Flavonol-serum albumin complexation. Two-electron oxidation of flavonols and their complexes with serum albumin

Article abstract of DOI:10.1039/a810017i

Quercetin (3,3′,4′,5,7-pentahydroxyflavone) and quercetin derivatives (3-methylquercetin, isoquercitrin, rutin) are strong polyphenolic antioxidants abundant in plants and in the human diet. Recent investigations have shown that significant concentrations of albumin-bound quercetin conjugates are present in the plasma of humans fed a quercetin-rich diet. In this work, binding of quercetin and quercetin glycosides to bovine serum albumin (BSA) is quantitatively investigated by fluorescence spectroscopy. The strong fluorescence enhancement of quercetin upon binding points to the fact that a significant fraction of quercetin adopts a pyrylium-like structure in the complex. On the other hand, the observation of a very efficient quenching of tryptophan fluorescence by quercetin is consistent with a binding occurring in the IIA domain. Flavonoid-derived quinones may be formed upon quenching of reactive oxygen species by flavonoids (antioxidant activity). In this work, the quinones are conveniently formed upon periodate oxidation of the selected flavonoids in methanol and in aqueous buffers with and without BSA. A kinetic investigation by UV-visible spectroscopy shows that albumin-bound flavonoids are oxidized as quickly as free flavonoids. Interestingly, the quercetin quinone, which is merely detectable in the absence of BSA because of fast solvent addition, is efficiently stabilized in the complex by charge transfer interactions (pH 9). No evidence for quercetin-BSA conjugates could be found, thus showing that water addition (and subsequent degradation) remains the sole significant pathway of quinone transformation in the complex.

Full text of DOI:10.1039/a810017i

View Article Online / Journal Homepage / Table of Contents for this issue
Flavonol–serum albumin complexation. Two-electron oxidation of
flavonols and their complexes with serum albumin
Olivier Dangles,*a Claire Dufourb and Stephan Breta
a Université Claude Bernard-Lyon I, UPRES A - CNRS 5078, Bât. 303,
43, bld du 11 Novembre 1918, 69622 Villeurbanne, France. E-mail: dangles@univ-lyon1.fr
b Institut National de la Recherche Agronomique, Unité des Arômes et Substances Naturelles,
2, place Viala, 34060 Montpellier, France
Received (in Cambridge) 24th December 1998, Accepted 1st February 1999
Quercetin (3,3Ј,4Ј,5,7-pentahydroxyflavone) and quercetin derivatives (3-methylquercetin, isoquercitrin, rutin)
are strong polyphenolic antioxidants abundant in plants and in the human diet. Recent investigations have shown
that significant concentrations of albumin-bound quercetin conjugates are present in the plasma of humans fed
a quercetin-rich diet.
In this work, binding of quercetin and quercetin glycosides to bovine serum albumin (BSA) is quantitatively
investigated by fluorescence spectroscopy. The strong fluorescence enhancement of quercetin upon binding points to
the fact that a significant fraction of quercetin adopts a pyrylium-like structure in the complex. On the other hand,
the observation of a very efficient quenching of tryptophan fluorescence by quercetin is consistent with a binding
occurring in the IIA domain.
Flavonoid-derived quinones may be formed upon quenching of reactive oxygen species by flavonoids (antioxidant
activity). In this work, the quinones are conveniently formed upon periodate oxidation of the selected flavonoids in
methanol and in aqueous buffers with and without BSA. A kinetic investigation by UV–visible spectroscopy shows
that albumin-bound flavonoids are oxidized as quickly as free flavonoids. Interestingly, the quercetin quinone, which
is merely detectable in the absence of BSA because of fast solvent addition, is efficiently stabilized in the complex by
charge transfer interactions (pH 9). No evidence for quercetin–BSA conjugates could be found, thus showing that
water addition (and subsequent degradation) remains the sole significant pathway of quinone transformation in the
complex.
metries and binding constants). In the case of quercetin, the
possible binding site and structural changes occurring in the
flavonol nucleus upon binding are discussed.
Flavonoid quinones and quinonoid compounds could be
Introduction
Flavonoids are widely distributed plant polyphenols involved in
several biologically important mechanisms such as pigmen-
tation, nitrogen fixation and chemical defense.1 Many flavo-
formed in significant concentrations upon quenching of
noids also possess anti-cancer, anti-viral and anti-inflammatory
reactive oxygen species by flavonoids (antioxidant activity).
properties which may be the consequence of their ability to
Since quinones and quinonoid compounds may cause bio-
inhibit a broad range of enzymes and to act as powerful
polymer modifications,8 a kinetic investigation of flavonol
antioxidants.2 In particular, polyhydroxylated flavones and
oxidation by sodium periodate in the absence and in the
presence of BSA is carried out in order to outline the possible
consequences of flavonol–albumin binding on the formation
and reactivity of flavonol quinones.
flavonols (3-hydroxyflavones) display electron-rich highly con-
jugated nuclei allowing them to quickly react with toxic reactive
oxygen species (hydroxy, alkoxy and alkylperoxy radicals, sing-
let dioxygen, etc.)3 and efficiently inhibit lipid peroxidation.4
Being relatively abundant in the human diet, flavonoids could
Results and discussion
play an important role in the protective mechanisms against
cardiovascular diseases and other pathologies in which reactive
oxygen species are involved, as suggested by recent epidemi-
ological studies.5
Quercetin has been selected in this work because of its abund-
ance in plants and because structure–activity relationships con-
sistently rank it as one of the most powerful antioxidants in the
flavonoid family.2–4 Rutin, isoquercitrin and 3-methylquercetin
(Scheme 1) are also considered in order to outline the influence
Human serum albumin (HSA), the well-known carrier of
fatty acids in blood, has been shown to strongly bind quercetin
(3,3Ј,4Ј,5,7-pentahydroxyflavone) and other structurally related
flavonoids in vitro.6 In addition, significant concentrations of
quercetin conjugates were found in the plasma of rats fed a
quercetin-rich diet, a result recently confirmed in man.7 Such
conjugates(glucuronidesandsulfatesofquercetinand3Ј-methyl-
quercetin) were essentially bound to albumin and retained
significant antioxidant properties. Hence, the recognized pro-
tective effect of flavonoids on the vascular wall may be due to
their albumin complexes rather than their free forms.
In this work, binding of quercetin and quercetin glycosides to
bovine serum albumin (BSA) is quantitatively investigated by
fluorescence spectroscopy (determination of the stoichio-
Scheme 1
J. Chem. Soc., Perkin Trans. 2, 1999, 737–744
737

View Article Online
is part of the so-called IIA domain which is known to bind in its
hydrophobic cavity a variety of ligands10 including warfarin, a
coumarin structurally related to flavonoids. X-ray crystallo-
graphic data confirm that Trp 214 in HSA is strategically
located in the IIA domain to interact with bound ligands.12
The location of HSA-bound quercetin in the IIA domain was
demonstrated from competitive binding studies between quer-
cetin and warfarin.6c In this work, the observation of a very
efficient quenching of tryptophan fluorescence by quercetin is
consistent with this view.
3-Hydroxyflavone and its anion were recently shown to bind
HSA to low affinity and high affinity sites, respectively.13 The
efficient energy transfer from BSA to neutral 3-hydroxyflavone
was taken as evidence that the low affinity binding site was the
IIA domain. Alternatively, the much slower energy transfer
from BSA to the 3-hydroxyflavone anion suggested that the
high affinity binding site was the IIIA domain.
Similar analysis cannot be carried out in the case of the
quercetin–BSA complex because no spectral decomposition
between the different acid–base forms of quercetin occurs.
However, a spectroscopic titration of quercetin (5 × 10Ϫ5 mol
dmϪ3) between pH 5.0 and 8.2 in the presence of a constant
concentration of BSA (7.52 × 10Ϫ5 mol dmϪ3) yielded an
absorbance vs. pH curve that could be fitted assuming a single
proton transfer with pKa = 7.36 ( 0.03) at 25 ЊC and 0.2 mol
dmϪ3 ionic strength. Hence, at physiological pH, BSA-bound
quercetin is a mixture of two acid–base forms, probably neutral
quercetin and its monoanion. The quercetin monoanion itself
may be a mixture of tautomers formed upon deprotonation of
the most acidic 4Ј-OH and 7-OH groups.14 Since Scatchard
analysis applies to the quercetin–BSA complex with an occu-
pation number of ca. 1, it may be concluded that both neutral
quercetin and its anion bind to the same site.
The quercetin–BSA complexation can also be investigated by
fluorescence at pH 5.6 in spite of a much weaker BSA-induced
fluorescence enhancement of quercetin. Scatchard analysis of
the fluorescence titration curve yields: n = 1.09 ( 0.04), K = 81
( 14) × 103 dm3 molϪ1 at 25 ЊC. Hence, neutral quercetin and
its anions roughly display the same affinity for BSA.
It must be emphasized that binding to BSA may help investi-
gate the fluorescence properties of quercetin. Indeed, as a
5-hydroxyflavone, free quercetin fluoresces very weakly in both
aqueous and non aqueous solvents.15 For instance, we did not
observe any significant fluorescence of quercetin in aqueous
buffers, acetone or acetic acid. By contrast, quercetin displays a
very strong fluorescence emission band around 520 nm (exci-
tation at 450 nm) when dissolved in a 1 mol dmϪ3 HBr solution
in acetic acid. Similar observations have already been made
with flavones and flavonols15,16 and interpreted by the proton-
ation of O-4 in strongly acidic conditions and subsequent con-
version of the neutral flavonoids into highly fluorescent
pyrylium cations.
Binding to BSA in aqueous buffers is an alternative way to
enhance quercetin fluorescence. A comparison between both
experiments leads to the following observations:
(1) In spite of the very different conditions, the emission
spectra of quercetin in HBr–acetic acid and in a pH 7.4 phos-
phate buffer containing BSA (excitation at 450 nm) are remark-
ably similar and display a maximum in the range 520–530 nm
(Fig. 2, upper part).
(2) The excitation spectra of quercetin in HBr–acetic acid
and in the phosphate–BSA buffer (emission at 530 nm) are also
very similar and display a maximum in the range 450–460 nm.
(3) Whereas the excitation spectrum of quercetin coincides
with its absorption spectrum in HBr–acetic acid, striking differ-
ences are observed in the phosphate–BSA buffer. Indeed,
whereas BSA-bound quercetin displays an absorption maxi-
mum at 388 nm in the pH 7.4 phosphate buffer, it absorbs only
very weakly in the range 450–460 nm where the excitation
spectrum reaches its maximum (Fig. 2, lower part). Consist-
Fig. 1 Plot of the fluorescence intensity as a function of the total
quercetin concentration in a pH 7.4 buffer containing BSA (7.52 × 10Ϫ5
mol dmϪ3) (25 ЊC, ionic strength: 0.2 mol dmϪ3). Squares: emission at
530 nm (excitation wavelength: 450 nm). The solid line is the result of
the curve-fitting procedure (Scatchard analysis). Circles: emission at 340
nm (excitation wavelength: 295 nm).
of common substituents (methyl and glycosyl groups) on flavo-
noid oxidation and flavonoid–protein complexation.
Binding of quercetin and quercetin glycosides to BSA
The recent finding that significant concentrations of albumin-
bound quercetin conjugates are present in the plasma of
humans fed a quercetin-rich diet makes the quercetin–albumin
complexation a biologically relevant model for investigating
flavonoid–protein interactions. In this work, binding of querce-
tin to bovine serum albumin (BSA) is investigated by fluor-
escence spectroscopy. At pH 7.4, BSA was found to promote a
strong saturable enhancement of quercetin fluorescence emis-
sion around 530 nm (excitation at 450 nm) (Fig. 1) as reported
in the literature for HSA.6a Assuming n identical binding sites
with a microscopic binding constant K (Scatchard analysis9),
one obtains: n = 0.95 ( 0.04), K = 103 ( 16) × 103 dm3 molϪ1 at
25 ЊC. Hence, the stoichiometry of the quercetin–BSA complex
is 1:1.
In the same conditions, the BSA-induced fluorescence
enhancements of the quercetin-3-glycosides, rutin and isoquer-
citrin (at saturation of the BSA binding sites) are respectively
ca. 10 and 4 times as weak as for quercetin. Scatchard analysis
of the quercetin-3-glycoside–BSA complexation yields n = 2.66
( 0.21), K = 8.6 ( 1.1) × 103 dm3 molϪ1 in the case of rutin and
n = 3.18 ( 0.15), K = 14.5 ( 2.8) × 103 dm3 molϪ1 in the case of
isoquercitrin. Hence, the presence of a sugar moiety on the
3-position markedly weakens the flavonoid–BSA complex-
ation and results in the rather loose binding of 2–3 flavonoid
molecules to BSA.
Quercetin promotes a strong quenching of the BSA emission
at 340 nm (excitation at 295 nm) (Fig. 1). For quercetin concen-
trations ensuring complete complexation of BSA (quercetin–
BSA molar ratios higher than 2), the fluorescence emission of
BSA is virtually suppressed. This efficient quenching is accom-
panied by a weak increase in the fluorescence emission at 530
nm typical of quercetin. Since excitation at 295 nm is expected
to specifically involve tryptophan residues in proteins, this
observation bears evidence of energy transfer occurring within
the quercetin–BSA complex from an excited Trp residue to the
flavonoid chromophore.
Among the two Trp residues present in the BSA structure,
one (Trp 134) could be solvent-exposed whereas the other (Trp
212, Trp 214 in HSA) is buried inside the protein.10 However,
NMR investigations failed to detect signals characteristic of
Trp residues, thus indicating that both lack mobility.11 Trp 212
738
J. Chem. Soc., Perkin Trans. 2, 1999, 737–744

View Article Online
Table 1 Relative total energies of quercetin monoanions (A) deduced
from semi-empirical quantum mechanics calculations in vacuum
(HyperChem program, AM1 parametrization). The figures refer to the
positions of the negatively charged oxygen centers in the pyrylium
mesomeric forms (see Scheme 2)
Form
A4,7
0
A4,4Ј
ϩ5.6
A5,7
A4Ј,5
A3,7
A3,4Ј
A4Ј,7
E/kcal molϪ1
ϩ16.4 ϩ10.9 ϩ10.3 ϩ11.9 ϩ14.6
Fig. 2 Upper part: emission spectra of quercetin bound to BSA (solid
line, quercetin–BSA molar ratio = 2, pH 7.4 phosphate buffer) and in
HBr (1 mol dmϪ3)–acetic acid (broken line). Excitation wavelength: 450
nm. Lower part: excitation spectrum of BSA-bound quercetin (solid
line, emission wavelength: 530 nm, quercetin–BSA molar ratio = 2, pH
7.4 phosphate buffer), absorption spectra of quercetin bound to BSA
(dotted line, quercetin–BSA molar ratio = 2, pH 7.4 phosphate buffer)
and in HBr (1 mol dmϪ3)–acetic acid (broken line).
Scheme 2
O-4 (one pyrylium form out of 3 mesomeric forms). The charge
density on O-1 may be taken as a measure of the pyrylium
character of the central ring. Values derived from semi-
empirical quantum mechanics calculations are Ϫ0.14 for A3,7
and A3,4Ј and Ϫ0.16 for A5,7 and A4Ј,5. For comparison, values
calculated for quercetin and its pyrylium cation are Ϫ0.13 and
Ϫ0.09, respectively. We thus assume that A3,4Ј and A3,7 are the
major fluorescent tautomers of bound quercetin. Consistently,
no fluorescence enhancement could be observed upon binding
to HSA of luteolin (3Ј,4Ј,5,7-tetrahydroxyflavone), the 3-deoxy
analog of quercetin.6a
It must be noted that tautomerization may also occur in the
excited state through a fast intramolecular proton transfer from
a highly acidic OH group (e.g., OH-3 and OH-7) to the 4-keto
group as demonstrated, for instance, in the case of 3-hydroxy-
flavone and 7-hydroxyflavone.16,17 Such a phenomenon must
also take place in the case of quercetin to account for the inten-
sity of the emission band at 530 nm which is characteristic of
pyrylium tautomers. However, it cannot explain the differ-
ences observed between the absorption and excitation spectra.
Clearly, BSA allows the formation of significant, although
probably weak, concentrations of pyrylium-like tautomers in
the ground state.
ently, the fluorescence intensity at 530 nm is very weak when the
sample is excited at 388 nm.
The latter observation suggests that bound quercetin in the
ground state is actually a mixture of different forms with very
different fluorescence properties. In weakly acidic and neutral
aqueous buffers, free quercetin is essentially a mixture of non
fluorescent 4-oxo forms. Binding to BSA may favour highly
fluorescent pyrylium-like forms displaying a OH group on C-4
(4-hydroxy forms). This is consistent with the 11 nm batho-
chromic shift in the quercetin absorption band observed upon
binding. Such tautomers could form upon intramolecular
proton transfer along hydrogen bonds between OH-5 or OH-3
(donor) and O-4 (acceptor). In Table 1 are reported the calcu-
lated energies of tautomeric quercetin monoanions (Scheme 2)
including the normal 4-oxo tautomers (A4,7, A4,4Ј) and the most
probable 4-hydroxy tautomers (A3,7, A3,4Ј, A5,7, A4Ј,5, A4Ј,7).
Although much less stable (in their free form) than the normal
4-oxo tautomers, some highly fluorescent 4-hydroxy tautomers
could be specifically stabilized upon binding to BSA, for
instance, by hydrogen bonding between BSA and OH-4 (donor)
and/or the 3- or 5-oxyanions (acceptor).
The close similarity between the excitation and emission
spectra of BSA-bound quercetin with those of its free pyrylium
cation points to a strong pyrylium character in the fluorescent
tautomers of bound quercetin. Intramolecular proton transfer
from OH-3 to O-4 leads to 4-hydroxy tautomers A3,7 and A3,4Ј
which display a stronger pyrylium character (one pyrylium
form out of 2 mesomeric forms, see Scheme 2) than A5,7 and
A4Ј,5 formed upon intramolecular proton transfer from OH-5 to
Oxidation of quercetin and quercetin derivatives
Antioxidants are compounds which, at low concentrations, can
protect biomolecules (proteins, nucleic acids, polyunsaturated
lipids, sugars) from oxidative degradations.18 Flavonoids may
act in a variety of ways including direct quenching of the
reactive oxygen species and inhibition of enzymes involved
in the production of the reactive oxygen species and chelation
of low valent metal ions (Fe2ϩ, Cuϩ) able to promote radical
formation through Fenton-type reactions.2–4
Upon quenching of one-electron oxidants such as oxygen-
J. Chem. Soc., Perkin Trans. 2, 1999, 737–744
739

View Article Online
Table 2 Flavonol oxidation by NaIO4 in MeOH at 25 ЊC. Flavonol
concentration: 5 × 10Ϫ5 mol dmϪ3. NaIO4–flavonol molar ratio = 1
(quercetin) or 10 (3-O-methylquercetin). QH2 :flavonol, Q:flavonol
quinone
Flavonol
Quercetin
3-O-Methylquercetin
k1/dm3 molϪ1 sϪ1
k2/sϪ1
λmax(QH2)/nm
ε(Q) at λmax(QH2)/
dm3 molϪ1 cmϪ1
6.0 ( 0.2)a
4.4 ( 0.2) × 10Ϫ4 a
372
1.98 ( 0.02)a
b
359
5450 ( 80)
14500 ( 130)
a Determined at λmax(QH2). b No detectable solvent addition.
Fig. 3 Changes in the UV–visible spectrum of quercetin (5 × 10Ϫ5 mol
dmϪ3) after addition of sodium periodate (1 equiv) in methanol at
25 ЊC. Time interval: 0–2 h.
centered radicals, flavonoids are converted into aryloxy radi-
cals. From pulse radiolysis experiments, such radicals were
characterized by their UV–visible spectra and shown to quickly
decay through second-order kinetics (2k in the range 106–107
dm3 molϪ1 sϪ1).3 We have recently demonstrated that the
second-order decay of quercetin radicals actually leads to
quinones and quinonoid compounds which quickly form water
adducts.19 Hence, flavonoid-derived quinones and quinonoid
compounds could be produced in significant concentrations
upon flavonoid oxidation, especially in media of relatively low
water content e.g., in hydrophobic protein cavities and at the
surface of biological membranes. Since quinones and quin-
onoid compounds have well-documented oxidizing and elec-
trophilic properties which may eventually cause biopolymer
modifications through oxidation and/or covalent coupling,8
investigating the formation and reactivity of flavonoid quinones
in the presence of biopolymers may be a biologically relevant
issue.
Periodate oxidation in protic solvents is a very convenient
way to generate quinones from the parent flavonoids. Kinetic
investigations of flavonoid oxidation by sodium periodate have
been carried out in methanol and in aqueous buffers with and
without BSA.
Scheme 3
procedure, where A: absorbance at time t, A0: initial absorb-
ance, c: total quercetin concentration, ε1: molar absorption
coefficient of Q, ε2: molar absorption coefficient of QS2.
Values for k1 and k2 are reported in Table 2. The structure of
the quinone–methanol adduct is consistent with earlier reports
in the literature.21 In addition, CI mass analysis allowed detec-
tion of a quinone–methanol adduct (m/z = 333 (MHϩ)). The
quercetin quinone may be depicted as a mixture of two tauto-
meric o-quinone and p-quinonoid forms (Scheme 3).
3-Methylquercetin is oxidized by periodate 3 times as slowly
as quercetin. By contrast, the 3-methylquercetin quinone does
not undergo solvent addition and steadily accumulates as evid-
enced by the raising of a weak broad absorption in the range
420–550 nm. Such differences in reactivity can be interpreted as
follows:
(1) The 3-hydroxy-4-keto group of quercetin is a suitable site
for reaction with periodate. Indeed, even flavonols (3-hydroxy-
flavones) devoid of the catechol group quickly react with peri-
odate in alcohols. Thus, a privileged route in periodate
oxidation of quercetin may be periodate coordination with
the 3-hydroxy-4-keto group and subsequent formation of the
p-quinonoid form. With 3-methylquercetin however, the only
possible, and probably less favourable, route is periodate
coordination with the catechol nucleus and subsequent forma-
tion of the o-quinone.
Oxidation in methanol
Oxidation of quercetin with sodium periodate in methanol
results in a decrease of the UV absorption bands characteristic
of quercetin (λmax = 370, 255 nm) with a simultaneous increase
of a band with an absorption maximum at 293 nm (Fig. 3). In
addition, NMR analysis in CD3OD shows that all five aromatic
protons of the chromophore are strongly shielded upon oxid-
ation. Both UV–visible and NMR spectra point to restricted
electronic delocalization in the oxidized product in comparison
to quercetin.
The absorbance vs. time traces can be satisfactorily fitted
assuming a bimolecular oxidation of quercetin (QH2) leading
to a quinone (Q) (rate constant k1) that quickly adds two
solvent molecules (S) on positions 2 and 3 to give QS2 (rate
constant k2) (Scheme 3). Constant k1 essentially refers to the
breakdown of the quercetin–periodate adduct. As demon-
strated for catechol,20 fast kinetic techniques are required for
detecting the step of adduct formation.
The following equations (1)–(3) were used in the curve-fitting
Ϫd[QH2]/dt = k1[QH2][IO4] = Ϫd[IO4]/dt
d[Q]/dt = k1[QH2][IO4] Ϫ k2[Q]
(1)
(2)
(3)
(2) The quercetin quinone reacts with methanol through
its p-quinonoid form and adds solvent molecules on C-2
and C-3. Methylation of OH-3 prevented the o-quinone–p-
A = A0[QH2]/c ϩ ε1[Q] ϩ ε2[QS2]
740
J. Chem. Soc., Perkin Trans. 2, 1999, 737–744

View Article Online
Fig. 5 Upper part: changes in the UV–visible spectrum of quercetin
Fig. 4 Upper part: changes in the UV–visible spectrum of quercetin
(5 × 10Ϫ5 mol dmϪ3) after addition of sodium periodate (1 equiv) in a
pH 9.0 buffer (25 ЊC, ionic strength: 0.2 mol dmϪ3). Lower part: kinetic
traces at 690 (᭝), 402 (ᮀ) and 324 nm (᭺).
(5 × 10Ϫ5 mol dmϪ3) after addition of sodium periodate (1 equiv) in a
pH 9.0 buffer containing BSA (7.52 × 10Ϫ5 mol dmϪ3) (25 ЊC, ionic
strength: 0.2 mol dmϪ3). Lower part: kinetic traces at 723 (᭝), 406 (ᮀ)
and 324 nm (᭺).
quinone adduct (ca. 330 nm) which yielded the bimolecular rate
constant of quinone formation (rate constant k1). Values for k1
and k2 are reported in Table 3. At pH 9.0, periodate oxidation is
so fast that only a rough estimate of k1 can be given (1–3 × 104
dm3 molϪ1 sϪ1).
No attempt was made to isolate oxidized products because of
their well-known instability. Indeed, water addition on the
quinone is followed by opening of the pyran ring and sub-
sequent cleavage of the carbon chain.23
Periodate oxidation in the pH 9 phosphate buffer is slower
for rutin and 3-methylquercetin than for quercetin (Table 4) in
agreement with the investigation in methanol (Table 2). How-
ever, under those conditions, the rutin and 3-methylquercetin
quinones decay with first-order rate constants of the same
order of magnitude than that of the quercetin quinone.
quinonoid tautomerism leaving as the only possibility for the
3-methylquercetin quinone to react with the solvent through a
Michael addition on the C-2Ј, C-5Ј and/or C-6Ј positions of its
o-quinone nucleus. Such a reaction is probably very slow in
methanol and is not detected in the course of the experiment.
Rutin was not significantly oxidized by periodate in methanol
over a period of half an hour.
Oxidation in aqueous buffers
In the absence of BSA. Oxidation is much faster in aqueous
buffers (pH 5.5 to 9.0) than in methanol. The wavelengths of
absorption maximum for quercetin and the quinone–water
adduct gradually shift to higher values when the pH is
increased. This is consistent with deprotonation of the most
acidic phenolic OH groups. At pH 5.5, quercetin is a mixture of
neutral and monoanionic forms22 (λmax = 368 nm). At physio-
logical pH values (pH 7.4), it is essentially dianionic (λmax = 377
nm) but not significantly more prone to periodate oxidation. By
In the presence of BSA. Periodate oxidation of quercetin in
BSA containing aqueous buffers at pH 5.5 to 9.0 results in the
following changes in the UV–visible spectrum of the flavonoid
(Fig. 5):
contrast, at pH 9.0, quercetin is essentially trianionic (λmax
=
402 nm) and its periodate oxidation is much faster.
(1) A fast decrease in the absorption band characteristic of
quercetin (λmax ranging from 375 nm at pH 5.5 to 408 nm at pH
9) followed, at pH higher than 7, by a slow increase.
(2) A monotonous increase in the absorption band character-
istic of quinone–water adducts (λmax in the range 320–340 nm,
band masked by protein absorption at low pH).
(3) Above pH 7, a fast building-up of a new absorption band
with λmax ranging from 697 nm at pH 7.4 to 723 nm at pH 9. At
pH 7.4, this band remains very weak and disappears within 10
seconds whereas, at pH 9, it grows very intense and persists for
ca. 2 minutes.
The previous mechanism can be kept for fitting the
absorbance vs. time traces, i.e. a second-order step of quercetin
oxidation (rate constant k1) leading to low concentrations of
quinone which rapidly adds two solvent molecules (rate con-
stant k2). The values for k1 and k2 (Table 3) show that the kinet-
ics of both steps are weakly altered by the presence of BSA
Remarkably, a weak, rapidly decaying absorption band
around 690 nm is observed when oxidation is carried out at pH
7.4 and 9.0 (Fig. 4). This transient absorption (∆E ca. 1.8 eV)
is attributed to a quinone–quercetin charge transfer com-
plex. Charge transfer must occur primarily from the HOMO of
the donor (quercetin) to the LUMO of the acceptor (quinone).
The calculated energy differences between the HOMO of the
quercetin dianion (deprotonation of OH-4Ј and OH-7) and the
LUMO of the o-quinone and p-quinonoid 7-oxyanions are
respectively 1.64 and 1.85 eV and thus fall in the correct range
to account for a charge transfer process from the flavonoid to its
parent quinone.
The first-order decay of the charge transfer band at 690 nm
gives access to the rate constant of water addition on the quin-
one (rate constant k2). At pH 7.4 and 9.0, this value was then
used in the curve-fitting of the kinetic trace at λmax of the water–
J. Chem. Soc., Perkin Trans. 2, 1999, 737–744
741

View Article Online
Table 3 Quercetin oxidation by NaIO4 in aqueous buffers (25 ЊC, ionic strength: 0.2 mol dmϪ3). Quercetin concentration: 5 × 10Ϫ5 mol dmϪ3
NaIO4–quercetin molar ratio = 1
.
pH
5.5
7.4
9.0
e
k1/dm3 molϪ1 sϪ1
33.4 ( 0.4) × 102 b
0.060 ( 0.003)b
67.1 ( 1.5) × 102 b
0.111 ( 0.005)b
20.7 ( 0.8) × 102 c
0.63 ( 0.02)d
k2/sϪ1
0.379 ( 0.003)c
e
k1 a/dm3 molϪ1 sϪ1
k2 a/sϪ1
20.0 ( 0.8) × 102 c
0.087 ( 0.007)d
11.83 ( 0.09) × 10Ϫ3 c
a In the presence of BSA (7.52 × 10Ϫ5 mol dmϪ3). b Determined at λmax(quercetin). c Determined at λmax(quinone–water adduct). d Determined from
the first-order decay of the charge transfer band. e In the range 1–3 × 104 dm3 molϪ1 sϪ1
.
Table 4 Flavonol oxidation by NaIO4 in a pH 9.0 phosphate buffer (25 ЊC, ionic strength: 0.2 mol dmϪ3). Flavonol concentration: 5 × 10Ϫ5 mol
dmϪ3 for quercetin and 3-methylquercetin, 3.7 × 10Ϫ5 mol dmϪ3 for rutin
Flavonol
Quercetin
3-Methylquercetin
Rutin
k1/dm3 molϪ1 sϪ1
k2/sϪ1
1–3 × 104
30.4 ( 0.8) × 102
0.22 ( 0.01)
556 ( 10)
40.6 ( 0.6) × 102
0.167 ( 0.004)
86 ( 8) × 102
0.22 ( 0.01)
0.379 ( 0.003)
k1 a/dm3 molϪ1 sϪ1
k2 a/sϪ1
1–3 × 104
11.83 ( 0.09) × 10Ϫ3
0.197 ( 0.006)
a In the presence of BSA (7.52 × 10Ϫ5 mol dmϪ3).
except at pH 9.0 where BSA (7.52 × 10Ϫ5 mol dmϪ3) slows down
the addition of water on the quinone by a factor ca. 30.
Remarkably, the oxidation of bound quercetin is as fast as, if
not faster than, the oxidation of free quercetin (in the absence
of BSA). Thus, the 3Ј,4Ј-dihydroxy group and/or the 3-hydroxy-
4-keto group of bound quercetin seem to remain accessible to
oxidizing agents such as the periodate anion.
(3) The deprotonation of the quinone. Deprotonation of
the o-quinone 7-oxyanion on OH-3 and deprotonation of the
p-quinonoid 7-oxyanion on OH-3Ј both lead to two mesomeric
structures which outline the large electron delocalization in the
quinone dianion (Scheme 4). Hence, the second deprotonation
At pH 5.5 and 7.4, the quinone remains a very minor species
that does not significantly accumulate during the kinetic runs.
By contrast, at pH 9.0, the quinone reaches much higher con-
centrations as evidenced by the formation of an intense absorp-
tion band at 723 nm which is attributed to a quinone–BSA
charge transfer complex (half-life ca. 1 min). In addition, a
weaker absorption band at 504 nm also appears which may be
characteristic of the free quinone.
Scheme 4
of the quercetin quinone may be primarily responsible for the
much higher stability of the quinone at pH 9 as shown by its
much slower reaction with water (Table 3).
Size exclusion chromatography on a G100 Sephadex gel
resulted in the complete recovery of unmodified BSA. No trace
of stable flavonoid–protein adducts could be detected. This is in
contrast to previous works reporting the formation of stable
covalent adducts between BSA and the o-quinone of caffeic
acid (generated by polyphenol oxidase catalyzed oxidation of
caffeic acid).26
Periodate oxidation of rutin and 3-methylquercetin in a pH
9.0 phosphate buffer was only slightly affected by the presence
of BSA (Table 4). In particular, BSA has no influence on the
kinetics of water addition on the quinones and no quinone–
protein charge transfer complex could be detected.
The strong enhancement in charge transfer and quinone
stability when the pH is raised from 7.4 to 9.0 may be a combin-
ation of effects:
(1) The deprotonation of residues (Lys, Cys, Tyr) that would
enhance the donor character of the protein. For instance,
charge transfer interactions between the thiolate group of a
cysteine residue and the oxidized form of a flavine coenzyme
have been demonstrated in the catalytic mechanism of gluta-
thione reductase.24 If quercetin is assumed to remain in the IIA
domain upon oxidation, no free Cys residue is available for
charge transfer interaction. By contrast, in addition to Tyr 261,
up to 12 Lys residues10 (from Lys 203 to Lys 284) may lose a
proton on their side chain and act as donor centers.
(2) A pH-dependent conformational change in the protein
allowing more favourable quinone–donor molecular contacts in
the complex. Deprotonation of Lys residues is actually known
to promote a conformational change in BSA from the N form
to the B form.10 In the B form which prevails at pH 9, the helical
content is slightly lower than in the N form which prevails
around neutrality. The conformational change between the N
and B forms could place the quinone nucleus in tight contact
with Trp 212 in the IIA domain, thus favouring charge transfer
interactions. Hence, Trp 212 may act as a donor not only in the
energy transfer process occurring in the quercetin–BSA com-
plex but also in the charge transfer process occurring in the
quinone–BSA complex. Whereas the former process can take
place over large distances extending up to 6–8 nm,25 the latter
process requires significant HOMO (donor)–LUMO (acceptor)
overlapping, typically occurring for donor–acceptor distances
lower than 0.8 nm. Hence, the strong quinone–BSA charge
transfer band observed at pH 9 may be evidence of a close
contact between the quinone nucleus and Trp 212.
Substitution of OH-3 in flavonols has the following
consequences:
(1) It may markedly weaken the binding to BSA as demon-
strated in this work in the case of the quercetin-3-glycosides
rutin and isoquercitrin. Because of the large structural similar-
ity between 3-substituted flavonols and their quinones, loose
binding to BSA may be postulated for the quinones themselves
in a way that does not allow suitable donor–acceptor molecular
contacts in the complexes.
(2) It holds the corresponding quinones in an o-quinone
structure that could be much less prone to charge transfer inter-
actions than the p-quinonoid nucleus of the quercetin quinone.
Conclusion
The large changes induced by the quercetin–BSA complexation
in the fluorescence properties of both partners are consistent
with a 1:1 binding occurring in the IIA domain and allowing
742
J. Chem. Soc., Perkin Trans. 2, 1999, 737–744

View Article Online
the formation of significant concentrations of highly fluor-
escent pyrylium-like quercetin tautomers. Quercetin-3-glyco-
sides display lower affinities for BSA than quercetin.
water. After a few minutes, a part of the reaction mixture (5
cm3) was laid on the top of a G100 Sephadex columm (mass of
gel: 10 g) conditioned with a pH 5.5 phosphate buffer. After
elution with the same buffer, the samples containing unmodi-
fied BSA (based on its UV spectrum) were gathered. The
amount of BSA was determined spectroscopically and found to
be the same (within experimental error) as the initial amount.
Periodate oxidation in pH 5.5–9.0 aqueous buffers is as fast
for the flavonol–BSA complexes as for the free flavonols and
leads to quinones which quickly add water molecules and are
then degraded. At pH 9, water addition on the quinone is much
slower and large concentrations of quinone in charge transfer
interaction with BSA can be detected in the first step of the
reaction. However, even under such conditions, no evidence of
BSA–quinone covalent adducts could be found, thus showing
that water addition remains the sole significant pathway of
quinone deactivation. This observation suggests that flavonoid
quinones that may form upon quenching of reactive oxygen
species by flavonoids (antioxidant activity) are innocuous,
rapidly degraded compounds rather than potentially damaging
electrophiles or oxidizing agents.
Methanol–quercetin quinone adduct
A 10Ϫ2 mol dmϪ3 solution of quercetin in CD3OD was added
with an equivalent amount of sodium periodate. After stirring
1
for a few hours, the H-NMR spectrum was recorded (300
MHz, 27 ЊC): δ (ppm) = 7.14 (d, J = 2.2 Hz, H-2Ј), 7.02 (dd,
J = 8.5, 2.2 Hz, H-6Ј), 6.79 (d, J = 8.5 Hz, H-5Ј), 5.99 (d, J = 1.8
Hz, H-8), 5.95 (d, J = 1.8 Hz, H-6). Mass (CI, sample in
CH3OH): m/z = 333 (MHϩ, mono-adduct).
Data analysis
Experimental
The curve-fittings of absorbance vs. time plots were carried out
on a Pentium 120 PC using the Scientist program (MicroMath,
Salt Lake City, Utah, USA). Beer’s law and sets of differential
kinetic equations with initial conditions on concentrations were
input in the model. Curve-fittings were achieved through least
square regression and yielded optimized values for the para-
meters (kinetic rate constants, molar absorption coefficients).
Standard deviations are reported.
Materials
Quercetin, rutin and BSA (fraction V, MW = 66500 g molϪ1)
were from Sigma-Aldrich. Isoquercitrin was from Extra-
synthese (Genay, France). 3-Methylquercetin was a gift from
Professors B. Voirin and M. Jay (Université Claude Bernard-
Lyon I). The following buffers were prepared: 0.2 mol dmϪ3
acetate buffer (pH 5.5), 0.02 mol dmϪ3 Na2HPO4 (pH 7.4), 0.05
mol dmϪ3 Na2HPO4–0.05 mol dmϪ3 NH4Cl buffer (pH 9.0). In
all buffers, the ionic strength is 0.2 mol dmϪ3 (adjusted by NaCl
addition when needed).
Calculations
Semi-empirical quantum mechanics calculations were run at
zero kelvin in vacuum on a Pentium 90 PC using the Hyper-
Chem program (Autodesk, Sausalito, California, USA) with
the AM1 parametrization.
Absorption spectra
Spectra were recorded on a Hewlett-Packard 8453 diode-array
spectrometer equipped with a magnetically stirred quartz cell
(optical pathlength: 1 cm). The temperature in the cell was kept
at 25 ЊC by means of a water-thermostated bath.
Acknowledgements
We thank Professor B. Roux (Université Claude Bernard-Lyon
I) for the access to the fluorescence spectrometer and Professors
B. Voirin and M. Jay (Université Claude Bernard-Lyon I) for
the gift of a sample of 3-methylquercetin.
Fluorescence spectra
Fluorescence spectra were recorded on a BioLogic spectrometer
using a quartz cell thermostated at 25 ЊC.
References
Quercetin–BSA complexation
1 The Flavonoids, Advances in Research since 1986, ed. J. B. Harborne,
Chapman and Hall, London, 1994.
Two parts of a solution of BSA (7.25 × 10Ϫ5 mol dmϪ3) in a pH
7.4 phosphate buffer were respectively added with methanol
and a solution of quercetin in methanol (quercetin concen-
tration after dilution in the buffer = 2.5 × 10Ϫ4 mol dmϪ3). The
final methanol concentration was 4% in both solutions. Both
solutions were then mixed in different ratios in order to cover
the quercetin concentration range 0–2.5 × 10Ϫ4 mol dmϪ3. The
binding constant and number of binding sites on BSA were
calculated from the plot of the fluorescence intensity at 530
nm (excitation at 450 nm) as a function of the total quercetin
concentration according to the Scatchard method.9
2 (a) E. Middleton Jr. and C. Kandaswami, in The Flavonoids,
Advances in Research since 1986, ed. J. B. Harborne, Chapman and
Hall, London, 1994, p. 619; (b) Flavonoids in Health and Disease,
eds. C. A. Rice-Evans and L. Packer, M. Dekker, New York, 1997;
(c) P. Cos, L. Ying, M. Calomme, J. P. Hu, K. Cimanga, B. Van Poel,
L. Pieters, A. J. Vlietinck and D. Vanden Berghe, J. Nat. Prod., 1998,
61, 71; (d) J. R. S. Hoult, M. A. Moroney and M. Paya, Methods
Enzymol., 1994, 234, 443; (e) I. Morel, G. Lescoat, P. Cillard and
J. Cillard, Methods Enzymol., 1994, 234, 437; ( f ) W. F. Hodnick,
D. L. Duval and R. S. Pardini, Biochem. Pharmacol., 1994, 47,
573; (g) N. Cotelle, J.-L. Bernier, J.-P. Catteau, J. Pommery, J.-C.
Wallet and E. M. Gaydou, Free Radical Biol. Med., 1996, 20, 35;
(h) G. Cao, E. Sofic and R. L. Prior, Free Radical Biol. Med., 1997,
22, 749.
Periodate oxidation
3 (a) S. Steenken and P. Neta, J. Phys. Chem., 1982, 86, 3661;
(b) W. Bors and M. Saran, Free Radical Res. Commun., 1987, 2, 289;
(c) W. Bors, W. Heller, C. Michel and M. Saran, Methods Enzymol.,
1990, 186, 343; (d) W. Bors, C. Michel and M. Saran, Methods
Enzymol., 1994, 234, 420; (e) S. V. Jovanovic, S. Steenken, M. Tosic,
B. Marjanovic and M. G. Simic, J. Am. Chem. Soc., 1994, 116, 4846;
( f ) S. V. Jovanovic, S. Steenken, Y. Hara and M. G. Simic, J. Chem.
Soc., Perkin Trans. 2, 1996, 2497; (g) W. Bors and M. Saran, Free
Radical Res. Commun., 1987, 2, 289.
4 (a) N. Salah, N. J. Miller, G. Paganga, L. Tijburg, G. P. Bolwell and
C. Rice-Evans, Arch. Biochem. Biophys., 1995, 322, 339; (b) J. A.
Vinson, J. Jang, Y. A. Dabbagh, M. M. Serry and S. Cai, J. Agric.
Food Chem., 1995, 43, 2798; (c) J. Torel, J. Cillard and P. Cillard,
Phytochemistry, 1986, 25, 383; (d) M. Foti, M. Piattelli, M. T.
Baratta and G. Ruberto, J. Agric. Food Chem., 1996, 44, 497; (e)
In a typical experiment, 50 mm3 of a freshly prepared 2 × 10Ϫ3
mol dmϪ3 solution of flavonoid in methanol was diluted into
1.9 cm3 of buffer (with or without BSA) placed in the spec-
trometer cell and thermostated at 25 ЊC. At time zero, 50 mm3
of a 2 × 10Ϫ3 mol dmϪ3 solution of sodium periodate in water
was added.
Size exclusion chromatography
To 50 mg of BSA dissolved in 10 cm3 of a pH 9.0 phosphate
buffer were successively added 200 mm3 of a freshly prepared
2.5 × 10Ϫ3 mol dmϪ3 solution of quercetin in methanol and 100
mm3 of a 5 × 10Ϫ3 mol dmϪ3 solution of sodium periodate in
J. Chem. Soc., Perkin Trans. 2, 1999, 737–744
743

View Article Online
17 (a) M. Itoh, K. Tokumura, Y. Tanimoto, Y. Okada, H. Takeuchi,
K. Ioku, T. Tsushida, Y. Takei, N. Nakatani and J. Terao, Biochim.
Biophys. Acta, 1995, 1234, 99.
K. Obi and I. Tanaka, J. Am. Chem. Soc., 1982, 104, 4146; (b) M.
Itoh and T. Adachi, J. Am. Chem. Soc., 1984, 106, 4320.
18 (a) A. T. Diplock, in Free Radical Damage and its Control,
New Comprehensive Biochemistry, vol. 28, eds. C. A. Rice-Evans
and R. H. Burdon, Elsevier, Amsterdam, 1994, p. 113; (b) R. A.
Larson, Naturally Occurring Antioxidants, Lewis Publishers, CRC
Press LLC, Boca Raton, 1997; (c) C. A. Rice-Evans, in Free Radical
Damage and its Control, New Comprehensive Biochemistry, vol. 28,
eds. C. A. Rice-Evans and R. H. Burdon, Elsevier, Amsterdam,
1994, p. 131; (d) H. Sies, Angew. Chem., Int. Ed. Engl., 1986, 25,
1058.
19 O. Dangles, C. Dufour and G. Fargeix, J. Chem. Soc., Perkin
Trans. 2, submitted.
20 S. W. Weidman and E. T. Kaiser, J. Am. Chem. Soc., 1966, 88, 5820.
21 (a) M. Utaka and A. Takeda, J. Chem. Soc., Chem. Commun., 1985,
1824; (b) M. A. Smith, R. A. Webb and L. J. Cline, J. Org. Chem.,
1965, 30, 995.
5 (a) M. G. L. Hertog, E. J. M. Feskens, P. C. H. Hollman, M. B.
Katan and D. Kromhout, The Lancet, 1993, 342, 1007; (b) D. S.
Leake, In Phytochemistry of Fruit and Vegetables, eds. F. A. Tomas-
Barberan and R. J. Robins, Clarendon Press, Oxford, 1997, p. 287.
6 (a) C. Manach, C. Morand, O. Texier, M.-L. Favier, G. Agullo,
C. Demigné, F. Régérat and C. Rémésy, J. Nutr., 1995, 125, 1911;
(b) C. Manach, O. Texier, F. Régérat, G. Agullo, C. Demigné and
C. Rémésy, Nutr. Biochem., 1996, 7, 375; (c) D. W. Boulton, U. K.
Walle and T. Walle, J. Pharm. Pharmacol., 1998, 50, 243.
7 C. Manach, C. Morand, V. Crespy, C. Demigné, O. Texier,
F. Régérat and C. Rémésy, FEBS Lett., 1998, 426, 331.
8 (a) M. G. Peters, Angew. Chem., Int. Ed. Engl., 1989, 28, 555;
(b) Y. J. Abul-Hajj, K. Tabakovic and I. Tabakovic, J. Am. Chem.
Soc., 1995, 117, 6144.
9 K. A. Connors, Binding Constants, The Measurement of Molecular
Complex Stability, Wiley, New York, 1987.
10 T. Peters, Adv. Protein Chem., 1985, 37, 161.
11 P. J. Sadler and A. Tucker, Eur. J. Biochem., 1992, 205, 631.
12 X. M. He and D. C. Carter, Nature, 1992, 358, 209.
13 A. Sytnik and I. Litvinyuk, Proc. Natl. Acad. Sci. USA, 1996, 93,
12959.
22 G. M. Escandar and L. F. Sala, Can. J. Chem., 1991, 69, 1994.
23 (a) A. Nishinaga, T. Tojo, H. Tomita and T. Matsuura, J. Chem.
Soc., Perkin Trans. 1, 1979, 2511; (b) C. Tournaire, M. Hocquaux,
I. Beck, E. Oliveros and M.-T. Maurette, Tetrahedron, 1994, 50,
9303.
14 P. K. Agrawal and H.-J. Schneider, Tetrahedron Lett., 1983, 24, 177.
15 O. S. Wolfbeis, M. Begum and H. Geiger, Z. Naturforsch., 1984, 39b,
231.
16 (a) R. Schipfer, O. S. Wolfbeis and A. Knierzinger, J. Chem. Soc.,
Perkin Trans. 2, 1981, 1443; (b) O. S. Wolfbeis, A. Knierzinger and
R. Schipfer, J. Photochem., 1983, 21, 67.
24 V. Massey and S. Ghisla, Ann. N. Y. Acad. Sci., 1974, 227, 446.
25 B. Valeur, in Molecular Luminescence Spectroscopy, Part 3, Vol. 77,
ed. S. G. Schulman, Wiley and Sons, New York, 1993, p. 25.
26 W. S. Pierpoint, Biochem. J., 1969, 112, 619.
Paper 8/10017I
744
J. Chem. Soc., Perkin Trans. 2, 1999, 737–744