Anthocyanins and Flavonols from the Flowers of Ranunculus Cultivars (Ranunculaceae) and Their Color Expression

Abstract

Isolation, purification and identification of anthocyanins and flavonols were carried out on flowers of Ranunculus cultivars. Three anthocyanins and 11 flavonols were characterized by chemical and spectroscopic techniques. The anthocyanins were identified as cyanidin 3-O-sambubioside, cyanidin 3-O-(6''-malonylsambubioside) and delphinidin 3-O-(6''-malonylsambubioside). The flavonols were identified as 3-O-glucosides and 3-O-sophorosides of kaempferol and quercetin, and their acylated compounds with malonic acid. Flower colors were divided into six groups, Red-Purple, White, Yellow-Orange, Orange, Red, and Violet groups using the Royal Horticultural Society Colour Chart. The absorption maxima of buffer solutions containing anthocyanins and flavonols isolated in this study were measured to understand the effect of intermolecular copigmentation between these compounds on flower color. The results showed that by addition of 3-O-(6''-malonylglucoside) of kaempferol or quercetin, the absorption maximum of cyanidin 3-O-(6''-malonylsambubioside) or delphinidin 3-O-(6''-malonylsambubioside) shifted bathochromically and exhibited a closer absorption maximum to fresh flower petals than anthocyanin alone. This indicates that the intermolecular copigment effect between anthocyanins and flavonols is responsible for the flower color expression in Ranunculus cultivars.

Introduction

The genus Ranunculus (Ranunculaceae) comprises ca. 600 species (Aslam et al., 2012). It exhibits a variety of flower colors and shapes, and production of cut flowers is expected to respond to changing consumer preferences (Mayoli et al., 2009). A survey in 2017 found that 44.7% of 188 cut flower growers in North America produce Ranunculus (Loyola et al., 2019). Although breeding efforts have led to increased petal numbers and stronger stems in Ranunculus cultivars, and other characteristics have also been modified to make it a more suitable cut flower product (Kenza et al., 2000), cultivars with blue petals have not been seen in the market. Given the popularity of blue flowers in recent years, a fundamental understanding of the pigments in its petals is important for further breeding of Ranunculus cultivars with new flower colors.

In 1996, the flower pigments of R. asiaticus were initially reported by Toki et al. Their study revealed the presence of four anthocyanin glycosides, 3-O-sambubiosides and 3-O-(6''-malonylsambubiosides) of delphinidin and cyanidin (Toki et al., 1996). Furthermore, in a 2019 study by Takamura and Morino, the main anthocyanins in the petals of four Temari series Ranunculus cultivars were determined. They found that acylated cyanidin 3-O-sambubioside was the predominant anthocyanin in the petals of three reddish Ranunculus cultivars, while acylated delphinidin 3-O-sambubioside was the main anthocyanin in the petals of a violet Ranunculus cultivar. Simultaneously, they also found that the primary flavonol aglycones in the flowers of four Temari series Ranunculus cultivars were kaempferol and quercetin (Takamura and Morino, 2019).

These studies offer valuable insights into the flower pigments of Ranunculus, particularly anthocyanins. However, there is currently a lack of research investigating the role of flavonoids other than anthocyanins, such as flavonols and flavones, in the expression of flower color in Ranunculus cultivars. It is known that flavonols and flavones in flower cells, which are pale-yellow or almost colorless compounds to the human eye, can cause the bluing of anthocyanin by coexisting with anthocyanin, a phenomenon called intermolecular co-pigmentation (Asen et al., 1972; Mizuno et al., 2013).

In this study, to obtain fundamental knowledge needed for breeding new colored Ranunculus cultivars, the chemical structures and contents of flavonoids that affect flower color in 12 Ranunculus cultivars with different flower colors were investigated. Furthermore, the intermolecular copigment effect between anthocyanins and flavonols in the flowers of Ranunculus was also examined.

Materials and Methods

Plant materials

Fresh petals of 12 Ranunculus cultivars were used in this study, i.e., five Elegance series cultivars (Takii & Co., Ltd., Japan), ‘Elegance Pink’, ‘Elegance White’, ‘Elegance Yellow’, ‘Elegance Orange’, and ‘Elegance Red’, and five Etude series cultivars (Miyoshi & Co., Ltd., Japan), ‘Etude Light Pink’, ‘Etude White’, ‘Etude Gold’, ‘Etude Orange’, and ‘Etude Red’. ‘Mistral Blackjack’ (Miyoshi & Co., Ltd.) were purchased from Yoshida Engei (Ibaraki Pref.) as cut flowers between December 2021 and February 2022. Another cultivar ‘Cottun’ (Aya Engei, Miyazaki Pref.) was purchased from the Japan Agricultural Cooperatives in Takachiho in March 2022. High performance liquid chromatography (HPLC) analysis was performed on all cultivars. The petals of ‘Elegance Orange’, ‘Mistral Blackjack’, and a mixture of ‘Elegance Red’ and ‘Etude Red’ were used to isolate anthocyanins and flavonols.

Measurement of flower colors

Flower colors of 11 Ranunculus cultivars (12 cultivars excluding ‘Etude Light Pink’) were recorded based on the Royal Horticultural Society Colour Chart (RHSCC), sixth edition, and by measuring CIE L*a*b* chromaticity values with a color reader (CR-10; Konica Minolta, Inc., Japan) (Gonnet, 1995, 1998).

Measurement of flower petals’ absorption spectra

The visible absorption spectra (400–700 nm) of the fresh petals of 11 Ranunculus cultivars (all cultivars except for ‘Etude Light Pink’) were measured directly on a UV-2600 spectrophotometer (Shimadzu, Japan) equipped with an integrated sphere ISR-2600 (Shimadzu). These spectra were measured with transmitted light.

Extraction, isolation, and identification of flavonoids

1)  Extraction

Fresh petals (0.2 g) from each cultivar were torn into pieces and immersed in 70% methanol aq. containing 5% HCOOH for quantitative HPLC analysis in triplicate. Petals were randomly selected from different individuals. Fresh petals of ‘Elegance Orange’ (37.2 g), ‘Elegance Red’ (9.3 g), ‘Etude Red’ (21.3 g) and ‘Mistral Blackjack’ (134.8 g) were extracted in methanol containing 5% HCOOH for isolation and identification of anthocyanins and flavonols. They were kept for one week in a refrigerator.

2)  Pigment analysis

Analytical HPLC was performed using a Shimadzu HPLC system with SPD-20A UV-Vis detector (190–700 nm) and an InertSustain AQ-C18 column (5 μm particle material, I.D. 4.6 × 250 mm, GL Sciences, Japan), at a flow rate of 1.0 mL·min⁻¹. The solvents A (5% HCOOH aq.) and B (90% MeCN aq., 5% HCOOH) were used for the mobile phase under a gradient condition. The gradient condition was as follows: a linear gradient from 10 to 50% solvent B for 35 min, and 50% solvent B for 10 min.

Quantitative analysis of all cultivars was performed. The calibration curves were prepared with authentic samples of rutin and cyanidin 3-O-rutinoside (Tokiwa Phytochemical Co., Ltd., Japan) to calculate the total flavonol and anthocyanin contents, respectively. These values were expressed as nmol per mg fresh weight.

3)  Isolation

Based on HPLC chromatographical data of the crude extracts, flower petals of ‘Elegance Orange’, ‘Mistral Blackjack’ and the mixture of ‘Elegance Red’ and ‘Etude Red’ were used for isolation. The evaporated extracts were applied to a Diaion HP-20SS column chromatography (Mitsubishi Chemical, Japan) and eluted with 280 mL 0.5% trifluoroacetic acid (TFA) aq., 280 mL 10% MeCN containing 0.5% TFA and 280 mL 25% MeCN containing 0.5% TFA.

The separated fractions were applied to preparative HPLC with a Shimadzu HPLC system equipped with an SPD-20A UV-Vis detector using Inertsil ODS-4 (5 μm particle material, I.D. 10 × 250 nm, GL Sciences) at a flow rate of 3.0–3.5 mL·min⁻¹, detection at 350 nm for flavonols and 530 nm for anthocyanins, and an eluent of 10–25% MeCN containing 5% HCOOH. The fractions that were not sufficiently isolated were subsequently applied to a Sephadex LH-20 gel filtration column for chromatography and eluted with 70% MeOH.

4)  Identification

The pigments isolated by preparative HPLC were injected into HPLC and their purity was confirmed. The molecular weights of the isolated compounds were measured by a Shimadzu liquid chromatography-mass spectrometry (LC-MS) system equipped with an SPD-20A UV-Vis detector (190–700 nm) using Inertsil ODS-4 (3 μm particle material, I.D. 2.1 × 100 mm, GL Sciences) at a flow rate of 0.2 mL·min⁻¹, eluent with 7–25% MeCN containing 5% HCOOH, ElectroSpray Ionization⁺ (ESI⁺) 4.5 kV and ESI⁻ 3.5 kV, 250°C. Based on the molecular weight of each compound and the absorption characteristics, the isolated compounds were compared with corresponding authentic samples by HPLC and thin-layer chromatography (TLC). The authentic samples of quercetin 3-O-sophoroside, quercetin 3-O-glucoside and kaempferol 3-O-glucoside were isolated from the leaves of Asarum yakusimense (Iwashina et al., 2005), the aerial parts and fruits of Osyris alba (Iwashina et al., 2008) and the fronds of Cyrtomium falcatum subsp. australe (Iwashina et al., 2006), respectively. The authentic samples of two anthocyanins, cyanidin 3-O-sambubioside and cyanidin 3-O-(6''-malonylsambubioside), were isolated from the corollas and calyces of Aeschynanthus cultivars (Iwashina et al., 2021). The above authentic samples were deposited in the Flavonoid Collection of the National Museum of Nature and Science, Japan. The authentic samples of kaempferol and quercetin were purchased from Extrasynthese (France).

Anthocyanin A2 was identified by nuclear magnetic resonance (NMR). [JEOL ECZ400S in CD₃OD-TFA (9:1) in tetra methyl silane (TMS) as 0 ppm, 400 MHz as ¹H NMR and 100 MHz as ¹³C NMR)] including ¹H and ¹³C NMR, HMQC, HMBC, COSY and NOESY. Flavonols F3, F5, F7, F9, and F10 were also identified by NMR [Bruker AV-600 in DMSO‑d₆ at 600 MHz (¹H NMR) and 150 MHz (¹³C NMR)] including ¹H and ¹³C NMR, HSQC, HMBC, COSY, and NOESY.

UV-Vis spectra were recorded on an UV-2600 spectrophotometer (Shimadzu). Anthocyanins (A1–A3) were measured at a wavelength at 220 to 700 nm after dissolving in methanol containing 0.01% (w/v) HCl. After adding three drops of 5% (w/v) aluminium chloride (AlCl₃) solution, absorption spectra were measured again. For flavonols (F1–F11), each component was dissolved in MeOH and measured at a wavelength of 220–500 nm. Furthermore, five reagents, NaOMe, AlCl₃, AlCl₃/HCl, NaOAc and NaOAc/H₃BO_3, were added to each sample according to Mabry et al. (1970).

TLC of the isolated anthocyanins (A1–A3) and flavonols (F1, F3, F5, F6, and F8–10) was performed using the solvent systems, BAW (n-BuOH:HOAc:H₂O = 4:1:5, upper phase), BuHCl (n-BuOH:2N HCl = 1:1, upper phase), AHW (HOAc:HCl:H₂O = 15:3:82) and 1% HCl for anthocyanins, and BAW, 15% HOAc and BEW (n-BuOH:EtOH:H₂O = 4:1:2.2) for flavonols. The spot colors were observed under UV light (365 nm) with and without ammonia vapor exposure.

Determination of the fresh flower pressed juice pH

The pH of the fresh petal pressed juice obtained from one cultivar selected from each flower group, namely, ‘Etude White’, ‘Etude Gold’, ‘Etude Orange’, ‘Etude Red’, ‘Mistral Blackjack’, and ‘Cottun’, was determined using a twin pH meter AS-212 (HORIBA Ltd., Japan). The measurements were performed in triplicate.

Deacylation and acid hydrolysis

Deacyl treatment of A3 and F2 was performed by the addition of 2 M HCl and the mixture was left for one week at room temperature. These solutions were then compared by HPLC with their assumed deacylated compounds, A1 and F1, respectively.

Acid hydrolysis of crude extracts (three replicates) of 12 cultivars was performed by adding 3.5 mol·L⁻¹ HCl (1 mL) and heating at 100°C for 30 min. After cooling, hydrolysates were filtered using Sep-Pak Plus Light C18 (Waters Corporation, USA). The fractions containing aglycones were analyzed by direct HPLC comparisons with authentic samples. The authentic samples of kaempferol and quercetin were purchased from Extrasynthese, while pelargonidin, cyanidin and delphinidin were purchased from Tokiwa Phytochemical Co., Ltd. The distributions of flavonol and anthocyanin aglycones in petals were calculated based on HPLC results.

In vitro reconstruction of flower color

Before color reconstruction, we investigated the chemical structures of major anthocyanins and flavonols, as well as the ratio of total anthocyanin content to total flavonol content and the pH of pressed petal juice in one delphinidin-type cultivar, ‘Cottun’ and one cyanidin-type cultivar, ‘Etude Red’.

The quantitative analysis results indicated that ‘Cottun’ primarily contains A2 and F5 as the main anthocyanin and flavonol, while ‘Etude Red’ is characterized by A3 and F7 as the main anthocyanin and flavonol (Table 2). The lyophilized powders of anthocyanins A2 and A3 from ‘Mistral Blackjack’, along with flavonol F5 from ‘Elegance Orange’, and F7 from ‘Elegance Red’ and ‘Etude Red’, were used for in vitro color reconstruction. Additionally, F10 isolated from ‘Elegance Orange’, ‘Elegance Red’, and ‘Etude Red’ was also utilized.

Quantitative experiments demonstrated that the ratios of total anthocyanins to total flavonols in ‘Cottun’ and ‘Etude Red’ were both close to 1:4 (Fig. 5). Consequently, the anthocyanin-to-flavonol ratios were set at 1:4 in experimental pairs.

In the control pair, two sets were established: A2 and A3. The experimental pairs consisted of A2+F10, A2+F7, A2+F5, A3+F10, and A3+F7. In all pairs, the anthocyanin concentration was set to 0.5 mM, while the flavonol concentration was set to 2 mM. The anthocyanin powders were dissolved in 2 μL DMSO and then diluted with a certain amount of McIlvaine buffer to give a concentration of 0.5 mM. McIlvaine buffer was utilized at a pH of 5.5, close to that of the pressed juice pH of ‘Cottun’ and ‘Etude Red’ (pH 5.4). The visible absorption data (400–700 nm) of these solutions were measured using a UV-2600 spectrophotometer (Shimadzu) with Nano Stick-S (Scinco, Korea) according to Mizuno et al. (2021).

Furthermore, to investigate how the ratio of anthocyanin to flavonol affects the color expression, we conducted supplementary experiments to observe the absorption maxima of the anthocyanin and flavonol mixture at a ratio of 1:8 in McIlvaine buffer with a pH of 5.4. The experiment pairs were as follows: A2+F10, A2+F7, A3+F10, and A3+F7.

Results

Analysis of flower colors

Distribution of flower colors in eleven Ranunculus cultivars is shown in Figure 1. Eleven cultivars were divided into six groups, i.e., Red-Purple, White, Yellow-Orange, Orange, Red, and Violet groups, based on the RHSCC (Table 1). The hue angles, represented by arctan (b*/a*), were 345.40–358.34° for the Red-Purple group, 75.22–82.77° for the Yellow-Orange group, 62.30–68.09° for the Orange group, and 20.75–25.27° for the Red group. In terms of L*, the Red-Purple group ranged from 28.97 ± 1.57 to 71.87 ± 2.55, the White group ranged from 79.17 ± 0.67 to 87.97 ± 0.19, the Yellow-Orange group ranged from 66.60 ± 0.25 to 78.47 ± 0.35, the Orange group ranged from 57.07 ± 0.52 to 67.87 ± 0.66 and the Red group ranged from 32.13 ± 0.61 to 38.70 ± 1.00.

Fig. 1

Distribution of flower colors in Ranunculus cultivars (CIE L*a*b* chromaticity diagram.).

Table 1

Flower color and spectral data of Ranunculus cultivar fresh petals.

Based on absorption maxima of fresh petals, Red-Purple, Yellow-Orange, Orange, Red, and Violet group cultivars exhibited absorption maxima within the 530 to 560 nm range (Table 1), indicating the presence of anthocyanins. Additionally, absorption maxima of the Yellow-Orange, Orange, and Red group cultivars were observed between 400 and 500 nm. The spectral data suggested that ‘Elegance Yellow’, ‘Elegance Orange’, ‘Elegance Red’, ‘Etude Gold’, ‘Etude Orange’, and ‘Etude Red’ contain carotenoids as flower pigments. No or weak absorption maxima were detected in the White group cultivars under visible wavelength, indicating the absence or low level anthocyanin content in these cultivars.

Identification of anthocyanins and flavonols

Three anthocyanins (A1: 6.7 mg, A2: 32.1 mg, and A3: 5.4 mg) and eleven flavonols (F1: trace, F2: trace, F3: 1.1 mg, F4: trace, F5: 37.6 mg, F6: trace, F7: 29.2 mg, F8: 1.6 mg, F9: trace, F10: 8.4 mg, and F11: trace) were isolated and characterized. The chemical structures of the anthocyanins and flavonols isolated from the flowers of Ranunculus cultivars are shown in Figures 2 and 3.

Fig. 2

Chemical structures of anthocyanins isolated from the flowers of Ranunculus cultivars. A1: cyanidin 3-O-sambubioside, A2: delphinidin 3-O-(6''-malonylsambubioside), A3: cyanidin 3-O-(6''-malonylsambubioside).

Fig. 3

Chemical structures of flavonols isolated from the flowers of Ranunculus cultivars. F1: quercetin 3-O-sophoroside, F2: quercetin 3-O-(malonylsophoroside), F3: kaempferol 3-O-sophoroside, F5: kaempferol 3-O-(6''-malonylsophoroside), F6: quercetin 3-O-glucoside, F7: quercetin 3-O-(6''-malonylglucoside), F8: kaempferol 3-O-glucoside, F9: kaempferol 3-O-(6'''-malonylsophoroside), F10: kaempferol 3-O-(6''-malonylglucoside), F11: quercetin.

Table 2

Distribution of anthocyanins and flavonols from the petals of Ranunculus cultivars.

A1 Cyanidin 3-O-sambubioside (Sambicyanin): TLC: Rf 0.45 (BAW), 0.21 (BuHCl), 0.16 (1% HCl), 0.53 (AHW). UV/Vis λmax (nm): 0.01% HCl-MeOH 281, 334, 380, 530; E₄₄₀/E_max 21%; +AlCl₃ bathochromic shift. LC-MS: m/z 581 [M]⁺, 287 [M−glucosyl−xylosyl]⁺. HPLC: tR 10.8 min.

A2 Delphinidin 3-O-(6''-malonylsambubioside): TLC: Rf 0.45 (BAW), 0.29 (BuHCl), 0.30 (1% HCl), 0.66 (AHW). UV/Vis λmax (nm): 0.1% HCl-MeOH 279, 347, 541; E₄₄₀/E_max 18%; +AlCl₃ bathochromic shift. LC-MS: m/z 683 [M]⁺, 303 [M−glucosyl−xylosyl−malonyl]⁺. HPLC: tR 12.5 min. ¹H and ¹³C NMR, Table 3.

Table 3

The ¹H and ¹³C NMR data of anthocyanins and flavonols isolated from the petals of Ranunculus cultivars.

A3 Cyanidin 3-O-(6''-malonylsambubioside): TLC: Rf 0.49 (BAW), 0.35 (BuHCl), 0.30 (1% HCl), 0.66 (AHW). UV/Vis λmax (nm): 0.01% HCl-MeOH 269, 356, 531; E₄₄₀/E_max 22%; +AlCl₃ bathochromic shift. LC-MS: m/z 667 [M]⁺, 287 [M−glucosyl−xylosyl−malonyl]⁺. HPLC: tR 14.4 min.

F1 Quercetin 3-O-sophoroside (Baimaside): TLC: Rf 0.64 (BAW), 0.63 (15% HOAc), and 0.65 (BEW); color UV (365 nm): dark purple, UV/NH₃: yellow. UV: λmax (nm) MeOH 257, 357; +NaOMe 273, 327, 408 (inc.); +AlCl₃ 274, 430; +AlCl₃/HCl 269, 297, 371, 398; +NaOAc 266, 306, 374; +NaOAc/H₃BO₃ 266, 380. LC-MS: m/z 649 [M+H+Na]⁺, 625 [M−H]⁻, 303 [M−diglucosyl+H]⁺. HPLC: tR (min) 14.2.

F2 Quercetin 3-O-malonylsophoroside: LC-MS: m/z 735 [M+H+Na]⁺, 711 [M−H]⁻, 303 [M−diglucosyl−malonyl+H]⁺, 625 [M−malonyl−H]⁻. HPLC: tR (min) 15.4.

F3 Kaempferol 3-O-sophoroside (Sophoraflavonoloside): TLC: Rf 0.80 (BAW), 0.69 (15%HOAc), 0.78 (BEW); color UV (365 nm): dark purple, UV/NH₃: yellow. UV: λmax (nm) MeOH 266, 348; +NaOMe 275, 324, 397 (inc.); +AlCl₃ 274, 303, 348, 395; +AlCl₃/HCl 274, 301, 345, 393; +NaOAc 274, 306, 380; +NaOAc/H₃BO₃ 266, 352. LC-MS: m/z 633 [M+H+Na]⁺, 609 [M−H]⁻, 449 [M−monoglucosyl+H]⁺, 287 [M−diglucosyl+H]⁺. HPLC: tR 16.2 min. ¹H and ¹³C NMR, Table 3.

F4 Quercetin 3-O-malonylpentosylhexoside: UV: λmax (nm) MeOH 260, 306, 361; +NaOMe 260, 306, 413 (inc.); +AlCl₃ 273, 415; +AlCl₃/HCl 268, 296, 361, 394; +NaOAc 270, 306, 406; +NaOAc/H₃BO₃ 267, 306, 396. LC-MS: m/z 683 [M+H]⁺, 681 [M−H]⁻, 551 [M−pentosyl+H]⁺, 303 [M−malonyl−glucosyl−pentosyl+H]⁺, 637 [M−CO₂−H]⁻. HPLC: tR 16.8 min.

F5 Kaempferol 3-O-(6''-malonylsophoroside): TLC: Rf 0.80 (BAW), 0.79 (15%HOAc), 0.76 (BEW); color UV (365 nm): dark purple, UV/NH₃: yellow. UV: λmax (nm) MeOH 265, 350; +NaOMe 277, 316, 392 (inc.); +AlCl₃ 275, 302, 347, 396; +AlCl₃/HCl 274, 301, 346, 392; +NaOAc 273, 305, 375; +NaOAc/H₃BO₃ 266, 354. LC-MS: m/z 697 [M+H]⁺, 695 [M−H]⁻, 535 [M−glucosyl+H]⁺, 287 [M−malonyl−diglucosyl+H]⁺, 651 [M−CO₂−H]⁻. HPLC: tR 17.4 min. ¹H and ¹³C NMR, Table 3.

F6 Quercetin 3-O-glucoside (Isoquercitrin): TLC: Rf 0.74 (BAW), 0.31 (15%HOAc), and 0.76 (BEW); color UV (365 nm): dark purple, UV/NH₃: yellow. LC-MS: m/z 465 [M+H]⁺, 463 [M−H]⁻, 303 [M−glucosyl+H]⁺. HPLC: tR 18.0 min.

F7 Quercetin 3-O-(6''-malonylglucoside): UV: λmax (nm) MeOH 257, 359; +NaOMe 273, 328, 411 (inc.); +AlCl₃ 274, 423; +AlCl₃/HCl 268, 298, 371, 399; +NaOAc 272, 326, 379; +NaOAc/H₃BO₃ 262, 380. LC-MS: m/z 551 [M+H]⁺, 549 [M−H]⁻, 303 [M−malonylglucosyl+H]⁺, 505 [M−CO₂−H]⁻. HPLC: tR 19.1 min. ¹H and ¹³C NMR, Table 3.

F8 Kaempferol 3-O-glucoside (Astragalin): TLC: Rf 0.87 (BAW), 0.40 (15% HOAc), 0.88 (BEW); color UV (365 nm): dark purple, UV/NH₃: yellow. UV: λmax (nm) MeOH 265, 349; +NaOMe 276, 322, 398 (inc.); +AlCl₃ 274, 302, 347, 396; +AlCl₃/HCl 274, 301, 346, 392; +NaOAc 273, 301, 373; +NaOAc/H₃BO₃ 302, 340. LC-MS: m/z 449 [M+H]⁺, 447 [M−H]⁻, 287 [M−glucosyl+H]⁺. HPLC: tR 20.6 min.

F9 Kaempferol 3-O-(6'''-malonylsophoroside): TLC: Rf 0.79 (BAW), 0.57 (15%HOAc), 0.72 (BEW); color UV (365 nm): dark purple, UV/NH₃: yellow. LC-MS: m/z 697 [M+H]⁺, 695 [M−H]⁻, 287 [M−diglucosyl−malonyl+H]⁺, 651 [M−CO₂−H]⁻. HPLC: tR (min) 21.0. UV: λmax (nm) MeOH 265, 306, 347; +NaOMe 275, 326, 401 (inc.); +AlCl₃ 275, 302, 345, 395; +AlCl₃/HCl 273, 301, 345, 392; +NaOAc 273, 305, 371; +NaOAc/H₃BO₃ 266, 306, 350. ¹H and ¹³C NMR, Table 3.

F10 Kaempferol 3-O-(6''-malonylglucoside): TLC: Rf 0.95 (BAW), 0.51 (15%HOAc), and 0.92 (BEW); color UV (365 nm): dark purple, UV/NH₃: yellow. UV: λmax (nm) MeOH 265, 350; +NaOMe 276, 322, 396 (inc.); +AlCl₃ 275, 303, 348, 396; +AlCl₃/HCl 274, 301, 345, 394; +NaOAc 273, 306, 371; +NaOAc/H₃BO₃ 266, 353. LC-MS: m/z 535 [M+H]⁺, 533 [M−H]⁻, 287 [M−malonylglucosyl+H]⁺, 489 [M−CO₂−H]⁻. HPLC: tR 22.1 min. ¹H and ¹³C NMR, Table 3.

F11 Quercetin: LC-MS: m/z 303 [M+H]⁺, 301 [M−H]⁻. HPLC: tR (min) 27.4.

LC-MS of A1 and A3 indicated that they were likely cyanidin pentosylhexoside and cyanidin malonylpentosylhexoside; the UV-Vis absorption spectral data of A1 and A3 showed that they were 3-substituted anthocyanins. Finally, the retention times of A1 and A3 agreed with those of authentic samples from corollas and calyces of Aeschynanthus (Iwashina et al., 2021) by HPLC. Thus, A1 and A3 were identified as cyanidin 3-O-sambubioside and cyanidin 3-O-(6''-malonylsambubioside), respectively.

A2 was assumed to be delphinidin malonylpentosylhexoside based on LC-MS results. UV-Vis absorption spectral properties indicated that A2 was a 3-substituted delphinidin. The final identification of A2 was performed by ¹H and ¹³C NMR (Table 3). Proton and carbon signals were assigned by COSY, NOESY, HMQC, and HMBC (Table 3). The coupling constant of hexose is J = 7.8 Hz, and a NOESY correlation was observed between the H-2 and H-4 of glucose (Fig. 4). These facts indicated that the hexose was a β-D-glucopyranose. HMBC correlation was detected between the C-3 carbon signal at δ_C 145.40 and the glucosyl anomeric proton signal at δ_H 5.46. These results indicated that glucose is attached to the 3-position of delphinidin. HMBC correlation was also observed between the C-2 carbon signal of glucose at δ_C 82.49 and the anomeric proton signal of xylose at δ_H 4.73, indicating that the disaccharide is xylosyl-(1→2)-glucose (sambubiose). Furthermore, HMBC correlation was observed between glucosyl H-6a, H-6b proton signals at δ_H 4.33 and 4.54, and the carbonyl carbon signal of malonic acid at δ_C 168.77 (Fig. 3). These results indicated that a malonyl moiety was attached to the 6-position of 3-glucose. Thus, A2 is identified as delphinidin 3-O-[β-D-xylopyranosyl-(1→2)-β-D-(6''-malonylglucopyranoside)].

Fig. 4

Important cross peaks of HMBC (solid arrows) and NOESY (dashed arrows) in delphinidin 3-O-(6''-malonylsambubioside) (A, A2), kaempferol 3-O-sophoroside (B, F3), kaempferol 3-O-(6''-malonylsophoroside) (C, F5), quercetin 3-O-(6''-malonylglucoside) (D, F7), kaempferol 3-O-(6'''-malonylsophoroside) (E, F9), and kaempferol 3-O-(6''-malonylglucoside) (F, F10).

F1, F4, and F7 generated a fragment ion peak at m/z 303 [M+H]⁺ on LC-MS, and showed essentially the same UV-spectral properties, indicating that they were pentahydroxyflavones, assumed to be 3-substituted quercetin. Among them, F1 was compared with an authentic sample from the leaves of Asarum yakusimense (Iwashina et al., 2005) by HPLC, and they had the same retention times. Thus, F1 was identified as quercetin 3-O-sophoroside. F4 was assumed to be quercetin malonylpentosylhexoside and F7 was assumed to be quercetin malonylhexoside by LC-MS. Based on its UV-spectral properties, F7 was assumed to be 3-substituted quercetin. The final identification of F7 was made by ¹H and ¹³C NMR (Table 3). There were HMBC correlations between glucosyl H-6a and H-6b proton signals at δ_H 4.03 and 4.24 and the carbonyl carbon signal of malonic acid at δ_C 166.75 (Fig. 4). The coupling constant of the glucosyl anomeric proton signal was J = 7.2 Hz, showing that the glucose was in β-D-pyranose form, and a NOESY correlation was observed between the H-1 and H-3 proton signals of glucose (Fig. 4), indicating that the hexose was a β-D-glucopyranose. Therefore, F7 was identified as quercetin 3-O-β-D-(6''-malonylglucopyranoside). F7 has been reported from the aerial parts of Salicornia europaea (Geslin and Verbist, 1985).

The LC-MS result indicated that F2 was a malonylated compound of F1. The retention time of the hydrolysate of F2 was the same as that of F1 on HPLC. These results confirmed that F2 was quercetin 3-O-malonylsophoroside. However, the position at which malonic acid is attached could not be determined for a trace amount of F2.

F6 was characterized as a quercetin hexoside by LC-MS, and it was finally determined as quercetin 3-O-glucoside by HPLC comparison of the retention time with that of an authentic sample from the aerial parts and fruits of Osyris alba (Iwashina et al., 2008).

F11 exhibited a retention time of over 25 minutes on HPLC, and a molecular ion peak at m/z 301 [M−H]⁻ on LC-MS, suggesting that it was quercetin itself. An HPLC comparison with authentic quercetin confirmed that F11 was quercetin.

F8 was presumed to be a kaempferol glucoside according to LC-MS results. It was compared with the authentic sample from the fronds of Cyrtomium falcatum subsp. australe (Iwashina et al., 2006) by HPLC, and finally identified as kaempferol 3-O-glucoside.

Identification of F3, F5, F9, and F10 were mainly performed with LC-MS and NMR. Compound F3 was assumed to be kaempferol 3-O-dihexoside based on LC-MS and UV-spectra. HMBC correlations were shown between the C-2 carbon signal of 3-glucose at δ_C 82.34 and anomeric proton signal of another glucose at δ_H 4.61 (Fig. 4). The glucose coupling constants were J = 6.6 and 7.8 Hz, respectively, and a NOESY correlation was shown between their H-1 and H-5 signals (Fig. 4). These results indicated that these two glucoses were in β-D-pyranose form. Thus, F3 was identified as kaempferol 3-O-β-D-glucopyranosyl-(1→ 2)-β-D-glucopyranoside (kaempferol 3-O-sophoroside).

Although F5 and F9 showed the same molecular ion peak at m/z 697 [M+H]⁺ and fragment ion peak at m/z 287 [M−86−162−162+H]⁺ on LC-MS, their retention times were different. HMBC correlations of F5 were shown between the anomeric proton signal of one glucose at δ_H 5.58 and the C-3 carbon signal of kaempferol at δ_C 132.57, the anomeric proton signal of another glucose at δ_H 4.60 and C-2 carbon signal of 3-glucose at δ_C 81.94, and the carbonyl carbon signal of malonic acid at δ_C 166.27 and C-6 of 3-glucose at δ_H 3.92 and 4.17 (Fig. 4). On the other hand, an HMBC correlation with F9 occurred between the carbonyl carbon signal of malonic acid at δ_C 168.23 and H-6a and H-6b proton signals of 2''-glucose at δ_H 3.41 and 3.71 (Fig. 4). Thus, F5 and F9 were identified as kaempferol 3-O-[β-D-glucopyranosyl-(1→2)-β-D-(6''-malonylglucopyranoside)] and kaempferol 3-O-[β-D-(6'''-malonylglucopyranosyl)-(1→2)-β-D-glucopyranoside)], respectively. 7-Glycosylated F9 has been reported in the flowers of Papaver nudicaule (Tatsis et al., 2013). However, this is the first time that two different types of malonylated kaempferol 3-O-sophoroside have been isolated and identified separately.

F10 was presumed to be a kaempferol malonylglucoside based on LC-MS results. An HMBC correlation occurred between the C-3 carbon signal of kaempferol at δ_C 132.95 and the glucosyl anomeric proton signal at δ_H 5.36 (Table 3), showing that glucose is attached to the 3-position of kaempferol. Another HMBC correlation occurred between the H-6a and H-6b proton signals of glucose at δ_H 4.00 and 4.18 and the carbonyl carbon signal of malonic acid at δ_C 166.38. The coupling constant of glucose was J = 7.2 Hz (Table 3; Fig. 4), showing that glucose was in the β-D-pyranose form. Finally, F10 was identified as kaempferol 3-O-β-D-(6''-malonylglucopyranoside). This compound has been reported in the overground sporophytes of Equisetum species (Veit et al., 1995).

The same anthocyanins have been isolated from the petals of Ranunculus asiaticus (Toki et al., 1996). F1 and F6 have been isolated from the leaves of aquatic Ranunculus (subgenus Batrachium) species (Gluchoff-Fiasson et al., 1997), while F3 and F11 have been isolated from a batch of Ranunculus sceleratus L. (Cao et al., 2022).

Analysis of flavonoids

The main anthocyanins of ‘Elegance Pink’, ‘Elegance Orange’, ‘Elegance Red’, ‘Etude Light Pink’, ‘Etude Orange’, and ‘Cottun’ were cyanidin derivatives, specifically, A1 and A3 (Table 2). On the other hand, the main anthocyanin in ‘Cottun’ and ‘Mistral Blackjack’ was a delphinidin derivative, specifically, A2 (Table 2).

Variations in flavonoid composition were observed in two cultivars, i.e., ‘Mistral Blackjack’ and ‘Etude Gold’. In ‘Mistral Blackjack’, F5 was detected in samples 2 and 3, but was not found in sample 1. F7 was detected in samples 1 and 2, but was not found in sample 3. In ‘Etude Gold’, F4 was detected in samples 1 and 3, but was not found in sample 2 (Table 2).

Quantitative analysis of flavonols and anthocyanins in 12 cultivars

Total anthocyanin and flavonol contents from the flowers of 12 cultivars were calculated as nmol per mg fresh petal weight (Fig. 5). Among these cultivars, ‘Mistral Blackjack’ exhibited the most elevated level of anthocyanin content at 8.15 nmol·mg⁻¹, whereas ‘Etude White’ had the lowest level at 0.02 nmol·mg⁻¹. ‘Etude Red’ demonstrated the highest flavonol content at 19.29 nmol·mg⁻¹, while ‘Cottun’ exhibited the lowest at 5.60 nmol·mg⁻¹. Notably, both ‘Elegance Red’ and ‘Etude Red’, belonging to the Red Group, showed comparatively high levels of both anthocyanin and flavonol contents.

Fig. 5

Total anthocyanin and flavonol contents of 12 Ranunculus cultivars (n = 3) (Error bars represent SE.).

Determination of the pH of fresh flower pressed juice

The mean pH values of pressed petal juice (n = 3) were as follows: ‘Etude White’ 5.5, ‘Etude Gold’ 5.2, ‘Etude Orange’ 5.4, ‘Etude Red’ 5.4, ‘Mistral Blackjack’ 5.4, and ‘Cottun’ 5.4. Despite belonging to the same Etude series, ‘Etude White’, ‘Etude Gold’, ‘Etude Orange’, and ‘Etude Red’ exhibited distinct variations in petal pH.

Deacylation and acid hydrolysis

The hydrolysis product of A3 exhibited a peak on HPLC with the same retention time as A1, confirming that A3 was an acylated substance of A1. The same was true for F2, confirming that F2 was an acylated substance of F1.

As for petal extract hydrolysis, only two types of flavonol aglycone were detected in all cultivars. The aglycone proportions in petal extracts are shown below (calculated as peak area in HPLC analysis, kaempferol: quercetin): ‘Elegance Pink’ 46.5%: 15.7%, ‘Elegance White’ 53.9%: 6.5%, ‘Elegance Yellow’ 60.6%: 9.3%, ‘Elegance Orange’ 58.2%: 14.4%, ‘Elegance Red’ 2.3%: 69.6%, ‘Etude Light Pink’ 69.0%: 10.9%, ‘Etude White’ 64.4%: 3.2%, ‘Etude Gold’ 63.4%: 9.9%, ‘Etude Orange’ 71.6%: 9.5%, ‘Etude Red’ 10.8%: 67.8%, ‘Cottun’ 57.8%: 5.2%, and ‘Mistral Blackjack’ 8.5%: 47.8%. All the values were the average of three experiments. The results showed that the main flavonols present in White, Yellow, Yellow-Orange, and Violet group flowers were kaempferol derivatives. On the other hand, the main flavonols in ‘Elegance Red’ and ‘Etude Red’ (Red group) were quercetin derivatives. As for flowers in Red-Purple group, ‘Elegance Pink’ contained kaempferol derivatives as the main flavonols, while ‘Mistral Blackjack’ contained quercetin derivatives as the main flavonols.

For anthocyanins, three types of aglycones were detected, that is, cyanidin, delphinidin and pelargonidin. In ‘Elegance Pink’, ‘Elegance Orange’, ‘Elegance Red’, ‘Etude Light Pink’, ‘Etude Orange’, and ‘Etude Red’, only cyanidin was detected. Both cyanidin and delphinidin-type aglycone were detected in ‘Mistral Blackjack’ and ‘Cottun’. In ‘Etude Orange’, both pelargonidin type and cyanidin type aglycones were detected.

In vitro reconstruction of petal color

The direct visible absorption measurements of ‘Cottun’ and ‘Etude Red’ petals are shown in Figure 6A, B. The absorption maximum of ‘Cottun’ was 542.8 nm (shoulder: 562.4 nm) while the maxima of ‘Etude Red’ were 451.0 nm, 485.2 nm, and 534.0 nm (shoulder: 553.0 nm).

Fig. 6

Absorption spectra of fresh flower petals of ‘Cottun’, ‘Etude Red’, and of the reconstruction solutions. The visible absorption curves of ‘Cottun’ fresh petals (A), 0.5 mM A2, and a mixture of 0.5 mM A2 with 2 mM F7 or F10 or F5 (B), fresh petals of ‘Etude Red’ (C), 0.5 mM A3 and a mixture of 0.5 mM A3 with 2 mM F7 or F10 (D). The triangles on spectral lines represent an absorption maximum or a shoulder. A2: delphinidin 3-O-(6''-malonylsambubioside), A3: cyanidin 3-O-(6''-malonylsambubioside), F5: kaempferol 3-O-(6''-malonylsophoroside), F7: quercetin 3-O-(6''-malonylglucoside), F10: kaempferol 3-O-(6''-malonylglucoside).

The absorption spectrum of A2 displayed an absorption maximum at 533.5 nm (shoulder: 564.5 nm). The addition of F5 or F10 to A2 produced an absorption maximum at 543.0 nm (shoulder: 564.5 nm), which showed a similarity with the absorption curve of ‘Cottun’ (Fig. 6A). The addition of F7 to A2 resulted in a bathochromic shift of the absorption maximum to 544.0 nm (shoulder: 569.0 nm), indicating a bluer color.

The absorption spectrum of A3 displayed an absorption maximum at 542.0 nm. The absorption maximum remained 542.0 nm after the addition of F10. Subsequently, when F7 was added to A3, and the absorption maximum shifted to 547.0 nm, which was the nearest to the shoulder of ‘Etude Red’ (553.0 nm) among all the color reconstruction solutions (Fig. 6B).

When the ratio of anthocyanin and flavonol increased to 1:8 in pH 5.4, A2+F10 produced an absorption maximum at 558.0 nm (shoulder: 564.5 nm), A2+F7 produced an absorption maximum at 562.0 nm (shoulder: 564.5 nm), A3+F10 produced an absorption maximum at 546.5 nm, and A3+F7 produced an absorption at 547.5 nm. All the absorption maxima shifted bathochromically compared to their 1:4 corresponding pairs. Especially, the 1:8 pairs consisting of A2 showed bathochromic shifts of 15.0 nm and 18.0 nm compared to their corresponding 1:4 pairs, while 1:8 pairs consisting of A3 showed bathochromic shifts of 4.5 nm and 0.5 nm compared to their corresponding 1:4 pairs.

Discussion

For Yellow-Orange group and Orange group flowers, specifically ‘Elegance Yellow’, ‘Etude Gold’, ‘Elegance Orange’, and ‘Etude Orange’, absorption maxima were detected around 400–500 nm, indicating that carotenoids play a role in their color expression.

The same major flavonol was observed in ‘Elegance Yellow’ and ‘Etude Gold’, and the same major flavonol and anthocyanin components were observed in ‘Elegance Orange’ and ‘Etude Orange’, ‘Elegance Pink’ and ‘Etude Light Pink’, ‘Elegance Red’ and ‘Etude Red’ by pigment analysis. It can therefore be concluded that factors other than flavonoid components, such as carotenoid contents, pH and/or flavonoid contents are responsible for their subtle color distinction.

In a previous report, flavones, luteolin and apigenin, were observed in wild species of Ranunculus (Cao et al., 2022). However, flavonoids other than flavonol and anthocyanin were not detected in this study. We speculate that introducing pigments from wild species of Ranunculus into horticultural lines could potentially alter their color traits.

In a color reconstruction experiment, the absorption maximum of delphinidin 3-O-(6''-malonylsambubioside) shifted to the long wavelength side with the addition of kaempferol 3-O-(6''-malonylsophoroside) or kaempferol 3-O-(6''-malonylglucoside). The resulting spectral curve closely resembled that of ‘Cottun’. Previous studies have also reported that flavonol can induce a copigment effect on anthocyanins (Cao et al., 2023). This indicates that intermolecular copigmentation is responsible for the color expression of ‘Cottun’. In addition, the observed effect on flower color due to the differing glycosylation patterns of flavonols between F5 and F10 was minimal (Fig. 6). Furthermore, when kaempferol 3-O-(6''-malonylsophoroside) or kaempferol 3-O-(6''-malonylglucoside) was changed to quercetin 3-O-(6''-malonylglucoside), the absorption maxima shifted even further toward the bathochromic side, resulting in a bluer color. This suggests that quercetin derivatives induce a stronger intermolecular copigment effect with delphinidin than kaempferol derivatives. We speculate that for strategic breeding of blue-flowered Ranunculus cultivars, introducing quercetin derivatives would be effective.

When cyanidin 3-O-(6''-malonylsambubioside) was mixed with quercetin 3-O-(6''-malonylglucoside) or kaempferol 3-O-(6''-malonylglucoside), a bathochromic shift was observed in comparison with cyanidin 3-O-(6''-malonylsambubioside) alone. Especially, when quercetin 3-O-(6''-malonylglucoside) was added to cyanidin 3-O-(6''-malonylsambubioside), an absorption maximum nearest to the shoulder of ‘Etude Red’ was observed. From this result, it can be deduced that kaempferol and quercetin can induce intermolecular copigmentation with cyanidin, and intermolecular copigmentation also plays a role in the color expression of Red Group flowers. Although a mixture of the main flavonol and anthocyanin in ‘Etude Red’, specifically quercetin 3-O-(6''-malonylglucoside) and cyanidin 3-O-(6''-malonylsambubioside), failed to produce a similar absorption curve, we speculate that the presence of carotenoids in ‘Etude Red’ affected the absorption spectra, evidenced by an absorption maximum of carotenoid (451.0 nm).

When the ratio of anthocyanin to flavonol increased to 1:8, the solution showed a further bathochromic shift compared to that of corresponding 1:4 pairs, meaning that the color became bluer. This indicates that a change in the anthocyanin to flavonol ratio can change the color expression. It is also worth noting that even at a ratio of 1:8, delphinidin evoked a robust copigment effect, as evidenced by a bathochromic shift exceeding 10 nm. In contrast, cyanidin demonstrated a comparatively weaker copigment effect, with minimal changes in the shift observed.

In black flowers of ‘Mistral Blackjack’, inter-cultivar variation was observed, although variation in the flower color was not detected (Tables 1 and 2). It was shown that the flowers had an extremely high anthocyanin content. We speculated that not the intermolecular copigment, but the high anthocyanin content was responsible for its color expression. Coloring of black flowers due to high accumulation of anthocyanins was also reported in other species. For example, high accumulation of multiple anthocyanins in the petals of black hollyhock was reported to be responsible for its black color expression (Hosaka et al., 2012). In a blackish-flowered cultivar of Catharanthus roseus, high accumulation of one peonidin derivative was observed (Deguchi et al., 2020). In this study, the black color expression of ‘Mistral Blackjack’ was mainly due to the high accumulation of delphinidin derivatives. This suggests that high accumulation of cyanidin derivatives in flowers could be anticipated in the breeding new black-flowered Ranunculus cultivars.

Acknowledgements

We sincerely thank Mrs. Mari Sato of the Ibaraki South Agricultural and Forestry Office for introducing us to the Ranunculus farmers in Ibaraki Prefecture. We also thank the main provider of plant materials for this experiment, Mr. Kaoru Yoshida of Yoshida Engei. It was thanks to their support that this research was made possible.

Literature Cited

•  Asen,  S.,  R. N.  Stewart and  K. H.  Norris. 1972. Co-pigmentation of anthocyanins in plant tissues and its effect on color. Phytochemistry 11: 1139–1144.
•  Aslam,  M. S.,  B. A.  Choudhary,  M.  Uzair and  A. S.  Ijaz. 2012. The genus Ranunculus: a phytochemical and ethnopharmacological review. Int. J. Pharm. Pharm. Sci. 4: 15–22.
•  Cao,  S.,  M.  Hu,  L.  Yang,  M.  Li,  Z.  Shi,  W.  Cheng,  Y.  Zhang,  F.  Chen,  S.  Wang and  Q.  Zhang. 2022. Chemical constituent analysis of Ranunculus sceleratus L. Using ultra-high-performance liquid chromatography coupled with quadrupole-orbitrap high-resolution mass spectrometry. Molecules 27: 3299. DOI: 10.3390/molecules27103299.
•  Cao,  Y.,  Q.  Xia,  J.  Chen and  Z.  Jin. 2023. Copigmentation effect of flavonols on anthocyanins in black mulberry juice and their interaction mechanism investigation. Food Chem. 399: 133927. DOI: 10.1016/j.foodchem.2022.133927.
•  Deguchi,  A.,  F.  Tatsuzawa and  K.  Miyoshi. 2020. A blackish-flowered cultivar of Catharanthus roseus accumulates high concentrations of a novel anthocyanin with a unique feature of aggregation in weak acid solutions. Dyes Pigm. 173: 108001. DOI: 10.1016/j.dyepig.2019.108001.
•  Geslin,  M. and  J. F.  Verbist. 1985. Flavonoides de Salicornia europaea. J. Nat. Prod. 48: 111–113.
•  Gluchoff-Fiasson,  K.,  J. L.  Fiasson and  H.  Waton. 1997. Quercetin glycosides from European aquatic Ranunculus species of subgenus Batrachium. Phytochemistry 45: 1063–1067.
•  Gonnet,  J.-F. 1995. A colorimetric look at the RHS chart—Perspectives for an instrumental determination of colour codes. J. Hort. Sci. 70: 191–206.
•  Gonnet,  J.-F. 1998. Colour effects of co-pigmentation of anthocyanins revisited—1. A colorimetric definition using the CIELAB scale. Food Chem. 63: 409–415.
•  Hosaka,  H.,  T.  Mizuno and  T.  Iwashina. 2012. Flavonoid pigments and color expression in the flowers of black Hollyhock (Alcea rosea ‘nigra’). Bull. Natl. Mus. Nature Sci. Ser. B 38: 69–75.
•  Iwashina,  T.,  J.  Kitajima and  S.  Matsumoto. 2006. Flavonoids in the species of Cyrtomium (Dryopteridaceae) and related genera. Biochem. Syst. Ecol. 34: 14–24.
•  Iwashina,  T.,  J. A.  López-Sáez and  J.  Kitajima. 2008. Flavonoids from Osyris alba. Biochem. Syst. Ecol. 36: 146–147.
•  Iwashina,  T.,  J.  Kitajima,  T.  Shiuchi and  Y.  Itou. 2005. Chalcones and other flavonoids from Asarum sensu lato (Aristolochiaceae). Biochem. Syst. Ecol. 33: 571–584.
•  Iwashina,  T.,  S.  Rahayu,  K.  Sugahara,  T.  Mizuno,  C.  Tsutsumi and  D.  Widyatmoko. 2021. Acylated pelargonidin and cyanidin 3-sambubiosides from the flowers of Aeschynanthus species and cultivars. Phytochemistry 192: 112956. DOI: 10.1016/j.phytochem.2021.112956.
•  Kenza,  M.,  N.  Umiel and  A.  Borochov. 2000. The involvement of ethylene in the senescence of ranunculus cut flowers. Postharvest Biol. Technol. 19: 287–290.
•  Loyola,  C. E.,  J. M.  Dole and  R.  Dunning. 2019. North American specialty cut flower production and postharvest survey. HortTechnology 29: 338–359.
• Mabry, T. J., K. R. Markham and M. B. Thomas. 1970. The Systematic Identification of the Flavonoids, Springer, New York.
•  Mayoli,  R. N.,  D. K.  Isutsa and  G. O.  Tunya. 2009. Growth of Ranunculus cut flower under tropical high altitude conditions. 1: Effects of GA3 and shade. Afr. J. Hort. Sci 2: 13–28.
•  Mizuno,  T.,  T.  Yabuya,  J.  Kitajima and  T.  Iwashina. 2013. Identification of novel C-glycosylflavones and their contribution to flower colour of the Dutch iris cultivars. Plant Physiol. Biochem. 72: 116–124.
•  Mizuno,  T.,  K.  Sugahara,  C.  Tsutsumi,  M.  Iino,  S.  Koi,  N.  Noda and  T.  Iwashina. 2021. Identification of anthocyanin and other flavonoids from the green–blue petals of Puya alpestris (Bromeliaceae) and a clarification of their coloration mechanism. Phytochemistry 181: 112581. DOI: 10.1016/j.phytochem.2020.112581.
•  Takamura,  T. and  T.  Morino. 2019. Coloration and pigmentation in petals of some cyanic Ranunculus cultivars. Acta Hortic. 1237: 251–258.
•  Tatsis,  E. C.,  H.  Böhm and  B.  Schneider. 2013. Occurrence of nudicaulin structural variants in flowers of Papaveraceous species. Phytochemistry 92: 105–112.
•  Toki,  K.,  M.  Takeuchi,  N.  Saito and  T.  Honda. 1996. Two malonylated anthocyanidin glycosides in Ranunculus asiaticus. Phytochemistry 42: 1055–1057.
•  Veit,  M.,  C.  Beckert,  C.  Höhne,  K.  Bauer and  H.  Geiger. 1995. Interspecific and intraspecific variation of phenolics in the genus Equisetum subgenus Equisetum. Phytochemistry 38: 881–891.