Amic Et Al-Curr Med Chem 2007 14 827-845

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/233639875
SAR and QSAR of the antioxidant activity of ﬂavonoids
Article in Current Medicinal Chemistry · March 2007

DOI: 10.2174/092986707780090954
CITATIONS READS
307 7,043
6 authors, including:
Dragan Amic Drago Beslo

University of Osijek University of Osijek
80 PUBLICATIONS 3,312 CITATIONS 35 PUBLICATIONS 1,684 CITATIONS
SEE PROFILE SEE PROFILE
Vesna Rastija Bono Lucic

University of Osijek Ruđer Bošković Institute
48 PUBLICATIONS 1,305 CITATIONS 91 PUBLICATIONS 2,655 CITATIONS
SEE PROFILE SEE PROFILE
Some of the authors of this publication are also working on these related projects:
dbeslo@pfos.hr View project
CroArtScia2011 View project
All content following this page was uploaded by Vesna Rastija on 15 May 2014.
The user has requested enhancement of the downloaded file.

Current Medicinal Chemistry, 2007, 14, 827-845 827
SAR and QSAR of the Antioxidant Activity of Flavonoids

Dragan Ami*,1, Duanka Davidovi-Ami1, Drago Belo1, Vesna Rastija1, Bono Lui2 and
Nenad Trinajsti*,2
1
Faculty of Agriculture, The Josip Juraj Strossmayer University, P.O. Box 719, HR-31107 Osijek, Croatia
2
The Rugjer Bokovi Institute, P.O. Box 180, HR-10002 Zagreb, Croatia
Abstract: Flavonoids are a group of naturally occurring phytochemicals abundantly present in fruits, vegetables, and bev-
erages such as wine and tea. In the past two decades, flavonoids have gained enormous interest because of their beneficial
health effects such as anti-inflammatory, cardio-protective and anticancer activities. These findings have contributed to
the dramatic increase in the consumption and use of dietary supplements containing high concentrations of plant flavon-
oids. The pharmacological effect of flavonoids is mainly due to their antioxidant activity and their inhibition of certain en-
zymes. In spite of abundant data, structural requirements and mechanisms underlying these effects have not been fully un-
derstood. This review presents the current knowledge about structure-activity relationships (SARs) and quantitative struc-
ture-activity relationships (QSARs) of the antioxidant activity of flavonoids. SAR and QSAR can provide useful tools for
revealing the nature of flavonoid antioxidant action. They may also help in the design of new and efficient flavonoids,
which could be used as potential therapeutic agents.
Keywords: SAR, QSAR, flavonoids, antioxidant activity, free radical scavenging, metal chelation, enzyme inhibition, prooxi-
dant activity, mechanisms, modelling, molecular descriptors.
INTRODUCTION and more water soluble, permitting its storage in the cell
vacuole. Structural diversity in each flavonoid family arises
Flavonoids are a group of naturally occurring polypheno-
from the various hydroxylation, methoxylation, and glycosy-
lic compounds ubiquitously found in the plant kingdom [1,
lation patterns of ring substitution.
2]. They are widespread in vegetables, fruits, flowers, seeds,
and grains [3]. The exact role of these secondary metabolites Health Effects of Flavonoids
is still unclear, but it is known that flavonoids are important
Research in the field of flavonoids has increased since
for the survival of a plant in its environment; they regulate
the discovery of the French paradox [10]. The French have
plant growth, inhibit or kill many bacterial strains, inhibit
one of the lowest incidences of coronary heart disease de-
major viral enzymes, and destroy some pathogenic protozo-
spite their high consumption of saturated fat and smoking
ans [4]. Also, they act in plants as visual attractors, feeding
habits similar to those of other countries. This can be ex-
repellents, photoreceptors, and as protection against UV
radiation [5-7]. Thus far, approximately 9000 different fla- plained by their moderate and regular consumption of red
wine. The wide range of health effects of flavonoids has
vonoids have been identified and they form the largest group
attracted great attention in recent years, especially because of
of naturally occurring polyphenols [8]. This list is constantly
the broad prevalence of these compounds in many common
growing owing to the enormous structural diversity associ-
fruits and vegetables [11]. Flavonoids are important con-
ated with these compounds.
stituents of the human diet, and their daily intake, depending
Chemical Structure of Flavonoids on diet, can range from several hundred mg up to 1-2 g [4,
12]. Besides their relevance in plants, it has been shown that
Basic structure of flavonoid has a flavan nucleus consist-
flavonoids are pharmacologically active in humans [4]. Con-
ing of two benzene rings combined by an oxygen-containing
sumption of fruits, vegetables and certain beverages, such as
pyran ring [9]. The various classes of flavonoids differ in
tea and red wine, is associated with many health promoting
their level of oxidation of the C ring of the basic 4-oxo-
effects [13-15]. Flavonoids have recently been identified as a
flavonoid (2-phenyl-benzo--pyrone) nucleus. Common sub-
major cancer-preventive component of our diet [16]. Re-
families of flavonoids are flavones, flavanes, flavonols, duced risk of coronary heart disease and premature aging is
flavanols (catechins), anthocyanidins, and isoflavones (Fig.
also associated with a regular diet rich in flavonoids [17]. In
1).
addition, flavonoids exhibit a wide range of biological activi-
Flavonoids usually occur as glycosides in plants because ties, including anti-inflammatory, antiviral, antibacterial,
the effect of glycosylation renders the flavonoid less reactive antiulcer, antiosteoporotic, antiallergic, and antihepatotoxic
actions [18, 19]. These activities are mainly attributed to
their powerful antioxidant activity [20, 21] and/or modula-
*Address correspondence to this author (DA) at the Faculty of Agriculture, tion of enzymatic activities [22, 23]. These effects, along
The Josip Juraj Strossmayer University, P.O. Box 719, HR-31107 Osijek, with epidemiological studies and animal models, have led to
Croatia; Tel: ++385 31 224 200; Fax: ++385 31 207 017; E-mail:
damic@pfos.hr and to this author (NT) at the Rugjer Bokovi Institute, the hypothesis that dietary flavonoids may be potential can-
P.O. Box 180, HR-10002 Zagreb, Croatia; Tel: ++385 1 468 00 95; Fax: didates for use as drugs in illnesses such as cancer, athero-
++385 1 468 02 45; E-mail: trina@irb.hr sclerosis, cardiovascular and coronary heart diseases, diabe-
0929-8673/07 $50.00+.00 © 2007 Bentham Science Publishers Ltd.

828 Current Medicinal Chemistry, 2007, Vol. 14, No. 7 Ami et al.
flavan nucleus 4-oxo-flavonoid nucleus

3' 2-phenyl-benzo-γ-pyrone
basic flavonoid structure 4' O
2'
8 B
O 2 5'
7
A 6'
C
6 3 OH OH
O
5 4 OH OCH3
HO O HO O
OH Hesperetin (flavanone)
OH OH O
(+)-Catechin (catechin, flavanol)
OH OH
HO O HO O
OH
Apigenin (flavone)
OH O OH O
Taxifolin (flavanonol)
OH OCH3
OH OH
HO O HO O
OCH3
+
OH OH
OH O OH
Quercetin (flavonol) Malvidin (anthocyanidin)
OCH3
OH
HO O
HO O
OCH3
+
HO OH
O OH O
O OH OH
OH
Genistein (isoflavone)
Malvidin 3-glucoside (anthocyanin) OH
Fig. (1). Structural formulae of the main sub-classes of flavonoids.
tes, AIDS, heart ailments, ulcer formation, bacterial infec- cological properties. The activity of flavonoids is closely
tions, mutagenesis, neurodegenerative diseases such as Park- linked to their structure. They are not equally physiologically
inson’s and Alzheimer’s diseases, and arthritis, as well as active, presumably because of the presence of different sub-
premature body aging [24]. Consumers and food manufac- stitutions on the carbon atoms of the basic flavonoid struc-
tures have become increasingly interested in flavonoids for ture and differences in lipid solubility. Despite the fact that
their potential beneficial role in prevention of the above- flavonoids generally occur as glycosides, their bioactivity is
listed diseases. For example, flavonoids are major functional attributed to aglycon structures, rather than to sugar moieties.
components of propolis and honey, which have been used Differentiation of flavonoids is not easy because thousands
since ancient times. Today, hundreds of herbal supplements of them share a common phenyl-benzo--pyrone skeleton,
containing flavonoids are available on the market. However, and they differ from one another only in the position and
the safety of these products is questionable [25]. Exposure to number of hydroxyl and/or methoxyl groups, as well as the
increased levels of flavonoids can adversely affect human position and number of different saccharides involved in
health due to prooxidant and promutagen activities of these glycosylation. Acylation may often occur at various posi-
compounds [26]. Therefore, careful evaluation of the bio- tions of the flavonoid nucleus as well as of the glycosyl
logical activity of flavonoids seems to be important for a residues. These diverse substitution patterns make the fla-
proper determination of their safety. vonoids an ideal object of QSAR studies [27-29].
Flavonoids and QSAR Despite the enormous interest in flavonoids and other
polyphenolic compounds as potential protective agents
The multiple activities of flavonoids as well as their against the development of human diseases, real contribu-
structural diversity make this class of compounds a rich tions of such compounds to health maintenance and mecha-
source for modelling lead compounds with targeted pharma-
SAR and QSAR of the Antioxidant Activity of Flavonoids Current Medicinal Chemistry, 2007, Vol. 14, No. 7 829
nisms through which they act are still unclear [30]. QSARs nolics, such as flavonoids, have antioxidant capacities that
represent an attempt to correlate physicochemical or struc- are much stronger than those of vitamins C and E [47].
tural descriptors of a set of structurally related compounds
with their biological (pharmacological, toxicological or eco- Structural Criteria for the Antioxidant Action of Flavon-
oids
logical) activities or physical properties (quantitative struc-
ture-property relationship, QSPR) [31, 32]. Molecular de- Intensity of the antioxidant activity of a flavonoid
scriptors usually include parameters accounting for elec- strongly depends on its chemical structure. There is a great
tronic properties, hydrophobicity, topology, and steric ef- deal of discussion and contradiction regarding the structure-
fects. Activities include chemical measurements and biologi- antioxidant activity relationships of flavonoids [48, 49].
cal assays. A crucial factor in advancing QSAR is to find However, it is well-accepted that the antioxidant activity of
information-rich descriptors for a molecule or a fragment. flavonoids is markedly influenced by the number and posi-
Once developed, QSARs provide predictive models of bio- tion of hydroxyl groups on the B and A rings, and by the
logical activity and may shed light on the mechanism of extent of conjugation between the B and C rings [50-59].
action.
On the basis of many previous and recent findings [21,
A number of studies have been conducted to elucidate the 34, 60-66], it seems that favourable general structural re-
structural requirements of flavonoids for their biological quirements for effective radical scavenging and/or the anti-
activities in order to predict the potency of these compounds oxidative potential of flavonoids follow the famous three
with regard to the targeted activity and to direct the synthesis Bors’ criteria [33]:
of more potent analogues [33-39]. Both dietary and synthetic
a) The o-dihydroxy (3’,4’-diOH, i.e., catechol) structure
flavonoids could be subjected to clinical trials in order to
in the B ring, which confers high stability to the fla-
evaluate their activities. The flavonoid QSAR enables pre-
vonoid phenoxyl radicals via hydrogen bonding or by
diction of the activities of many other untested flavonoids,
and directs the synthesis of flavonoid compounds with expanded electron delocalization;
higher potency for potential clinical application. b) The C2-C3 double bond (in conjugation with the 4-
oxo group), which determines the coplanarity of the
FLAVONOIDS AS ANTIOXIDANTS heteroring and participates in radical stabilization via
The best-described activity of flavonoids is their capacity electron delocalization over all three ring systems;
to act as antioxidants [21, 40-46]. Flavonoids may exert c) The presence of both 3-OH and 5-OH groups for the
antioxidative effects as free radical scavengers, hydrogen- maximal radical scavenging capacity and the strong-
donating compounds, singlet oxygen quenchers, and metal est radical absorption.
ion chelators, properties attributed to the phenolic hydroxyl
groups attached to ring structures [40]. Free radicals are Moreover, an additional criterion could be added:
constantly generated in our body for specific metabolic pur- d) In the absence of o-dihydroxy structure in the B ring,
poses. Examples of oxygen free radicals include singlet hydroxyl substituents in a catechol structure on the A
oxygen (1O2), superoxide (O2•), alkyl peroxyl (ROO •), ring are able to compensate and become a larger de-
alkoxyl (RO•), and hydroxyl (HO•). Among other functions, terminant of flavonoid antiradical activity [67-74].
free radicals are involved in energy production, regulation of
cell growth, and intercellular signalling. However, when an According to van Acker et al. [48], the basic flavonoid
imbalance between free radical generation and body defence structure does not seem to be essential for good antioxidant
mechanisms occurs, free radicals can attack lipids in cell activity. It becomes important only when the catechol moiety
membranes, proteins in tissues and enzymes, and DNA, to is not present. In addition, glycosylation of flavonoids
induce oxidations, which cause membrane damage, protein mostly decreases their antioxidant activity. Blocking the
modifications, and DNA damage. This oxidative damage is hydroxy group at the C-3 position or removing the 3-OH
considered to play a causative role in a series of human ill- group decreases antioxidative properties of flavonoids. Fig. 2
nesses such as cancer, heart disease, and premature body summarizes the structural criteria that modulate the antioxi-
aging. Humans possess a wide array of antioxidant physio- dant activity of flavonoids.
logical defences to scavenge free radicals, chelate metal ions Mechanisms of the Antioxidant Action of Flavonoids
involved in their formation and repair damage. Diets rich in
polyphenols contribute to these defences as well. Many phe- Mechanisms of the antioxidant action of flavonoids can
include direct scavenging of reactive free radicals, chelating
OH
3'
3'
4' OH
2' OH 2' 4'
8 B HO 8
O
B 5'
HO 7 O 2
5' 7 2
6'
A C 3
6'
A C
6 6
3
4 OH 5 4 OH
5
OH O OH O
Fig. (2). Structural features of flavonoids with high antioxidant activity.

of trace metal ions involved in free radical formation, inhibi- radical cation results from the initial radical cation B and the
tion of enzymes involved in free radical production, and proton transfer from C3-OH to the C-4 carbonyl group.
regeneration of membrane-bound antioxidants such as -
A number of flavonoids efficiently chelate trace metal
tocopherol [24, 52, 75].
ions, such as Fe2+ and Cu+ that play an important role in
The antioxidant action of flavonoids can arise from direct oxygen metabolism and free radical formation [21]. Free
scavenging of reactive oxygen species. It is generally con- iron(II) and copper(I) help the formation of reactive oxygen
sidered that the primary mechanism of the radical scaveng- species, as exemplified by the reduction of hydrogen perox-
ing activity of flavonoids is hydrogen atom donation. These ide (Fenton reaction) with generation of the highly aggres-
antioxidants may also act by single-electron transfer [76]. sive hydroxyl radical:
Structural requirements for the H-donating antioxidant activ- H2O2 + Fe2+ (Cu+) HO• + OH + Fe3+ (Cu2+)
ity include ortho-dihydroxy substitution in the B ring, C2-C3
double bond, and C-4 carbonyl group in the C ring [33, 45]. The proposed binding site for trace metal ions to flavon-
In the hydrogen atom transfer mechanism, hydroxy groups oids is the 3’,4’-diOH moiety in the B ring. In addition, C-3
donate hydrogen to a radical, stabilizing it and giving rise to and C-5 OH groups and the 4-carbonyl group also contribute
a relatively stable flavonoid phenoxyl radical (Fig. 3). The to metal ion chelation (Fig. 5).
flavonoid phenoxyl radical may react with a second radical Besides scavenging free radicals directly and chelating
(RO•), acquiring a stable quinone structure. transition metal ions by masking their prooxidant actions,
H
OH hydrogen-bond stabilized O
OH flavonoid phenoxyl radical O
HO O HO O
RO ROH
OH OH
OH O OH O
H
O O
O stable quinone structure O
HO O HO O
RO ROH
OH OH
OH O OH O
Fig. (3). Mechanism of antioxidant action of 3',4'-diOH flavonoids.
The electron donation mechanism may be valid for the flavonoids also behave as antioxidants through inhibition of
monohydroxyflavones, where hydrogen atom donation by prooxidant enzymes. This mechanism seems to be responsi-
other hydroxyl moieties is not an option. For 3-OH and/or 5- ble for their in vivo effects [77].
OH hydroxyflavones, the strong hydrogen bond of their OH Non-antioxidant mechanisms of flavonoid action, such as
moiety with the oxygen atom of the C-4 carbonyl group may modulation of signalling pathways and gene expression,
prevent not only their efficient deprotonation, but also their could also contribute to protective properties of flavonoids
radical scavenging action by means of hydrogen atom dona- [37, 45, 66].
tion. The proposed mechanism of the antioxidant action of
C3-OH or C5-OH hydroxyflavones is shown in Fig. 4. SAR AND QSAR OF THE ANTIOXIDANT ACTION
OF FLAVONOIDS
Structure A is the parent neutral molecule of 3-
hydroxyflavone, B is the initial radical cation (resulting from Number and Position of OH Groups
electron abstraction from the neutral molecule), and C is its Despite the fact that numerous authors have investigated
more stable tautomeric form. The tautomeric form C of the the antioxidant activity of flavonoids, the relationship be-
A B C
O O O
-e-
+ +
O O O
O O O
H H
H
Fig. (4). Mechanism of antioxidant action of C3-OH or C5-OH hydroxyflavones [69].
Men+ data set studied, the most effective free radical scavengers
HO
were flavonoids with the 3’,4’-dihydroxy substitution pattern
OH on the B ring and/or a hydroxyl group in the C-3 position.
The presence of a C2-C3 double bond in the C ring does not
HO O
seem to be a prerequisite for high antiradical activities, while
the presence of a 5-OH group enhances radical scavenging.
OH Dugas et al. [82] studied the influence of the number and
HO O Men+ position of OH and/or OCH3 groups on the peroxyl radical-
Men+ scavenging capacity of 7 flavonoids. The results of that SAR
study suggest that it is not the number but the position of OH
Fig. (5). Binding sites for trace elements [21]. and OCH3 groups that is essential for the antioxidant activ-
tween their structure and antioxidant potency is not quite ity.
clear [62, 78, 79]. Until recently, elucidated SARs of flavon- SAR studies of the antioxidant activity of anthocyanins
oids have only been descriptive, not explanatory by means of and their aglycons indicated that the activity increased with
QSARs [33, 48, 70]. Earlier SAR studies reported somewhat the number of hydroxyl groups on the B ring [83, 84]. Sub-
controversial statements regarding the role of the number stitution of the hydroxyl groups on the B ring with methoxyl
and position of OH groups in the antioxidant activity of groups resulted in decreasing the antioxidant activity. De-
flavonoids. For example, Chen et al. [51] emphasized that pending on the anthocyanidin, different glycosylation pat-
antioxidant activity of natural flavonoids is governed by the terns either enhanced or diminished the antioxidant power.
number and location of their aromatic hydroxyl groups. In Generally, anthocyanidins are better antioxidants than their
contrast, Foti et al. [68] found that in determining the level corresponding glycosidic forms, the anthocyanins.
of antioxidant activity of flavonoids, the number of hydroxyl
groups is of negligible importance, and so is their position - Reaction rate constants of the superoxide anion radical
either in ring A or ring B. Instead, they found that high activ- (O2•) scavenging by plant flavonoids were determined by
ity is associated with flavonoids possessing an ortho- Taubert et al. [85]. Analyzing the relations between O2•
dihydroxy system in the B ring as well as with an “unnatu- scavenging kinetics and structural features of flavonoids,
ral” ortho-hydroxylation pattern in ring A (synthetic 6,7- and some descriptive SARs were outlined. The substituents at
7,8-dihydroxyflavones). The report by Haenen et al. [80] ring B determined the superoxide scavenging kinetics,
indicated that the catechol group in ring B and the C-3 OH whereas substituents at rings A and C had little impact on
group give the highest contribution to the scavenging activity O2• scavenging rate constants. Flavonoids with the ortho-
of flavonoids. trihydroxy (pyrogallol) group were revealed as the most
rapid superoxide scavengers, followed by flavonoids with
Seven years ago, Lien et al. [62] established QSARs of the ortho-dihydroxy (catechol) group. Substitution of the
Trolox equivalent antioxidant capacities (TEACs) for 42 neighbouring OH groups at ring B by methoxyl groups
different flavonoids. They found that TEACs are mainly caused a marked decrease in O2• scavenging kinetics. Inter-
governed by the number and position of hydroxyl groups estingly, neither the existence of a C2-C3 double bond nor
(nOH) on the flavonoid ring system: the existence of OH groups at C-3 and C-5 and a keto group
TEAC = 0.454(±0.088) nOH + 0.402(±0.453) at C-4 revealed the necessary structural features for superox-
ide scavenging. It seems that pyrogallol and catechol moie-
n = 42 r2 = 0.729 s = 0.767 F1,40 = 107.33 p < 0.0005 ties are the main sites of the superoxide attack, resulting in
By adding another indicator variable (I), being the sum of the formation of flavonoid phenoxyl radicals that may be
the following indicators: presence of the 2,3-double bond (I stabilized by the mesomeric equilibrium to ortho-
= 1) or two of 3,5,7-OH groups (I = 1) or two of 3’,4’,5’-OH semiquinone structures without involvement of oxygen sub-
groups (I = 1), or absence of the above situations (I = 0), an stituents at C-3, C-4, and C-5 in charge delocalization. Fig. 6
improved equation was derived: shows a possible mesomeric equilibrium of the flavonoid
phenoxyl radical. The semiquinone structures incorporating
TEAC = 0.441(±0.079) nOH + 0.498(±0.293) I – 0.320(±0.588) the oxonium ion are also presented, which has been reported
n = 42 r2 = 0.792 s = 0.681 F2,39 = 74.01 p < 0.0005 to be the most stable mesomeric structures [86] (see also the
note in ref. 86).
In a follow-up of the Lien et al. study, Ami et al. [38]
derived a QSAR model for predicting the free radical scav- Santos and Mira [87] studied protection against the per-
enging activity (RSA) for 28 flavonoids from the data pub- oxynitrite oxidation of dihydrorhodamine by 13 flavonoids.
lished by Burda and Oleszek [81]: Correlation was observed between the number of hydroxyl
groups and the oxidation efficiency (r = 0.80). This means
RSA = 3.954(±3.556) + 75.950(±3.631) I3’,4’-diOH or 3-OH +
that flavonoids with a large number of hydroxyl groups are
8.499(±3.877) I5-OH
more effective in preventing oxidation by peroxynitrite.
n = 28 r = 0.974 s = 9.5 F = 230.7
The QSAR study of Rasulev et al. [88] emphasized the
where I3’,4’-diOH or 3-OH and I5-OH represent indicator variables. role of OH groups such as catechol moiety in the B ring and
If a particular flavonoid possesses 3’,4’-diOH or 3-OH moi- 3-OH group. The authors studied the inhibition of lipid per-
ety, then the value 1 is ascribed to the indicator variable I3’,4’- oxidation (antioxidant activity) using a set of 27 flavonoids.
diOH or 3-OH, elsewhere 0; similarly, if the flavonoid bears a 5- Numerous molecular descriptors were calculated by the
OH group, the value 1 is ascribed to I5-OH, elsewhere 0. In the DRAGON program. To this descriptor pool, a number of
OH OH
O O
+
HO O O O
O O
OH OH
OH O OH O
OH O
O O
+ +
O O O O
O O
H OH OH
OH O OH O
Fig. (6). Mesomeric equilibriums of the flavonoid phenoxyl radical [86].
quantum-chemical descriptors, as well as several indicator In the study of Di Majo et al. [90], the crocin bleaching
variables, were added. Genetic algorithm and multiple linear method was used to determine the antioxidant capacity of
regression analysis were used to select the most important nine glycosylated flavanones and the related aglycons. The
molecular descriptors and to generate QSAR models. The results from this work demonstrate that the 3’,4’-dihydroxy
best QSAR model developed is as follows: substitution, in the aglycone form, does not greatly influence
the antioxidant activity. To the contrary, in the glycosylate
ILPO = 0.561(±0.059) IOH – 0.036(±0.016) μ – 0.239 (±0.072) IGlc +
forms, the 3’,4’-catechol structure noticeably increases the
0.273 (±0.081)
antioxidant power, while O-methylation decreases the anti-
n = 27 r = 0.933 q2 = 0.821 s = 0.146 F = 51.42 oxidant activity. The kind of sugar in the C-7 position and
SPRESS = 0.171 the position of the methoxyl group (C-3’ or C-4’) perturbs
where ILPO is the antioxidant activity expressed in percent- the planarity of the flavanone phenoxyl radicals and influ-
ages of inhibition of lipid peroxidation, IOH is the indicator ences the ability to delocalize electrons.
variable denoting the presence (IOH = 1) or absence (IOH = 0) Another very recently published SAR study of flavonoids
of the 3’,4’-dihydroxy moiety in the B ring or the OH group highlighted the role of ortho-dihydroxy groups. Namely, Cai
at the C-3 position, μ is the dipole moment, and IGlc is the et al. [70] investigated the radical scavenging activity of 100
indicator variable that denotes the presence (IGlc = 1) or ab- phenolic compounds (17 phenolic acids, 41 flavonoids, 6
sence (IGlc = 0) of the O-glucose group and/or the presence of tannins, 9 stilbenes, 9 lignans and 18 quinones) isolated from
the OCH3 group at the C-3’ position in the B ring. Several traditional Chinese medicinal plants. The set of flavonoids
QSAR models using the topological descriptor PJI3 (Petijean encompassed 5 flavanols, 11 flavonols, 5 chalcones, 9 fla-
shape index) were also developed. One of the favourable vones, 5 flavanones and 6 isoflavones. The tested flavonoids
four-descriptor QSAR models is as follows: exhibited a wide variation of the radical scavenging activity.
ILPO = 0.526(±0.074) IOH + 0.287(±0.352) PJI3 – 0.029(±0.018) μ Differences in the radical scavenging activity were attributed
0.262 (±0.078) IGlc + 0.028 (±0.313) to the structural differences in hydroxylation, glycosylation
and methoxylation. The ortho-dihydroxy groups in the basic
n = 27 r = 0.935 q2 = 0.808 s = 0.147 F = 38.16 flavonoid structure were the most important structural fea-
SPRESS = 0.181 ture of high activity. Flavonoids without any hydroxyl group
The obtained QSAR models show that the presence of had no radical scavenging capacity. Besides the ortho-
hydroxyl or O-R’ groups in relevant positions, the magnitude dihydroxy groups in the B ring or in the A ring, the required
of the dipole moment, and the shape of the molecule play an structural criteria of high radical scavenging activity among
important role in the inhibition of lipid peroxidation by fla- the investigated flavonoids included the 3-hydroxy group or
vonoids. the 3-galloyl group in the C ring, and the C2-C3 double bond
in conjugation with C-4 carbonyl group in the C ring. Glyco-
In an SAR investigation of the tocopherol-regeneration sylation of the hydroxyl groups diminished the antiradical
reaction by catechins, Mukai et al. [89] showed that reaction capacity of the flavonoids.
rates increased remarkably with increasing the anionic char-
acter of catechins, that is, the electron-donating capacity of Pirker et al. [91] studied the antioxidant behaviour of
catechins. The mono anion from the catechol B and resorci- luteolin and kaempferol. Antioxidant activity under the in-
nol A rings and the dianion form from the pyrogallol B and vestigated conditions of these two flavonoids, differing only
G rings show the highest activity for the free radical scav- in the position of one OH group, was similar. However, the
enging. It has been found that catechins exert high activity in mechanisms of action were completely different. Whereas
vitamin E regeneration. the catechol moiety of luteolin stabilizes the radical anion,
the initial phenoxyl radical formed by the oxidation of
kaempferol is unstable. The authors concluded that the bio- This relationship also suggested two important character-
logical activities of kaempferol are likely to be determined istics determining the antioxidant activity, namely the ease
by the action of its oxidation products. of oxidation and the lipophilicity.
In conclusion, the number of OH groups on the flavonoid Hotta et al. [97] investigated the radical scavenging ac-
nucleus, and especially their position, implicate multiple tivity of 34 natural polyphenolic antioxidants (14 flavonoids
possible meanings. For example, the increased number of and 20 non-flavonoids) by electrochemical and spectropho-
OH groups could be related to the increased ability of H tometric measurements. The radical scavenging activity
atom abstraction or electron donating capacity and increased (EC50 – the ratio of the antioxidant concentration necessary
scavenging of free radicals. Flavonoid phenoxyl radicals to decrease the initial DPPH concentration by 50 % to the
formed by abstraction of the H atom are stabilized by hydro- initial DPPH concentration) was measured by the DPPH
gen bonding: thus, favourable position of OH groups, like method. The electrochemical parameters of antioxidants (Epa
the catechol moiety in the B ring or 3,5-diOH substitution in – the anodic peak potential, and Ipa – the anodic peak cur-
conjunction with the C-4 keto group, could be a prerequisite rent) were measured by cyclic voltammetry, and the n value
for the stability of flavonoid phenoxyl radicals. Further, (i.e., the number of electrons involved in the oxidation of a
stabilization of flavonoid phenoxyl radicals to semiquinone polyphenolic antioxidant) was determined by flow-column
structures is achieved by suitable arrangement of OH groups. electrolysis. In addition to EC50, the average stoichiometric
number (nDPPH) of DPPH in reactions with each antioxidant
Oxidation/Reduction Potential and the Number of In- was evaluated. DPPH scavenging activities were correlated
volved Electrons (n Value)
with electrochemical parameters of antioxidants. The linear
Earlier SAR reports based on experimentally measured correlation between the DPPH radical scavenging activities
oxidation/reduction potentials of flavonoids offer evidence (1/EC50) and oxidation potentials (expEpa) was poor:
that the catechol moiety in the B ring is the antioxidant ac- 1/EC50 = 6.55 expEpa + 13.8 r = 0.73
tive moiety [43, 53, 92]. The half-peak oxidation potential A certain improvement was achieved by introducing Ipa
(Ep/2) of flavonoids has been proposed as a suitable parame-
as an additional variable:
ter for evaluating the scavenging activity [48]. This assumes
that both the electrochemical oxidation FlOH FlO• + e 1/EC50 = 5.60 expEpa + 0.294 Ipa + 9.47 r = 0.86
+ H+ and the hydrogen atom donating reaction FlOH The n value of polyphenols has been generally found to
FlO• + H• involve breaking of the same OH bond [40]. increase with the electrolysis time. Moreover, for some
Yang et al. [93] estimated the antioxidant activity of 23 polyphenols, the n value may exceed the number of OH
flavonoids from their oxidation potentials. They derived the groups [98]. This suggests that some chemical reactions
QSAR equation: (e.g., dimerization), following oxidations of a polyphenol,
regenerate the oxidizable OH moieties in the oxidation prod-
IC50(μM) = 30.36 + 151.50 E(V) – 12.63 log P r = 0.852 uct. The n values determined at a lower flow rate show a
where IC50 represents the concentration for 50 % inhibition higher correlation with their DPPH scavenging activities:
of lipid peroxidation, E represents the half-wave potential 1/EC50 = 1.67 n + 0.50 r = 0.94
of the first oxidation wave measured by flow-through col- The nDPPH values determined by the DPPH method were
umn electrolysis, and log P represents the octanol/water generally very close to the n values. It seems that subsequent
partition coefficient calculated by software. The potential of chemical reactions most probably enhance the antioxidant
flavonoids was shown to be strongly dependent on their activities of the polyphenols. The authors concluded that the
structure [92, 94]. The antioxidant activity of flavonoids is n values should provide important information about the
inversely proportional to their E, i.e., the lower the E of antioxidant activity of polyphenols. These findings suggest
flavonoids, the higher is their antioxidant activity [93, 95]. that electrochemical properties of flavonoids contribute to
Lipophilicity of flavonoids (log P) is an important factor of their antioxidant activity, and thus the n values of flavonoids
their antioxidant activity in biological systems. In another can be used as descriptors of their antioxidant activities.
study, Yang et al. [96] disclosed a relationship between the
electrochemical oxidation of catechins and their antioxidant Fujisawa et al. [99] estimated the n value (number of
activity in microsomal lipid peroxidation. The following moles of peroxy radicals trapped by one mole of flavonoid)
quantitative relationship was obtained to describe the anti- using both kinetic measurements and theoretical calcula-
oxidant activity of catechins: tions. For example, PM3 calculation produced an n value of
4 for catechin (experimental value was 3.54), suggesting
log IC50(μM) = 1.56 + 2.49 E(V) – 0.29 log P r = 0.907 formation of the ortho-quinone product (Fig. 7).
Catechin (n = 4) OH O
OH O
H H
HO O HO O
- 4 H+
H - 4 e-
OH O
OH OH
Fig. (7). Catechin (n = 4) and the corresponding fully oxidized ortho-quinone product [99].
The obtained results indicate that the antioxidative larly good. This indicates that descriptors accounting for
mechanisms of flavonoids are not simple, but multivariate, other driving forces of the antioxidant activity of flavonoids
and dependent on the n value. Radical scavenging mecha- should be considered.
nisms for catechin, quercetin and hesperetin were proposed.
However, differences in reactivity towards various types of Heat of Formation of the Flavonoid Radical (Hf)
radicals may result in different experimentally determined n Possible explanations for some experimental antioxidant
values [100, 101]. activities of flavonoids could be derived from molecular
Firuzi et al. [102] evaluated antioxidant activities of 18 parameters related to electron distribution and structure, for
flavonoids by the ferric reducing antioxidant power (FRAP) example, the difference in heat of formation between the
assay. Oxidation potentials of flavonoids were determined by flavonoid and its radical, Hf. The Hf of a given radical
cyclic voltammetry. Good correlation was found between represents the heat of formation difference between the par-
FRAP values and anodic oxidation potential (r = 0.907). ent flavonoid and the appropriate radical, which results from
Hydroxyl groups and especially the catechol moiety, 3-OH, the abstraction of a hydrogen atom from an assigned OH
and the C2-C3 double bond appeared to be the most impor- group [43]. The Hf represents the relative stability of a
tant factors in determining high antioxidant activity. The possible phenoxyl radical with respect to its parent flavon-
correlation of nOH with oxidation potentials (r = 0.960) oid, and enables comparison between the alternative posi-
was slightly better than with FRAP values (r = 0.908). tions within an individual flavonoid, as well as between
different flavonoids. Therefore, the calculation of Hf for the
Nagai et al. [103] performed a kinetic study of the reaction FlOH FlO• + H•, regardless of the flavonoid sub-
quenching reaction of singlet oxygen (1O2) by 8 flavonoids. class or substitution pattern, enables the search for a favour-
The result suggests that flavonoids may contribute to the able molecule with high activity. The lower the Hf value,
protection from oxidative damage in foods and plants by the more stable the phenoxyl radical, and consequently the
quenching 1O2. The overall rate constants (kQ) for the reac- more active the antioxidant. Van Acker et al. [105] consid-
tion of 1O2 with flavonoids increase as the number of OH ered Hf as probably the best molecular descriptor for mod-
groups substituted to the flavone skeleton (i.e., the total elec- elling the antioxidant activity. Following this statement,
tron-donating capacity of flavonoids) increases. The exis- Zhang [106] calculated Hf using different semiempirical
tence of catechol or pyrogallol structure in the B ring is es- methods. The AM1 (Austin Model 1) method was found to
sential for the 1O2 quenching by flavonoids. It was found that be best suited for Hf calculation [107]. Linear correlation
log kQ correlates with peak oxidation potentials measured by was found between log k3/k1 (relative rate constants of scav-
Hotta et al. [97]. Flavonoids that have lower Epa values show enging free radicals) and Hf:
higher reactivities. For flavonoids with the C2-C3 double
bond, log kQ correlates well with Epa (r = 0.99). However, log k3/k1 = 14.6491 – 0.0955 Hf
flavonoids without the C2-C3 double bond deviate from the n = 15 r = 0.9491
correlation line. Quenching rates of 1O2 by catechins have
been studied recently [104], and a slightly lower correlation In another study, Zhang and Chen [108] elucidated activ-
ity differences of 10 flavonoid antioxidants. They found a
between log kQ and Epa was obtained (r = 0.88). Further,
linear correlation between Hf and the logarithm of relative
log kQ values of flavonoids correlate well (r = 0.91) with the
antioxidant efficiency (log RAE, r = 0.7523), and no corre-
energy of the highest occupied molecular orbital, EHOMO.
lation with EHOMO.
Flavonoids that have higher EHOMO values show higher reac-
tivity with singlet oxygen. The result is reasonable, because Vaya et al. [109] investigated the relationship of struc-
flavonoids having higher EHOMO values will show a lower tures of 20 flavonoids to in vitro inhibition of the low-
ionization potential, i.e., lower oxidation potential. Wave- density lipoprotein (LDL) oxidation. Linear correlation was
lengths of absorption maximum (max) in the UV-vis absorp- found between the calculated Hf values and the experimen-
tion spectra of studied flavonoids increased with increasing tal values of antioxidant activity. The following QSAR
the number of OH groups substituted to the flavone skeleton. model results:
Good correlation (r = 0.96) was observed between log kQ and % inhibition = 270.1 – 6.55 H f
1/max, indicating that flavonoids with higher max values
show faster 1O2 quenching rates. n = 20 r = 0.883
Butkovi et al. [72] found that logarithms of reaction rate Calculated heat of formation data (Hf) indicated that the
constants with stable free radicals correlate well with the donation of a hydrogen atom from the OH group at C-3 was
reduction potential of the flavonoids. They studied antiradi- the most likely result, followed by that of an OH from ring
cal activities of 12 flavonoids by measuring the reaction B.
kinetics and stoichiometric factors. Their results confirmed Modak et al. [110] studied structure-antioxidant activity
the stoichiometric factors of 1, 2, and 3 for flavonoids with relationships of flavonoids using Hf and spin densities.
one, two, and three hydroxyl groups in the B ring, respec- They stated that it is not possible to set forth a unique de-
tively. For the present series of flavonoids, SAR indicated scriptor for correlating the antioxidant activity. The most
the importance of multiple OH substitutions and conjugation. active flavonoids possess hydroxyl groups at C-4’ and/or C-
The results presented in this section indicate that oxida- 3’, for which the lowest Hf values were obtained. The pres-
tion potentials (Ep/2, E and Epa) and n values could be used ence of unsaturation at C2-C3 allows resonance stabilization
with some success as descriptors in constructing QSAR of formed radicals according to the analysis of spin density
models. However, even in combination with other descrip- maps.
tors, the predictive power of models generated is not particu-
Sadeghipour et al. [111] examined the antioxidant effects ther by hydrogen atom transfer, for which the calculation of
of flavonoids on the peroxynitrite oxidation reaction. The BDE is relevant, or by single-electron transfer, for which the
ability of 11 flavonoids with different OH substitutions to calculation of IP is relevant. A lower BDE value is usually
inhibit peroxynitrite-induced nitration of tyrosine was inves- attributed to a higher ability to donate a hydrogen atom from
tigated using Hf. Also, the heat of the hydrogen transfer the hydroxyl group and thereby scavenge free radicals. A
reaction from the flavonoid to the tyrosyl radical was calcu- relatively high value of IP decreases the electron-transfer
lated (Hf = (Hf(flavonoid) Hf(tyrosine)), i.e., the heat for the rate between antioxidant and oxygen, and thus reduces the
reaction of tyrosyl radical repair by flavonoids: TyO• + pro-oxidative potency of the antioxidant. In an attempt to
FlOH TyOH + FlO •. Good correlation was observed design an optimum synthetic antioxidant, e.g., for a given
between the calculated Hf and in vivo inhibition effects of biological role, Wright et al. [115] suggested that BDE and
flavonoids against tyrosine nitration. Using linear regression IP are excellent primary descriptors of the antioxidant activ-
analysis, the QSAR model for predicting flavonoids’ inhibi- ity. This was supported by the recent SAR study on rational
tory activity was made: design of phenolic and flavonoid antioxidants by Zhang et
al. [116]. The study revealed that the catechol moiety in ring
inhibition(%) = 10.108 Hf r2 = 0.9056
B of flavonoids has the advantage of a relatively low BDE
Rackova et al. [112] investigated the influence of 19 value for OH.
flavonoid structure-related parameters on the lipid peroxida-
Recent studies indicate that flavonoid derived anions are
tion inhibition of a set of 12 flavonoids. The best developed
more active than neutral molecules to scavenge free radicals
QSAR models for the antioxidant activity (pIC50) include the
[69, 117]. Martins et al. [118] found that the antioxidant
following molecular descriptors: hydration energy (EHYDR),
activity of flavonoids is comparable to the ease of deprotona-
Hf, and energy of the lowest unoccupied molecular orbital
tion i.e. to their acidity. Dissociation constants, absolute
(ELUMO):
hardness, partition coefficient and binding energy may be
pIC50 = 0.0319 EHYDR + 3.46 used as descriptors for the relationship between the acidity of
n = 12 r = 0.747 p < 0.005 s = 0.227 hydroxyl groups and the biological activity of flavonoids
[119]. Zhang and Wang [120, 121] pointed out that it is not
pIC50 = 0.035 EHYDR + 0.012 Hf + 2.99 the H atom abstraction but the proton coupled electron trans-
n = 12 r = 0.756 p < 0.022 s = 0.235 fer reaction that is responsible for the enhanced radical scav-
enging activity of the anionic form. Therefore, to select or
pIC50 = 0.033 EHYDR + 0.29 ELUMO + 3.72 rationally design novel antioxidants, the proton dissociation
n = 12 r = 0.759 p < 0.021 s = 0.234 process should be taken into consideration, especially in
polar systems [121, 122].
The highest (absolute) values of EHYDR were obtained for
the most potent flavonoids possessing the highest number of McPhail et al. [57] determined the stoichiometry and
OH groups, while the lowest (absolute) values of EHYDR were kinetics of the hydrogen-donating ability of 15 flavonoids by
attributed to flavonoids that exerted low antioxidant activity electron spin resonance spectroscopy. The second-order rate
[112] (see also the note in ref. 112). The authors assumed constants (k2) of the reduction of galvinoxyl radical by fla-
that the parameter EHYDR reflects the hydrophilic properties vonoids, governed by the BDE value for OH, are highly
of flavonoids. dependent on the configuration of OH groups on the flavon-
oid B and C rings. To have high reaction rates and high reac-
Seyoum et al. [113] performed a SAR study where they tion stoichiometries, flavonoids must be capable of being
experimentally determined the DPPH radical scavenging oxidized to ortho-quinones or extended para-quinones.
activity of 52 flavonoids and calculated the Hf values asso- Moderately high correlation (r = 0.818) was found between
ciated with the formation of various flavonoids and related log(k2) and the reaction stoichiometry. This result highlights
simplified phenolic radicals. Isolated para-dihydroxyl group the importance of considering reaction kinetics, as well as
on either A or B ring, as an active hydrogen donating fea- stoichiometry, when assessing the antioxidant capacity of
ture, was suggested. Spin density of flavonoid radicals was flavonoids.
also analyzed. The authors concluded that the ease of hydro-
gen atom abstraction and the ease of termination of the fla- Using the semiempirical quantum chemical parametric
vonoid phenoxyl radicals could be responsible for the radical method 3 (PM3), Kondo et al. [123] have calculated not only
scavenging activity of flavonoids. However, there is no phenolic OH but also all of the BDEs for CH of catechins.
QSAR model to confirm this statement. This lack of model The calculated BDEs for CH for catechins at the C-2 posi-
is in accord with the suggestion that it is hard to believe that tion were unexpectedly low compared to BDEs of CH at
only one molecular descriptor, even assigned as “the best phenolic sites, suggesting that hydrogen at the C-2 position
molecular descriptor for modelling the antioxidant activity”, may be abstracted by free radicals. The authors proposed
could generate a good predictive QSAR model. Only one tentative antioxidative mechanisms of catechins based on
molecular descriptor could not embrace the manifold nature kinetic measurements and theoretical calculations. Zhang
of antioxidant processes. and Wang [124] ascribed the unexpectedly low BDEs for
CH in catechins to the inaccuracy of the quantum chemical
Bond Dissociation Energy (BDE) of the O H Group and method used. By the Gaussian-94 program, they recalculated
Ionization Potential (IP) the results of Kondo et al. [123] and found that the BDEs for
Wright et al. [114, 115] performed density functional CH in catechins are higher than the BDEs for OH in the B
theory (DFT) calculations to discern the activity of several ring. The obtained results indicated that the C-2 hydrogen is
classes of phenolic antioxidants. These antioxidants act ei- not more abstractable than catecholic hydrogens and that the
hydrogen abstraction from C-4’ OH is favoured. However, scriptor to characterize the stability of free radicals, because
the BDEs for CH at position C-2 in catechins are compara- the energy of a free radical can be efficiently decreased if
ble to the BDEs for OH at position C-3’, implying that the unpaired electrons are highly delocalized through the conju-
C-2 hydrogen can also participate in radical scavenging. gated system. Derivatives with the ortho-hydroxy-amino
Hence, a high level ab initio calculation is essential in this group show stronger antioxidant activity than derivatives
field. with a monohydroxy or ortho-dihydroxy group. The authors
concluded that the ortho-hydroxy-amino group can be used
A study of the effect of pH on antioxidant properties
as another potential functional group to synthesize novel
(TEAC assay) of a series of 22 hydroxyflavones was per-
antioxidants.
formed by Lemanska et al. [69]. They found that the pH
dependent antioxidant activity is related to hydroxyl moiety In summary, according to the current knowledge of the
deprotonation, resulting in an increase of the antioxidant radical scavenging processes of flavonoid antioxidants, H-
potential upon formation of deprotonated forms. Comparison atom donation is the dominant mechanism, which involves
of experimental results with the calculated BDE value for two pathways: (1) H-atom transfer, and (2) electron trans-
OH and IP of nondeprotonated and deprotonated forms of fer/proton transfer.
various hydroxyflavones indicates that especially the pa-
RO• + FlOH ROH + FlO• (1)
rameter reflecting the ease of electron donation, IP, is greatly
influenced by the deprotonation. These results point to the RO• + FlOH RO + FlOH•+ ROH + FlO• (2)
conclusion that upon deprotonation, the radical scavenging If the flavonoids tend to deprotonate, then flavonoid
capacity increases because the electron donation becomes anions should be considered. H-atom transfer (1) can be
easier. Taking into account that the mechanism of radical characterized by BDE of OH groups. Electron trans-
scavenging activity of the neutral form of flavonoids is gen- fer/proton transfer (2) can be measured by IP. The lower
erally considered to be hydrogen atom donation, this implies these parameters are, the stronger is the flavonoid radical
that not only the ease of radical scavenging, but also the scavenging ability [29]. Currently, besides the Hf, BDE of
mechanism of antioxidant activity may change upon flavon- the OH group and IP are the most considered primary de-
oid deprotonation. However, the absence of a QSAR be- scriptors of antioxidant activity. Lack of QSAR models of
tween the IP and TEAC values of neutral hydroxyflavones antioxidant activity of flavonoids using only BDE and IP
indicates that electron donation is not the only mechanism by additionally reflects the complex nature of antioxidant proc-
which the neutral forms of hydroxyflavones may act as anti- esses.
oxidants.
Miscellaneous Molecular Descriptors
Russo and coworkers [125, 126] evaluated the antioxi-
dant activity of a series of polyphenolic molecules com- Heijnen et al. [129] studied the peroxynitrite scavenging
monly present in the Mediterranean diet. These authors activity of substituted phenols and several flavonoids. Good
found that the most efficient hydrogen-donor systems are QSAR models were found between the peroxynitrite scav-
characterized by the vicinal dihydroxy moiety, as confirmed enging activity of substituted phenols and the Hammett , or
by their low BDE values. Radicalization of their hydroxyl the EHOMO. However, no unambiguous QSAR has been ob-
groups gives rise to phenoxyl radicals, in which the odd tained for flavonoids. Instead, two relatively independent
electron appears to be delocalized over the whole molecule pharmacophores were identified, located on either the
and stabilized by hydrogen bonds. On the basis of the com- catechol group (3’,4’-diOH) in ring B or on three OH groups
puted BDE and IP values, taxifolin, luteolin and epicatechin (3,5,7-triOH) in the AC ring. In the AC ring, the 3-OH group
are expected to act as hydrogen donors. Luteolin, apigenin was the reactive centre and the reactivity of this group was
and kaempferol appear to be good candidates for the single- enhanced by electron donating groups at C-5 and/or C-7.
electron transfer mechanism. Their planar conformation and Theoretical calculation of the electronic properties and
the extended electron delocalization between adjacent reactivity of flavonoids offers valuable descriptors for QSAR
rings determine low IP values. analysis [130, 131]. Ghiotto et al. [132] proposed the new
In their recent SAR study, Trouillas et al. [76] investi- electronic index EH,H1, i.e., energy difference between
gated the specificity of the 3-OH group in the antioxidant HOMO and HOMO1, as the descriptor for modelling the
action of quercetin and taxifolin. The analysis of BDE val- antioxidant activity of flavonols. Linear regression analysis
ues, for all OH sites in these flavonoids, clearly shows the between TEAC and EH,H1 resulted in the following equa-
importance of the 3’,4’-diOH and 3-OH moieties only when tion:
the C2-C3 double bond is present. Distribution of spin densi- TEAC = 9.44797 – 8.82387 EH,H1
ties in radicals formed by H-removal from each OH site of
both flavonoids indicates that the 3-OH quercetin radical n=4 r = 0.972 s = 0.403
possesses a large spin density on the C-2 atom, which ex- Farkas et al. [133] developed the PLS (partial least
plains the C-ring opening process observed in flavonol deg- squares projection of latent structures) model for predicting
radation during metabolism [127]. the antioxidant activity of 36 flavonoids. They found that
Chen et al. [128] performed a theoretical examination of connectivity indices (variations of the Randi index ) and
two-dimensional topological indices play an important role
designed four ortho-hydroxy-amino derivatives of flavon-
in describing antioxidant activities. Surprisingly, the number
oids. The results revealed that the ortho-hydroxy-amino
of OH groups does not have a significant role in this PLS
group plays an important role in promoting the antioxidant
model.
properties of molecules because of its lowering effect on
BDE, IP, and spin density. Spin density is an important de-
Estrada et al. [134] developed QSAR models for predict- In another study employing a theoretical approach, Gupta
ing inhibitory activity of a dataset of 22 cinnamic acids and et al. [139] generated QSAR models for predicting the
flavonoid derivatives against peroxidation of linoleic acid. DPPH scavenging activity of 47 naturally occurring phenolic
The models are based on the Topological Sub-Structural antioxidants. The authors assumed that chemical reactivity is
Molecular Design (TOPS-MODE) approach. This approach primarily determined by the properties of bonds available in
permitted structural interpretation of the activity/property of a molecule (being related to the ability of bond breaking and
studied compounds in terms of bond contributions. Descrip- bond making). They developed molecular maps of atom-
tors accounting for hydrophobic/polarity, electronic and level properties (MOLMAPs) to represent the diversity of
steric features of the molecules were calculated and corre- chemical bonds existing in an antioxidant molecule. Coun-
lated with experimental values of the antioxidant activity terpropagation neural networks (CPG NNs) and random
expressed as log(IC50). The best model that was obtained forests were used to model the relationship between the
contained three descriptors and is given by the following MOLMAP descriptors of local bond properties and the cor-
equation: responding IC50 values. The IC50 values calculated by CPG
NNs by performing cross-validation correlated with the
log(1/IC50) = 1.863 – 0.544 ov + 2.536 4vp 85.906 6ring
experimentally observed IC50 values exhibiting a q2 of 0.712
n = 19 r = 0.9382 s = 0.820 F(3,15) = 36.7 and RMS error of 1.001. Antioxidants with the presence of
where ov is the valence connectivity index of order zero, catechol moiety contribute to high antioxidant activity.
p is the path valence connectivity index of order 4, and
4 v
Fan et al. [140] described an effective way to explore and
6
ring is the ring connectivity index of order 6. However, the visualize the SARs of a set of 31 flavonoids with antioxidant
shortcoming of that model is the absence of a clear physico- activity using structure-activity maps (SAMs). SAMs are
chemical meaning of the descriptors involved. By the virtual graphical maps plotting molecular descriptors such as the
structure generation procedure, the authors generated a set of number of non-hydrogen atoms and bonds in a molecule or
192 flavonoid compounds. Among them, 75 designed fla- the molecular similarity index against their biological activi-
vonoids, according to their calculated IC50 values, exhibit ties. SAMs can be used to identify important chemical struc-
more potent antioxidant activity than the most active flavon- tural features of flavonoids with antioxidant activity, and to
oid from the experimental dataset. Accuracy of such predic- determine the position and type of modification for improved
tions is questionable due to the very limited number of com- activity.
pounds used in the dataset (only 19 compounds were used
for model development), and because of limited accuracy of Erlejman et al. [141] studied the effects of 26 flavonoids
model (r = 0.9382). In addition, because the authors per- and related compounds on lipid oxidation, membrane fluid-
formed only fitting and gave only statistical parameters of ity, and membrane integrity. The results presented in this
fit, there is no evidence that such accuracy will be obtained work stress the importance of considering not only the
in prediction on an external dataset containing compounds of chemical structure of flavonoids per se, but also the nature of
higher antioxidant activity. The proposed model should be the interactions between these molecules and membranes
additionally tested. when estimating the flavonoid potential antioxidant capacity.
The ability of flavonoids to interact with membranes at the
Weber et al. [135] used chemometric methods (principal water-lipid interface should be regarded as another factor
component analysis (PCA), hierarchical cluster analysis contributing to the antioxidant activity of flavonoids.
(HCA), and k-nearest neighbour (KNN)) to build a model
able to find a relationship between electronic features of a set Knowledge of flavonoid interactions with membrane
of 25 flavonoid compounds and their antioxidant activity components seems to be necessary to predict the structure of
(TEAC). Quantum chemical calculations using the AM1 potential flavonoid based drugs for desired biological effects
method were employed for the evaluation of molecular de- [142]. For this reason, the lipophilicity of flavonoids is usu-
scriptors. Four electronic descriptors were related to the ally one of their most important pharmacological features,
antioxidant activity of the flavonoid compounds studied: and interactions with membranes play an essential role in
polarizability (), charge at carbon 3 (QC3), total charge at their biological activity. The C-3 position is an excellent
substituent 5 (QS5), and total charge at substituent 3’ choice for substitution to give the flavonoid an optimum
(QS3’). PCA resulted in the following equation: lipophilicity, allowing both easy application and uptake into
the membranes [53].
PC1 = 0.517 + 0.254 QC3 + 0.597 QS5 0.558 QS3’
Kanakis et al. [143] estimated the binding constants of
These variables were found to be responsible for the flavonoid-DNA adducts. Flavonoids are strong antioxidants
separation of more and less antioxidant flavonoids. The that prevent DNA damage. Low flavonoid concentration
results obtained with HCA and KNN agree with those from stabilizes the DNA duplex, whereas helix destabilization can
PCA. Terms in the above equation were explained as fol- occur when DNA is incubated for a long time with high
lows. The relevance of the atomic charge at C-3’ lies in the flavonoid content. Among the studied compounds, del-
fact that the oxidation occurs, preferably, at ring B [92, 136- phinidin with a positive charge induces a more stabilizing
138]. The substituent bonded to C-3 determines the planarity effect on a DNA duplex than quercetin and kaempferol.
of the flavonoid core and the stability of the flavonoid phe-
noxyl radical. The role of position C-5 is relevant only for As presented in this section, many descriptors accounting
compounds lacking hydroxyls on ring B. Polarizability can for electronic properties, topology, steric effects and hydro-
be related to the HOMOLUMO energy gap, since the elec- phobicity were used in the construction of QSAR models.
tronic distribution can be easily deformed if the LUMO is Among them, there are descriptors with a clear physical
close to the HOMO. meaning as well as those without a clear physical meaning. It
seems reasonable to prefer construction of QSAR models gests that flavonoid-metal complexes are more effective
using the former. antioxidants than free flavonoids.
Metal Ions Chelation The ability of flavonoids to act as chelators of transition
metal ions was examined by Teixeira et al. [153]. In this
Examination of the antioxidant power of flavonoids is SAR study, acidity constants of catechin and taxifolin, and
usually carried out by determining their profile as chain-
formation constants of the corresponding copper(II) com-
breaking antioxidants, by evaluating their direct free radical-
plexes were investigated by potentiometry and/or spectro-
scavenging activity as hydrogen- or electron-donating com-
photometry. In addition, partition coefficient values (log P)
pounds [41]. However, this may not be the only mechanism
were determined. The authors pointed out that for successful
underlying the antioxidant activity of flavonoids. Another
QSAR determinations different molecular descriptors of
antioxidant mechanism may result from the ability of fla- flavonoids would be of interest: not only those calculated by
vonoids to chelate metal ions, rendering them inactive to
theoretical methods but also those based on experimental
participate in free radical generating reactions [21, 67, 95,
physicochemical data.
144]. An earlier attempt to establish the SAR of iron(II)
chelation by flavonoids [53] demonstrated that 3-OH in the Russo and coworkers [154,155] emphasized the role of
C ring and catechol moiety (3’,4’-diOH) in the B ring are anionic forms of flavonoids in metal chelation. Loss of a
more important for chelation than 5-OH. Catechol moiety in proton in the flavonoid molecule is crucial for its antioxidant
the B ring is shown to be important for copper(II) chelation activity, because the chelation often occurs through at least
[145]. Khokhar and Owusu Apenten [146] also emphasized one deprotonated ligand.
the role of vicinal OH groups (3’,4’ or 7,8 dihydroxy groups)
Inhibition of Prooxidant Enzymes
in iron binding, as well as the presence of C-5 and/or C-3
OH in conjunction with the C-4 keto group. Prooxidant enzymes, such as xanthine oxidase, lipoxy-
Mira and coworkers [147, 148] studied the interactions of genase, protein kinase C, cyclooxygenase, microsomal
monooxygenase, mitochondrial succinoxidase, and NADPH
flavonoids with iron and copper ions. Complexes with the
oxidase, are responsible for reactive oxygen species genera-
range of stoichiometries, of metalflavonoid, 1:1, 1:2, 2:2,
tion [4, 21, 22]. A better understanding of structural charac-
2:3 were observed. The 1:2 stoichiometry is, in general, the
teristics of flavonoids that promote prooxidant enzyme inhi-
preferred one. For flavones, the binding metal sites are pref-
bition may guide the development of flavonoid compounds
erably at 5-hydroxy and 4-oxo groups. Additionally, the
ortho-catechol group is also a chelating site. The SAR of as potential therapeutic agents. Here, we are focusing on a
survey of papers dealing with SAR and QSAR of xanthine
flavonoids related to reduction of iron(III) and copper(II)
oxidase and lipoxigenases inhibition by flavonoids.
ions was also investigated. Moridani et al. [149] showed that
the initial step in superoxide radical scavenging activity Xanthine Oxidase
involves a redox-active flavonoid-Fe3+ complex. Melidou et
The enzyme xanthine oxidase (XO) catalyzes the oxida-
al. [150] performed a SAR study on the protection against
nuclear DNA damage offered by flavonoids in cells exposed tion of hypoxanthine and xanthine to uric acid. Since accu-
mulation of excess uric acid in the body results in the painful
to hydrogen peroxide. The results presented in this work
disease gout (caused by crystallization of uric acid in the
support the notion that iron chelation is responsible for the
joints), there has been considerable interest in designing XO
DNA protection.
inhibitors. As reviewed by Cotelle [156], several authors
Engelmann et al. [151] determined the stability constants have investigated the influence of the substituent nature and
of ferric complexes with three differently substituted flavon- position on inhibition of XO by flavonoids. Hayashi et al.
oids. Hydroxy substitution pattern of three flavonoids was [157] measured the inhibitory activities of 103 flavonoids
chosen, each of which possesses one of the three most com- against XO. Their results indicate that the planar flavone
monly suggested sites for metal binding by flavonoids (see core (double bond between C-2 and C-3) and the presence of
Fig. 5). The stoichiometry for the Fe(III)-flavonoid complex free hydroxyl groups at C-5 and C-7 are important prerequi-
formation was 1:1 for all three flavonoids examined, and the sites for the activity.
preferred binding site was the 3’,4’-dihydroxy moiety.
According to Costantino et al. [158] and Cotelle et al.
The cyclic voltammetric data of de Souza and De Gio- [159], the OH group at position C-7 is fundamental for activ-
vani [152] show a considerable decrease of the oxidation ity. Flavones with both 7-OH substitution and a catechol or
potential of the Al(III) and Zn(II) complexes with flavon- pyrogallol moiety on the B ring are the most effective inhibi-
oids, compared to that of free flavonoids. This finding sug- tors of XO (Fig. 8A).
OH OH
A OH B OH
B B
HO O O
A C A C
OH OH
OH O O
Fig. (8). Structural requirements for: A) inhibition of xanthine oxidase [157,159]; B) lipoxygenase inhibition [173].
Anthocyanidin derivatives are less active than the corre- groups) and QA descriptor (the sum of charge densities at the
sponding flavones [160]. Rastelli et al. [161] proposed a A ring carbons for anhydrobase form A7). Using these de-
model of interaction of flavonoids with XO. Hydroxy fla- scriptors, the dissociation constant of the enzyme-inhibitor
vones are active in their dissociated form. Anionic oxygen at complex (KEI, in μM) was modelled. The models presented
C-7 has a very strong carbonyl character because of extended were developed using only 16 flavylium salts and contained
delocalization of the negative charge on the entire benzopy- three or four descriptors, and each model had a correlation
rone ring. It enables flavones to enter hydrophobic interac- coefficient with KEI higher than 0.98. However, after recon-
tions in the optimal region of the enzyme. sidering the modelling process, we concluded that, due to its
high experimental KEI value compared to experimental val-
Rastelli et al. [162] reported the dissociation constants
ues of other compounds in the dataset, compound no. 11 in
(pKEI) of enzyme-inhibitor complexes for 18 polyhydroxy-
lated and polymethoxylated flavone inhibitors of XO. Using Table 1 should be omitted [164]. After that, the best 4-
descriptor model contains 5OH, 7OH, I, and Q A as descrip-
binary (0/1) variables for possible hydroxyl (Hn) and
tors, having r = 0.98, rcv = 0.94 (correlation coefficient ob-
methoxyl (Mn) substituents in their n positions on the fla-
tained by performing leave-one-out cross-validation). How-
vone skeleton, the multiple linear regression analysis re-
ever, a model developed using only 15 compounds should
sulted in the following QSAR model:
not have more than two descriptors. Details of the best 2-
pKEI = 2.69(±0.30) + 1.88(±0.23) H7 + 1.22(±0.26) M4’ + descriptor model and its statistical parameters are:
1.06(±0.20) H5 + 0.71(±0.25) H4’ + 0.55(±0.18) H3’
KEI = 2986.0(±510.0) 39.8(±5.4) I + 495.9(±84.2) QA
n = 18 r2 = 0.93 s = 0.328 F = 30.46
n = 15 r = 0.91 rcv = 0.84 s = 7.65 F = 27.82
On the basis of additional molecular orbital calculations,
the authors suggest that the most active form in solution is Nagao et al. [165] evaluated the inhibitory effect of 25
flavonoids on XO. The obtained SAR revealed that the pla-
the C-7 anionic flavone form. Anions at the C-4’ or C-3’
nar flavones and flavonols with the C-7 OH group inhibited
hydroxyls are not suitable for interactions with XO.
XO activity at low concentrations, while the nonplanar fla-
The same group of investigators [158] synthesized and vonoids were less inhibitory. This is consistent with the
tested the XO inhibitory activity of seventeen 7- presented results of other authors. However, contrary to the
hydroxyflavones carrying different substituents at the C-4’ findings of Cotelle et al. [159], it appears that the catechol
position. Anionic form generated by the C-7 OH group rep- moiety of the B ring, which gives antioxidative potential to
resents the main species responsible for activity. An excel- flavonoids, was not related to XO inhibition. This inconsis-
lent correlation between molecular refractivity (MR) and tency was probably caused by the use of different assays.
log IC50 is found:
Ponce et al. [166] developed a SAR topological model
log IC50 = 0.07 MR + 4.51 for predicting the activity of 22 flavonoids as XO inhibitors.
n = 17 r = 0.96 F = 187.3 The inhibiting activity of XO by flavonoids is determined, to
a large extent, by their structural properties. The authors
MR is a descriptor frequently interpreted as reflecting found that the J2 charge index was the best one for predicting
drug-receptor dispersion interactions: it describes steric and inhibitory activity, showing the following statistics: r =
electronic properties of a molecule. The authors suggest that 0.8495, s = 0.617, F = 51.6. The developed model was able
C-4’ substituents are probably involved in dispersion interac- to classify the studied flavonoids into four groups according
tions with the enzyme. to their activity on XO (inactive, low, significant, or high).
In a study of Cos et al. [163], SARs of flavonoids, as Lin et al. [167] assessed the inhibition of XO by six fla-
inhibitors of XO and as scavengers of the superoxide radical, vonoids. The obtained results emphasized the presence of the
produced by the action of this enzyme, were investigated. C2-C3 double bond, OH groups at C-5 and C-7 and the car-
The obtained results support the findings of a study of Haya- bonyl group at C-4 as prerequisites for XO inhibition. The
shi et al. [157]. Namely, these authors suggested that the authors presented a 3D molecular model of flavonoids bind-
planar flavonoid structure is important for XO inhibition. ing to the active site of XO.
The C2-C3 double bond makes the B ring coplanar with
rings A and C due to conjugation. Hydroxyl groups at C-5 Van Hoorn et al. [168] described a method for predicting
and C-7 also contribute to the inhibitory activity, while the the IC50 values of flavones based upon an individual contri-
presence of the hydroxyl group at C-3 slightly decreases the bution factor dependent on the location of a hydroxyl moiety
inhibitory activity. For the superoxide scavenging activity, in the flavone skeleton. The strongest contribution towards
the hydroxyl group at C-3’ and at C-3 were essential. Fla- XO inhibition results from introduction of C-5 or C-7 OH
vonoids with both XO inhibitory activity and additional groups to a planar flavone core.
superoxide scavenging activity could be of advantage as Using molecular mechanics (MM+) and AM1 methods,
potentially applicable compounds against gout. the influence of electronic, steric and hydrophobic properties
Ami et al. [164] developed simple QSAR models of of seven flavonoid compounds on the inhibition of XO was
inhibitory activity of flavylium salts on XO. The data set of studied by da Silva et al. [169]. The obtained results showed
descriptors contained Hansch’s hydrophobicity parameter that geometric properties of flavonoids are the most impor-
( values for substituents in positions C-7 and C-5: 7O- and tant ones for XO inhibition. When the torsion angle between
5OH with preferably ionized C-7 OH group; 5O- and 7OH C and B rings is larger than 27o, the flavonoid compound is
with preferably ionized C-5 OH group), indicator variable I not able to inhibit the enzyme. Small size of the flavonoid
(value 1 for the presence, and 0 for the absence of 5,7-diOH
and high hydrophobicity (log P value) enhance the inhibitory tipotent antioxidants. Multipotent antioxidants are of great
activity. interest for the treatment of diseases in which multiple pa-
toghenetic factors are implicated, such as free radicals, metal
A recent SAR study of XO inhibition by 26 flavonoids
ions and prooxidant enzymes [177].
isolated from different plant species [170] is consistent with
the statements reported by Hayashi et al. [157]. Prooxidant Activity of Flavonoids
Lipoxygenases Although the ability of flavonoids to act as antioxidants
Lipoxygenases (LOXs) are prooxidant enzymes that has been demonstrated, there is an increasing evidence of
catalyze enzymatic lipid peroxidation. Additionally, they are their prooxidant activity [42, 156, 178-182]. The presence of
a source of free radicals initiating nonenzymatic lipid per- a different number of OH groups in the B ring of flavonols
oxidation and other oxidative processes involved in the for- may contribute to their antioxidant activity, as well as to
mation of atherosclerotic lesions [171-173]. Consequently, their toxicity, and may play an important role in their po-
inhibition of LOXs may contribute to the universal antioxi- tency for biological action such as angiogenesis and im-
dant activities of flavonoids. mune-endothelial cell adhesion, which, respectively, are
important processes in the development of cancer and athe-
Sadik et al. [174] performed a SAR study of inhibition of rosclerosis [183]. Bors et al. [184] have suggested that the
15-lipoxygenases (15-LOXs) by flavonoids. Planar flavonoid stability of flavonoid phenoxyl radical is sometimes ques-
structure with delocalized electrons (achieved by the C2-C3 tionable and may give rise to prooxidant action. This might
double bond) seems to be a prerequisite for inhibitory activ- help explain the occasional, unpredictable relationships
ity. The following structural features were found to enhance sometimes observed between the structure of some flavon-
the inhibitory potency: (a) presence of a catechol moiety in oids and their antioxidant activities. Prooxidant activity is
the B or A ring, (b) 4-carbonyl group in the C ring (Fig. 8B). thought to be directly proportional to the total number of
In contrast, the presence of the 3-OH group in the C ring hydroxyl groups [54].
diminished the inhibition of 15-LOX. While the presence of
the catechol moiety is favourable, an excessive number of The flavonoid phenoxyl radical could interact with oxy-
OH groups lowers the hydrophobicity of flavonoids and gen, generating quinones and superoxide anion, rather than
diminishes intercalation in the hydrophobic cavity at the terminating the chain reaction (Fig. 9). This reaction may
active site of the enzyme. Because flavonoids possess LOX take place in the presence of high levels of transient metal
inhibitory activities as well as free radical scavenging activi- ions and may be responsible for the undesired prooxidant
ties, they could be promising compounds, as therapeutic effect of flavonoids [21].
agents, against cardiovascular diseases. The same flavonoid compound could behave both as an
Inhibition of 15-LOX by quercetin and quercetin mono- antioxidant and a prooxidant. Van Acker et al. [53] pointed
glucosides was reported by da Silva et al. [175]. The ob- out that the most active antioxidants are likely to be prooxi-
tained results indicate that quercetin glycoside as well as its dants. This should be taken into account when designing
aglycone are capable of inhibiting the LOX induced LDL therapeutically interesting flavonoids: compounds for thera-
oxidation more efficiently than vitamins C and E. Quercetin, peutic application may not be those with the highest antioxi-
quercetin 3-glucoside, and quercetin 7-glucoside exhibited a dant activity, but those with optimum (high) antioxidant
higher inhibitory effect than quercetin 4’-glucoside. The activities.
authors concluded that the catechol structure in the B ring Ohshima et al. [185] studied the antioxidant and prooxi-
largely contributes to the inhibition of the 15-LOX induced dant effects of 18 flavonoids and related phenolic com-
LDL oxidation. pounds on the DNA damage induced by NO, peroxynitrite,
Redrejo-Rodriguez et al. [176] performed a theoretical, and nitroxyl anion. Most of the tested flavonoids inhibited
semiempirical SAR study of LOX inhibition by four flavon- DNA strand breakage. Only flavonoids having an ortho-
oids. The obtained results showed a clear relation between trihydroxyl group either in the B ring or in the A ring acted
the planar character of flavonoids and the potency of these as prooxidants.
compounds as LOX inhibitors. Also, the ELUMO and delocali- Yoshino et al. [186] found that flavonoids having a
zation of LUMO orbital are related to inhibitory potency. catechol structure in the A or B ring exert prooxidant activ-
The ability of flavonoids to inhibit prooxidant enzymes ity. Oxidation of catechol by a copper(II) ion bound to DNA
as well as scavenge free radicals and chelate metal ions can generate reactive oxygen species responsible for the site-
makes them very promising compounds for designing mul- specific DNA damage. The mechanism of the DNA damage
H
O O
O O
O O O O
H O2 O2 H
-H+
O O
O O O O
H H
H H
Fig. (9). Prooxidant activity of flavonoids.
by quercetin was proposed [187]. Reacting with Cu(II) quer- vivo situation, because in vitro activity of flavonoids does
cetin induced extensive DNA damage, but kaempferol and not contain information about the flavonoid solubility, ab-
luteolin induced little DNA damage, even in the presence of sorption, metabolism and degradation in the colon, and
Cu(II). El Amrani et al. [188] studied the oxidative DNA transport properties (e.g., whether flavonoids cross the
cleavage induced by an iron(III) flavonoid complex. For the blood-brain barrier, and in which forms) [44]. Only flavon-
cleavage mechanism, the authors proposed the mechanism of oids possessing suitable transport, metabolic, absorption
electron transfer in which Fe(III) takes one electron of the properties, as well as flavonoids that will not undergo struc-
carbon atom vicinal to the 3-hydroxy-4-keto moiety generat- tural modifications can reach tissue in an acceptable high
ing the iron(II) flavonoid complex (Fig. 10). Then, the fer- concentration and exert activity similar to flavonoids in vi-
rous complex binds to DNA, which results in the formation tro.
of reactive oxygen species and DNA cleavage. As emphasized in a recent study by Haenen et al. [59],
determination of structure-activity relationships of flavon-
oids is subject to many competing variables. Stability of the
O O
antioxidant, antioxidant effects of reaction products, concen-
tration of the antioxidant and concentration of the reactive
O O species should always be evaluated critically. In vitro studies
can be much simpler and controllable compared to in vivo
O O
Fe(III) Fe(II) studies for screening antioxidants for the structure-activity
and elucidation of the mode of action. The problem is that in
Fig. (10). Proposed route for generation of the ferrous complex
vitro systems usually deviate from in vivo situations. Con-
from the Fe(III) flavonoid complex [188].
struction of a valid structure-activity relationship may be
Nemeikaite-eniene et al. [189] studied the prooxidant complicated by the fact that more that one moiety within the
cytotoxicity of 13 hydroxybenzenes and 6 flavonoids in FLK flavonoids can independently exert a potent activity. In con-
and HL-60 cells. These authors demonstrated that the redox trast, despite the fact that some moieties (for example, the
potential of phenoxyl radical/phenol couple may serve as an catechol group in ring B and the C-3 OH group) have an
important tool to predict the prooxidant cytotoxicity of poly- important role in metal chelating and in radical scavenging,
phenols. SAR for metal chelating need not be identical to SAR for
radical scavenging. In this case, the overall SAR (comprising
CONCLUDING REMARKS both metal chelating and radical scavenging) is a mixed one.
The potential health promoting properties of flavonoids Even more factors are important in the in vivo activity of
result from their capability to decrease the risk of develop- antioxidants, e.g. lipophilicity, activity of metabolism prod-
ment of degenerative diseases such as coronary heart dis- ucts and binding to proteins. As concluded by Haenen et al.
eases, some forms of cancer, rheumatoid arthritis, and Park- [59], in vitro studies reveal the initial features of a series of
inson’s and Alzheimer’s diseases. These medicinal actions of antioxidants. These data can be transferred to the more com-
flavonoids are mostly ascribed to their antioxidant activity, plex in vivo situation. Detailed investigations should be done
scavenging of free radicals, metal ions chelation, enzyme of the actual fate of flavonoid compounds in vivo, to be able
inhibition, modulation of gene expression and interaction to predict their effects in living cells. In particular, specific
with the cell signalling pathways. mechanisms explaining how these compounds act need to be
discovered.
Despite many efforts, the unequivocal relationships be-
tween the antioxidant activity of the flavonoids and their Future research should be focused on at least two major
chemical structures are still not discovered. The construction areas:
of SAR/QSAR of antioxidant actions of flavonoids is not an 1. Development and application of specific and une-
easy task. The results of a number of studies, analyzed and quivocal methods for measuring the biological activ-
discussed in the present review, provide a solid rationale for ity of an enlarged pool of flavonoids. Uniform ana-
continuing efforts to improve QSAR models of flavonoid lytical methods are needed to allow data from differ-
antioxidant activities. Common unfavourable characteristics ent sources to be compared. This is most apparent in
of available models for modelling the antioxidant activity of the area of antioxidant research, which has been ham-
flavonoids are (1) a relatively small number of compounds in pered by the lack of suitable methods to measure the
datasets used for model building, (2) validation of developed antioxidant activity [59, 192];
models usually based only on the fit of statistical parameters,
and (3) a relatively large number of parameters (descriptors) 2. Improvement of the methods used in QSAR model-
involved in models compared to the total number of com- ling, as well as the methods for designing and using
pounds used for models development. Due to these charac- suitable molecular descriptors. In fact, the best set
teristics, such models have strong limitations and are not of descriptors may be achieved by careful selection of
appropriate for real prediction of the antioxidant activity of several different types of molecular descriptors, de-
flavonoids, i.e., predictions by such models cannot be reli- scribing the driving forces related to the activity of
able [190, 191]. An additional common characteristic of flavonoids. The interpretability of chosen descriptors,
almost all developed antioxidant activity models of flavon- i.e., their chemical meaning, is desirable. This may
oids is that they are based on in vitro experimental values. help direct the synthesis and selection of new drugs
We must be aware that models developed based on in vitro with potential therapeutic applications for the treat-
antioxidant activity of flavonoids cannot be applied to an in ment of a wide range of free radical-induced diseases.
Synthesis of favourable flavonoid compounds with HCA = Hierarchical cluster analysis

physiological and pharmacological activities could be
HOMO = Highest occupied molecular orbital
achieved by different approaches. Classical in vitro chemical
synthesis is one of them [193-196]. An alternative in vivo IC50 = Concentration for 50 % inhibition
approach is to produce flavonoids using plant cell culture IP = Ionization potential
strategies [197]. Also, microbial and enzymatic transforma-
tions of flavonoids can be used to convert abundant flavon- Ipa = Anodic peak current
oids to desired products [198]. Another in vivo approach to KNN = k-nearest neighbor
synthesizing favourable flavonoid compounds is the genetic
modification of flavonoid biosynthesis. Flavonoid biosynthe- kQ = Overall rate constant
sis is one of the best characterized natural product pathways LDL = Low-density lipoprotein
in plants, and is therefore an excellent target for metabolic
engineering [199-201]. Manipulation of flavonoid biosynthe- log P = Octanol/water partition coefficient
sis can be approached via manipulation of pathway genes, LOX = Lipoxygenase
modification of the expression of regulatory genes, and gen-
LUMO = Lowest unoccupied molecular orbital
eration of novel enzymatic specificities [202]. This may
result in the plant’s ability to synthesize flavonoid bioactive MOLMAP = Molecular maps of atom-level properties
natural products for human health.
n value = Number of electrons involved in the oxida-
Reviewed SARs and QSARs demonstrate that the anti- tion
oxidant activities of flavonoids, i.e. free radical-scavenging,
PCA = Principal component analysis
metal ion chelating and enzyme inhibitor activities, are de-
pendent on suitable structure motifs. In particular, the QSAR = Quantitative structure-activity relationship
catechol moiety makes important contributions, not only to QSPR = Quantitative structure-property relationship
scavenging free radicals and sequestering metal ions but also
to inhibiting prooxidant enzymes. Integration of radical RSA = Radical scavenging activity
scavenging, metal ion chelating and enzyme inhibiting ac- SAM = Structure-activity map
tivities into a single structure make flavonoids very promis-
ing multifunctional antioxidants. Flavonoids therefore seem SAR = Structure-activity relationship
to be a cornerstone in the exciting field of rational design of TEAC = Trolox equivalent antioxidant capacity
naturally occurring multipotent antioxidants.
XO = Xanthine oxidase
ACKNOWLEDGEMENTS
REFERENCES
This work was supported in part by Grants 0079025
[1] Andersen, Ø.M.; Markham, K.R. Eds. Flavonoids: Chemistry,
(D.A., D.D.A., D.B., V.R.) and 0098034 (B.L., N.T.) Biochemistry and Applications, CRC Press: Boca Raton, 2005.
awarded by the Ministry of Science, Education and Sports of [2] Grotewold, E. Ed. The Science of Flavonoids, Springer: New York,
the Republic of Croatia. The authors gratefully acknowledge 2006.
the anonymous reviewers for their helpful comments on the [3] USDA Database for the Flavonoid Content of Selected Foods,
2003. http://www.nal.usda.gov/fnic/foodcomp/Data/Flav/flav.html
manuscript. We also thank Dr. Wolf Bors for correspon- [4] Havsteen, B.H. Pharmacol. Ther., 2002, 96, 67.
dence and help. [5] Harborne, J.B. In Plant Flavonoids in Biology and Medicine:
Biochemical, Pharmacological, and Structure-Activity Relation-
ABBREVIATIONS ships; Cody, V.; Middleton, Jr. E.; Harborne, J.B., Eds.; Alan R.
Liss: New York, 1986; pp. 15-24.
AM1 = Austin Model 1 [6] Harborne, J.B.; Williams, C.A. Phytochemistry, 2000, 55, 481.
BDE = Bond dissociation energy [7] Treutter, D. Environ. Chem. Lett., 2006, 4, 147.
[8] Williams, C.A.; Grayer, R.J. Nat. Prod. Rep., 2004, 21, 539.
DFT = Density functional theory [9] Aherne, S.A.; O’Brien, N.M. Nutrition, 2002, 18, 75.
[10] Renaud, S.C.; de Lorgeril, M. Lancet, 1992, 339, 1523.
Hf = Heat of formation of the flavonoid radical [11] King, A.; Young, G. J. Am. Diet. Assoc., 1999, 99, 213.
[12] Beecher, G.R. J. Nutr., 2003, 133, 3248S.
DPPH = 1,1-diphenyl-2-picrylhydrazyl [13] Hollman, P.C.H.; Katan, M.B. Food Chem. Toxicol., 1999, 37, 937.
E = Half-wave potential of the first oxidation [14] Kaur, C.; Kapoor, H.C. Int. J. Food Sci. Technol., 2001, 36, 703.
[15] Van Duyn, M.A.S.; Pivonka, E. J. Am. Diet. Assoc., 2000, 100,
wave 1511.
EHOMO = Energy of the highest occupied molecular [16] Le Marchand, L. Biomed. Pharmacother., 2002, 56, 296.
[17] Mojiova, G.; Kuchta, M. Physiol. Res., 2001, 50, 529.
orbital [18] Di Carlo, D.; Mascolo, N.; Izzo, A.A.; Capasso, F. Life Sci., 1999,
EHYDR = Hydration energy 65, 337.
[19] Cushnie, T.P.T.; Lamb, A.J. Int. J. Antimicr. Agents, 2005, 26, 343.
ELUMO = Energy of the lowest unoccupied molecular [20] Shahidi, F.; Wanasundara, P.K.J.P.D. Crit. Rev. Food Sci. Nutr.,
orbital 1992, 32, 67.
[21] Pietta, P.-G. J. Nat. Prod., 2000, 63, 1035.
Epa = Anodic peak potential [22] Middleton, Jr. E.; Kandaswami, C. In The Flavonoids: Advances in
Research Since 1986; Harborne, J.B., Ed.; Chapman and Hall:
Ep/2 = Half-peak oxidation potential London, 1994; pp. 619-652.
[23] Bors, W.; Heller, W.; Michel, C.; Stettmaier, K. In Handbook of
FRAP = Ferric reducing antioxidant power assay Antioxidants; Cadenas, E.; Packer, L., Eds.; Marcel Dekker, Inc.:
New York, 1996; pp. 409-466.
[24] Nijveldt, R.J.; van Nood, E.; van Hoorn, D.E.C.; Boelens, P.G.; van [60] Bors, W.; Heller, W.; Michel, C.; Saran, M. In Free Radicals and
Norren, K.; van Leeuwen, P.A.M. Am. J. Clin. Nutr., 2001, 74, the Liver, Csomos, G.; Feher, J.; Eds.; Springer-Verlag: Berlin,
418. 1992; pp. 77-95.
[25] Ross, J.A.; Kasum, C.M. Annu. Rev. Nutr., 2002, 22, 19. [61] Croft, K.D. Ann. N. Y. Acad. Sci., 1998, 854, 435.
[26] Skibola, C.F.; Smith, M.T. Free Radic. Biol. Med., 2000, 29, 375. [62] Lien, E.J.; Ren, S.; Bui, H.-H.; Wang, R. Free Radic. Biol. Med.,
[27] Trinajsti, N.; Nikoli, S.; Lui, B.; Ami, D. Acta Pharm., 1996, 1999, 26, 285.
46, 249. [63] Bors, W.; Michel, C.; Stettmaier, K. Methods Enzymol., 2001, 335,
[28] Voskresensky, O.N.; Levitsky, A.P. Curr. Med. Chem., 2002, 9, 166.
1367. [64] Bors, W.; Michel, C. Ann. N.Y. Acad. Sci., 2002, 957, 57.
[29] Zhang, H.-Y. Curr. Comp.-Aided Drug Des., 2005, 1, 257. [65] Heijnen, C.G.M.; Haenen, G.R.M.M.; van Acker, F.A.A.; van der
[30] Halliwell, B.; Rafter, J.; Jenner, A. Am. J. Clin. Nutr., 2005, 81, Vijgh, W.J.F.; Bast, A. Toxicol. In Vitro, 2001, 15, 3.
268S. [66] Soobrattee, M.A.; Neergheen, V.S.; Luximon-Ramma, A.;
[31] Livingstone, D.J. J. Chem. Inf. Comput. Sci., 2000, 40, 195. Arouma, O.I.; Bahorun, T. Mutat. Res., 2005, 579, 200.
[32] Kontogiorgis, A.C.; Pontiki, A.E.; Hadjipavlou-Litina, D. Mini [67] Arora, A.; Nair, M.G.; Strasburg, G.M. Free Radic. Biol. Med.,
Rev. Med. Chem., 2005, 5, 563. 1998, 24, 1355.
[33] Bors, W.; Heller, W.; Michel, C.; Saran, M. In Methods in Enzy- [68] Foti, M.; Piattelli, M.; Baratta, M.T.; Ruberto, G. J. Agric. Food.
mology; Packer, L.; Glazer, A.N. Eds.; Academic Press: San Diego, Chem., 1996, 44, 497.
CA, 1990; Vol. 186, pp. 343-355. [69] Lemanska, K.; Szymusiak, H.; Tyrakowska, B.; Zielinski, R.;
[34] Benavente-Garcia, O.; Castillo, J.; Marin, F.R.; Ortuno, A.; Del Soffers, A.E.M.F.; Rietjens, I.M.C.M. Free Radic. Biol. Med.,
Rio, J.A. J. Agric. Food Chem., 1997, 45, 4505. 2001, 31, 869.
[35] Huang, X.; Liu, T.; Gu, J.; Luo, X.; Ji, R.; Cao, Y.; Xue, H.; Wong, [70] Cai, Y.-Z.; Sun, M.; Xing, J.; Luo, Q.; Corke, H. Life Sci., 2006,
J.T.-F.; Wong, B.L.; Pei, G.; Jiang, H.; Chen, K. J. Med. Chem., 78, 2872.
2001, 44, 1883. [71] Chen, J.-W.; Zhu, Z.-Q.; Hu, T.-X., Zhu, D.-Y. Acta Pharmacol.
[36] de Wet, H.; McIntosh, D.B.; Conseil, G.; Baubichon-Cortay, H.; Sin, 2002, 23, 667.
Krell, T.; Jault, J.-M.; Daskiewicz, J.-B.; Barron, D.; Di Pietro, A. [72] Butkovi, V.; Klasinc, L.; Bors, W. J. Agric. Food Chem., 2004,
Biochemistry, 2001, 40, 10382. 52, 2816.
[37] Williams, R.J.; Spencer, J.P.E.; Rice-Evans, C. Free. Radic. Biol. [73] Furusawa, M.; Tanaka, T.; Ito, T.; Nishikawa, A.; Yamazaki, N.;
Med., 2004, 36, 838. Nakaya, K.-i.; Matsuura, N.; Tsuchiya, H.; Nagayama, M.; Iinuma,
[38] Ami, D.; Davidovi-Ami, D.; Belo, D.; Trinajsti, N. Croat. M. J. Health Sci., 2005, 51, 376.
Chem. Acta, 2003, 76, 55. [74] Sroka, Z. Z. Naturforsch., 2005, 60c, 833.
[39] Wei, D.G.; Yang, G.-F.; Wan, J.; Zhan, C.-G. J. Agric. Food [75] Liu, W.; Guo, R. J. Agric. Food Chem., 2005, 53, 2890.
Chem., 2005, 53, 1604. [76] Trouillas, P.; Marsal, P.; Siri, D.; Lazzaroni, R.; Duroux, J.-L.
[40] Rice-Evans, C.A.; Miller, N.J.; Paganga, G. Trends Plant Sci., Food Chem., 2006, 97, 679.
1997, 2, 152. [77] Guzik, T.J.; Harrison, D.G. Drug Discov. Today, 2006, 11, 524.
[41] Rice-Evans, C.A.; Miller, N.J. In Flavonoids in Health and Dis- [78] Jovanovic, S.V.; Steenken, S.; Tosic, M.; Marjanovic, B.; Simic,
ease; Rice-Evans, C.A.; Packer, L., Eds.; Marcel Dekker: New M.G. J. Am. Chem. Soc., 1994, 116, 4846.
York, 1998, pp. 199-219. [79] Bors, W.; Heller, W.; Michel, C.; Saran, M. In: Antioxidants in
[42] Heim, K.E.; Tagliaferro, A.R.; Bobilya, D.J. J. Nutr. Biochem., Therapy and Preventive Medicine; Emerit, I. et al., Eds.; Plenum
2002, 13, 572. Press: New York, 1990; pp. 165-170.
[43] van Acker, S.A.B.E., de Groot, M.J.; van den Berg, D.-J.; Tromp, [80] Haenen, G.R.M.M.; Paquay, J.B.G.; Korthouwer, R.E.M.; Bast, A.
M.N.J.L.; den Kelder, G.D.-O.; van der Vijgh, W.J.F.; Bast, A. Biochem. Biophys. Res. Commun., 1997, 236, 591.
Chem. Res. Toxicol., 1996, 9, 1305. [81] Burda, S.; Oleszek, W. J. Agric. Food Chem., 2001, 49, 2774.
[44] Rice-Evans, C. Curr. Med. Chem., 2001, 8, 797. [82] Dugas Jr., A.J.; Castaneda-Acosta, J.; Bonin, G.C.; Price, K.L.;
[45] Rice-Evans, C. Why do we Expect Flavonoids to Function as Fischer, N.H.; Winston, G.W. J. Nat. Prod., 2000, 63, 327.
Antioxidants in vivo? 2002 http://www.medicine.uiowa.edu/frrb [83] Seeram, N.P.; Nair, M.G. J. Agric. Food Chem., 2002, 50, 5308.
/SRFRS/www/SFRS2002/pdf/SFRS2002RiceFlavonoids.pdf [84] Kähkönen, M.P.; Heinonen, M. J. Agric. Food Chem., 2003, 51,
[46] Bors, W.; Michel, C.; Stettmaier, K. Flavonoids and their free 628.
radical reactions, 2001 http://www.medicine.uiowa.edu/frrb/ [85] Taubert, D.; Breitenbach, T.; Lazar, A.; Censarek, P.; Harlfinger,
VirtualSchool/Bors-Flavonoids.pdf S.; Berkels, R.; Klaus, W.; Roesen, R. Free Radic. Biol. Med.,
[47] Prior, R.L.; Cao, G. J. AOAC Int., 2000, 83, 950. 2003, 35, 1599.
[48] van Acker, S.A.B.E.; Bast, A.; van der Vijgh, W.J.F. In: Flavon- [86] Bors, W.; Michel, C. Free Radic. Biol. Med., 1999, 27, 1413.
oids in Health and Disease; Rice-Evans, C.A.; Packer, L., Eds.; (Note: On page 1604 of the ref. 85 and on page 1422 of ref. 86, the
Marcel Dekker: New York, 1998, pp. 221-251. authors proposed mesomeric equilibrium incorporating the oxo-
[49] Silva, M.M.; Santos, M.R.; Caroco, G.; Rocha, R.; Justino, G.; nium structure incorrectly drawn. We are grateful to Dr. Wolf Bors
Mira, L. Free Radic. Res., 2002, 36, 1219. for providing us this information and for advice and discussion re-
[50] Sichel, G.; Corsaro, C.; Scalia, M.; Di Bilio, A.J.; Bonomo, R.P. garding mesomeric oxonium structures.)
Free Radic. Biol. Med., 1991, 11, 1. [87] Santos, M.R.; Mira, L. Free Radic. Res., 2004, 38, 1011.
[51] Chen, Z.Y.; Chan, P.T.; Ho, K.Y.; Fung, K.P.; Wang, J. Chem. [88] Rasulev, B.F.; Abdullaev, N.D.; Syrov, V.N.; Leszczynski, J.
Phys. Lipids 1996, 79, 157. QSAR Comb. Sci., 2005, 24, 1056.
[52] Rice-Evans, C.A.; Miller, N.J.; and G. Paganga, G. Free Radic. [89] Mukai, K.; Mitani, S.; Ohara, K.; Nagaoka, S.-I. Free Radic. Biol.
Biol. Med., 1996, 20, 933. Med., 2005, 38, 1243.
[53] van Acker, S.A.B.E.; van den Berg, D.-J.; Tromp, M.N.J.L.; Grif- [90] Di Majo, D.; Giammanco, M.; La Guardia, M.; Tripoli, E.; Giam-
fioen, D.H.; van Bennekom, W.P.; van der Vijgh, W.J.F.; Bast, A. manco, S.; Finotti, E. Food Res. Int., 2005, 38, 1161.
Free Radic. Biol. Med., 1996, 20, 331. [91] Pirker, K.F.; Stolze, K.; Reichenauer, T.G.; Nohl, H.; Goodman, B.
[54] Cao, G.; Sofic, E.; Prior, R.L. Free Radic. Biol. Med., 1997, 22, A. Free Radic. Res., 2006, 40, 513.
749. [92] Jovanovic, S.V.; Steenken, S.; Hara, Y.; Simic, M.G. J. Chem.
[55] Okawa, M.; Kinjo, J.; Nohara, T.; Ono, M. Biol. Pharm. Bull., Soc., Perkin Trans. 2, 1996, 2497.
2001, 24, 1202. [93] Yang, B.; Kotani, A.; Arai, K.; Kusu, F. Anal. Sci., 2001, 17, 599.
[56] Yokozawa, T.; Chen, C.P.; Dong, E.; Tanaka, T.; Nonaka, G.-I.; [94] Yang, B.; Arai, K.; Kusu, F. Electrochem., 2001, 69, 519.
Nishioka, I. Biochem. Pharmacol., 1998, 56, 213. [95] van Acker, S.A.B.E.; van Balen, G.P.; van den Berg, D.-J.; Bast,
[57] McPhail, D.B.; Hartley, R.C.; Gardner, P.T.; Duthie, G.G. J. Agric. A.; van der Vijgh, W.J.F. Biochem. Pharmacol., 1998, 56, 935.
Food Chem., 2003, 51, 1684. [96] Yang, B.; Kotani, A.; Arai, K.; Kusu, F. Chem. Pharm. Bull., 2001,
[58] Woodman, O.L.; Meeker, W.F.; Boujaoude, M. J. Cardiovasc. 49, 747.
Pharmacol., 2005, 46, 302. [97] Hotta, H.; Nagano, S.; Ueda, M.; Tsujino, Y.; Koyama, J.; Osakai,
[59] Haenen, G.R.M.M.; Arts, M.J.T.J.; Bast, A.; Coleman, M.D. T. Biochim. Biophys. Acta, 2002, 1572, 123.
Environ. Toxicol. Phar., 2006, 21, 191.
[98] Hotta, H.; Sakamoto, H.; Nagano, S.; Osakai, T.; Tsujino, Y. [139] Gupta, S.; Matthew, S.; Abreu, P.M.; Aires-de-Sousa, J. Bioorg.
Biochim. Biophys. Acta, 2001, 1526, 159. Med. Chem., 2006, 14, 1199.
[99] Fujisawa, S.; Ishihara, M.; Kadoma, Y. SAR QSAR Envir. Res., [140] Fan, W.; Lin, X.; Hsieh, Y.-W.; Lin, B.; Baker, J. W.; Tsai, C.-c.
2002, 13, 617. 2005 IEEE Computational Systems Bioinformatics Conference –
[100] Fujisawa, S.; Kadoma, Y. Chemosphere, 2006, 62, 71. Workshops, 2005, pp. 267-268.
[101] Roginsky, V.; Lissi, E.A. Food Chem., 2005, 92, 235. [141] Erlejman, A.G.; Verstraeten, S.V.; Fraga, C.G.; Oteiza, P.I. Free
[102] Firuzi, O.; Lacanna, A.; Petrucci, R.; Marrosu, G.; Saso, L. Bio- Radic. Res., 2004, 38, 1311.
chim. Biophys. Acta, 2005, 1721, 174. [142] Hendrich, A. B. Acta Pharmacol. Sin., 2006, 27, 27.
[103] Nagai, S.; Ohara, K.; Mukai, K. J. Phys. Chem. B, 2005, 109, 4234. [143] Kanakis, C.D.; Tarantilis, P.A.; Polissiou, M.G.; Diamantoglou, S.;
[104] Mukai, K.; Nagai, S.; Ohara, K. Free Radic. Biol. Med., 2005, 39, Tajmir-Riahi, H.A. J. Biomol. Struct. Dyn., 2005, 22, 719.
752. [144] Thompson, M.; Williams, C.R.; Elliot, G.E.P. Anal. Chim. Acta,
[105] van Acker, S.A.B.E.; Koymans, L.M.H.; Bast, A. Free Radic. Biol. 1976, 85, 375.
Med., 1993, 15, 311. [145] Brown, J.E.; Khodr, H.; Hider, R.C.; Rice-Evans, C.A. Biochem. J.,
[106] Zhang, H.-Y. J. Am. Oil Chem. Soc., 1998, 75, 1705. 1998, 330, 1173.
[107] Zhang, H.-Y. J. Am. Oil Chem. Soc., 1999, 76, 745. [146] Khokhar, S.; Owusu Apenten R.K. Food Chem., 2003, 81, 133.
[108] Zhang, H.-Y.; Chen, D.-Z. Acta Biochim. Biophys. Sin., 2000, 32, [147] Fernandez, M.T.; Mira, M.L.; Florencio, M.H.; Jennings, K.R. J.
317. Inorg. Biochem., 2002, 92, 105.
[109] Vaya, J.; Mahmood, S.; Goldblum, A.; Aviram, M.; Volkova, N.; [148] Mira, L.; Fernandez, M.T.; Santos, M.; Rocha, R.; Florencio, M.H.;
Shaalan, A.; Musa, R.; Tamir, S. Phytochemistry, 2003, 62, 89. Jennings, K.R. Free Radic. Res., 2002, 36, 1199.
[110] Modak, B.; Contreras, M.L.; Gonzalez-Nilo, F.; Torres, R. Bioorg. [149] Moridani, M.Y.; Pourahmad, J.; Bui, H.; Siraki, A.; O’Brien, P.J.
Med. Chem. Lett., 2005, 15, 309. Free Radic. Biol. Med., 2003, 34, 243.
[111] Sadeghipour, M.; Terreux, R.; Phipps, J. Toxicol. In Vitro, 2005, [150] Melidou, M.; Riganakos, K.; Galaris, D. Free Radic. Biol. Med.,
19, 155. 2005, 39, 1591.
[112] Rackova, L.; Firakova, S.; Kostalova, D.; Stefek, M.; Sturdik, E.; [151] Engelmann, M.D.; Hutcheson, R.; Cheng, I.F. J. Agric. Food
Majekova, Bioorg. Med. Chem., 2005, 13, 6477. (Note: On page Chem., 2005, 53, 2953.
6481, the authors gave a wrong description of the correlation be- [152] de Souza, R.F.V.; De Giovani, W.F. Spectrochim. Acta Part A,
tween EHYDR and nOH, as well as between EHYDR and pIC50, because 2005, 61, 1985.
they omitted the fact that EHYDR values are negative. Their state- [153] Teixeira, S.; Siquet, C.; Alves, C.; Boal, I.; Marques, M.P.; Borges,
ment is correct only for absolute EHYDR values, which is taken into F.; Lima, J.L.F.C.; Reis, S. Free Radic. Biol. Med., 2005, 39, 1099.
account in this review.) [154] Leopoldini, M.; Russo, N.; Toscano M. J. Agric. Food Chem.,
[113] Seyoum, A.; Asres, K.; El-Fiky, F.K. Phytochemistry, 2006, 67, 2006, 54, 3078.
2058. [155] Leopoldini, M.; Russo, N.; Chiodo, S.; Toscano M. J. Agric. Food
[114] Wright, J.S.; Carpenter, D.J.; McKay, D.J.; Ingold, K.U. J. Am. Chem., 2006, 54, 6343.
Chem. Soc., 1997, 119, 4245. [156] Cotelle, N. Curr. Top. Med. Chem., 2001, 1, 569.
[115] Wright, J.S.; Johnson, E.R.; DiLabio, G.A. J. Am. Chem. Soc., [157] Hayashi, T.; Sawa, K.; Kawasaki, M.; Arisawa, M.; Shimizu, M.;
2001, 123, 1173. Morita, N. J. Nat. Prod., 1988, 51, 345.
[116] Zhang, H.-Y.; Sun, Y.-M.; Wang, X.-L. Chem. Eur. J., 2003, 9, [158] Costantino, L.; Rastelli, G.; Albasini, A. Eur. J. Med. Chem., 1996,
502. 31, 693.
[117] Zielonka, J.; Gebicki, J.; Grynkiewicz, G. Free Radic. Biol. Med., [159] Cotelle, N.; Bernier, J.-L.; Catteau, J.-P.; Pommery, J.; Wallet, J.-
2003, 35, 958. C.; Gaydou, E.M. Free Radic. Biol. Med., 1996, 20, 35.
[118] Martins, H.F.P.; Leal, J.P.; Fernandez, M.T.; Lopes, V.H.C., Cor- [160] Costantino, L.; Rastelli, G.; Albasini, A. Pharmazie, 1995, 50, 573.
deiro, M.N.D.S. J. Am. Soc. Mass Spectrom., 2004, 15, 848. [161] Rastelli, G.; Costantino, L.; Albasini, A. J. Am. Chem. Soc., 1997,
[119] Mielczarek, C. Eur. J. Pharm. Sci., 2005, 25, 273. 119, 3007.
[120] Zhang, H.-Y.; Wang, L.-F. J. Biomol. Struct. Dyn., 2005, 22, 483. [162] Rastelli, G.; Costantino, L.; Albasini, A. Eur. J. Med. Chem., 1995,
[121] Wang, L.-F.; Zhang, H.-Y. Bioorg. Chem., 2005, 33, 108. 30, 141.
[122] Ji, H.-F.; Zhang, H.-Y.; Shen, L. Bioorg. Med. Chem. Lett., 2006, [163] Cos, P.; Ying, L.; Calomme, M.; Hu, J.P.; Cimanga, K.; Van Poel
16, 4095. B.; Pieters, L.; Vlietinck, A.J.; Vanden Berghe, D. J. Nat. Prod.,
[123] Kondo, K.; Kurihara, M.; Miyata, N.; Suzuki, T.; Toyoda, M. Arch. 1998, 61, 71.
Biochem. Biophys., 1999, 362, 79. [164] Ami, D.; Davidovi-Ami, D.; Belo, D.; Lui, B.; Trinajsti, N.
[124] Zhang, H.-Y.; Wang, L.F. J. Am. Oil Chem. Soc., 2002, 79, 943. J. Chem. Inf. Comput. Sci., 1998, 38, 815.
[125] Leopoldini, M.; Pitarch, I.P.; Russo, N.; Toscano, M. J. Phys. [165] Nagao, A.; Seki, M.; Kobayashi, H. Biosci. Biotechnol. Biochem.,
Chem. A, 2004, 108, 92. 1999, 63, 1787.
[126] Leopoldini, M.; Marino, T.; Russo, N.; Toscano, M. J. Phys. Chem. [166] Ponce, A.M.; Blanco, S.E.; Molina, A.S.; Garcia-Domenech, R.;
A, 2004, 108, 4916. Galvez, J. J. Chem. Inf. Comput. Sci., 2000, 40, 1039.
[127] Marfak, A.; Trouillas, P.; Allais, D.P.; Calliste, C.A.; Cook- [167] Lin, C.-M.; Chen, C.-S.; Chen, C.-T.; Liang, Y.-C.; Lin, J.-K.
Moreau, J.; Duroux, J.-L. Biochim. Biophys. Acta, 2004, 1670, 28. Biochem. Biophys. Res. Commun., 2002, 294, 167.
[128] Chen, W.; Guo, P.; Song, J.; Cao, W.; Bian, J. Bioorg. Med. Chem. [168] Van Hoorn, D.E.C.; Nijveldt, R.J.; Van Leeuwen, P.A.M.; Hofman,
Lett., 2006, 16, 3582. Z.; M’Rabet, L.; De Bont, D.B.A.; Van Norren, K. Eur. J. Pharma-
[129] Heijnen, C.G.M.; Haenen, G.R.M.M.; Vekemans, J.A.J.M.; Bast, col., 2002, 451, 111.
A. Environ. Toxicol. Pharmacol., 2001, 10, 199. [169] da Silva, S.L.; da Silva, A.; Honorio, K.M.; Marangoni, S.; Toya-
[130] Erkoc, S.; Erkoc, F.; Keskin, N. J. Mol. Struct. (Theochem.), 2003, ma, M.H.; da Silva, A.B.F. J. Mol. Struct., (Theochem.), 2004, 684,
631, 141. 1.
[131] Mendoza-Wilson, A.M.; Glossman-Mitnik, D. J. Mol. Struct. [170] Montoro, P.; Braca, A.; Pizza, C.; De Tommasi, N. Food Chem.,
(Theochem.), 2005, 716, 67. 2005, 92, 349.
[132] Ghiotto, R.C.T.; Lavarda, F.C.; Ferreira, F.J.B. Int. J. Quantum [171] Brash, A.R. J. Biol. Chem., 1999, 274, 23679.
Chem., 2004, 97, 949. [172] Schewe, T. Biol. Chem., 2002, 383, 365.
[133] Farkas, O.; Jakus, J.; Heberger, K. Molecules, 2004, 9, 1079. [173] Schewe, T.; Sies, H. Research monographs: Flavonoids and
[134] Estrada, E.; Quincoces, J.A.; Patlewicz, G. Mol. Div., 2004, 8, 21. prooxidant enzymes. http://www.uniklinik-duesseldorf.de/img/ejbfile/
[135] Weber, K.C.; Honorio, K.M.; da Silva, S.L.; Mercadante, R.; da Research_monographs.pdf?id=280
Silva, A.B.F. Int. J. Quantum. Chem., 2005, 103, 731. [174] Sadik, C.D.; Sies, H.; Schewe, T. Biochem. Pharmacol., 2003, 65,
[136] Pannala, A.S.; Chan, T.S.; O’Brien, P.J.; Rice-Evans, C.A. Bio- 773.
chem. Biophys. Res. Commun., 2001, 282, 1161. [175] da Silva, E.L.; Tsushida, T.; Terao, J. Arh. Biochem. Biophys.,
[137] Zhang, H. Sci. China Ser. B, 1999, 42, 106. 1998, 349, 313.
[138] Zhang, H.-Y.; Wang, L.-F.; Sun, Y.-M. Bioorg. Med. Chem. Lett., [176] Redrejo-Rodriguez, M.; Tejeda-Cano, A.; del Carmen Pinto, M.;
2003, 13, 909. Macias, P. J. Mol. Struct., (Theochem.), 2004, 674, 121.
[177] Zhang, H.-Y.; Yang, D.-P.; Tang, G.-Y. Drug Discov. Today, 2006, [189] Nemeikaite-eniene, A.; Imbrasaite, A.; Sergediene, E.; enas, N.
11, 749. Arch. Biochem. Biophys., 2005, 441, 182.
[178] Awad, H.M.; Boersma, M.G.; Boeren, S.; van Bladeren, P.J.; [190] Jorgensen, W.L. J. Chem. Inf. Model., 2006, 46, 937.
Vervoort, J.; Rietjens, I.M.C.M. Chem. Res. Toxicol., 2001, 14, [191] Maggiora, G.M. J. Chem. Inf. Model., 2006, 46, 1535.
398. [192] Frankel, E.N.; Meyer, A.S. J. Sci. Food Agric., 2000, 80, 1925.
[179] Galati, G.; O’Brien, P.J. Free Radic. Biol. Med., 2004, 37, 287. [193] Iacobucci, G.A.; Sweeny, J.G. Tetrahedron, 1983, 39, 3005.
[180] Hodnick, W.F.; Ahmad, S.; Pardini, R.S. In Flavonoids in the [194] van Acker, F.A.A.; Hageman, J.A.; Haenen, G.R.M.M.; van der
Living System; Manthey, J.A.; Buslig, B., Eds.; Plenum Press, New Vijgh, W.J.F.; Bast, A.; Menge, W.M.P.B. J. Med. Chem., 2000,
York, 1998; pp. 131-150. 43, 3752.
[181] Metodiewa, D.; Jaiswal, A.K.; Cenas, N.; Dickancaite, E.; Segura- [195] Cotelle, N.; Vrielynck, L.; Nowogrocki, G.; Cotelle, P.; Vezin, H.
Aguilar, J. Free Radic. Biol. Med., 1999, 26, 107. J. Phys. Org. Chem., 2004, 17, 226.
[182] Galati, G.; Sabzevari, O.; Wilson, J.X.; O’Brien, P.J. Toxicology, [196] Kondo, T.; Oyama, K.-i.; Nakamura, S.; Yamakawa, D.; Tokuno,
2002, 177, 91. K.; Yoshida, K. Org. Lett., 2006, 16, 3609.
[183] Kim, J.-D.; Liu, L.; Guo, W.; Meydani, M. J. Nutr. Biochem., [197] Lila, M.A.; In Plant Pigments and their Manipulation; Davies,
2006, 17, 165. K.M., Ed.; CRC Press: Boca Raton, 2004; pp. 248-274.
[184] Bors, W.; Michel, C.; Schikora, S. Free Radic. Biol. Med., 1995, [198] Das, S.; Rosazza, J.P.N. J. Nat. Prod., 2006, 69, 499.
19, 45. [199] Winkel-Shirley, B. Plant Physiol., 2001, 126, 485.
[185] Ohshima, H.; Yoshie, Y.; Auriol, S.; Gilibert, I. Free Radic. Biol. [200] Davies, K.M.; Schwinn, K.E. In Flavonoids: Chemistry, Biochem-
Med., 1998, 25, 1057. istry and Applications; Andersen, Ø.M.; Markham, K.R., Eds.;
[186] Yoshino, M.; Haneda, M.; Naruse, M.; Murakami, K. Mol. Genet. CRC Press: Boca Raton, 2005; pp. 143-218.
Metab., 1999, 68, 468. [201] Lepiniec, L.; Debeaujon, I.; Routaboul, J.-M.; Baudry, A.; Pourcel,
[187] Yamashita, N.; Tanemura, H.; Kawanishi, S. Mutat. Res., 1999, L.; Nesi, N.; Caboche, M. Annu. Rev. Plant Biol., 2006, 57, 405.
425, 107. [202] Dixon, R.A.; Steele, C.L. Trends Plant Sci., 1999, 4, 394.
[188] El Amrani, F.B.A.; Perello, L.; Real, J.A.; Gonzalez-Alvarez, M.;
Alzuet, G.; Borras, J.; Garcia-Granda, S.; Montejo-Bernardo, J. J.
Inorg. Biochem., 2006, 100, 1208.
Received: September 1, 2006 Revised: November 23, 2006 Accepted: November 23, 2006
View publication stats

Amic Et Al-Curr Med Chem 2007 14 827-845

Uploaded by

Copyright:

Available Formats

Amic Et Al-Curr Med Chem 2007 14 827-845

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Amic Et Al-Curr Med Chem 2007 14 827-845

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

SAR and QSAR of the antioxidant activity of ﬂavonoids

Article in Current Medicinal Chemistry · March 2007

Dragan Amic Drago Beslo

SEE PROFILE SEE PROFILE

Vesna Rastija Bono Lucic

SEE PROFILE SEE PROFILE

dbeslo@pfos.hr View project

CroArtScia2011 View project

The user has requested enhancement of the downloaded file.

SAR and QSAR of the Antioxidant Activity of Flavonoids

0929-8673/07 $50.00+.00 © 2007 Bentham Science Publishers Ltd.

flavan nucleus 4-oxo-flavonoid nucleus

Fig. (2). Structural features of flavonoids with high antioxidant activity.

Fig. (6). Mesomeric equilibriums of the flavonoid phenoxyl radical [86].

Synthesis of favourable flavonoid compounds with HCA = Hierarchical cluster analysis

View publication stats

You might also like