied to a column of SP Sepharose FF (Cytiva; column volume, 100 ml) equilibrated with buffer A. The column was washed with 300 ml buffer A, then with 300 ml buffer A containing 0.5 M NaCl and finally with 600 ml buffer A containing 1.0 M NaCl. The active fractions were identified by assay method I, combined and dialyzed against 0.01 M potassium phosphate, pH 7.5. The enzyme solution was centrifuged at 10,000×g for 20 min.
Step 4: Foresight CHT type I—The enzyme solution was applied to a column of Foresight CHT type I (Bio-Rad Laboratories, Tokyo, Japan; column volume, 5 ml) equilibrated with 0.01 M potassium phosphate, pH 7.5. After washing the column with the equilibration buffer, the enzyme activity was eluted with a linear gradient (0–0.25 M) of potassium phosphate, pH 7.5 (total gradient volume, 50 ml) at the flow rate of 2.0 ml min-1 using an ÄKTA start apparatus (GE Healthcare Japan, Tokyo, Japan). The active fractions were identified by assay method I, combined and concentrated to 0.5 ml using an Amicon Ultra-15 centrifugal filter unit (Ultracel-10 kDa).
Step 5: Superdex 200 10/300 GL—The enzyme solution was applied to a column of Superdex 200 10/300 GL (GE Healthcare Japan), which was equilibrated with buffer A containing 0.15 M NaCl and eluted at the flow rate of 0.2 ml min-1 using an ÄKTA purifier apparatus (GE Healthcare Japan). The active fractions were identified by assay method I and combined. The buffer was replaced by 0.01 M potassium phosphate, pH 7.5, by repeated concentrations and dilutions by ultrafiltration.
Step 6: Bio-Scale CHT2-I—The enzyme was subjected to chromatography on a Bio-Scale CHT2-I column (Bio-Rad Laboratories; column volume, 2 ml) equilibrated with 0.01 M potassium phosphate, pH 7.5. After washing the column with the equilibration buffer, the enzyme activity was eluted with a linear gradient (0–0.5 M) of NaCl in the equilibration buffer (total gradient volume, 40 ml) at the flow rate of 0.4 ml min-1 using an ÄKTA purifier apparatus. The active fractions were identified by assay method I, combined and concentrated by ultrafiltration. The buffer was replaced by buffer A by repeated concentration and dilution by ultrafiltration.
Step 7: Mono S 5/50—The enzyme was subjected to chromatography on a Mono S 5/50 column (GE Healthcare Japan; column volume, 1 ml) equilibrated with buffer A. After washing the column with the equilibration buffer, the enzyme activity was eluted with a linear gradient (0–0.175 M) of NaCl in the equilibration buffer (total gradient volume, 16 ml) at the flow rate of 0.5 ml min-1 using an ÄKTA purifier apparatus. The active fractions were identified by assay method I, combined and concentrated by ultrafiltration.
Step 8: Superdex 200 10/300 GL—The enzyme was subjected to chromatography on Superdex 200 10/300 GL essentially as described above (step 5). The active fractions were identified by assay method I, combined and concentrated by ultrafiltration, followed by rechromatography on Superdex 200 10/300 GL.
Purification from cream corollas.
Step 1: Preparation of crude extract—Cream corollas (stage 1, 130 g; the involucre was removed) were suspended in 500 ml buffer A containing 150 mM NaCl and disrupted in a Waring blender. The homogenate was filtered through a cheesecloth and the filtrate was centrifuged at 5,600×g for 20 min, followed by centrifugation at 5,600×g for 20 min. The supernatant was extensively dialyzed against buffer A. After dialysis, the enzyme solution was centrifuged at 7,700×g for 20 min, followed by passage through filters (Advantec No. 1).
Step 2: DEAE Sephadex A-25—The filtrate was applied to a column of DEAE Sephadex A-25 (Cytiva, Tokyo, Japan; column volume, 200 ml) equilibrated with buffer A. The column was washed with buffer A. The CarS activity was detected in the flow-through fraction, as assayed by assay method II.
Step 3: SP Sepharose FF—The flow-through fraction was applied to a column of SP Sepharose FF (Cytiva; column volume, 100 ml) equilibrated with buffer A. The column was washed with 300 ml of buffer A. The CarS activity was detected in the flow-through fraction, as assayed by assay method II.
Step 4: Ammonium sulfate fractionation—The flow-through fraction was subjected to ammonium sulfate precipitation (90% saturation). The resulting precipitate was recovered by centrifugation at 7,700×g for 30 min and dissolved in buffer A, followed by centrifugation. The supernatant was extensively dialyzed against buffer A.
Step 5: CM Sepharose FF—The enzyme solution was applied to a column of CM Sepharose FF (Cytiva; column volume, 100 ml) equilibrated with buffer A. The column was washed with 300 ml of buffer A and then with 200 ml of buffer A containing 0.1 M NaCl. The active fractions were identified by assay method II, combined and the buffer was replaced by 0.01 M potassium phosphate, pH 7.0, by repeated concentrations and dilutions by ultrafiltration using an Amicon Ultra-15 centrifugal filter unit (Ultracel-10 kDa).
Step 6: Bio-Scale CHT2-I—The enzyme was subjected to chromatography on a Bio-Scale CHT2-I column (Bio-Rad Laboratories; column volume, 2 ml) equilibrated with 0.01 M potassium phosphate, pH 7.5. After washing the column with the equilibration buffer, the enzyme activity was eluted with a linear gradient (0–0.4 M) of NaCl in the equilibration buffer (total gradient volume, 20 ml) at the flow rate of 1.0 ml min-1 using a BioLogic DuoFlow 10 System equipped with a Model 2110 fraction collector (Bio-Rad Laboratories). The active fractions were identified by assay method II, combined and concentrated by ultrafiltration.
Step 7: Superdex 200 10/300 GL—The enzyme solution was applied to a column of Superdex 200 10/300 GL (Cytiva), which was equilibrated with buffer A containing 0.15 M NaCl and eluted at the flow rate of 0.3 ml/min using a BioLogic DuoFlow 10 system. The active fractions were identified by assay method II.
Protein chemical analysis
SDS-PAGE was conducted in accordance with the method of Laemmli (1970). The proteins on the gels were visualized by silver staining. The amino acid sequences of the tryptic peptides (see the ‘Results’ section) were determined with an EASY-n LC1000 system equipped with the Q Exactive Plus apparatus (Thermo Fisher Scientific, Tokyo, Japan).
CarS reactions under N2-substituted conditions
CarS (CtPOD1) (6.4 pkat in 2 μl) and PC (1 mM, 5 μl) were separately placed on the bottom of rubber-sealed flat-bottomed 13-ml glass vials. For CarS reaction in the presence of H2O2, 1 μl of 50 mM H2O2 was also placed in the bottom of the vials. The vials were N2-substituted. Subsequently, 100 μl of N2-substituted 100 mM sodium citrate, pH 5.0, which was prepared by bubbling of N2 gas for 1 h, was injected into the N2-substituted vial to generate a homogeneous reaction mixture, followed by incubation at room temperature for 10 min. The reaction was stopped by injecting 500 μl methanol into the vials. The resulting mixture was subjected to HPLC analysis (System B). CarS reactions under aerobic conditions in the absence or presence of H2O2 were carried out in the same manner as described above except that all N2-substitution procedures were omitted.
Transcriptome analysis
Total RNA was extracted from the corollas (stage 3) of an orange cultivar using kits [Fruit-mate for RNA Purification (TakaRa Bio, Shiga, Japan) and RNAiso plus (TakaRa Bio)]. The extracted RNA was treated with DNase I Recombinant, RNase free (Roche, Basel, Switzerland). Preparation of cDNA sequencing libraries using the SureSelect Strand-Specific RNA Library Preparation Kit (Agilent Technologies, Santa Clara, CA, USA) and their pair-end sequencing using the Illumina HiSeq 2500 system were outsourced to Kazusa DNA Research Institute, Chiba, Japan. A total of 57,265,702 raw reads (28,632,851 raw read pairs) were subjected to quality control by clipping adapter sequences, trimming low-quality bases and removing overly short reads using Trimmomatic (Bolger et al. 2014), with the following custom parameters: LEADING: 30, TRAILING: 30 and MINLEN: 50. The remaining high-quality 28,192,385 read pairs were assembled using Trinity v2.8.4 (Haas et al. 2013) with the following custom parameter: min_contig_length: 300. The resultant 54,102 assembled transcripts were functionally annotated by predicting the protein sequences using TransDecoder.LongOrfs v5.5.10, and homologous protein information was assigned using a Protein BLAST (blastp) search against the Swiss-Prot database.
cDNA cloning of CtPOD1 and related CtPOD paralogs
Total RNA was prepared from corollas (stages 1–5) of orange safflower ‘Orange Maruba-shu’ using a FruitMate RNA purification kit and RNAiso plus (TaKaRa Bio). Equal amounts (1 μg) of total RNA from each stage were combined and used for cDNA synthesis, which was performed using the Prime Script II 1st-Strand cDNA Synthesis kit (Takara Bio) in accordance with the manufacturer’s guidelines. Using the cDNA library as a template, CtPOD1 cDNA was amplified by PCR with KOD FX Neo DNA polymerase (TOYOBO, Osaka, Japan) using the primers CtPOD1_UTR_Fw and CtPOD1_UTR_Rv (first PCR; for nucleotide sequences of PCR primers, see Supplementary Table S3). The PCR product was then subjected to second PCR using the primers CtPOD1_CDS_Fw and CtPOD1_CDS_Rv (Supplementary Table S3). The thermal cycling conditions (for first and second PCRs) were as follows: 94°C for 2 min; 35 cycles of 98°C for 10 s, 58°C for 30 s and 68°C for 50 s. The PCR product was gel-purified and cloned into the pGEM-T Easy vector (Promega, Tokyo, Japan). The recombinant plasmid was used to transform E. coli DH5 α cells. The transformant E. coli cells were selected and grown at 37°C overnight in liquid Luria–Bertani medium supplemented with 50 μg ml-1 ampicillin. The plasmids were recovered from the cells using the Fast Gene Plasmid Mini Kit (Nippon Genetics, Tokyo, Japan). The DNA insert was sequenced to confirm the nucleotide sequence of CtPOD1. The cDNAs of other CtPODs (CtPOD2, CtPOD3, CtPOD4 and CtPOD5) were obtained essentially as described above except that different sets of PCR primers were used (see Supplementary Table S3).
Quantitative real-time RT-PCR
Total RNA was extracted from individual organs of safflower plants (see Fig. 6) as previously described (Funaki et al. 2015). The extracted RNA was used as the template for reverse transcription, which was conducted using the ReverTra Ace qPCR RT Master Mix (TOYOBO). The CtPOD transcripts were analyzed by quantitative real-time PCR using the THUNDERBIRD SYBR qPCR Mix (TOYOBO), gene-specific primers (e.g. RT-CtPOD1_Fw and RT-CtPOD1_Fw for CtPOD1; see Supplementary Table S3 for nucleotide sequences) and the Eco Real-Time PCR System (Illumina, Tokyo, Japan). The thermal cycling conditions were 95°C for 1 min, followed by 40 cycles of 95°C for 15 s and 60°C for 30 s. Each PCR was followed by a melting curve analysis to confirm that only one PCR product was amplified. Gene transcripts were quantified after normalization against the safflower actin gene (CtACT) transcript levels. The nucleotide sequence of the CtACT gene was identified in the transcriptome data obtained in this study by using that of the soybean actin gene (Phytozome v12.1, gene locus Glyma.05G188800) as a query and was used to design the PCR primers RT-CtACT_Fw and RT-CtACT_Rv (Supplementary Table S3). A standard curve was generated using a serial dilution of the pGEM-T Easy-based construct for each gene (see above).
Subcellular localization analysis
The coding sequences lacking the translation termination codon of CtPOD1 and CtPOD2 were amplified by PCR using the pGEM-T Easy-based constructs (see above) as templates and the corresponding primer sets (see Supplementary Table S3). The fragments were subcloned into the pDOE-13 vector (Gookin and Assmann 2014; The Arabidopsis Biological Resource Center, Columbus, OH, USA) as described previously (Mameda et al. 2018). The fusion protein was transiently expressed with mTurquoise2 in N. benthamiana leaves, and the fluorescence in the transformed leaves was observed using a TCS-SP8 laser scanning confocal microscope (Leica, Mannheim, Germany) as described previously (Mameda et al. 2018).
Supplementary Data
Supplementary Data are available at PCP online.
Data Availability
The nucleotide sequences for genes underlying this article are available in DNA Data Bank of Japan at https://www.ddbj.nig.ac.jp/index-e.html and can be accessed with accession numbers LC634056 (for CtPOD1), LC634057 (for CtPOD2), LC634058 (for CtPOD3), LC644555 (for CtPOD4), LC644556 (for CtPOD5) and DRA012178 [for transcriptome data of orange safflower corollas (stage 3)]. The other data underlying this article are available in the article and in its online supplementary material. Additional data related to this article may be shared on reasonable request to the corresponding author.
Acknowledgements
We thank Hiroyuki Kato, Kato Engei, Sendai, Japan, for cultivating safflower plants and Miho Hosoya, Tohoku University, for her support for pigment analyses. We thank Robert McKenzie, PhD, from Edanz (http://jp.edanz.com/ac), for editing the English text of a draft of this manuscript.
