Article pubs.acs.org/JAFC
Metal Complex Pigment Involved in the Blue Sepal Color Development of Hydrangea Kin-ichi Oyama,† Tomomi Yamada,‡ Daisuke Ito,‡ Tadao Kondo,‡ and Kumi Yoshida*,‡ †
Research Center for Materials Science and ‡Graduate School of Information Science, Nagoya University, Chikusa, Nagoya 464-8601, Japan S Supporting Information *
ABSTRACT: Anthocyanins exhibit various vivid colors from red through purple to blue and are potential sources of food colorants. However, their usage is restricted because of their instability, especially as a blue colorant. The blue sepal color of Hydrangea macrophylla is due to a metal complex named “hydrangea-blue complex” composed of delphinidin 3-O-glucoside, 1, 5O-caffeoylquinic acid, 2, and/or 5-O-p-coumaroylquinic acid, 3, as copigments, and Al3+ in aqueous solution at approximately pH 4.0. However, the ratio of each component ins not stoichiometric, but is fluctuates within a certain range. The hydrangea-blue complex exists only in aqueous solution, exhibiting a stable blue color, but attempts at crystallization have failed; therefore, the structure remains obscure. To clarify the basis of the character of the hydrangea-blue pigment and to obtain its structural information, we studied the mixing conditions to reconstruct the same blue color as observed in the sepals. In highly concentrated sodium acetate buffer (6 M, pH 4.0) we could measure 1H NMR of both the hydrangea-blue complex composed of 1 (5 mM), 2 (10 mM), and Al3+ (10 mM) and a simple 1−Al3+ complex. We also recorded the spectra of complexes composed with structurally different anthocyanins and copigments. Comparison of those signals indicated that in the hydrangea-blue complex 1 might be under equilibrium between chelating and nonchelating structures having an interaction with 2. KEYWORDS: aluminum complex, blue sepal color, delphinidin 3-O-glucoside, Hydrangea macrophylla, 5-O-caffeoylquinic acid
■
INTRODUCTION
hydrangea sepals can easily change from red through purple to blue upon alteration of the cultivation conditions and/or transplanting.3−6 All sepal colors are developed by a single anthocyanin, delphinidin 3-O-glucoside, 1, and the same three copigment components, 5-O-caffeoylquinic acid, 2, 5-O-pcoumaroylquinic acid, 3, and 3-O-caffeoylquinic acid, 4 (Figure 2).7−11 Thus, the molecular mechanism of color variation has garnered interest and has been studied for more than 100 years. Soil acidity has previously been correlated with blue color development, and the involvement of Al3+ was clarified in the early 20th century;4,5 in acidic soils (pH 10 °C) of the Botanical Garden, Nagoya University Museum, before use. Materials. Delphinidin 3-O-glucoside, 1, cyanidin 3-O-glucoside, 5, and pelargonidin 3-O-glucoside, 6, were isolated from the seed coat of Phaseolus coccineus, Glycine max, and Phaseolus vulgaris, respectively.43 5O-Caffeoylquinic acid, 2, 5-O-p-coumaroylquinic acid, 3, 5-Ocinnamoylquinic acid, 7, 5-O-(3-phenylpropionyl)quinic acid, 8, and 5-O-(3′,4′-dimethoxycinnamoyl)quinic acid, 9, were synthesized according to our procedures.21 Chlorogenic acid (3-O-caffeoylquinic acid), 4, was purchased (Tokyo Chemical Industry Co., Ltd. (TCI), Japan). AlNH4(SO4)2·12H2O was purchased (Nacalai Tesque, Inc., Japan). All commercially available reagents and solvents were used directly without further purification. 1 H NMR Measurement of Al3+ Complex of 5-O-Caffeoylquinic acid, 2. 2 was dissolved in D2O and dried up in vacuo. To the residue were added D2O and deuterated AlNH4(SO4)2 (0.5 and 1.0 equiv to 2) at the concentration of 2 as 5 mM. The pH of the mixture was adjusted to be 3.6 (pD 4.0) by addition of CD3COOD and/or NaOD−D2O. After the addition of t-BuOH as an internal standard, the mixture was filtered by using a cartridge (pore size, 0.45 μm), and 1H NMR was recorded. Procedure of Blue Pigment Reconstruction. Trifluoroacetic acid (TFA) salt of 1 (1.77 mg, 3 μmol, TFA salt), 2 (1−3 equiv to 1), and a trivalent metal salt (1−4 equiv to 1) were dissolved in a deuterated buffer solution (206 μL of CD3COOD, 175 μL of 2 M NaOD, and 214 μL of D2O), and then the solution was adjusted to pH 4.0 (pD 4.4) by the addition of 2 M NaOD. The visible absorption spectrum and CD were measured in a quartz cell (0.1 mm path length). For the 1H NMR measurements, t-BuOH was added as an internal standard.
Figure 4. Visible absorption spectra and CD of the mixtures of 1 coexisting with copigment 2 and Al3+ for reconstruction of hydrangeablue complex. (A) Spectra in various concentrated buffer solutions: purple line, 1 M; blue line, 6 M; orange line, 10 M; green line, 11.6 M; red line, in acetic acid. (B) Spectra of the various combination mixtures in 6 M deuterated acetate buffer: red line, without any copigment or metal ion; purple line, 1 (5 mM) and Al3+ (5 mM); black line, 1 (5 mM) and 2 (10 mM); blue line, 1 (5 mM), 2 (10 mM), and Al3+ (10 mM).
■
RESULTS AND DISCUSSION Complexation of Copigment with Al3+. In the hydrangeablue complex, it was suggested that copigment 5-O-caffeoylquinic acid, 2, also might coordinate to Al3+ in aqueous solution. Therefore, we first measured 1H NMR of 2 coexisting with Al3+ in D2O at pD 4.0 (Figure 3). By addition of 0.5 equiv of Al3+ (2.5 mM) to 2 (5 mM) the signals became complicated (Figure 3B), and when the amount of Al3+ was increased to 1.0 equiv (5 mM) to 2 (5 mM), the signals converged to a simpler one (Figure 3C). Compared with the spectra of Figure 3A,C, the spectrum of Figure 3B (0.5 equiv of Al3+ to 2) seemed to be just a 1:1 mixture of the spectra of Figure 3A and 3C. Therefore, it was concluded that addition of 1 equiv of Al3+ might give 2−Al3+ complex completely and addition of 0.5 equiv of Al3+ might give a half amount of 2−Al3+ with remaining noncomplexed 2. In the spectrum of 2−Al3+ (Figure 3C) signals attributable to caffeoyl residue shifted upfield. The signals of quinic acid (H-2a,b and H6a,b) also shifted slightly compared with those of H-3,4,5. These results indicated that the chelating position might be at the catechol part of the caffeoyl residue and 1-OH and 1-COOH of quinic acid. In contrast, 5-O-p-coumaroylquinic acid, 3, and 3-Ocaffeoylquinic acid, 4, did not show such a phenomenon under coexisting with Al3+. Thus, the formation of Al3+ complex of copigment in D2O at pH 4.0 (pD 4.4) was a special phenomenon
Figure 5. 1H NMR of the various combinations of 1, copigment 2, and Al3+ in 6 M deuterated acetate buffer: (A) 1 (5 mM); (B) 1 (5 mM) and Al3+ (5 mM); (C) 2 (5 mM); (D) 2 (5 mM) and Al3+ (10 mM); (E) 1 (5 mM) and 2 (10 mM); (F) 1 (5 mM), 2 (10 mM), and Al3+ (10 mM).
observed only with 2, which has a catechol moiety on the axial acyl substituent. Reconstruction Conditions for the Same Blue Color as Hydrangea Sepals. We surveyed the combination of components and mixing conditions.15,18−20,42 To analyze the structure and spatial arrangement of components in the hydrangea-blue complex, measurements of visible absorption spectrum, CD, and 1H NMR should give useful information. To evaluate whether the same complex as the hydrangea-blue complex is reproduced or not, those spectra were compared. At C
DOI: 10.1021/acs.jafc.5b02368 J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Article
Journal of Agricultural and Food Chemistry
at 527 nm) and in CD, a negative exciton-type Cotton effect was observed, indicating anticlockwise stacking of the chromophores (Figure 4B, red line). When 1 equiv of Al3+ was added to 1, the visible absorption spectrum exhibited a bathochromic shift (λmax at 558 nm) with a negative exciton-type Cotton effect (Figure 4B, purple line), indicating that the anthocyanidin chromophores still self-associate in an anticlockwise stacking manner in the Al3+ complexed state. The mixture of 1 and 2 exhibited no bathochromic shift but a small effect on increase in absorbance with the same exciton-type Cotton effect (Figure 4B, black line). The visible spectrum of blue complex pigment composed of 1, 2, and Al3+ showed very high intensity at its λmax compared with the other mixture mentioned above (Figure 4B, blue line). In the CD, the exciton-type negative Cotton effect disappeared. This indicates that no self-association exists in the solution of hydrangea-blue complex and that copigmentation between 1 and 2 became stronger than the mixture of 1 and 2. The arrangement of each component resulted in the increase of population of the blue anhydrobase anion form. To optimize the reconstruction condition suitable for 1H NMR measurements, the replacement of the metal ions with other ions of the same valency and copigments with unnatural synthetic ones was carried out. We reconstructed the blue complex by mixing various trivalent metal ions (Sc3+, Ga3+, In3+, La3+, Yb3+, and Bi3+); however, only Ga3+ gave a blue solution with a low absorbance (λmax 580 nm), and other ions did not give stable blue solutions. This result indicated that Al3+ is essential for the formation of hydrangea-blue complex. We previously reported that the aromatic plane of the 5-acyl group of 2 is also essential for reconstruction of the hydrangea-blue complex.17,19,20 To re-examine these phenomena in the 6 M buffer conditions, reconstruction experiments using other copigments, 4, 7, 8, and 9, were carried out (Figure 1). It was confirmed that the conjugated system of the 5-O-acyl group in the copigment was crucial for the reconstruction of the natural-type hydrangeablue complex. Furthermore, the oxygen atoms at the aromatic ring of the cinnamate residue may not be directly involved in coordination to Al3+ but instead assist in the stabilization of the blue complex pigment. 1 H NMR Analysis of Blue Complex Pigment. Using the optimized reconstruction conditions, the hydrangea-blue complex was analyzed by 1H NMR. In the NMR spectrum the signals of sugar residues were very complicated and thus not assignable; therefore, only the signals at lower field attributable to aromatic protons were analyzed. To obtain structural information on hydrangea-blue complex, solutions of 1, 2, 1−Al3+ complex, a mixture of 1 and 2, and 2 and Al3+ at pD 4.4 were measured and the signals were compared (Figure 5). In addition, various 2D NMR experiments were carried out. The 1H NMR spectrum of 1 showed typical signals of the flavylium cation form of delphinidin nucleus (Figure 5A); in the aromatic region very broad H-4 (8.67 ppm) and H-2′ and H-6′ (7.60 ppm) peaks as well as two sets of H-6 and H-8 peaks were observed. However, the intensity of the visible absorption spectrum is low (Figure 4B, red line), and several broad signals (Figure 5A, black arrows) suggested that 1 may exist under an equilibrium mixture of the pseudobase−flavylium cation− quinonoidal base form.44,45 Very broad signals at 7.0, 6.6, and 6.1 ppm could be assigned as signals of the pseudobase.44,45 In a dilute solution (6 M) could solubilize both the hydrangea-blue complex and simple 1−Al3+ complex. In addition, in 6 M buffer, the 2−Al3+ complex did not form and the signals of 2 became simple. Therefore, we optimized the concentration of the buffer and the ratio of each component, 1, 2, and Al3+, by reproduction experiments (Figure 4). The concentration of anthocyanin 1 was fixed at 5 mM, which was nearly the same as that in the sepal cells,18,42 and other conditions were evaluated with the obtained spectra by the similarity between the visible absorption spectrum and the CD to those of the colored cells, which specifically included the value of λmax, the absorbance, and the curve shape (Figure 1).16−18 As shown in Figure 4A, all of the spectra of 1 (5.0 mM) with 1 equiv of 2 and 4 equiv of Al3+ with differing concentrations of buffer (1−11.6 M) exhibited a blue color with a λmax of approximately 580 nm, but in acetic acid, the mixture was red (λmax 533 nm). However, the CD in a 1 M buffer solution showed a negative exiton-type Cotton effect, indicating the different arrangement of anthocyanins: existence of self-association of anthocyanins, 1, with an anticlockwise stacking manner. After optimization experiments, the ratio of 1, 2, and Al3+ to 1 was determined to be 1:2:2 (Figure 4B, blue line). Using the optimized condition, visible absorption spectrum and CD were compared to analyze the structure of hydrangeablue complex (Figure 4B). 1 in the buffer solution was red (λmax D
DOI: 10.1021/acs.jafc.5b02368 J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Article
Journal of Agricultural and Food Chemistry
supramolecular structure may not be a static structure but instead a fluctuating structure in equilibrium state. The combination of 1, 2, and Al3+ may construct a very unique and special coordination structure, and this imparts a very elusive character to hydrangea blue.
indicates that the equilibrium constant between the pseudobase−flavylium cation−quinonoidal base form may change depending on concentration. In previous studies, a simple Al3+ complex of 1 was not watersoluble and the 1H NMR spectrum was recorded in organic solvent such as CD3OD.19,40,41 In our study we found 6 M acetate buffer as a new solvation system and could measure the 1−Al3+ complex in aqueous solution for the first time. In the 1H NMR spectrum of 1 with 1 equiv of Al3+, the additional signals to Figure 5A appeared, and those signals were assigned as protons of the 1−Al3+ complex (Figure 5B). The signal of H-4 (Figure 5B, blue numbers) was determined by NOESY correlation between the anomeric proton of the sugar. The NOESY correlation between H-4 (red) and H-4 (blue) and between H2′,6′ (red) and H-2′,6′ (blue) suggested that 1 may attain equilibrium of the flavylium cation−quinonoidal base−quinonoidal base anion by adding 1 equiv of Al3+ to 1. Figure 5C,D shows the 1H NMR of copigment 2. The signals of 2 were very sharp and easily assignable (Figure 5C). The addition of Al3+ did not induce remarkable spectral changes in 2 (Figure 5D), indicating that 2 equiv of Al3+ in 6 M acetate buffer had no effect. It was concluded that Al3+ may not coordinate with 2, although in a diluted buffer 2 forms a complex with Al3+ and gave a complex signal. The results of mixing 1 (5 mM) and 2 (10 mM) are shown in Figure 5E. In the 1H NMR spectrum the signals for H-4, H-2′, H-6′, H-6, and H-8 of 1 shifted slightly upfield, but no such shift was observed with the signals of 2. This upfield shift could be the copigmentation effect by hydrophobic interaction between the chromophore of 1 and the caffeic acid ring of 2. This indicated that 1 and 2 may interact with one another, but this interaction is not strong. The 1H NMR spectrum of the hydrangea-blue complex composed with 1 (5 mM), 2 (10 mM), and Al3+ (10 mM) showed a completely different spectrum attributable to the signals of delphinidin chromophore (Figure 5F). There were no similar signals found in Figure 5A (flavylium cation form) and Figure 5B (Al3+ complex of 1). However, the signals of 2 were almost the same as found in Figure 5C−E. Concerning the signals of delphinidin chromophore, new, broad signals appeared at approximately 8.0, 7.1, 6.8, 6.6, and 5.9 ppm (Figure 5F, black arrows). Among these signals, a broad peak at 6.6 ppm disappeared after 1 day. It was hard to find the signals attributable to H-6 and H-8; however, the protons of 5-OH free anthocyanin rapidly changed with D in D2O.45 Using this phenomenon, we assigned the signals. Thus, this signal was assigned to H-6 and/or H-8 of 1. As mentioned above, the major signals of 2 were sharp and remained unchanged, but the newly appeared signals at 5.9 and 7.1 ppm were assigned as H-α and H-β, respectively, by COSY and the J values (16.0 Hz). These signals were also shifted upfield, indicating that the anisotropic effect caused by stacking with 1 may occur. For the study of blue coloration of hydrangea we investigated reconstruction conditions and obtained the same blue complex, the hydrangea-blue complex, which was composed of 1, 2, and Al3+ in a ratio of 5:10:10 mM. In a 6 M deuterated acetate buffer, 1 H NMR spectra of the blue complex were measured and several aromatic protons were assigned. In the blue solution, multiple anthocyanin and copigment signals were observed, indicating that the complex structure is not fixed but instead under equilibrium. It was clarified that the blue anhydrobase anion form of delphinidin 3-glucoside, 1, is formed by complexation with Al3+ and the complex is stabilized by nonflavonoid copigment, 2, which also chelates to Al3+ (Figure 6). However, the whole
■
ASSOCIATED CONTENT
S Supporting Information *
1
H NMR of metal complex solutions (1−2−Al3+, 5−2−Al3+, and 6−2−Al3+) (Figure S1); difference in visible and CD spectra of the mixture of various trivalent metal ions with 1 and 2 (Figure S2); synthesis of 9 (Figure S3); difference in visible and CD spectra of the structure of the copigment mixture of 4, 7, 8, and 9 with 1 and Al3+ (Figure S4); 1H NMR spectrum of the mixed solution with delphinidin 3-O-glucoside, 1, and Al3+ (Figure S5); NOESY NMR spectrum of the mixed solution with delphinidin 3-O-glucoside, 1, and Al3+ (Figure S6); 1H NMR of timedependent change of delphinidin 3-O-glucoside, 1 (Figure S7); 1 H NMR of time-dependent change of the mixed solution with delphinidin 3-O-glucoside, 1, and Al3+ (Figure S8); 1H NMR of time-dependent change of the mixed solution with delphinidin 3O-glucoside, 1, 2, and Al3+ (Figure S9); COSY NMR spectrum of the mixed solution with delphinidin 3-O-glucoside, 1, copigment 2, and Al3+ (Figure S10); characterization of new compounds and experimental procedures. The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jafc.5b02368.
■
AUTHOR INFORMATION
Corresponding Author
*(K.Y.) Phone: +81-52-789-5638. Fax: +81-52-789-5638. Email:
[email protected]. Funding
This work was financially supported by the Ministry of Education, Culture, Sports, Science and Technology, Japan ((B) No. 24380062, Creative Scientific Research No. 16GS0206, and Global COE in Chemistry, Nagoya University). Notes
The authors declare no competing financial interest.
■ ■
ACKNOWLEDGMENTS We thank Kazushi Koga (Graduate School of Bioagricultural Sciences, Nagoya University) for the NMR measurement. ABBREVIATIONS USED CD, circular dichroism; NMR, nuclear magnetic resonance; TFA, trifluoroacetic acid; NOESY, nuclear Overhauser effect spectroscopy; COSY, correlation spectroscopy
■
REFERENCES
(1) Francis, F. J. Food colorants: anthocyanins. Crit. Rev. Food Sci. Nutr. 1989, 28, 273−314. (2) Jian He, J.; Giusti, M. M. Anthocyanins: natural colorants with health-promoting properties. Annu. Rev. Food Sci. Technol. 2010, 1, 163− 187. (3) Chenery, E. M. The problem of the blue hydrangea. J. R. Hortic. Soc. 1937, 62, 604−620. (4) Allen, R. C. Influence of aluminum on the flower color of Hydrangea macrophylla DC. Boyce Thompson Inst. 1943, 13, 221−242. (5) Ma, J. F.; Ryan, P. R.; Delhaize, E. Aluminium tolerance in plants and the complexing role of organic acids. Trends Plant Sci. 2001, 6, 273− 278. E
DOI: 10.1021/acs.jafc.5b02368 J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Article
Journal of Agricultural and Food Chemistry (6) Yoshida, K.; Ito, D.; Shinkai, Y.; Kondo, T. Change of color and components in sepals of chameleon hydrangea during maturation and senescence. Phytochemistry 2008, 69, 3159−3165. (7) Lawrence, W. J. C.; Price, J. R.; Robinson, G. M.; Robinson, R. Survey of anthocyanins. V. Biochem. J. 1938, 32, 1661−1667. (8) Robinson, G. M.; Robinson, R. The colloid chemistry of leaf and flower pigments and the precursors of the anthocyanins. J. Am. Chem. Soc. 1939, 61, 1606−1607. (9) Hayashi, K.; Abe, Y. Studien über anthocyane, XXIII. Papierchromatographische übersicht der anthocyane im pflanzeneich. I. Misc. Rep. Res. Inst. Nat. Resour. 1958, 29, 1−8. (10) Asen, S.; Siegelman, H. W.; Stuart, N. W. Anthocyanin and other phenolic compounds in red and blue sepals of Hydrangea macrophylla var. Merveille. Proc. Am. Soc. Hortic. Sci. 1957, 69, 561−569. (11) Asen, S.; Siegelman, H. W. Effect of aluminum on absorption spectra of the anthocyanin and flavonols from sepals of Hydrangea macrophylla var. Merveille. Proc. Am. Soc. Hortic. Sci. 1957, 70, 478−481. (12) Kinraide, T. Identity of the rhizotoxic aluminium species. Plant Soil 1991, 134, 167−178. (13) Takeda, K.; Kubota, R.; Yagioka, C. Copigments in the blueing of sepal colour of Hydrangea macrophylla. Phytochemistry 1958, 24, 1207− 1209. (14) Takeda, K.; Kariuda, M.; Itoi, H. Blueing of sepal colour of Hydrangea macrophylla. Phytochemistry 1958, 24, 2251−2254. (15) Takeda, K.; Yamashita, T.; Takahashi, A.; Timberlake, C. F. Stable blue complexes of anthocyanin-aluminium-3-p-coumaroyl- or 3caffeoyl-quinic acid involved in the blueing of Hydrangea flower. Phytochemistry 1990, 29, 1089−1091. (16) Yoshida, K.; Toyama-Kato, Y.; Kameda, K.; Kondo, T. Sepal color variation of Hydrangea macrophylla and vacuolar pH measured with a proton-selective microelectrode. Plant Cell Physiol. 2003, 44, 262−268. (17) Kato, Y. Chemical Studies on Blue Sepal-Color Development of Hydrangea by Using Molecular-Designed Co-pigments. Thesis, Nagoya University, 2006. (18) Ito, D.; Shinkai, Y.; Kato, Y.; Kondo, T.; Yoshida, K. Chemical studies on different color development in blue- and red-colored sepal cells of Hydrangea macrophylla. Biosci., Biotechnol., Biochem. 2009, 73, 1054−1059. (19) Kondo, T.; Toyama-Kato, Y.; Yoshida, K. Essential structure of co-pigment for blue sepal-color development of hydrangea. Tetrahedron Lett. 2005, 46, 6645−6649. (20) Toyama-Kato, Y.; Kondo, T.; Yoshida, K. Synthesis of designed acylquinic acid derivatives involved in blue color development of hydrangea and their co-pigmentation effect. Heterocycles 2007, 72, 239− 254. (21) Oyama, K.-I.; Watanabe, N.; Yamada, T.; Suzuki, M.; Sekiguchi, Y.; Kondo, T.; Yoshida, K. Efficient and versatile synthesis of 5-Oacylquinic acids with a direct esterification using a p-methoxybenzyl quinate as a key intermediate. Tetrahedron 2015, 71, 3120−3130. (22) Goto, T.; Kondo, T. Structure and molecular stacking of anthocyanins − flower color variation. Angew. Chem., Int. Ed. Engl. 1991, 30, 17−33. (23) Brouillard, R.; Dangles, O. Flavonoids and flower color. In The Flavonoids Advances in Research since 1986; Harborne, J. B., Ed.; Chapman & Hall: London, UK,1994; pp 565−588. (24) Yoshida, K.; Mori, M.; Kondo, T. Blue flower color development by anthocyanins: from chemical structure to cell physiology. Nat. Prod. Rep. 2009, 26, 884−915. (25) Yoshida, K.; Oyama, K.-I.; Kondo, T. Chemistry of flavonoids in color development. In Recent Advances in Polyphenol Research; Cheynier, V., Sarni-Manchado, P., Quideau, S., Eds.; Wiley: Chichester, UK, 2012; Vol. 3, pp 99−129. (26) Hayashi, K.; Abe, Y.; Mitsui, S. Blue anthocyanin from flowers of Commelina, the crystallisation and some properties thereof. Studies on anthocyanins. XXX. Proc. Jpn. Acad. 1958, 34, 373−378. (27) Mitsui, S.; Hayashi, K.; Hattori, S. Commelinin, a crystalline blue metallo-anthocyanin from the flowers of Commelina. Proc. Jpn. Acad. 1959, 35, 169−174.
(28) Kondo, T.; Yoshida, K.; Nakagawa, A.; Kawai, T.; Tamura, H.; Goto, T. Structural basis of blue-colour development in flower petals from Commelina communis. Nature 1992, 358, 515−518. (29) Shiono, M.; Matsugaki, N.; Takeda, K. Structure of commelinin, a blue complex pigment from the bue flowers of Commelina communis. Proc. Jpn. Acad., Ser. B 2008, 84, 452−456. (30) Bayer, E. Natürliche und synthetische Anthocyan-Metallkomplexe. Chem. Ber. 1958, 91, 1115−1122. (31) Kondo, T.; Ueda, M.; Tamura, H.; Yoshida, K.; Isobe, M.; Goto, T. Composition of protocyanin, a self-assembled supramolecular pigment from the blue cornflower Centaurea cyanus. Angew. Chem., Int. Ed. Engl. 1994, 33, 978−979. (32) Kondo, T.; Ueda, M.; Isobe, M.; Goto, T. A new molecular mechanism of blue color development with protocyanin, a supramolecular pigment from corn flower Centaurea cyanus. Tetrahedron Lett. 1998, 39, 8307−8310. (33) Shiono, M.; Matsugaki, N.; Takeda, K. Structure of the blue cornflower pigment. Nature 2005, 436, 791. (34) Takeda, K.; Yanagisawa, M.; Kifune, T.; Kinoshita, T.; Timberlake, C. F. A blue pigment complex in flowers of Salvia patens. Phytochemistry 1994, 35, 1167−1169. (35) Kondo, T.; Oyama, K.-I.; Yoshida, K. Chiral molecular recognition on formation of a metalloanthocyanin: a supramolecular metal complex pigment from blue flower of Salvia patens. Angew. Chem., Int. Ed. 2001, 40, 894−897. (36) Mori, M.; Kondo, T.; Yoshida, K. Cyanosalvianin, a supramolecular blue metalloanthocyanin, from petals of Salvia uliginosa. Phytochemistry 2008, 69, 3151−3158. (37) Shoji, K.; Miki, N.; Nakajima, N.; Momonoi, K.; Kato, C.; Yoshida, K. Perianth bottom-specific blue color development in tulip cv. Murasakizuisho requires ferric ions. Plant Cell Physiol. 2007, 48, 243− 251. (38) Tanaka, M.; Fujimori, T.; Uchida, I.; Yamaguchi, S.; Takeda, K. A malonylated anthocyanin and flavonols in blue Meconopsis flowers. Phytochemistry 2001, 56, 373−376. (39) Kumi Yoshida, K.; Kitahara, S.; Ito, D.; Kondo, T. Ferric ions involved in the flower color development of the Himalayan blue poppy Meconopsis grandis. Phytochemistry 2006, 67, 992−998. (40) Dangles, O.; Elhabiri, M.; Brouillard, R. Kinetic and thermodynamic investigation of the aluminium-anthocyanin complexation in aqueous solution. J. Chem. Soc., Perkin Trans. 2 1994, 2587− 2596. (41) Schreiber, H. D.; Jones, A. J.; Lariviere, C. M.; Mayhew, K. M.; Cain, J. B. Role of aluminium in red-to-blue color changes in Hydrangea macrophylla sepals. BioMetals 2011, 24, 1005−1015. (42) Ito, D. Chemical Studies on Sepal-Color Variation of Hydrangea by Using Single Cell-Analytical Method. Thesis, Nagoya University, 2009. (43) Yoshida, K.; Sato, Y.; Okuno, R.; Kameda, K.; Isobe, M.; Kondo, T. Structural analysis and measurement of anthocyanins from colored seed coats of Vigna, Phaseolus, and Glycine legumes. Biosci., Biotechnol., Biochem. 1996, 60, 589−593. (44) Cheminat, A.; Brouillard, R. PMR investigation of 3-O-(β-Dglucosyl)malvidin structural transformations in aqueous solutions. Tetrahedron Lett. 1986, 27, 4457−4460. (45) Yanamala, N.; Tirupula, K. C.; Balem, F.; Klein-Seetharaman, J. pH-Dependent interaction of rhodopsin with cyanidin-3-glucoside. 1. Structural aspects. Photochem. Photobiol. 2009, 85, 454−462.
F
DOI: 10.1021/acs.jafc.5b02368 J. Agric. Food Chem. XXXX, XXX, XXX−XXX
