Characterization of Acylated Anthocyanins in Callus Induced From Storage Root of Purple-Fleshed Sweet Potato, Ipomoea batatas L

Four anthocyanins were isolated from a highly pigmented callus induced from the storage root of purple-fleshed sweet potato (Ipomoea batatas L) cultivar Ayamurasaki. The anthocyanins were respectively identified as cyanidin 3-O-(2-O-(6-O-(E)-caffeoyl-β-D-glucopyranosyl)-β-D-glucopyranoside) -5-O-β-D-glucopyranoside, cyanidin 3-O-(2-O-(6-O-(E)-p -coumaroyl-β-D-glucopyranosyl)-6-O-(E)-caffeoyl-β-D-glucopyranoside)-5-O-β-D-glucopyranoside, cyanidin 3-O-(2-O-(6-O-(E)-p -coumaroyl-β-D-glucopyranosyl)-6-O-(E)-p-coumaroyl-β-D-glucopyranoside)- 5-O-β-D-glucopyranoside, and peonidin 3-O-(2-O-(6-O-(E)-p -coumaroyl-β-D-glucopyranosyl)-6-O-(E)-p-coumaroyl-β-D-glucopyranoside)-5-O-β-D-glucopyranoside by chemical and spectroscopic analyses. These anthocyanins were examined with respect to the stability in neutral aqueous solution as well as the radical scavenging activity against the 1,1-diphenyl-2-picrylhydrazyl (DPPH) radical. These acylated anthocyanins exhibited both higher stability and higher DPPH radical scavenging activity than corresponding nonacylated cyanidin and peonidin 3-O-sophoroside-5-O-glucosides.

Recently, from the storage root of the cultivar Ayamurasaki we have established a high-anthocyanin accumulating cell line. In the crude pigment extract, some cellline-specific anthocyanins have been detected which were different from those in the original storage root [15]. Previously, we have isolated and determined the molecular structures of the two anthocyanins, in which one was a known anthocyanin, cyanidin 3-sophoroside-5-glucoside, and the other was a new one, cyanidin 3-(p-coumaroyl) sophoroside-5-glucoside [5]. In sequence, we have isolated four more callus anthocyanins. The purpose of this study is to characterize the structures and the stability in a neutral aqueous solution and the antioxidative activity to utilize the pigment for food or other material.

Plant materials
A previously established high-anthocyanin-accumulating callus culture generated from the storage root of sweet potato cultivar Ayamurasaki has been used for this study [15]. Suspended cell cultures were initiated by transferring about 1.0 g (fresh weight) of callus to 50 mL of liquid medium in 250 mL Erlenmeyer flasks. Basal Murashige and Skoog (MS) medium [16] supplemented with 1.0 mg/L 2,4-D, and 3% sucrose has been used as a maintenance medium (MM). Medium pH has been adjusted to 5.8 before autoclaving. Subcultures have been done in 7-day intervals. The cultures were incubated on a rotary shaker (130 rpm) at 25 • C in the dark. For the purpose of present analysis, suspension cultures have been produced under two different medium conditions: MM and a high-anthocyanin producing medium (APM).
The APM was a modified MS medium with 9.4 mM KNO 3 , without NH 4 NO 3 , with 5% sucrose and nil growth regulators [17]. Medium pH has been adjusted to 5.8 before autoclaving and suspension cultures were incubated in 250 mL flasks. Five hundred mg of cell aggregates were placed in flasks containing 50 mL medium. The cultures were harvested after 7 days' growth on MM (200 g) and after 14 days on APM (199 g).

Extraction and isolation of anthocyanins
The aggregates were removed from the culture media, rinsed with distilled water, separated from the liquid by vacuum filtration, and weighed. A total amount of 399 g of fresh tissue was steeped in 15% CH 3 COOH (1 L) for one day and filtered. This operation was repeated three more times (1, 1, and 0.5 L). The combined crude extract (3.5 L) contained nineteen or more anthocyanins [18]. The extract solution was applied on two XAD-2000 resin columns (30 id × 385 mm), the columns were washed with water (3 L), eluted stepwise with 10%, 20%, 30%, 40%, 60%, or 70% EtOH all with 1% CH 3 COOH (1 L). Pigments containing 30%, 40%, 60%, and 70% EtOH fractions were combined and evaporated to dryness under reduced pressure. Subsequently, the residue was separated using ODS column (60 id × 320 mm) with 10%, 20%, 30%, 40%, 60%, or 70% EtOH all with 1% CH 3 COOH (1 L). HPLC analysis confirmed that 1 was contained in 20% EtOH fraction, 2 and 3 in 60% EtOH fraction, and 4 in 70% EtOH fraction, and each fraction was evaporated to dryness under reduced pressure. Finally, 1, 2, 3, and 4 were isolated on purification of the individual fractions by preparative HPLC monitoring at 310 nm. The elutions were evaporated to dryness, dissolved in minimum amount of TFA, precipitated with excess ether, and dried in a silica gel desiccator under reduced pressure.

Chemical analysis
Alkaline hydrolysis of isolated pigment was performed as follows. The pigment powder (3 mg) was dissolved in 2N NaOH, left for 15 minutes with a sealed cap, and then acidified with CH 3 COOH. The components in the reaction mixture were identified by analytical HPLC.  Table 1.  Table 1.  Table 1.  Table 1.

Stability test
Stability of anthocyanins 1, 2, 3, and 4 in neutral aqueous solution was compared with that of nonacylated anthocyanins cyanidin 3-O-sophoroside-5-O-glucoside (Cy3S5G) and peonidin 3-O-sophoroside-5-O-glucoside (Pn3S5G) isolated from original storage root according to a previously reported method [19]. Each anthocyanin TFA salt was dissolved in McIlvaine (pH 7.0, 0.1 M citrate-0.2 M phosphate) buffer solution to make 50 µM test solution, and the Vis spectra (400-700 nm) were measured automatically at appropriate time intervals. Based on the absorbance at λ vis.max of each spectrum, the residual color (%) was calculated as percent of the initial absorbance (= 100%). The stability of the anthocyanin could be evaluated on the basis of the half-life (t 1/2 ), defined as the time required to reach 50% residual color.

DPPH radical scavenging activity assay
Radical scavenging activity of 1, 2, 3, and 4 was tested according to the DPPH-colorimetric method developed by Yamaguchi et al [20] and compared with Cy3S5G, Pn3S5G, and α-tocopherol as natural antioxidants and BHT as a synthetic antioxidant. Each sample was dissolved in EtOH to 500 µM concentration. Sample solution (25 µL) was added 375 µL of EtOH, 350 µL of Tris-HCl buffer solution (pH 7.4, 0.1 M), and 250 µL of 500 µM DPPH-EtOH solution (to obtain a final sample concentration was 12.5 µM), and immediately shaken and then kept standing for 20 minutes in the dark at room temperature. The absorbance of residual DPPH in sample solution was measured at 520 nm. Initial and blank were measured without substrate and without DPPH, respectively. The DPPH radical scavenging activity (RS%) was calculated as RS% = 100(A i − A s + A b )/A i , in which A i , A s , and A b were the absorbances at 520 nm of initial, sample, and blank solutions, respectively. The experiment was conducted with four replicates.
The complete structures of 1, 2, 3, and 4 were established by 13 C and 1 H NMR analyses containing 2D pulse experiments such as a homonuclear double quantum filtered correlation spectroscopy (DQF-COSY), a total correlation spectroscopy (TOCSY), a heteronuclear singlequantum correlation (HSQC), a nuclear Overhauser and exchange spectroscopy (NOESY), and a heteronuclear multiple bond correlation spectroscopy (HMBC) technique. Assignment of 13 C-and 1 H-signals was summarized in Table 1. 1 H-signals in low magnetic field (δ H 6-9 ppm) shows characteristic aglycons and cinnamic acids. Anthocyanins 1, 2, and 3 and 4 have Cy and Pn moieties, respectively, due to the corresponding signals of benzopyrylium nucleus and 1, 3, 4-trisubstituted aromatic B-ring. Only 4 has an additional methoxyl signal at high magnetic field (δ H 3.90 ppm, δ C 56.19 ppm). The presence of trans (E)-caffeoyl residue in 1 and 2 spectra and (E)p-coumaroyl residue in 2, 3, and 4 spectra is confirmed with the 1, 3, 4-trisubstituted and 1, 4-disubstituted benzenes, respectively having the (E)-olefinic proton signals with large coupling constant (about J α,β = 16 Hz). In high magnetic field (δ H 3-6 ppm), the spectra also show all sugars of 1, 2, 3, and 4 to be β-D-glucopyranosyl configuration because of the resonances at lower magnetic field (δ H 4.78-5.68 ppm) of all anomeric protons and the large J values (J = 7.2-9.4 Hz) of the anomeric protons and the ring protons. As shown in Table 1   G a -2H (at δ H 3.95-4.06 ppm) and G a -2C (δ C 81.54-81.86 ppm) clearly shift to downfield more than G b -2H (at δ H 3.14-3.19 ppm) and G b -2C (δ C 74.65-74.86 ppm) or G c − 2H (at δ H 3.51-3.56 ppm) and G c -2C (δ C 73.19-73.39 ppm). On the basis of this data, we concluded that glucose b (G b ) links to glucose a (G a )-20H. On the other hand, NOESY and HMBC spectra gave more direct and certain data on the presence of β-D-G b (1 → 2) G a bond of the sophorosyl residue.
The connecting relations among an aglycon, three sugars, and acyl groups in 1, 2, 3, and 4 were confirmed by NOESY and HMBC measurements, for example, anthocyanin 2 shown in Figure 2(b). In the NOESY spectra of 1, 2, 3, and 4, three intensive NOE signals between aglycon-4H and G a -1H (aglycon-4H/G a -1H), G a -2H/G b -1H, and aglycon-6H/G c -1H indicated that G a , G b , and G c connected at aglycon-3-OH, at G a -2OH, and at aglycon-5OH through glycosyl bond, respectively. In the HMBC spectra of 1, 2, 3, and 4, the clear 1 H-13 C cross peaks between G a -1H and aglycon-3-carbon signals (G a -1H/aglycon-3C), G b -1H/G a -2C, G a -2H/G b -1C, and G c -1H/aglycon-5C verified the connections of G a /aglycon-3OH, G b /G a -2OH, and G c /aglycon-5OH, respectively. Moreover, the distinct correlation peaks between G b -6H and acyl-carbonyl carbon signals (about δ C 166.5 ppm) provided decisive proof that acylating acids were linked at G b -6OH. In 2, 3, and 4, the cross peak between G c -6H and acyl-carbonyl carbon signals (about δ C 166.8 ppm) also showed directly links aromatic acids and G c -6OH [3]. In conclusion, 1, 2 (Figure 2(a)). Only anthocyanin 4 is a new compound, while 1, 2, and 3 are known ones. Anthocyanins 1 and 2 have already been identified in Ipomoea cairica flowers [21], but the exact binding sites of the acyl residues were not established. Similarly, anthocyanins 2 and 3 have been identified in the pigment of Ipomoea asarifolia flower [22], and in the pigment of Ajuga reptans flower and the corresponding cell cultures [23], respectively. Pigment 1 is also found in original purple-fleshed sweet potato storage root (Figure 1(b)) or sweet potato leaf pigment as a minor component, while p-coumaroylated 2, 3, and 4 are confirmed to be cell-line-specific pigments.
These anthocyanins were evaluated for the antioxidant activity as scavenging ability against DPPH radical. As shown in Figure 4, activity of identically substituted but Cy-based 3 exhibited higher scavenging activity than Pn-based 4. The caffeoyl-p-coumaroylated 2 and monocaffeoylated 1 showed higher activity than di-pcoumaroylated 3. Therefore, anthocyanins with catechol group(s) in the aglycon and/or the acyl group(s) were thought to be very effective in radical scavenging. Moreover, since hydroxycinnamic acid acylation enhanced the activity (1, 2, 3 > Cy3S5G, and 4 > Pn3S5G), the acylation might intramolecularly and synergistically achieve the radical scavenging activity of aglycon in addition to their own activity [26].
The cell line generated from Ayamurasaki storage root accumulates high levels of anthocyanin pigments. Depending on medium conditions under which the tissue is produced, the total amount of anthocyanins is equal to that of the original storage root on an MM [15] or 2.5-fold higher on a high-anthocyanin producing medium [27].
Thus the callus pigment has a potential to be utilized as a high-quality natural food colorant/natural food ingredient with protective action against oxidative damage [18].