Covalently linked anthocyanin-flavonol pigments from blue Agapanthus flowers

Article abstract of DOI:10.1016/S0031-9422(99)00572-5

The structures of the two major anthocyanins in blue Agapanthus flowers have been determined to be a p-coumaroylated delphinidin diglycoside attached to a flavonol triglycoside via a succinic acid diester link. The structure has been determined unambiguously through degradation studies, glycosidic analysis and NMR experiments. These compounds represent unique examples of anthocyanin pigments where both types of co-pigment, an aromatic acyl group and a flavonoid co-pigment, are attached covalently to the anthocyanin. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0031-9422(99)00572-5

Phytochemistry 53 (2000) 575±579
www.elsevier.com/locate/phytochem
Covalently linked anthocyanin±¯avonol pigments from blue
Agapanthus ¯owers
S.J. Bloor*, R. Falshaw
Industrial Research Ltd, P. O. Box 31-310, Lower Hutt, New Zealand
Received 11 August 1999; received in revised form 29 October 1999
Abstract
The structures of the two major anthocyanins in blue Agapanthus ¯owers have been determined to be a p-coumaroylated
delphinidin diglycoside attached to a ¯avonol triglycoside via a succinic acid diester link. The structure has been determined
unambiguously through degradation studies, glycosidic analysis and NMR experiments. These compounds represent unique
examples of anthocyanin pigments where both types of co-pigment, an aromatic acyl group and a ¯avonoid co-pigment, are
attached covalently to the anthocyanin. # 2000 Elsevier Science Ltd. All rights reserved.
Keywords: Agapanthus praecox sp. orientalis; African lily; Liliaceae; [60'-O-(delphinidin 3-O-(60-O-p-coumaroyl-glucoside) 7-O-glucosyl)] [600-O-
(
kaempferol 3-O-glucoside, 7-O-xyloside 4'-O-glucosyl)]succinate; Delphinidin 3,7 diglucoside; Kaempferol 3,4'-diglucoside 7-xyloside; Kaemp-
ferol 3,7,4'-triglucoside; Succinate diester; Blue ¯ower colour
1. Introduction
was remarkably stable and had unusually high molecu-
lar weight.
Blue ¯owers are of considerable interest to ¯ower
colour chemists. More often than not the blue colour
arises from an anthocyanin in a pH 5±7 environment
stabilised by interaction with a colourless co-pigment
2. Results and discussion
(
Goto & Kondo, 1991; Bloor, 1997, 1999). This stabil-
The two major anthocyanin pigments, 1 and 2, were
puri®ed by a series of chromatography steps involving
initial clean-up of the extract with Diaion HP-20, fol-
lowed by CC using Sephadex G-25, cellulose, RP and
isation can be through intramolecular interaction of
the anthocyanin with an attached aromatic acyl group
or intermolecular interaction between an anthocyanin
and a co-pigment derived from the ¯avonoid pathway
®nally Toyopearl HW-40. The Sephadex G-25 step
was very useful in partitioning these high MW antho-
cyanins from other metabolites. The UV±Vis spectra
showed the presence of signi®cant absorption in the
300±350 nm region indicating a high degree of aro-
matic acylation.
(
such as a ¯avone or ¯avonol). Metal ions may also be
involved.
Agapanthus sp. or African lilies present a striking
summer display of white or blue ¯owers and the blue
owers have yielded a pigment showing another vari-
¯
The mass spectrum of 1 showed a molecular ion at
m/z 1597. HPLC analysis of the acid hydrolysis pro-
duct mixture from 1 or 2 showed the presence of p-
coumaric acid, kaempferol and delphinidin. Base hy-
drolysis of 1 gave three UV-absorbing units, a delphi-
nidin glycoside, p-coumaric acid and a ¯avonol
glycoside, 3. Similar treatment of the second major
anthocyanin, 2, gave the same delphinidin glycoside
ation on this anthocyanin±co-pigment theme. We
became interested in this ¯ower when our initial stu-
dies showed the pigment mixture from the blue ¯owers
*
Corresponding author. Tel.: +64-4-569-0576; fax: +64-4-569-
055.
E-mail address: s.bloor@irl.cri.nz (S.J. Bloor).
0
0031-9422/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0031-9422(99)00572-5

5
76
S.J. Bloor, R. Falshaw / Phytochemistry 53 (2000) 575±579
and p-coumaric acid but yielded a di€erent ¯avonol
glycoside, 4. The ¯avonol glycosides, 3 and 4, were
also present in signi®cant amounts in fresh extracts of
the ¯ower tissue. These two ¯avonols were puri®ed
and shown, by analysis of NMR spectra, to be kaemp-
ferol 3,4'-diglucoside 7-xyloside (3) and kaempferol
ments 1 and 2 are delphinidin 3-O-(6-O-p-coumaroyl
glucoside) 7-O-(6-succinyl glucoside) linked to kaemp-
ferol-3,4'-diglucoside 7-xyloside and kaempferol-3,7,4'-
triglucoside, respectively.
There only remained the question of the site on the
¯avonol glycoside to which the succinyl group was
attached. Once again some degradation studies
revealed which sugar of the ¯avonol triglycoside is
involved. The relative rates of acid hydrolysis of sugars
attached to ¯avonols has been well studied (Harborne,
1965). Under normal conditions a 3-sugar can be selec-
tively cleaved in the presence of 7 or 4' substituents.
Mild acid hydrolysis of the two Agapanthus ¯avonols 3
and 4 showed this was true for 4, i.e. mild acid treat-
ment yielded kaempferol 7, 4'-diglucoside, but for 3
the 7- xylose and the 3-glucoside are lost at about the
same rate so mild hydrolysis yields kaempferol-4'-glu-
coside. This selectivity allowed the determination of
the linkage sugar. Mild acid hydrolysis of the antho-
cyanin 2 gave as an initial hydrolysis product Ð an
anthocyanin with an extra absorption peak at 370 nm
(from HPLC on-line spectra) characteristic of the
anthocyanin linked to a ¯avonol lacking a substituent
at 3-OH (Mabry, Markham & Thomas, 1970). Further
hydrolysis yielded 5 and 6 and kaempferol 7,4' diglu-
coside. Therefore, the linkage is via the 7 or 4' sugar.
The mixture of products from mild hydrolysis of the
major pigment 1 was more complex. However, it was
3,4',7-triglucoside (4). The structure of 3, a new trigly-
coside, was deduced by comparison with reported
NMR data for 4 (Yoshida, Saito & Kadoya, 1987). A
NOE between the xylose H-1 and H-6,8 of the kaemp-
ferol con®rmed the placement of the xylose at the 7
position. The mass spectrum of the second anthocya-
nin, 2, (m/z 1628 i.e. 30 more than 1) showed that the
xyloside/glucoside interchange was probably the only
di€erence between the two pigments.
The delphinidin glycoside derived from base hy-
drolysis of either 1 or 2 was almost certainly delphini-
din 3,7-diglucoside. It showed co-chromatography
(
HPLC, TLC) with one of the pigments in a mixture
derived from partial degradation of Felicia anthocya-
nin (delphinidin 3-rutinoside 7-(6-O-malonylglucoside))
(
Bloor, 1999), and was clearly di€erent (HPLC, on-line
spectrum, TLC) from delphinidin 3-rutinoside or the
more common diglucoside, delphinidin 3,5-diglucoside.
Thus, pigments 1 and 2 must be assembled from an
anthocyanin and a ¯avonol with a covalent linkage.
The relative ease with which this covalent linkage
could be broken suggested the involvement of an ali-
phatic diacid. The most likely linkage was a succinic
acid unit as 100 mass units remained after subtraction
of the ¯avonol glycoside molecular weight, a p-cou-
maroyl group and a delphinidin diglucoside, from the
molecular mass of 1 (and allowing for linkage).
Table 1
NMR data for 1
1H shift (ppm, m, J )
13
C (ppm)
a
NOE
Although a certain amount of structural information
could be obtained from the NMR spectra of the puri-
Delphinidin
H-4
8.60, s
133.5
113.3
104.5
95.5
Glc-A H-1
®
ed pigments, we resorted to degradation studies to
H-2',6' 7.72, s
H-6
H-8
6.45, d, 2
Glc-B H-1
Glc-B H-1
cleave the molecules into more easily studied sections.
Partial acid hydrolysis of a crude mixture of 1 and 2
gave a mixture of two main anthocyanins, 5 and 6 (see
structure), as well as the two ¯avonol glycosides. The
HPLC on-line UV±Vis spectra of 5 and 6 showed they
had similar spectra and had clearly retained the p-cou-
maric acid group. NMR and mass spectra of puri®ed
6.77, d, 2
7.33, d, 16
6.16, d, 16
6.69, d, 8.5
7.16, d, 8.5
6.35, d, 2
6.39, d, 2
p-Coumaroyl H-7
H-8
146.8
114.6
116.6
131.1
101.0
96.0
H-3,5
H-2,6
H-6
H-8
Kaempferol
xyl H-1
xyl H-1
H-2',6' 8.13, d, 9
H-3',5' 7.11, d, 9
132.0
117.5
102.1
5
and 6 revealed 6 to be the succinylated version of 5.
The structure of 5 was determined to be delphinidin 3-
6-p-coumaroylglucoside) 7- glucoside. The key anome-
Glc-C H-1
Glc-A
Glc-B
Glc-C
Glc-D
Xyl
H-1
H-6
H-1
H-6
H-1
H-6
H-1
H-6
H-1
H-5
5.39, d, 7.6
(
4.55, d; 4.41, dd 64.2
4.85, d, 7.8 101.5
ric proton signal of the acylated glucose was shown to
be that on the 3-sugar by an NOE with the delphinidin
H-4 and linked via a TOCSY experiment to the C-6
methylene protons which resonated at 4.5±4.4 ppm
due to acylation by the p-coumaric acid. In 6, both
glucose 6-hydroxyls appeared to be acylated. Signals
due to an extra methylene unit were present in the
same region as the ®rst and could be linked (TOCSY)
to the anomeric proton of the glucose at the delphini-
din C-7 position (NOE with H-6, H-8). Thus, the pig-
4.38, d; 4.27, dd 65.2
4.91, d (obsc)
4.61, d; 3.83, dd 65.5
5.22, d, 7.6
3.66, 3.48
4.79, d, 6.6
3.98, 3.36
103.7
62.4
101.8
66.8
a
Only those carbons connected to protons of interest are pre-
sented. These carbons shifts were determined from a HMQC exper-
iment.

S.J. Bloor, R. Falshaw / Phytochemistry 53 (2000) 575±579
577
clear from relative retention times and on-line spectral
data that products consistent with the anthocyanin
being attached to the ¯avonol via a 4'-sugar were pre-
sent, ie. three such products, due to loss of 7-xylose or
stitution at the 6-position of some of the glucose
units. The other species detected were T-Xyl (19
mol%), T-Glc (27 mol%), and small amounts of 2-
Glc (5 mol%) and 2,6-Glc (i.e. 1,2,5,6-tetra-O-
acetyl-1-deuterio-3,4-di-O-methyl glucitol, 5 mol%).
The proportion of T-Glc relative to 6-Glc was
higher than expected for structure 1 (actual ratio
1:1.6, cf. theoretical value of 1:3). Additional T-Glc
may arise from partial cleavage of the succinate die-
ster even under the ``neutral'' conditions used. Evi-
dence for this was obtained by neutral methylation
of cyanidin 3,5-diglucoside 60,60 malyl diester from
carnation ¯owers (Bloor, 1998). This species should
yield only 6-Glc after ``neutral'' methylation but
instead both T-Glc and 6-Glc were obtained, in a
ratio of 1:1.4. This indicates that one half of the
diester is cleaved even under ``neutral'' methylation
conditions. If the results for 1 are adjusted to allow
for this, then the ratio of T-Glc:6-Glc is 1:3.5 close
to that expected for the structure shown for 1. The
major species observed by this analysis support the
structure of 1 obtained above from NMR and
degradation studies. The presence of minor amount
of other species may be due to structural variability
or contamination in the sample.
3-glucose or both sugars, were seen and the later two
had the additional absorption at 370 nm. These results
can only be explained if the succinyl group is linked to
the 4'-glucose of the kaempferol glycoside. The relative
rates of hydrolysis are 3-glucose > 7-xylose > hydroly-
sis of succinyl ester bond > 4'-glucose. This assign-
ment of linking sugar was supported by data from
NMR experiments on the major pigment, 1 (see Table
1). An NOE between one of the glucose anomeric pro-
tons and the kaempferol H-3',5' placed this glucose at
the 4' position of the kaempferol. This proton could
then be connected using a TOCSY experiment to the
two C-6 protons (3.83 and 4.61 ppm). These protons
are attached (HMQC) to a carbon resonating at 65.5
ppm, a shift consistent with succinyl acylation. The
other signals in the spectrum of 1 are very similar to
those seen in the NMR spectra of the separate parts, 6
and 3, of the molecule.
Thus, the structures of 1 and 2 are the succinate die-
sters of delphinidin 3-(6-p-coumaroylglucoside) 7-glu-
coside and kaempferol 3,4'-diglucoside 7-xyloside or
kaempferol 3,4',7-triglucoside respectively, linked via
the 6 positions of the glucoses at the 7 position of the
anthocyanin and 4' position of the ¯avonol.
Further con®rmation of the structure of 1 was pro-
vided by chemical analysis of the glycosidic com-
ponents. The types of sugar units present were
determined by constituent sugar analysis of 1 (by gc/
ms analysis of alditol acetate derivatives). Only xylose
and glucose were detected, as expected. The molar
ratio was 1:3.3, close to that anticipated for structure
These compounds represent unique examples of
anthocyanin pigments where both types of co-pigment,
an aromatic acyl group and a ¯avonoid co-pigment,
are attached covalently to the anthocyanin. Previously
only one compound where an anthocyanin is attached
to a ¯avonoid co-pigment has been reported (Toki,
Saito, Limura, Suzuki & Honda, 1994), although evi-
dence exists for the occurrence of these anthocyanins
in blue lupin (Takeda, Harborne & Waterman, 1993)
and orchids (Strack, Busch & Klein, 1989).
1
. The position of glycosidic and certain other substi-
tuents, such as sulfate ester groups, can be determined
by methylation analysis. This method has been used
previously to determine the position of glycoside
branching on a triglycoside from Daucus carota L.
3
. Experimental
3.1. General
È
anthocyanin (Glassgen, Hofmann, Emmerling, Neu-
mann & Seitz, 1992). However, under the traditional,
basic conditions used for methylation, acyl groups are
cleaved so no information on the substitution position
of such groups can be obtained. Traditional methyl-
ation analysis of anthocyanin 1 produced predomi-
nantly T-Glc (i.e. 1,5-di-O-acetyl-1-deuterio-2,3,4,6-
tetra-O-methyl glucitol, 75 mol%) corresponding to
terminal glucose, as expected. T-Xyl (i.e. 1,5-di-O-
acetyl-1-deuterio-2,3,4-tri-O-methyl xylitol, 12 mol%)
corresponding to terminal xylose was also detected.
Methylation can be undertaken under conditions
where acyl groups are stable (Prehm, 1980). This
NMR experiments were run at 500 MHz or at 300
MHz (75 MHz for C). Anthocyanin samples were
dissolved in 2% CF COOD in CD OD. Flavonols
1
3
3
3
were run in DMSO-d . Mass spectra were obtained
6
using a VG 70-250S (FAB) or VG Platform II (Elec-
trospray) instrument. RP HPLC analyses were per-
formed as described previously (Bloor, 1997) using a
Waters 600 solvent delivery system coupled to a
Waters 994 PDA detector. GC/EIMS was performed
using a Hewlett-Packard MSD 5970 instrument ®tted
with an HP Ultra-2 column (2 m  0.20 mm) at 1008C
(3 min), 358C/min to 1808C (2 min), 508C/min (15
min), and electron impact voltage of 70 eV. Helium
was used as the carrier gas as 50 ml/min and 8 psi
head pressure.
`
`neutral'' methylation technique was also applied to
. In this case, the predominant species obtained
was 6-Glc (44 mol%). This would indicate acyl sub-
1

5
78
S.J. Bloor, R. Falshaw / Phytochemistry 53 (2000) 575±579
3
.2. Isolation of anthocyanins and ¯avonols
mediately frozen. Base hydrolysis was performed using
N NaOH in the absence of air for 1 h. Samples were
analysed by HPLC.
2
Freshly picked blue petals of Agapanthus praecox sp.
orientalis ``blue'' from roadside plantings in Welling-
ton, NZ, were extracted by soaking in 0.1% aq. TFA.
The ®ltered extracts were absorbed onto a large Diaion
HP-20 column, washed with 0.1% aq. TFA and the
3
.4. Glycosidic analysis.
Methylated samples were prepared from 1 or cyani-
din 3,5-diglucoside 6',60 malyl diester from carnation
owers (Bloor, 1998), using basic (Stevenson & Fur-
pigment mixture then eluted with MeOH±TFA±H O
2
(
80:0.1:19.9), and freeze-dried after evaporation of the
¯
MeOH. The ¯avonols were separated from the antho-
cyanin mixture using Avicel CC: ¯avonols are eluted
neaux, 1991) or neutral (Prehm, 1980) methylation
conditions. Native anthocyanin 1 (0.05 mg) or these
methylated samples were hydrolysed in sealed tubes
with aqueous TFA (2 M, 0.25 ml) for 1 h at 1208C.
After cooling, the solvent was evaporated at 408C in a
stream of dry air. Two portions of toluene ( 00:5 ml)
were added and the samples evaporated to dryness to
remove residual traces of acid.
®
anthocyanins with HOAc:H O (15:85). The ¯avonols
rst with t-BuOH:HOAc:H O (3:1:1), followed by the
2
2
were further puri®ed using RP CC with a Merck
1
Lobar column. The anthocyanins were passed through
a Sephadex G-25 column, eluted with MeOH±TFA±
H O (30:0.1:69.9). The major pigment fraction was
2
freeze-dried and further separated using RP CC
1
Reduction was performed by adding aqueous
NaBD (0.25 ml, 15 mg/ml in aqueous 1M NH OH)
(
Final puri®cation was achieved with Toyopearl HW-
Merck Lobar , CH CN±TFA±H O (15:0.2:84.8)).
3 2
4
4
to the samples followed by heating at 508C for 1 h.
After cooling, acetone (0.25 ml) was added to quench
excess reductant and the samples were evaporated to
40F gel CC using MeOH±TFA±H O (40:0.1:59.9).
2
3
.3. Degradation experiments
dryness in a stream of dry air at 408C. CH CN ꢀ00:4
3
ml) was then added and evaporated. The resulting aldi-
tols were acetylated using glacial HOAc (0.04 ml),
EtOAc (0.2 ml), Ac O (0.6 ml) and HClO4 (60%,
Partial acid hydrolysis was performed by dissolving
samples in 1N TFA and heating on a steam bath. Ali-
quots were removed after 1, 2, 5 and 10 min and im-
2
0.023 ml). After 15 min at room temperature, water (2

S.J. Bloor, R. Falshaw / Phytochemistry 53 (2000) 575±579
579
1
ml) and 1-methylimidazole (0.04 ml) were added to
decompose the excess Ac O. The mixture was
E283=528 ˆ 1:0: ESMS 773.1 (C H O = 773.2). H-
36
37 19
NMR; 8.80 (s, H-4), 7.82 (2H, s, H-2',6'), 7.40 (d, J ˆ
16 Hz, p-coum H-7), 7.26 (2H, d, J ˆ 8:6 Hz, p-coum
H-2,6), 7.15 (s, H-8), 6.77 (2H, d, J ˆ 8:6 Hz, p-coum
H-3,5), 6.63 (s, H-6), 6.24 (d, J ˆ 16 Hz, p-coum H-8),
5.44 (d, J ˆ 7:8 Hz, glc-3 H-1), 5.06 (d, J ˆ 5:8 Hz,
glc-7 H-1), 4.53 (dd, J ˆ 12, 1.7 Hz, glc-3 H-6a), 4.39
(dd, J ˆ 12, 8 Hz, glc-3 H-6b).
2
extracted with CH Cl (1 ml), and the organic layer
2
2
was washed with water (4 ml), aqueous Na CO (0.5
2
3
M, 4 ml) and water (4 ml). Each time, the phases were
separated by low speed centrifugation and the aqueous
phase drawn o€ and discarded. The organic layer was
then removed to a clean tube and evaporated at 408C
under a stream of dry air. The residue was dried by
addition and evaporation of CH CN ꢀ00:4 ml) evap-
3
3
(
.9. Delphinidin 3-O-(60-O-p-coumaroyl-glucoside) 7-O-
60'-O-succinyl-glucoside), 6
orated at 408C under a stream of dry air. The resulting
(
partially methylated) alditol acetates were dissolved in
acetone (0.05 ml). Derivatised samples were analysed
by GC/EIMS. Partially methylated alditol acetates
were identi®ed by comparison of their retention times
with those of authentic standards and mass spectral
data (Carpita & Shea, 1988).
UV±Vis (HPLC on-line) 283, 310(sh ), 526 E283=526 ˆ
1
1
: ESMS 873.1. H-NMR; 8.83 (s, H-4), 7.82 (2H, s,
H-2',6'), 7.39 (d, J ˆ 16 Hz, p-coum H-7), 7.26 (2H, d,
J ˆ 8:6 Hz, p-coum H-2,6), 7.16 (s, H-8), 6.79 (2H, d,
J ˆ 8:6 Hz, p-coum H-3,5), 6.57 (s, H-6), 6.23 (d, J ˆ
1
5
6 Hz, p-coum H-8), 5.46 (d, J ˆ 7:8 Hz, glc-3 H-1),
.18 (d, J ˆ 5:8 Hz, glc-7 H-1), 4.6-4.4 (3H, m, glc-3
3.5. (60'-O-(Delphinidin 3-O-(60-O-p-coumaroyl-
glucoside) 7-O-glucosyl)) (600-O-(kaempferol 3-O-
glucoside, 7-O-xyloside, 4'-O-glucosyl))succinate, 1
H-6a,6b, glc-7 H-6a), 4.34 (dd, J ˆ 12, 8 Hz, glc-7 H-
6
b), 2.60 and 2.46 (each 2H, m, succinyl CH ±CH ).
2 2
UV±Vis (HPLC on-line) 274, 283, 313, 541 nm,
+
E313=541 ˆ 1:23: LR FABMS(+) 1598 (MH ), 1619
1
(M + Na). ESMS 1597.6 (C H O = 1597.39). H-
NMR and C-NMR data, see Table 1.
7
2
77 41
13
Acknowledgements
Thanks to Herbert Wong for NMR spectroscopy.
This work was funded by the New Zealand Foun-
dation for Research Science and Technology (Contract
CO-8804).
3
.6. (60'-O-(Delphinidin 3-O-(60-O-p-coumaroyl-
glucoside) 7-O-glucosyl)) (600-O-(kaempferol 3,7-di-O-
glucoside, 4'-O-glucosyl))succinate, 2
UV±Vis (HPLC on-line) 274, 283, 313, 541 nm,
E313=541 ˆ 1:20: ESMS 1627.6.
References
3.7. Kaempferol 3,4'-di-O-glucoside-7-O-xyloside, 3
Bloor, S. J. (1997). Phytochemistry, 45, 1399.
Bloor, S. J. (1998). Phytochemistry, 49, 225.
UVlmax(MeOH) 270, 322, 345; +NaOMe 284, 380;
AlCl3 274, 300, 342; +AlCl ±HCl 278, 300, 338;
3
+
+
(
Bloor, S. J. (1999). Phytochemistry, 50, 1395.
Carpita, N. C., & Shea, E. M. (1988). In C. J. Bierman, & G. D. Mc
Ginnis, Analysis of Carbohydrates by GLC and MS (p. 157).
Florida: CRC Press.
NaOAc
270,
322,
345.
ESMS
742.8
1
C H O =742.2). H-NMR; 8.20 (2H, d, J = 8.9
32 38 20
Hz, H-2',6'),7.20 (2H, d, J = 8.9 Hz, H-3',5'), 6.84 (d,
J ˆ 1:7 Hz, H-8), 6.45 (d, J ˆ 1:7 Hz, H-6), 5.49 (d,
J ˆ 7:1 Hz, H-1 of 3-glc), 5.11 (m, H-1 of xyl), 5.04 (d,
È
Glassgen, W. E., Hofmann, R., Emmerling, M., Neumann, G. D., &
Seitz, H. U. (1992). J. Chromatogr, 598, 81.
Goto, T., & Kondo, T. (1991). Angew. Chem. Int. Ed. Eng, 30, 17.
Harborne, J. B. (1965). Phytochemistry, 4, 107.
13
J ˆ 6:9 Hz, H-1 of 4'-glc). C-NMR; 177.8 (C-4),
1
1
1
62.8 (C-7), 160.9 (C-5), 159.4 (C-4'), 156.1 (C-9),
56.1 (C-2), 134.1 (C-3), 130.7 (C-2',6'), 123.6 (C-1'),
15.8 (C-3',5'), 105.9 (C-10), 99.4 (C-6), 94.5 (C-8);
Mabry, T. J., Markham, K. R., & Thomas, M. B. (1970). The
Systematic Identi®cation of Flavonoids. New York: Springer-
Verlag.
Prehm, P. (1980). Carbohydr. Res, 78, 372.
Stevenson, T. T., & Furneaux, R. H. (1991). Carbohydr. Res, 210,
sugars: glc-3; 100.0, 73.3, 76.5, 69.7,77.1, 60.9. xyl-7;
100.8, 74.2, 75.3, 69.2, 65.8. glc-4'; 100.2, 73.0, 76.6,
277.
7
0.0, 77.6, 60.7.
Strack, D., Busch, E., & Klein, E. (1989). Phytochemistry, 28, 2127.
& Waterman, P. G. (1993).
Takeda, K., Harborne, J. B.,
Phytochemistry, 34, 421.
3.8. Delphinidin 3-O-(60-O-p-coumaroyl-glucoside) 7-O-
Toki, K., Saito, N., Iimura, K., Suzuki, T., & Honda, T. (1994).
Phytochemistry, 36, 1181.
glucoside, 5
Yoshida, T., Saito, T., & Kadoya, S. (1987). Chem. Pharm. Bull, 35,
97.
UV±Vis (HPLC on-line) 283, 310(sh ), 528 nm,