Water-soluble flavonol (= 3-hydroxy-2-phenyl-4H-1-benzopyran-4-one) derivatives: Chemical synthesis, colouring, and antioxidant properties

Article abstract of DOI:10.1002/(SICI)1522-2675(20000216)83:2<428::AID-HLCA428>3.0.CO;2-J

Water-soluble derivatives of rutin, a very common glycoside of quercetin (= 3,3',4',5,7-pentahydroxyflavone = 2-(3,4-dihydroxyphenyl)-3,5,7- trihydroxy-4H-1-benzopyran-4-one) and a potent plant antioxidant from the flavonol family, were synthesized by simple chemical procedures aimed at introducing carboxy or sulfo groups at the sugar moiety (Scheme 1). Such derivatives form stable molecular complexes with malvin, a polyphenolic pigment from the anthocyanin family, and thereby prove to be very effective in the enhancement (hyperchromism) and variation (bathochromism) of natural colours. The H₂O-solubilizing carboxylate and sulfate groups are shown to deeply modify the enthalpy-entropy balance of the pigment-flavonol complexation (copigmentation). A molecular interpretation of the complexation-induced bathochromic shift in the pigment VIS band is proposed. Finally, the H₂O-soluble rutin derivatives are shown to retain the high antioxidant ability of rutin as evidenced by their efficient trapping of the coloured radical DPPH (=2,2-diphenyl-1-(2,4,6-trinitrophenyl)hydrazyl).

Full text of DOI:10.1002/(SICI)1522-2675(20000216)83:2<428::AID-HLCA428>3.0.CO;2-J

428
Helvetica Chimica Acta ± Vol. 83 (2000)
Water-Soluble Flavonol (ˆ 3-Hydroxy-2-phenyl-4H-1-benzopyran-4-one)
Derivatives: Chemical Synthesis, Colouring, and Antioxidant Properties
by Bertrand Alluisa), Nathalie PeÂrola), Hakima El hajjib), and Olivier Danglesa)*
a
Â
Â
Ã
) Universite Claude Bernard-Lyon I, UMR-CNRS 5078, Laboratoire des Polyphenols, Bat. 303, 43, bld du 11
Novembre 1918, F-69622 Villeurbanne (e-mail: dangles@univ-lyon1.fr)
b
Â
Â
Â
) Universite Cadi Ayyad, Faculte des Sciences et Techniques-Gueliz, Av. Abdelkrim Khattabi, B.P. 618,
Marrakech, Maroc
Water-soluble derivatives of rutin, a very common glycoside of quercetin (ˆ 3,3',4',5,7-pentahydroxyflavone ˆ
2-(3,4-dihydroxyphenyl)-3,5,7-trihydroxy-4H-1-benzopyran-4-one) and a potent plant antioxidant from the
flavonol family, were synthesized by simple chemical procedures aimed at introducing carboxy or sulfo groups at
the sugar moiety (Scheme 1). Such derivatives form stable molecular complexes with malvin, a polyphenolic
pigment from the anthocyanin family, and thereby prove to be very effective in the enhancement (hyper-
chromism) and variation (bathochromism) of natural colours. The H2O-solubilizing carboxylate and sulfate
groups are shown to deeply modify the enthalpy-entropy balance of the pigment-flavonol complexation
(copigmentation). A molecular interpretation of the complexation-induced bathochromic shift in the pigment
VIS band is proposed. Finally, the H2O-soluble rutin derivatives are shown to retain the high antioxidant ability
of rutin as evidenced by their efficient trapping of the coloured radical DPPH (ˆ2,2-diphenyl-1-(2,4,6-
trinitrophenyl)hydrazyl).
Introduction. ± Flavonols (ˆ 3-hydroxyflavones ˆ 3-hydroxy-2-phenyl-4H-1-ben-
zopyran-4-ones) are an important class of flavonoids (polyphenols) abundant in plants
and in human diet [1]. Most of their biological activities are attributed to their affinity
for a wide range of proteins, including enzymes, and their antioxidant properties [2].
On the other hand, recent epidemiological studies have suggested that flavonoid-rich
diets could help prevent the development of cardiovascular diseases, a major cause of
mortality in occidental countries [3]. Flavonols also efficiently take part in natural
colour expression by forming molecular complexes with polyphenolic plant pigments
(anthocyanins), a phenomenon known as copigmentation [4]. Although flavones and
flavonols are the most efficient anthocyanin copigments reported so far [5], their
efficiency in stabilizing natural colours is limited by their poor solubility (of the order of
3
10 m for glycosides and much lower for aglycones) in slightly acidic to neutral aqueous
solutions modeling the natural medium. On the other hand, enzymes from the intestinal
microflora are known to promote extensive hydrolysis of the glycosidic bonds of
flavonol glycosides in the intestine and subsequent degradation of the aglycones into
rapidly excreted phenolic acids [6]. However, the quercetin conjugates (glucuronides
and sulfates of quercetin and its 3'-methyl ether) found in biologically significant,
although relatively low (0.1 ± 1 mm), concentrations in the plasma of human volunteers
after consumption of a meal rich in quercetin glycosides retain high antioxidant
activities [7]. Synthesis of H2O-soluble flavonol conjugates [8] or chemical modifica-
tions of flavonol glycosides may reduce intestinal degradation for a better absorption
and a higher plasma concentration.

Helvetica Chimica Acta ± Vol. 83 (2000)
429
This work describes some simple chemical modifications of rutin (ˆ 3-O-(a-l-rham-
nopyranosyl-1,6-b-d-glucopyranosyl)quercetin ˆ 3-{[6-O-(6-deoxy-a-l-mannopyrano-
syl)-b-d-glucopyranosyl]oxy}-2-(3,4-dihydroxyphenyl)-5,7-dihydroxy-4H-1-benzopyran-
4-one; Scheme 1) on its sugar moiety (introduction of sulfate and carboxylate groups)
aimed at increasing its water solubility and eventually its bioavailability. The influence
of the chemical modifications on the colouring and antioxidant properties is discussed.
Particular attention will be given to succinylated rutin derivatives since succinyl groups
are found in some naturally occurring flavonoid glycosides [1a].
Results and Discussion. ± 1. Synthesis (see Scheme 1). Rutin is one of the most
common flavonols in plants. Being commercially available and cheap, it is an ideal
material for chemical modifications aimed at modulating its colouring and antioxidant
properties. In the first step, the 1,2-cis-diol group of the rhamnose moiety is selectively
protected by a benzylidene group upon treatment of rutin by benzaldehyde dimethyl
acetal in DMF with p-toluenesulfonic acid (TsOH) as a catalyst. Subsequent
acetylation of the unprotected OH groups (Ac2O, Py, DMAP ˆ N,N-dimethylpyridin-
4-amine) afforded 1 as a mixture of epimers. Long reaction times were usually required
to ensure complete acetylation. In particular, OH C(5) (d(H) ca. 12.5 ppm) reacted
slowly, probably because of the H-bond it forms with the 4-keto group. Upon oxidative
ring opening of the benzylidene acetal (NBS(ˆ N-bromosuccinimide), benzoyl
peroxide, BaCO3, CHCl3/H2O), 2 was obtained as a single regioisomer in agreement
with [9]. Incomplete acetylation of the polyphenolic nucleus invariably resulted in low
yields. Treatment of 2 with SO3 ´ pyridine in DMF and subsequent deacetylation
(MeOH, MeONa) gave rutin 3'''-sulfate (sodium salt) (3) after elution on a Dowex ion-
‡
exchange resin (Na ) and purification on C18 silica gel. Alternatively, rutin was treated
by the ethyl acetoacetate dimethyl acetal (prepared upon refluxing ethyl acetoacetate
with trimethyl orthoformate in MeOH with TsOH as a catalyst) in DMF with catalytic
amounts of TsOH. Saponification (KOH, EtOH, reflux) and purification on C18 silica
gel afforded 4 as a mixture of epimers.
Treatment of rutin with succinic anhydride (6 equiv.) and DMAP (0.5 equiv.) in
pyridin at 708 resulted in a quantitative yield of rutin 2'',3'',4'',2''',3''',4'''-hexakis(hy-
drogen succinate) (5). No acylation of the phenolic OH groups occurred under these
conditions. As expected, 5 is highly soluble in H2O and aqueous buffers (except in
strongly acidic solutions). When the experiment was repeated at 08 with lower
concentrations of succinic anhydride (3 equiv.) and DMAP (ca. 0.2 equiv.), a mixture 6
of three rutin bis(hydrogen succinates) in a ca. 3 :3 :2 molar ratio was isolated after
1
chromatography on C18 silica gel. The H-NMR spectrum of 6 suggests that each
regioisomer is acylated at O C(4'''), the second succinyl (ˆ 3-carboxy-1-oxopropyl)
group being located either at O C(3''') or at one of the glucose O-atoms. Finally,
acylation at 08 with 1.5 equiv. of succinic anhydride and ca. 0.1 equiv. of DMAP
afforded rutin 4'''-(hydrogen succinate) (7) in 21% yield. Under mildly acidic
conditions (pH 3.5), the H2O solubility of rutin derivatives 3 ± 7 is higher than
10 2 m, i.e., at least 10 times higher than that of rutin.
2. Thermodynamic Investigation of Copigmentation. Flavonol glycosides such as
rutin are abundant in the vacuoles of fruits and flowers where polyphenolic pigments
(anthocyanins) are stored. Flavonols and anthocyanins form stable molecular

430
Helvetica Chimica Acta ± Vol. 83 (2000)
Scheme 1. Chemical Synthesis of Rutin Derivatives 3 ± 5
5
6
7
R ˆ R' ˆ R'' ˆ R''' ˆ C(ˆO)CH2CH2COOH
R ˆ C(ˆO)CH2CH2COOH, R'' ˆ H, R',R''' ˆ 1C(ˆO)CH2CH2COOH, 3 H
R ˆ C(ˆO)CH2CH2COOH, R' ˆ R'' ˆ R''' ˆ H
complexes upon stacking of their polyphenolic nuclei. Thereby, water addition to the
anthocyanin chromophore (Scheme 2) and subsequent colour loss are at least partially
prevented [4][5a ± d].

Helvetica Chimica Acta ± Vol. 83 (2000)
431
Scheme 2
The efficiency of rutin derivatives 3 ± 7 in copigmentation was tested with malvin
(ˆ 3,5-bis(b-d-glucopyranosyloxy)-7-hydroxy-2-(4-hydroxy-3,5-dimethoxyphenyl)-1-
benzopyrylium chloride; Scheme 2), one of the most common naturally occurring
anthocyanins. In strongly acidic solution (pH 0.9), where malvin is in a pure flavylium
form, copigmentation is manifested by strong bathochromic shifts in the VIS-
absorption band of the pigment. For a copigment/pigment molar ratio of 200, the
bathochromic shift reaches 40 nm in the case of 4, 6, and 7 (Fig. 1,b), thus
demonstrating the efficiency of H2O-soluble flavonols in colour variation. Similar
bathochromic shifts were previously observed upon addition of quercetin 5'-sulfonate,
a H2O-soluble analog of quercetin, to solutions of anthocyanins and synthetic flavylium
ions [10]. Under mildly acidic conditions (pH 3.5), closer to those of natural media,
malvin is predominantly in the hemiacetal form (H2O addition at C(2), see Scheme 2),
and the solutions are poorly coloured. Addition of 4, 6, or 7 (copigment/pigment molar
ratio 200) strongly intensifies the colour (Fig. 1,a) as the result of a shift in the
flavylium-hemiacetal equilibrium induced by the selective stacking of the copigment on
the planar flavylium nucleus. Relative hyperchromic shifts ((A A0)/A0 at a given
wavelength; A0, A ˆ VIS absorbance in the absence and in the presence of the
copigment, resp.) as high as 20 were measured at room temperature. The VIS
absorbance gradually decreases when the temperature is raised as a consequence of the
partial (endothermic) dissociation of the copigmentation complex. At pH 3.5 where the
free flavylium ion is in negligible concentration, the relative hyperchromic shift can be
directly related to the binding constant K according to Eqn. 1 [5a,b] (rˆ ratio of the
molar absorption coefficients of malvin in its bound and free forms; c ˆ total copigment
concentration). At the isosbestic point of the free and bound flavylium ions
(determined at pH ca. 1 where the pigment is in pure flavylium form), r is equal to 1
and Eqn. 1 can be converted to Eqn. 2.
A
A0
ˆ r c K
DS0
ꢀ1†
ꢀ2†
A0
A0
A
DH0
RT
ln
ˆ
c A0
R
Plots of ln ((A A0)/c A0) as a function of 1/T are linear (correlation coefficients
>0.995, Fig. 2) for all copigments in the temperature range investigated (15 ± 458) and

432
Helvetica Chimica Acta ± Vol. 83 (2000)
Fig. 1. Copigmentation of malvin (5 ´ 10 5m) by rutin derivative 6 (10 2 m). a) At pH 3.5 (0.2m acetate buffer),
VIS spectra of malvin/6 at 15.5 (1), 25.0 (2), 35.0 (3) and 44.28 (4); VIS spectrum of malvin alone at 25.38 (5).
b) At pH 0.9 and 258, VIS spectrum of malvin alone (2) and of malvin/6 (2)
Fig. 2. Vanꢀt Hoff plots for copigmentation of malvin (5 ´ 10 5 m) by rutin (10 3 m) and rutin derivative 7 (10 2 m)
at pH 3.5 (0.2m acetate buffer)

Helvetica Chimica Acta ± Vol. 83 (2000)
433
Table 1. Thermodynamic Parameters for Copigmentation between Malvin and Rutin and Rutin Derivatives 3 ± 7
Deduced from Temperature Variations in the Range 15 ± 458 (0.2m acetate buffer, pH 3.5, copigment/pigment
molar ratio 200, unless otherwise specified.)
Rutin
37.2 (Æ 0.3)
59.0 (Æ 0.8)
2750
3
4
5
6
7
1
DH0 [kJ mol
]
33.0 (Æ 0.2)
16.7 (Æ 0.5)
17.1 (Æ 0.7)
13.6 (Æ 0.3)
7.9 (Æ 0.3)
DS0 [J K 1 mol
K [m 1]a)
]
62.0 (Æ 0.5) ‡ 8.4 (Æ 1.8) ‡ 0.5 (Æ 2.2) ‡ 16.7 (Æ 1.2) ‡ 35.3 (Æ 0.9)
1
350
19
2320
41
1060
22
1820
41
1700
40
Dlmax [nm]b)
8
a) Determined at 258 from DH0 T DS0 ˆ RT ln K. b) Determined at pH ca. 1 with the following copigment/
pigment molar ratios: 200 (3, 4, 6, 7), 50 (5), and 10 (rutin).
give accurate values for the temperature-independent (Vanꢀt Hoff behaviour) enthalpy
and entropy changes of copigmentation (Table 1). The fact that the Vanꢀt Hoff law is
still obeyed with copigments 4 (2 epimers) and 6 (3 regioisomers) suggests that the
different isomers display roughly the same affinity for malvin.
The ROESY spectra of rutin and rutin derivatives 3 and 5 (500 MHz, (D6)DMSO)
all display long-range correlations between the aglycone and the rhamnose residue. For
instance, H C(5') (ring B) is in NOE relationship with H C(1) (Rha) in rutin, with
H
C(3) (Rha) in 3, and with H C(6) (Rha) and H C(2) (Rha) and/or H C(3)
(Rha) in 5. Moreover, H C(6') (ring B) displays NOE connectivities with H C(1)
(Glc), H C(2) (Glc), and most rhamnose protons in 5. Hence, rutin and its derivatives
display favourable conformations in which the flavonol moiety and the terminal sugar
residue (Rha) are in molecular contact. The hydrophobic stacking of the polyphenolic
nuclei occurring in copigmentation is thus expected to reduce the solvation of the
carboxylate and sulfate groups introduced on the sugar moiety for H2O-solubility
purposes.
The hydration of a neutral group is essentially structure-making (DS0 < 0) because
of the ordering of the solvation H2O molecules to preserve extensive H-bonding
between them. On the other hand, the hydration of an ionic group is the sum of a
structure-making neutral term and of an electrostatic term, which is structure-making
for the strongly held H2O molecules of the first solvation shell and structure-breaking
(DS0 > 0) for the H2O molecules of the second solvation shell [11]. The electrostatic
term is wholly structure-breaking for large ions. Hence, the partial desolvation of a
large ionic group, beside being endothermic, should not strongly contribute to the
overall entropy because of compensating structure-making and structure-breaking
terms. This is verified in the copigmentation of malvin by 3: the strong destabilization of
the copigmentation complex (in comparison with the malvin ´ rutin complex) is mainly
due to a reduction in exothermicity1). Most remarkable is the large influence of the
single hydrogen succinate moiety of 7, which lowers the exothermicity by ca.
1
30 kJ mol and turns copigmentation into an entropically favourable process. The
1
)
In spite of its lower affinity toward anthocyanins, 3 remains as efficient as rutin in binding hard metal ions.
Hence, from the plots of the absorbance of the aluminum chelates (at the wavelength of the absorption
maximum) as a function of the total aluminum concentration and with a simple theoretical treatment
developed elsewhere [12], rutin and 3 were shown to form 1:1 complexes with Al3‡ (0.2m acetate buffer, pH
5.0, 258). Values for the apparent binding constant (defined as [AlL]/([Al][L], L being the flavonol ligand)
1
were estimated to be 14.5 (Æ1.5) ´ 103 and 15.9 (Æ1.7) ´ 103 m for rutin and 3, respectively.

434
Helvetica Chimica Acta ± Vol. 83 (2000)
entropic effect may reflect the partial desolvation of the structure-making neutral
carboxy group. In their association with malvin, copigments 4 ± 6 display thermody-
namic behaviours which are intermediate between that of anion 3 and that of 7
(assumed under neutral form). This suggests that the carboxy groups in 4 ± 6 are
partially dissociated under the conditions of the copigmentation experiments (pH 3.5).
Alternatively, the carboxylate groups in 4 ± 6 and the sulfate group in 3 may develop
favourable electrostatic interactions with the positively charged flavylium nucleus of
malvin and thereby maintain a higher exothermicity. Anyway, copigmentation of
malvin by the easily available highly H2O-soluble rutin derivatives 4 ± 7 is much less
temperature-dependent (and above 408 eventually stronger, see Fig. 2) than copig-
mentation by rutin, a potential advantage for the use of such copigments as colour
enhancers in anthocyanin-containing food and beverages.
3. Investigation of Copigmentation by Fluorescence Spectroscopy. As evidenced
here for the malvin ´ rutin couple, fluorescence spectroscopy seems a very sensitive
method for the quantitative investigation of copigmentation, especially in the case of
poorly soluble copigments such as rutin for which large copigment/pigment molar
ratios (a condition for accurate investigations by UV/VIS spectroscopy) cannot be
achieved. Under strongly acidic conditions (pure flavylium form), malvin displays a
broad weak fluoresence emission band around 600 nm (excitation at 540 nm) [13].
Addition of rutin promotes a very efficient quenching of malvin fluorescence at
relatively low copigment/pigment molar ratios (2 ± 10). Addition of quercetin 5'-
sulfonate to solutions of synthetic flavylium ions also promotes an efficient quenching
of flavylium fluorescence, a phenomenon shown to correlate with the ability of the
copigment to inhibit the photodegradation of the pigment [14]. Eqns. 3 and 4 can be
easily derived from simple solution chemistry, assuming complete extinction of
fluorescence in the copigmentation complex (F, F0 ˆ fluorescence intensity in the
presence and in the absence of copigment, resp.; [L] ˆ free copigment concentration;
Lt ˆ total copigment concentration; K ˆ pigment ´ copigment binding constant; c ˆ to-
tal pigment concentration).
The curve-fitting procedure (rˆ 0.995) of the experimental F vs. Lt plot (Fig. 3)
1
using Eqns. 3 and 4 gives a K value of 2560 (Æ100) m , in good agreement with the
value obtained by UV/VIS spectroscopy.
F0
F ˆ
ꢀ3†
ꢀ4†
1 ‡ K ‰LŠ
K c
Lt ˆ ‰LŠ ꢀ1 ‡
†
1 ‡ K ‰LŠ
4. Copigmentation and Colour Variation. Stacking interactions between anthocya-
nins (flavylium ions) and polyphenolic copigments compete with the nucleophilic
addition of H2O on the chromophore and thereby lower the concentration of the
colorless forms at a given pH (colour intensification). Simultaneously, the pigment VIS-
absorption band is shifted toward longer wavelengths, thus turning the typically red-
orange colours of free anthocyanin solutions into deep red and purple colours (colour
variation) [4][5]. Although still poorly understood, the latter phenomenon plays a key

Helvetica Chimica Acta ± Vol. 83 (2000)
435
Fig. 3. Plot of the fluorescence intensity of malvin at 600 nm (excitation at 540 nm) as a function of the total rutin
concentration (pH 0.9, 258). The solid line is the result of the curve-fitting procedure according to Eqns. 3 and 4.
role in colour expression by plants. In this work for instance, it is shown that a potent
copigment (high binding constant) in large concentrations (high H2O solubility)
actually promotes spectacular colour changes (Fig. 1).
Solvatochromic shifts in UV/VIS spectroscopy are typically interpreted by differ-
ences in solvation properties between ground and excited states of solutes. For instance,
enhancement in (or weakening of) H-bonding in the excited state explains why an
absorption band is bathochromically (or hypsochromically, resp.) shifted when going
from a nonprotic to a protic solvent [15]. By analogy, we propose that the copigment-
induced bathochromic shift in the pigment VIS band reflects stronger pigment-
copigment interactions in the excited state of the pigment. Van der Waals and
electrostatic solute-solute interactions as well as the hydrophobic effect are the major
driving forces in p-stacking complexation in H2O [16]. Although van der Waals
interactions and the hydrophobic effects both favour maximal overlap, this is generally
not so in the case of electrostatic interactions since repulsion between p electrons is
maximized when p electron-rich planar solutes are arranged face-to-face with no offset
in the complex. In the case of copigmention by flavones and flavonols (the most potent
anthocyanin copigments), such arrangements clearly result in an electrostatic
mismatch since both electron-rich rings A, both electron-rich rings B, and both
electron-poor rings C are stacked on one another. This unfavourable electrostatic
contribution may well be alleviated by a significant offset between the pigment and
copigment planes in the complex (at the expense of a small loss in van der Waals
interactions and hydrophobic-effect contribution).
Semi-empirical quantum-mechanics calculations (AM1 parametrization) in the
ground and first excited state (singlet) of malvidin (malvin aglycone) suggest that the
electronic transition of lowest energy (VIS band) is accompanied by a strong transfer of

436
Helvetica Chimica Acta ± Vol. 83 (2000)
electron density from ring B to ring C, in agreement with the direction of the transition
dipole calculated for a series of anthocyanidins (anthocyanin aglycones) in a previous
detailed investigation [17]. The electron transfer occurring within the malvidin nucleus
upon absorption of a VIS photon is best evidenced by summing the atomic charges
borne by the ring-B and -C atoms in the ground and excited states. These ring charges
are displayed in Scheme 3, together with those for the copigment (quercetin) in the
ground state. A quinonoid mesomeric form of the pigment with a positively charged O-
atom at C(4') is chosen to depict the pigment in the excited state since it is more con-
sistent with an electron-poor ring B and an electron-rich ring C. Alternatively, the represen-
tation of the probability densities for HOMO and LUMO of ground-state malvidin
(Fig. 4) confirms that transferring one electron from HOMO to LUMO results in a
large increase of p-electron density on ring C with a concomitant decrease on ring B.
Hence, in the course of the electronic transition in the VIS range, the p-electron
densities on the flavylium rings B and C are changed so as to allow more favourable
electrostatic interactions with the corresponding rings in the copigment. Consequently,
the pigment-copigment contact must be stronger in the excited state. A simple thermo-
dynamic cycle allows estimation of the increase in the pigment-copigment binding energy
(DDG0) when the pigment is promoted from the ground to the first excited state (Fig. 5).
From the experimentally determined binding constant of the malvin ´ rutin complex
in the ground state, it can be estimated that more than 95% of the malvin flavylium ion
Scheme 3

Helvetica Chimica Acta ± Vol. 83 (2000)
437
Fig. 4. Probability densities for a) HOMO and b) LUMO of ground-state malvidin. HyperChem program, AM1
parametrization
Fig. 5. Estimation of the increase in the pigment (P) ´ copigment (CP) binding energy (DDG0) on excitation
is bound at pH < 1 (no colorless forms) for a copigment/pigment molar ratio of 200.
Since such a ratio cannot be reached with rutin because of its poor H2O solubility, the
value for the wavelength of the absorption maximum of the copigmentation complex
can be estimated with a highly H2O-soluble rutin derivative such as 4 or 6 whose affinity
for malvin is not significantly depressed by derivatization. The plot of the wavelength of
the flavylium absorption maximum as a function of copigment 6 concentration actually
shows a clear plateau around 560 nm at large copigment/pigment molar ratios (Fig. 6).
From the thermodynamic cycle (Fig. 5), one gets Eqn. 5 (DG0, DG*0 ˆ copigmentation
free enthalpy change in the ground and first excited states, resp.; K, K* ˆ pigment ´
copigment binding constant in the ground and first excited state, resp.; NA ˆ Avogadro
constant; h ˆ Planck constant; c ˆ speed of light; R ˆ gas constant).
ꢀ
ꢁ
*
K
1
lP
1
lPCP
DG0
DG0à ˆ RT ln
ˆ NA h ꢀnP
nPCP† ˆ NA h c
ꢀ5†
k
Energy differences between the vibrational ground states of the electronic ground
and first excited states should be considered in the thermodynamic cycle. Hence, a
calculation based on wavelengths of absorption maximum for the pigment and its
copigmentation complex (lP, lPCP) should be corrected for the fact that vibrational

438
Helvetica Chimica Acta ± Vol. 83 (2000)
Fig. 6. Plot of the wavelength of the absorption maximum of malvin (5 ´ 10 5m) as a function of the total
concentration of rutin derivative 6 (pH 0.9, 258)
excited states are actually populated in the electronic excited states (Franck-Condon
transitions)2). However, such corrections must be similar for the flavylium ion in its
free and bound forms and are thus expected to cancel out at least partially in Eqn. 5. In
the case of the malvin ´ rutin couple, lP and lPCP are equal to 520 and ca. 560 nm,
1
respectively, thus giving DDG0 ˆ 16.4 kJ mol and K*/K ˆ 760. Hence, this simple
model suggests that the binding of potent flavonol copigments to anthocyanins
(flavylium ions) may be two to three orders of magnitude stronger when anthocyanins
are in their first electronic excited state.
5. Antioxidant Properties. Flavonoids can exert their antioxidant properties via
several mechanisms including quenching of reactive O-species (hydroxyl, alkoxyl, and
peroxyl radicals) and inhibition of enzymes involved in their production, chelation of
transition-metal ions able to promote radical formation (Fenton reaction), and
regeneration of membrane-bound antioxidants such as a-tocopherol (vitamin E)
[2][7][18].
DPPH (ˆ2,2-diphenyl-1-(2,4,6-trinitrophenyl)hydrazyl; 8) is a highly coloured
commercially available radical widely used for estimating the ability of antioxidants to
trap potentially damaging one-electron oxidants/H-atom-abstracting agents [19]. In
particular, antioxidants can be characterized by their stoichiometry n, i.e. the number of
DPPH molecules reduced per molecule of antioxidant. Hence, the antioxidant can be
regarded as a source of n H-atoms that will convert DPPH to the corresponding
hydrazine. When monitoring the H-transfer reactions from a flavonoid antioxidant to
2
)
A better estimate of the energy differences between ground and excited states should be NA h(nabs ‡ nem)/2,
nabs and nem being the frequencies of absorption and emission maxima. However, such a correction cannot
be applied to the bound flavylium ion since copigmentation by rutin leads to a complete quenching of
flavylium fluorescence, thereby preventing the determination of nem
.

Helvetica Chimica Acta ± Vol. 83 (2000)
439
DPPH (8) by UV/VIS spectroscopy, two steps can be distinguished: a first step during
which the DPPH VIS absorbance (lmax 516 nm in MeOH) quickly decays (typical time
interval in MeOH: 60 s) and a second step during which the DPPH VIS absorbance
slowly decays over ca. 1 h to a final constant value. The fast step essentially refers to
abstractions of the most labile H-atoms (OH C(3), OH C(3'), OH C(4') in the case
of quercetin [20]), whereas the slow step reflects the remaining activity in the
oxidation-degradation products. A total stoichiometry (ntot) can be determined in a
static way from the overall amplitude of the kinetic run by Eqn. 6 (Af ˆ VIS absorbance
corresponding to the plateau at the end of the kinetic run (complete conversion of the
antioxidant into inert oxidation-degradation products); A0 ˆ initial absorbance (just
after addition of the antioxidant); c ˆ initial antioxidant concentration; c0 ˆ initial
DPPH concentration).
ꢀ
ꢁ
c0
c
Af
A0
ntot
ˆ
1
ꢀ6†
Of course, the initial DPPH/antioxidant molar ratio c0/c must be higher than ntot for
Eqn. 6 to apply. A kinetic stoichiometry n can also be estimated from the curve-fitting
of the kinetic traces featuring the decay of the DPPH band during the fast step. To that
purpose, an antioxidant of stoichiometry n is simply modeled as n independent sub-
units AH (typically, flavonoid OH groups bearing most labile H-atoms), which all
transfer a H-atom to DPPH with the same second-order rate constant k. This
procedure, already applied to quercetin, rutin [20], and some gallotannins [21], also
gives satisfying curve fits (r> 0.995) for rutin derivatives 5 ± 7 and consistent values for
n and k at different DPPH/flavonoid molar ratios (Table 2). The k value provides a
quantitative estimate of the overall reactivity of the antioxidant toward DPPH (8)
during the fast step of trapping. Rutin bis(hydrogen succinate) 6 and rutin 4'''-
(hydrogen succinate) (7) react with DPPH as rapidly as (and eventually, more rapidly
than) rutin. However, a higher degree of succinylation significantly slows the rate of
trapping. Stoichiometries close to 2 (fast step) are consistent with the formation of
quinone intermediates that have been trapped as solvent or benzenesulfinate adducts
in the case of quercetin and rutin [20]. To form quinones, aryloxy radicals must further
react with DPPH or with themselves (disproportionation). The latter process is more
probable since a fast second-order decay of polyphenolic aryloxyl radicals in H2O (2k
1
in the range 106 ± 107m 1 s ) has already been reported in the case of flavonoids from
pulse-radiolysis experiments [22] and, in the case of caffeic acid, from electrochemical
investigations [23]. Total stoichiometries ca. 2 ± 3 times larger than those determined by

440
Helvetica Chimica Acta ± Vol. 83 (2000)
Table 2. H-Abstraction from Rutin and Rutin Derivatives 5 ± 7 by DPPH (8) in MeOH at 258. For definition of
parameters n, ntot, and k, see text.
b
DPPH [equiv.]
k [m 1 s 1] a)
n a)
ntot
)
5
5
6
6
7
7
6
8
4
8
4
6
8
4
330 (Æ 7)
276 (Æ 5)
1.76 (Æ 0.01)
1.92 (Æ 0.01)
2.63 (Æ 0.01)
2.55 (Æ 0.01)
1.93 (Æ 0.01)
1.92 (Æ 0.01)
2.23 (Æ 0.01)
2.44 (Æ 0.01)
±
5.8
±
895 (Æ 9)
736 (Æ 10)
2220 (Æ 20)
1850 (Æ 20)
904 (Æ 17)
932 (Æ 5)
4.2 (4.3 c))
±
4.8 (4.7 d))
Rutin
Rutin
5.9
±
a) From the curve fitting of the kinetic traces (typical time intervals: 60 s), see [20]. b) From Eqn. 6; final DPPH
absorbance recorded 1 h after addition of the antioxidant. c) DPPH/6 molar ratio 6. d) DPPH/7 molar ratio 8.
the kinetic method (fast step) are obtained from Eqn. 6 (Table 2), thus showing that
quinones are actually intermediates and further react with DPPH. Quinone-DPPH
reactions could occur via direct H-abstraction from OH C(7) or, more probably, after
intermediate steps of solvent addition [20] and/or dimerization, which allow the
regeneration of a benzene-1,2-diol moiety (ring B) liable to transfer its phenolic H-atoms
to DPPH. Interestingly, although abstraction of H-atoms from 5 by DPPH is slower
than from rutin, the total number of DPPH radicals trapped (ntot) is the same for both
compounds. Hence, total or partial succinylation of the sugar moiety of rutin does not
seem to significantly alter its antioxidant ability and could eventually prolong it owing
to an improved bioavailability.
Experimental Part
General. Malvin (HPLC-pure; Extrasynthese, Genay, France), rutin, and DPPH (95%; Aldrich) were used
as received. UV/VIS Spectra: Hewlett-Packard-8453 diode-array spectrometer, equipped with a magnetically
stirred quartz cell (optical pathlength ˆ 1 cm) thermostated by a water bath; lmax in nm.
Fluorescence spectra: BioLogic spectrometer with a thermostated quartz cell.
NMR Spectra: 300-MHz- and 500-MHz-Bruker apparatus; at 278; chemical shifts d in ppm, coupling
1
constants J in Hz; H signals assigned from 1D- and 2D-COSY experiments, 13C signals assigned from DEPT,
HMBC, and HSQC experiments (3) or from DEPT and comparisons with [24] (4, 5, and 7). FC ˆ Flash
chromatography.
5,7-Bis(acetyloxy)-2-[3,4-bis(acetyloxy)phenyl]-3-{{2,3,4-tri-O-acetyl-6-O-[4-O-acetyl-6-deoxy-2,3-O-(phe-
nylmethylene)-a-l-mannopyranosyl]-b-d-glucopyranosyl}oxy}-4H-1-benzopyran-4-one (1). A soln. of rutin
(5 g, 8.2 mmol) and TsOH (156 mg, 0.82 mmol) in DMF (20 ml) is dried (H2O traces) upon co-evaporation with
toluene (10 ml). Benzaldehyde dimethyl acetal (3.70 ml, 24.6 mmol) and toluene (15 ml) are added, and the
mixture is heated at 408 under reduced pressure (Rotavap) for the gradual removal of MeOH formed in the
course of the reaction (co-evaporation with toluene). Portions of toluene (10 ml) are further added and the co-
evaporation repeated until complete disappearence of rutin on TLC. The orange oil thus obtained is poured into
CH2Cl2 (100 ml). The yellow precipitate (rutin benzylidene acetal) is filtered, washed with Et2O, and dried.
Rutin benzylidene acetal (1.15 g, 1.64 mmol) is then dissolved in pyridin (14 ml), and Ac2O (7.8 ml, 82 mmol)
and DMAP (40 mg, 0.32 mmol) are successively added at 08. The mixture is stirred at r.t. for 36 h, then at 508 for
24 h. After evaporation of pyridin, the product is precipitated in AcOEt/petroleum ether, washed with
petroleum ether, and dried: 82% (from rutin) of 1 (two epimers, ca. 1:1 molar ratio). Beige powder. TLC (silica
gel, AcOEt/petroleum ether 1:1): Rf 0.62, 0.70. UV/VIS (CHCl3): 302, 261. 1H-NMR (CDCl3, 300 MHz): 7.93,
7.91 (2d, J ˆ 2.2, 2 H, H C(2')); 7.83, 7.77 (2dd, J ˆ 2.2, 8.5, 2 H, H C(6')); 7.5 ± 7.3 (m, 12 H, H C(5'), Ph);
7.26, 7.18 (2d, J ˆ 2.2, 2 H, H C(8)); 6.84, 6.81 (2d, J ˆ 2.2, 2 H, H C(6)); 6.12, 5.70 (2s, 2 H, PhCH); 5.65, 5.51
(2d, J ˆ 8.1, 2 H, H C(1'')); 5.27 (m, 2 H, H C(3'')); 5.17 (m, 2 H, H C(2'')); 5.01 (m, 2 H, H C(4'')); 4.89 (m,

Helvetica Chimica Acta ± Vol. 83 (2000)
441
2 H, H C(4''')); 4.82, 4.75 (2s, 2 H, H C(1''')); 4.38, 4.21 (2dd, J ˆ 5.2, 8.4, 2 H, H C(3''')); 4.01, 3.95 (2d, J ˆ
5.2, 2 H, H C(2''')); 3.6 ± 3.5 (m, 8 H, H C(5''), 2 H C(6''), H C(5''')); 2.44 ± 1.99 (8s, 48 H, 8 Ac); 1.01, 0.98
(2d, J ˆ 6.2, 6 H, Me(6''')).
5,7-Bis(acetyloxy)-2-[3,4-bis(acetyloxy)phenyl]-3-{[2,3,4-tri-O-acetyl-6-O-(4-O-acetyl-2-O-benzoyl-6-deoxy-
a-l-mannopyranosyl-b-d-glucopyranosyl]oxy}-4H-1-benzopyran-4-one (2). To a soln. of 1 (1 g, 1 mmol) in
CHCl3 (100 ml) are successively added N-bromosuccinimide (430 mg, 2.5 mmol), BaCO3 (460 mg, 2.4 mmol),
H2O (104 ml, 6 mmol), and benzoyl peroxide (26 mg, 0.1 mmol). The mixture is refluxed under N2 for 4 h, then
cooled, washed with H2O, dried (MgSO4), and concentrated. Chromatography (silica gel, AcOEt/petroleum
ether 7:3) affords 2 (800 mg, 76%). Yellow powder. TLC (silica gel, AcOEt/petroleum ether 1 :1): Rf 0.48. UV/
VIS (CHCl3): 301. 1H-NMR (CDCl3, 300 MHz): 8.04 ± 7.96 (m, H C(6'), H C(2'), 2 Ho (Bz)); 7.57 (d, J ˆ 7.4,
Hp (Bz)); 7.44 (t, J ˆ 7.4, 2 Hm (Bz)); 7.35 (d, J ˆ 8.5; H C(5')); 7.28 (d, J ˆ 2.2; H C(8)); 6.83 (d, J ˆ 2.2;
H
H
H
C(6)); 5.68 (d, J ˆ 7.7, H C(1'')); 5.31 (t, J ˆ 9.6, H C(3'')); 5.20 (dd, J ˆ 9.6, 7.7, H C(2'')); 5.08 (d, J ˆ 4.4,
C(2''')); 5.03 (t, J ˆ 9.6, H C(4'')); 4.90 (t, J ˆ 9.9, H C(4''')); 4.67 (s, H C(1''')); 3.96 (dd, J ˆ 9.9, 4.4,
C(3''')); 3.65 ± 3.60 (m, H C(5''), H C(6''), H C(5''')); 3.28 (dd, J ˆ 11.0, 5.1, H C(6'')); 2.45 ± 2.02 (8s,
8 Ac); 1.03 (d, J ˆ 5.9, Me(6''')).
3-{[6-O-(6-Deoxy-3-O-sulfo-a-l-mannopyranosyl)-b-d-glucopyranosyl]oxy}-2-(3,4-dihydroxyphenyl)-5,7-
dihydroxy-4H-1-benzopyran-4-one Sodium Salt (3). A soln. of 2 (500 mg, 1.8 mmol) and SO3 ´ pyridin complex
(155 mg, 0.9 mmol) in DMF (6 ml) is heated at 508 under N2 for 12 h. After evaporation, MeOH (5 ml) and 1m
MeONa in MeOH (5 ml) are added. The mixture is stirred under N2 for 12 h in the dark, then neutralized with a
‡
‡
Dowex resin (H form), filtered, and evaporated. A dil. soln. of the residue in H2O is passed over a Dowex Na
ion-exchange resin and then evaporated. FC (C18 silica gel, H2O/MeCN 95 :5) affords 3 (100 mg, 30%). Yellow
1
powder. UV/VIS (H2O): 352, 264. H-NMR ((D6)DMSO, 300 MHz): 7.57 (dd, J ˆ 8.1, 2.2, H C(6')); 7.52 (d,
J ˆ 2.2, H C(2')); 6.87 (d, J ˆ 8.1, H C(5')); 6.36 (d, J ˆ 1.5, H C(8)); 6.17 (d, J ˆ 1.5, H C(6)); 5.17 (d, J ˆ
7.4, H C(1'')); 4.38 (br. s, H C(1''')); 4.05 (dd, J ˆ 8.8, 3.7, H C(3''')); 3.67 (br. s, H C(2''')); 3.36 ± 2.99 (m,
H
C(2''), H C(3''), H C(4''), H C(5''), 2 H C(6''), H C(4'''), H C(5''')); 0.95 (d, J ˆ 5.9, Me(6''')).
13C-NMR ((D6)DMSO, 75 MHz): 178.1 (C(4)); 165.0 (C(7)); 161.7 (C(5)); 157.8 (C(8a)); 157.4 (C(2)); 149.1
(C(4')); 145.4 (C(3')); 134.4 (C(3)); 122.7 (C(6')); 122.0 (C(1')); 116.9 (C(2')); 116.1 (C(5')); 104.8 (C(4a)); 102.7
(C(1')); 102.0 (C(1''')); 99.5 (C(6)); 94.6 (C(8)); 77.7 (C(3''')); 77.1 (C(5'')); 76.4 (C(2'')); 74.8 (C(4''')); 71.1
(C(2''')); 70.7 (C(3'')); 69.6 (C(4'')); 69.1 (C(5''')); 68.7 (C(6'')); 18.3 (C(6''')). ES-MS (neg. mode): 711.1 ([M
‡
‡
H] ), 689.1 ([M Na] ).
3-{{6-O-[2,3-O-(2-Carboxy-1-methylethylidene)-6-deoxy-a-l-mannopyranosyl]-b-d-glucopyranosyl}oxy}-2-
(3,4-dihydroxyphenyl)-5,7-dihydroxy-4H-1-benzopyran-4-one (4). As described for 1, with rutin (0.5 g,
0.8 mmol), TsOH (16 mg, 0.08 mmol), DMF (10 ml), and toluene (10 ml), then ethyl acetoacetate dimethyl
acetal (1.2 g, 7 mmol) and toluene (10 ml), and finally portions of toluene (10 ml). After evaporation, 1m KOH
‡
in EtOH (10 ml) is added. The mixture is heated at 608 under N2 for 1.5 h, then acidified with a Dowex resin (H
form) to pH ca. 1, filtered, and evaporated. Chromatography (C18 silica gel, H2O/MeCN 4 :1) affords 4 (2
epimers, ca. 1:1 molar ratio; 472 mg, 85%). UV/VIS (H2O): 351, 255. TLC (silica gel, AcOEt/butanone/
HCO2H/H2O 14 :2 :1 :1): Rf 0.5. 1H-NMR (CD3OD, 500 MHz): 7.71 (d, J ˆ 2.3, 2 H, H C(2')); 7.64 (dd, J ˆ 8.6,
2.3, 2 H, H C(6')); 6.89 (d, J ˆ 8.6, 2 H, H C(5')); 6.45 (s, 2 H, H C(8)); 6.25 (s, 2 H, H C(6)); 5.21, 5.17 (2d,
J ˆ 7.4, 2 H, H C(1'')); 4.82, 4.80 (2s, 2 H, H C(1'''); 4.01 ± 3.80 (m, 6 H, H C(6''), H C(2'''), H C(3'''));
3.53 ± 3.14 (m, 14 H, H C(2''), H C(3''), H C(4''), H C(5''), H C(6''), H C(4'''), H C(5''')); 2.76, 2.68
(2d, J ˆ 14.3, 2 H, C(Me)CH2); 2.59, 2.53 (2d, J ˆ 13.9, 2 H, C(Me)CH2); 1.62, 1.40 (2s, 6 H, C(Me)CH2); 1.11,
1.10 (2d, J ˆ 6.6, 6 H, Me(6''')). 13C-NMR (CD3OD, 300 MHz): 178.3 (C(4)); 172.5 (CO2H); 164.9 (C(7)); 161.9
(C(5)); 158.2 (C(8a)); 157.4 (C(2)); 148.8 (C(4')); 144.8 (C(3')); 134.5 (C(3)); 122.4 (C(6')); 122.0 (C(1')); 116.7
(C(2')); 115.0 (C(5')); 108.2 (C(Me)CH2); 104.6 (C(4a)); 103.3 (C(1'')); 98.9 (C(1''')); 97.7 (C(6)); 93.9 (C(8));
78.8 (C(3'')); 77.1 (C(5'')); 76.4 (C(2'')); 76.4 (C(4''')); 76.0 (C(2''')); 74.7 (C(3''')); 74.4 (C(4'')); 70.5 (C(5''));
66.9 (C(6'')); 65.7 (?); 47.2, 46.6 (C(Me)CH2); 23.8, 23.2 (C(Me)CH2); 16.5 (Me(6''')). ES-MS (neg. mode):
‡
693.1 ([M H] ).
2-(3,4-Dihydroxyphenyl)-5,7-dihydroxy-3-{{2,3,4-tris-O-(3-carboxy-1-oxopropyl)-6-O-[2,3,4-tris-O-(3-car-
boxy-1-oxopropyl)-6-deoxy-a-l-mannopyranosyl]-b-d-glucopyranosyl}oxy}-4H-1-benzopyran-4-one (5). A soln.
of rutin (500 mg, 0.82 mmol), succinic anhydride (821 mg, 8.2 mmol), and DMAP (50 mg, 0.41 mmol) in
pyridine (20 ml) is heated at 708 under N2 for 24 h. After removal of pyridine and addition of H2O (20 ml), the
‡
mixture is acidified with Dowex (H form) to pH 2, filtered, washed with CHCl3 (3 Â 30 ml), and lyophilized.
FC (C18 silica gel, H2O ! H2O/MeCN 4 :1) affords 5 (934 mg, 95%). Yellow powder. UV/VIS (MeOH): 356,
257. TLC (silica gel, AcOEt/butanone/HCO2H/H2O 20 :2 :1:1): Rf 0.53. 1H-NMR ((D6)DMSO, 300 MHz):
12.57 (1s, OH C(5)); 7.65 (d, J ˆ 2.2, H C(2')); 7.44 (dd, J ˆ 2.2, 8.8, H C(6')); 6.83 (d, J ˆ 8.8, H C(5')); 6.41

442
Helvetica Chimica Acta ± Vol. 83 (2000)
(d, J ˆ 1.5, H C(8)); 6.19 (d, J ˆ 1.5, H C(6)); 5.64 (d, J ˆ 8.1, H C(1'')); 5.38 (t, J ˆ 9.6, H C(3'')); 5.15 (t,
J ˆ 9.6, H C(2'')); 4.97 (t, J ˆ 9.6, H C(4'')); 4.90 (d, J ˆ 2.9, H C(2''')); 4.87 (dd, J ˆ 10.3, 2.9, H C(3'''));
4.74 (t, J ˆ 10.3, H C(4''')); 4.44 (s, H C(1''')); 3.7 ± 3.5 (m, H C(5''), 2 H C(6''), H C(5''')); 2.41 (s,
6 CH2CH2COOH); 0.94 (d, J ˆ 5.9, Me(6''')). 13C-NMR ((D6)DMSO, 75 MHz): 177.6 (C(4)); 174.5 ± 174.0
(CO2H); 172.4 ± 171.8 (COCH2CH2COOH); 165.1 (C(7)); 162.0 (C(5)); 157.3 (C(8a)); 157.1 (C(2)); 149.7
(C(4')); 145.6 (C(3')); 133.8 (C(3)); 121.8 (C(6')); 121.2 (C(1')); 117.2 (C(2')); 116.2 (C(5')); 104.7 (C(4a)); 99.8
(C(1'')); 99.5 (C(1''')); 98.3 (C(6)); 94.5 (C(8)); 73.2 (C(3'')); 73.0 (C(5'')); 72.3 (C(2'')); 71.1 (C(4''')); 70.8
(C(3''')); 69.4 (C(2''')); 69.1 (C(4'')); 67.4 (C(6'')); 66.7 (C(5''')); 29.7 ± 29.3 (CH2CH2COOH); 17.8 (Me(6''')).
‡
FAB-MS (pos. mode): 1211.4 ([M ‡ H] ).
3-{{6-O-[3,4-Bis-O-(3-carboxy-1-oxopropyl)-6-deoxy-a-l-mannopyranosyl]-b-d-glucopyranosyl}oxy}- and
3-{{2- or 3- or 4-O-(3-Carboxy-1-oxopropyl)-6-O-[4-O-(3-carboxy-1-oxopropyl)-6-deoxy-a-l-mannopyrano-
syl]-b-d-glucopyranosyl}oxy}-2-(3,4-dihydroxyphenyl)-5,7-dihydroxy-4H-1-benzopyran-4-one (6; isomer ratio
ca. 3 :2 :2). Succinic anhydride (98 mg, 0.98 mmol) is added to a soln. of rutin (200 mg, 0.33 mmol) and DMAP
(8.6 mg, 0.07 mmol) in pyridine (10 ml) at 08 under N2. After stirring for 6 h at 08, pyridine is evaporated and
MeOH (10 ml) added. The mixture is acidified with Dowex and filtered, the filtrate evaporated and the residue
subjected to chromatography (C18 silica gel, H2O ! H2O/MeOH 65 :35): 6 (3 regioisomers; 75 mg, 28%). Beige
solid. UV/VIS (MeOH): 358, 257. TLC (silica gel, AcOEt/butanone/HCO2H/H2O 20 :2 :1:1): Rf 0.25. 1H-NMR
((D6)DMSO, 300 MHz): 12.58, 12.56, 12.53 (3s, 3 OH C(5)); 7.53 ± 7.49 (m, 3 H C(2'), 3 H C(6')); 6.85 (d,
J ˆ 8.1, 3 H C(5')); 6.39 (s, 3 H C(8)); 6.20 (s, 3 H C(6)); 5.52, 5.46 (2d, J ˆ 8.1, 2 H C(1'')); 5.44 ± 5.35 (m,
1 H C(1''), 1 H(Glc)); 4.83 (m, 3 H C(4''')); 4.65 (t, J ˆ 9.5, 1 H(Glc)); 4.59 (dd, J ˆ 9.5, 3.7, 1 H C(3'''));
4.44, 4.41, 4.39 (3s, 3 H C(1''')); 3.75 ± 3.05 (other H signals); 2.50 (m, CH2CH2COOH); 1.02, 0.99, 0.74 (3d, J ˆ
‡
5.9, 3 Me(6''')). FAB-MS (pos. mode): 811.6 ([M ‡ H] ).
3-{{6-O-[4-O-(3-Carboxy-1-oxopropyl)-6-deoxy-a-l-mannopyranosyl]-b-d-glucopyranosyl}oxy}-2-(3,4-dihy-
droxyphenyl)-5,7-dihydroxy-4H-1-benzopyran-4-one (7). Succinic anhydride (49 mg, 0.49 mmol) is added to a
soln. of rutin (200 mg, 0.33 mmol) and DMAP (4 mg, 0.03 mmol) in pyridine (10 ml) at 08 under N2. After
stirring for 7 h at 08, pyridine is evaporated and MeOH (10 ml) added. The mixture is acidified with Dowex and
filtered, the filtrate evaporated, and the residue subjected to chromatography (C18 silica gel, H2O ! H2O/MeCN
70 :30): 7 (50 mg, 21%). Beige solid. UV/VIS (MeOH): 359, 257. TLC (silica gel, AcOEt/butanone/HCO2H/
H2O 20 :2 :1:1): Rf 0.22. 1H-NMR ((D6)DMSO, 300 MHz): 7.55 ± 7.46 (m, H C(2'), H C(6')); 6.84 (d, J ˆ 8.1,
H
H
H
C(5')); 6.37 (s, H C(8)); 6.17 (s, H C(6)); 5.42 (d, J ˆ 7.4, H C(1'')); 4.81 (t, J ˆ 8.8, H C(4''')); 4.39 (s,
C(1''')); 3.72 ± 3.05 (m, H C(2''), H C(3''), H C(4''), H C(5''), 2 H C(6''), H C(2'''), H C(3'''),
C(5''')); 2.41 (s, CH2CH2COOH); 1.00 (d, J ˆ 5.9, Me(6''')). 13C-NMR ((D6)DMSO, 75 MHz): 178.0 (C(4));
175.3 (CO2H); 172.9 (COCH2CH2COOH); 165.9 (C(7)); 162.0 (C(5)); 157.3 (C(8a), C(2)); 149.7 (C(4')); 145.9
(C(3')); 134.0 (C(3)); 122.0 (C(6')); 121.7 (C(1')); 117.1 (C(2')); 116.2 (C(5')); 104.4 (C(4a)); 101.6 (C(1''),
C(1''')); 99.8 (C(6)); 94.6 (C(8)); 78.7 (C(3'')); 76.3 (C(5'')); 72.7 (C(2''), C C(4''')); 71.4 (C(3''')); 71.2 (C(2'''));
69.2 (C(4'')); 68.5 (C(6'')); 67.4 (C(5''')); 31.3 (CH2 (Suc)); 30.6 (CH2 (Suc)); 18.6 (Me(6''')). FAB-MS (pos.
‡
mode): 711.2 ([M ‡ H] ).
Semiempirical Quantum-Mechanics Calculations. Run in vacuo on a Pentium-90 PC with the HyperChem
program (Autodesk, Sausalito, California, USA) and the AM1 parametrization. Atomic charges in the ground
state and probability densities for HOMO and LUMO were obtained after previous geometry optimization.
Atomic charges in the excited state were the result of so-called ꢁsingle-point calculationsꢀ (without geometry
optimization).
Data Analysis. The curve fittings were carried out on a Pentium-120 PC (Scientist program, MicroMath, Salt
Lake City, Utah, USA) through least-square regression. Optimized values for the parameters and the
corresponding standard deviations are reported.
Copigmentation Experiments by UV/VIS Spectroscopy. A given copigment is added to a 5 ´ 10 5 m soln. of
malvin in a 0.2m acetate buffer at pH 3.5 (copigment/pigment molar ratio 20 ± 200, depending on the copigment
solubility). An aliquot of the soln. is placed in the spectrometer cell, and UV/VIS spectra are recorded every 58
in the temp. range 15 ± 458. The soln. is kept ca. 30 min at the selected temp. before each measurement to allow
the structural transformations of the pigment (H2O addition) to reach complete equilibrium. The experiment is
repeated with the soln. of malvin without copigment. Both solns. are then acidified to pH 0.9 ± 1.0 with conc. HCl
(no significant dilution) for the determination of the isosbestic point of malvin (flavylium form) and its
copigmentation complex.
Copigmentation Experiments by Fluorescence Spectroscopy. Aliquots (10 ml) of a 2.5 ´ 10 2 m soln. of rutin in
MeOH are added to 2 ml of a 5 ´ 10 5 m soln. of malvin at pH 0.9 ± 1.0 (acetate buffer acidified with conc. HCl
soln.). Emission spectra are recorded after each addition (excitation at 540 nm).

Helvetica Chimica Acta ± Vol. 83 (2000)
443
H-Abstraction by DPPH (8). To 2 ml of freshly prepared 2 ´ 10 4 m 8 in MeOH placed in the spectrometer
cell, freshly prepared 10 3 m flavonol in MeOH (50 ± 100 ml) is added. Spectra are recorded every 0.5 s over 30 ±
90 s for the determination of parameters n and k. Kinetic runs over 1 h are used for the determination of ntot
.
REFERENCES
[1] a) ꢁThe Flavonoids, Advances in Research since 1986ꢀ, Ed. J. B. Harborne, Chapman and Hall, London,
1994; b) E. Haslam, ꢁPractical Polyphenolicsꢀ, University Press, Cambridge, 1998.
[2] ꢁFlavonoids in Health and Diseaseꢀ, Eds. C. A. Rice-Evans and L. Packer, M. Dekker, New York, 1997.
[3] M. G. L. Hertog, E. J. M. Feskens, P. C. H. Hollman, M. B. Katan, D. Kromhout, Lancet 1993, 342, 1007;
D. S. Leake, in ꢁPhytochemistry of Fruit and Vegetablesꢀ, Eds. F. A. Tomas-Barberan and R. J. Robins,
Clarendon Press, Oxford, 1997, p. 287.
[4] T. Goto, T. Kondo, Angew. Chem., Int. Ed. 1991, 30, 17; O. Dangles, Analusis 1997, 25, M 50.
[5] a) R. Brouillard, G. Mazza, Z. Saad, A.-M. Albrecht-Gary, A. Cheminat, J. Am. Chem. Soc. 1989, 111,
2604; b) O. Dangles, R. Brouillard, Can. J. Chem. 1992, 70, 2174; c) O. Dangles, R. Brouillard, J. Chem.
Soc., Perkin Trans. 2 1992, 247; d) T. V. Mistry, Y. Cai, T. H. Lilley, E. Haslam, J. Chem. Soc., Perkin Trans. 2
1991, 1287; e) S. Asen, R. N. Stewart, K. H. Norris, Phytochemistry 1972, 11, 1139.
[6] S. Baba, T. Furuta, M. Horie, H. Nakagawa, J. Pharm. Sci. 1981, 70, 780; S. Baba, T. Furuta, M. Fujioka, T.
Goromaru, J. Pharm. Sci. 1983, 72, 1155; V. D. Bokkenheuser, C. H. L. Shackleton, J. Winter, Biochem. J.
1987, 248, 953; J. Winter, L. H. Moore, V. R. Dowell, V. D. Bokkenheuser, Appl. Environ. Microbiol. 1989,
55, 1203; D. H. Kim, E. A. Jung, I. S. Sohng, J. A. Han, T. H. Kim, M. J. Han, Arch. Pharm. Res. 1998, 21, 17.
Â
Â
Â
Â
Â
[7] C. Manach, C. Morand, V. Crespy, C. Demigne, O. Texier, F. Regerat, C. Remesy, FEBS Lett. 1998, 426, 331.
[8] P. Calias, T. Galanopoulos, M. Maxwell, A. Khayat, D. Graves, H. N. Antoniades, M. dꢀAlarco, Carbohydr.
Res. 1996, 292, 83.
[9] R. W. Binkley, G. S. Goewey, J. C. Johnston, J. Org. Chem. 1984, 49, 992.
[10] J. G. Sweeny, M. M. Wilkinson, G. A. Iacobucci, J. Agric. Food Chem. 1981, 29, 563.
[11] M. H. Abraham, E. Matteoli, J. Liszi, J. Chem. Soc., Faraday Trans. 1 1983, 79, 2781.
[12] B. Alluis, O. Dangles, Helv. Chim. Acta 1999, 82, 2201.
[13] P. Figueiredo, F. Pina, L. Vilas-Boas, L. Maçanita, J. Photochem. Photobiol. 1990, 52, 411.
[14] M. Santhanam, R. R. Hautala, J. G. Sweeny, G. A. Iacobucci, Photochem. Photobiol. 1983, 38, 477.
[15] ꢁMolecular Interactions and Electronic Spectraꢀ, Eds. N. Mataga, T. Kubota, M. Dekker, New York, 1970;
ꢁSolvents and Solvent Effects in Organic Chemistryꢀ, Ed. C. Reichardt, VCH, Weinheim, 1990.
[16] C. A. Hunter, J. K. M. Sanders, J. Am. Chem. Soc. 1990, 112, 5525.
[17] W. E. Kurtin, P.-S. Song, Tetrahedron 1968, 24, 2255.
[18] P. Cos, L. Ying, M. Calomme, J. P. Hu, K. Cimanga, B. Van Poel, L. Pieters, A. J. Vlietinck, D. Vanden
Berghe, J. Nat. Prod. 1998, 61, 71; C. Rice-Evans, N. J. Miller, G. Paganga, Free Radical Biol. Med. 1996, 20,
933; G. Cao, E. Sofic, R. L. Prior, Free Radical Biol. Med. 1997, 22, 749; N. Cotelle, J.-L. Bernier, J.-P.
Catteau, J. Pommery, J.-C. Wallet, E. M. Gaydou, Free Radical Biol. Med. 1996, 20, 35; J. A. Vinson, J. Jang,
Y. A. Dabbagh, M. M. Serry, S. Cai, J. Agric. Food Chem. 1995, 43, 2798; S. V. Jovanovic, S. Steenken, M.
Tosic, B. Marjanovic, M. G. Simic, J. Am. Chem. Soc. 1994, 116, 4846; S. V. Jovanovic, S. Steenken, Y. Hara,
M. G. Simic, J. Chem. Soc., Perkin Trans. 2 1996, 2497.
[19] L. Valgimigli, J. T. Banks, K. U. Ingold, J. Lusztyk, J. Am. Chem. Soc. 1995, 117, 9966; C. Sanchez-Moreno,
J. A. Larrauri, F. Saura-Calixto, J. Sci. Food Agric. 1998, 76, 270; T. Yoshida, K. Mori, T. Hatano, T.
Okumura, I. Uehara, K. Komagoe, Y. Fujita, T. Okuda, Chem. Pharm. Bull. 1989, 37, 1919.
[20] O. Dangles, G. Fargeix, C. Dufour, J. Chem. Soc., Perkin Trans. 2 1999, 1387.
[21] P. Poitier, V. Maccario, M.-B. Giudicelli, Y. Queneau, O. Dangles, Tetrahedron Lett. 1999, 40, 3387.
[22] W. Bors, M. Saran, Free Radical Res. Commun. 1987, 2, 289; W. Bors, W. Heller, C. Michel, M. Saran,
Methods Enzymol. 1990, 186, 343.
[23] P. Hapiot, A. Neudeck, J. Pinson, H. Fulcrand, P. Neta, C. Rolando, J. Electroanal. Chem. 1996, 405, 169.
[24] ꢁCarbon-13 NMR of Flavonoidsꢀ, Ed. P. K. Agrawal, in ꢁStudies in Organic Chemistry,ꢀ Vol. 39, Elsevier,
Amsterdam, 1989; R. Nishida, T. Ohsugi, H. Fukami, S. Nakajima, Agric. Biol. Chem. 1990, 54, 1265.
Received November 15, 1999