Quercetin inhibits advanced glycation end product formation via chelating metal ions, trapping methylglyoxal, and trapping reactive oxygen species

Article abstract of DOI:10.1080/09168451.2017.1282805

Physiological concentration of Mg²⁺, Cu²⁺, and Zn²⁺ accelerated AGE formation only in glucosemediated conditions, which was effectively inhibited by chelating ligands. Only quercetin (10) inhibited MGO-mediated AGE formation as well as glucoseand ribose-mediated AGE formation among 10 polyphenols (1-10) tested. We performed an additional structure-activity relationship (SAR) study on flavanols (10, 11, 12, 13, and 14). Morin (12) and kaempherol (14) showed inhibitory activity against MGO-mediated AGE formation, whereas rutin (11) and fisetin (13) did not. These observations indicate that 3,5,7,4'-tetrahydroxy and 4-keto groups of 10 are important to yield newly revised mono-MGO adducts (16 and 17) and di-MGO adduct (18) having cyclic hemiacetals, while 3'-hydroxy group is not essential. We propose here a comprehensive inhibitory mechanism of 10 against AGE formation including chelation effect, trapping of MGO, and trapping of reactive oxygen species (ROS), which leads to oxidative degradation of 18 to 3,4-dihydroxybenzoic acid (15) and other fragments.

Full text of DOI:10.1080/09168451.2017.1282805

Bioscience, Biotechnology, and Biochemistry
ISSN: 0916-8451 (Print) 1347-6947 (Online) Journal homepage: http://www.tandfonline.com/loi/tbbb20
Quercetin inhibits advanced glycation end product
formation via chelating metal ions, trapping
methylglyoxal, and trapping reactive oxygen
species
Mohammad Nazrul Islam Bhuiyan, Shinya Mitsuhashi, Kengo Sigetomi &
Makoto Ubukata
To cite this article: Mohammad Nazrul Islam Bhuiyan, Shinya Mitsuhashi, Kengo Sigetomi
&
Makoto Ubukata (2017): Quercetin inhibits advanced glycation end product formation via
chelating metal ions, trapping methylglyoxal, and trapping reactive oxygen species, Bioscience,
Biotechnology, and Biochemistry, DOI: 10.1080/09168451.2017.1282805
To link to this article: http://dx.doi.org/10.1080/09168451.2017.1282805
Published online: 08 Feb 2017.
Submit your article to this journal
View related articles
View Crossmark data
Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=tbbb20
Download by: [University of Newcastle, Australia]
Date: 08 February 2017, At: 23:39

Bioscience, Biotechnology, and Biochemistry, 2017
Quercetin inhibits advanced glycation end product formation via chelating
metal ions, trapping methylglyoxal, and trapping reactive oxygen species
,†
,†
Mohammad Nazrul Islam Bhuiyan , Shinya Mitsuhashi , Kengo Sigetomi and Makoto Ubukata*
Division of Applied Bioscience, Graduate School of Agriculture, Hokkaido University, Sapporo, Japan
Received September 27, 2016; accepted December 29, 2016
http://dx.doi.org/10.1080/09168451.2017.1282805
2
+
2+
3,5–7,10,12–14)
Physiological concentration of Mg , Cu , and
(GO), has been proposed,
as far as we
2
+
Zn accelerated AGE formation only in glucose-
mediated conditions, which was effectively inhibited
by chelating ligands. Only quercetin (10) inhibited
MGO-mediated AGE formation as well as glucose-
and ribose-mediated AGE formation among 10
polyphenols (1–10) tested. We performed an addi-
tional structure-activity relationship (SAR) study on
flavanols (10, 11, 12, 13, and 14). Morin (12) and
kaempherol (14) showed inhibitory activity against
MGO-mediated AGE formation, whereas rutin (11)
and fisetin (13) did not. These observations indicate
that 3,5,7,4′-tetrahydroxy and 4-keto groups of 10
are important to yield newly revised mono-MGO
adducts (16 and 17) and di-MGO adduct (18) hav-
ing cyclic hemiacetals, while 3′-hydroxy group is not
essential. We propose here a comprehensive inhibi-
tory mechanism of 10 against AGE formation
including chelation effect, trapping of MGO, and
trapping of reactive oxygen species (ROS), which
leads to oxidative degradation of 18 to 3,4-dihydrox-
ybenzoic acid (15) and other fragments.
know, the comprehensive action mechanism of
flavonoids against AGE formation has not been fully
elucidated. In this article, we describe how quercetin,
one of the most abundant flavonols, efficiently inhibits
the formation of advanced glycation end products
(AGEs).
To determine the impact of metal ions on AGE for-
mation, we first performed the assays with or without
2
+
2+
2+
physiological concentration of Mg , Cu , and Zn in
glucose-, ribose-, and MGO-mediated AGE formations.
The structure-activity relationships of polyphenols were
determined using total 10 polyphenols: two catechin
monomers, (−)-epicatechin (1) and (−)-epigallocatechin
(2); a natural derivative of catechin, (−)-epigallocate-
chin-3-gallate (3); gallic acid (4), which is found both
as free acid and as part of hydrolysable tannin; a cate-
chin dimer, procyanidin B2 (5); a catechin trimer, pro-
cyanidin C1 (6); two isoflavons, genistein (7) and
daizein (8); a flavanone, hesperetin (9); and a flavonol,
quercetin (10). Quercetin (10) showed most potent inhi-
bitory effects in glucose-mediated or ribose-mediated
AGE formation assay, then we found a positive correla-
tion between the presence of chelating site and inhibi-
tory effects against glucose-mediated AGE formation.
Only quercetin (10) inhibited MGO-mediated AGE for-
mation, we therefore investigated the structure require-
ment for inhibitory effect of flavonol comparing 10
with rutin (11), morin (12), fisetin (13), and kaempferol
Key words: advanced glycation end products
(
(
AGEs); chelation effect; trapping ROS
reactive oxygen species); trapping
MGO (methylglyoxal); hemiacetal form
of di-MGO adduct
(14). Ketone at C-4 and 3,5,4′-triol of quercetin (10)
Polyphenols are ubiquitous secondary plant metabo-
lites, and their most important dietary sources are fruits,
vegetables, soybeans and beverages including green
and black tea, red wine, and beer. These ingredients
gather plenty of attention because they have many
potential effects for human health to treat diabetes,
allergy, asthma, cancer, cardiovascular disease, inflam-
seem to be essential for inhibition of MGO-mediated
AGE formation. Li et al. proposed the trapping mecha-
nism in which aldehyde moieties of one or two equiva-
lents of MGO were directly attacked by C-6 and C-8
of 10 to form quercetin mono-MGO adduct and querce-
tin di-MGO adduct base on LC-MS/MS analysis. We
suspected their proposed structures of MGO adducts
and isolated sufficient amounts of these compounds by
inhibiting degradation under oxygen-free conditions.
The true structures of MGO adducts were finally deter-
mined to be cyclic hemiacetals (16, 17 and 18) by
NMR analysis including NMR chemical shift calcula-
tions for each possible structure. We propose a new
10)
1
)
mation, and viral infectious disease. Especially, the
effectiveness of flavonoids for inhibition of advanced
glycation end product (AGE) formation has aroused
2
–11)
many researchers’ interest.
Although the restrictive
mechanism, that flavonoids can scavenge such reactive
carbonyl species as methylglyoxal (MGO) and glyoxal
*
Corresponding author. Email: m-ub@for.agr.hokudai.ac.jp
†
These authors contribute equally in this work.
©
2017 Japan Society for Bioscience, Biotechnology, and Agrochemistry

2
M.N.I. Bhuiyan et al.
inhibitory mechanism of 10 against MGO-mediated
AGE formations.
mixture was immediately frozen and stored at –83 °C
until use as an analytical control. The MGO mixture
was incubated at 37 °C for 16 h. The resultant mixture
was freeze-dried for 36 h, then allowed to melt for
Materials and methods
12 h, and finally freeze-dried again for another 5 days.
Chemicals and reagents.
3,4-Dihydroxybenzoic
The freeze-dried sample was extracted four times with
MeOH, and the combined solution was filtrated with
0.2 μm filter. The filtrate was concentrated in vacuo,
dissolved in DMSO, and purified by preparative ODS
C18 column (20 × 250 mm, 5 μm, Mightysil) using a
Hitachi HPLC system equipped with a diode array
detector. Millipore water with 0.1% AcOH and HPLC
grade MeOH with 0.1% AcOH were used for elution
as mobile phase B and mobile phase A, respectively.
Flow rate was 5 mL/min, and the isocratic HPLC con-
ditions were as follows: 0–30 min, 10% A and 90% B
followed by 15 min re-equilibration of the column with
UV detection at 254 nm. The injection volume was
200 μL, and 4.2 mg of 15, 0.2 mg of 16, 0.5 mg of 17,
and 1.2 mg of 18 were isolated, and each sample was
stored at –20 °C until use for nuclear magnetic
resonance (NMR) and electrospray ionization mass
spectrometry (ESI-MS) analyses.
acid and argon gas were purchased from Tokyo Chemi-
cal Industry (Tokyo, Japan) and Air Water Inc. (Osaka,
Japan), respectively. Methylglyoxal (MGO) solution
(
~40% in H O), bovine serum albumin (BSA),
2
dimethyl sulfoxide (DMSO), diethylenetriaminepen-
taacetic acid (DTPA), phosphate buffer saline (PBS, pH
7
.4), MeOH-d , DMSO-d , CuSO ZnSO4, MgSO4,
4 6 4,
and 96-well clear bottom plates were purchased from
Sigma-Aldrich (St. Louis, MO, USA). HPLC grade sol-
vents were from Kanto Chemical Co. Inc. (Tokyo,
Japan), quercetin dihydrate and other reagents were
from Wako Co. (Osaka, Japan). All chemicals were of
analytical grade.
Glucose-, ribose-, and methylglyoxal (MGO)-medi-
ated AGE formation assay.
and fluorescence detection of protein adducts were per-
formed in vitro as previously described method with
slight modifications. BSA-ribose or BSA-MGO assay
were conducted by essentially the same method except
for using ribose or MGO instead of glucose. In brief,
00 μL of reaction mixture containing 0.8 mg/mL
BSA, 200 mM ribose or 5 mM MGO with or without
.5 μL of 20 mM compound in 50 mM phosphate buf-
BSA-glucose glycation
2
)
Quantitative analysis of the products derived from
10 in MGO-mediated AGE formation. Quercetin (10)
(10.2 mg) was dissolved in 300 mL reaction solution
consisting of 50 mM phosphate buffer (pH 7.4) con-
5
taining 0.8 mg/mL BSA, 0.7 mM MgSO
CuSO 0.25 μM ZnSO The resultant solution
(300 mL) was divided into three equal parts (100 mL
each). Seventy-seven μL of H O was added instead of
, 0.75 μM
4
2
4
,
.
4
fer (pH 7.4) was prepared, and the reaction was
allowed to proceed at 37 °C for 24 h for BSA-ribose or
2
1
6 h for BSA-MGO assay. Trace metals, CuSO4
MGO solution to the first sample (100 mL), and the
control reaction was referred to as Condition A. This
sample was immediately frozen at –83 °C. Second sam-
ple (100 mL) was used for the conventional experiment
with 77 μL of 40% MGO and the reaction was referred
to as Condition B. The ratio of 10 to MGO was 1:50.
This sample was also immediately frozen at –83 °C.
Third sample (100 mL) was used for the experiment
with 77 μL of 40% MGO (10 mM) under oxygen-free
conditions and the reaction was referred to as Condi-
tion C. The ratio of 10 to MGO was 1:50. Removal of
oxygen from the reaction mixture was conducted by
gas change with argon using freeze-pump-thaw tech-
nique. This sample was immediately frozen and stored
at –83 °C until use. Each sample was incubated at
37 °C for 16 h, kept in a freezer at –83 °C for 40 min
to quench the reaction, and then lyophilized. The
freeze-dried sample was extracted four times with
MeOH, and the combined solution was filtrated with
0.2 μm filter. The filtrate was concentrated in vacuo,
and stored in freezer at –20 °C until analysis. HPLC
analysis of each MeOH extract was performed on a
Hitachi HPLC system equipped with a diode array
detector. Oxidative product 15, MGO-quercetin adducts
16, 17, 18, and starting material 10 were quantified
with UV detection at 254 nm using analytical ODS
C18 column (4.6 × 250 mm, 5 μm, Mightysil). HPLC
conditions were the same as those for preparative
HPLC except for flow rate, 1 mL/min and injection
volume, 20 μL.
(
0.75 μM), ZnSO4 (0.25 μM), and MgSO4 (0.7 mM)
were added in each reaction mixture to evaluate the
impact of metal ions on AGE formation. DTPA
(
(
0.1 mM and 1 mM) and 100 mM of aminoguanidine
AG, 1 mM) were used for positive controls as chelat-
ing agent and glycation inhibitor, respectively. The
fluorescence intensity (ex. 360 nm, em. 465 nm) of
AGE-related protein was measured using GENios FL
Fluorescent Microplate Reader (Tecan, Mannedorf,
Switzerland). The inhibitory activity was calculated by
subtracting the quenching effect for the apparent inhibi-
tory activity. One way-analysis of variance (ANOVA),
Tukey-Kramer method, and paired t-test were imple-
mented to determine statistical significance in a soft-
ware package, KaleidaGraph, ver 4.1 (Synergy
Software, Reading, PA, USA).
Large-scale preparation of reaction products derived
from quercetin (10) in MGO-mediated AGE formation
assay.
liter reaction solution consisting of 50 mM phosphate
buffer (pH 7.4) containing 0.8 mg/mL BSA, 0.7 mM
MgSO , 0.75 μM CuSO , and 0.25 μM ZnSO . To the
Quercetin (10) (34 mg) was dissolved in 1
4
4
4
resultant solution, 770 μL of H O (control mixture) or
2
7
70 μL of 40% MGO (MGO mixture) was added. The
ratio of 10 to MGO was 1:50 in which conditions were
the most effective to obtain sufficient amounts of the
reaction products in our experiment. The control

Action mechanism of quercetin as AGE inhibitor
3
NMR and ESI-MS analyses.
including H- H COSY, C DEPT-135 (distortionless
enhancement by polarization transfer), HSQC
heteronuclear single quantum correlation) and HMBC
heteronuclear multiple bond correlation) of 15 in
NMR measurements
Structure–activity relationships (SARs) on fourteen
different polyphenols (1–14)
1
1
13
To analyze structure–activity relationships (SARs)
between inhibitory activity against AGE-formation and
polyphenols, namely (–)-epicatechin (1), (–)-epigallo-
catechin (2), (–)-epigallocatechin gallate (3), gallic acid
(4), procyanidin B2 (5), procyanidin C1 (6), genistein
(7), daidzein (8), hesperetin (9), quercetin (10), and
four other flavonols (11–14), we evaluated the effects
of these polyphenols using three different types of
assay systems (Fig. 2). Since the inhibitory effect of
chelating agent was confirmed in glucose-mediated
AGE formation as shown in Fig. 1, inhibitory effects
on glucose-, ribose-, and MGO-mediated AGE forma-
tion were investigated using polyphenols having potent
(1, 2, 3, 4, 5, 6, 10), weak (7, 9), or no (8) chelating
unit as shown in Fig. 2.
(
(
MeOH-d , 16, 17, and 18 in DMSO-d were performed
4
6
with a Bruker AMX-500 (Bruker, MA, USA) at
GC-MS & NMR Laboratory, Faculty of Agriculture,
Hokkaido University. Chemical shift was reported in δ
ppm using tetramethylsilane as the internal standard,
and coupling constants (J) were given in hertz. All
ESI-MS spectra were recorded on LTQ-Orbitrap XL
ThermoScientific, MA, USA) using a Paradigm MS2
Michrom BioResources, CA, USA). NMR chemical
shifts of 16, 17, 18, and their isomers were predicted
using ChemNMR (Upstream Solutions, Zurich,
Switzerland) in ChemDraw Ultra, ver. 12.0
(
(
(
CambridgeSoft, MA, USA).
Such polyphenols having vicinal diol as 1, 2, 3, 4, 5,
6
, and 10 exhibited the significant inhibitory activities
on glucose- and ribose-mediated AGE formation,
whereas such polyphenols having C-4 keto and C-5
hydroxy groups as genistein (7) and hesperetin (9)
showed only moderate inhibitory activities in glucose-
mediated AGE formation (p < 0.01). Daidzein (8) lack-
ing the typical metal chelation site did not inhibit AGE
formation mediated by glucose (Fig. 2(a)).
Results and discussion
Impact of metal ions in AGE formation
Chelation is one of the fundamental mechanisms of
action occurring in wide range of polyphenols
may affect AGE formation in diabetes. AGE inhibitors
may act primarily as chelators, because metal-catalyzed
1
5,16)
that
These results suggested that chelating effects of
polyphenols were important for their inhibitory activity
on AGE formation, especially on glucose-mediated
AGE formation. Meanwhile, all these polyphenols
including daidzein (8) showed inhibitory activities
against ribose-mediated AGE formation (p < 0.01),
while quercetin (10) exhibited most potent inhibitory
activities against both glucose- and ribose-mediated
AGE formation (Fig. 2(a)). In addition, only 10 signifi-
cantly inhibited MGO-mediated AGE formation among
the above 10 polyphenols as indicated in Fig. 2(b).
Quercetin (10) was thus found to be the most potent
inhibitor against glucose-, ribose-, and MGO-mediated
AGE formations among the test compounds (1–10).
The flavonol 10 was therefore considered as a standard
AGE inhibitor in comparison with other flavonoids. To
determine the structural requirements of quercetin (10)
as the inhibitor of MGO-mediated AGE formation,
such other flavonols as rutin (11), morin (12), fisetin
(13), and kaempferol (14) were evaluated for their
effects on MGO-mediated AGE formation as indicated
in Fig. 2(b). The importance of C-3 hydroxy group and
C-5 hydroxy group of 10 was confirmed using rutin
(11) whose C-3 hydroxy was blocked by glycosylation
and fisetin (13) lacking C-5 hydroxy group. Both 11
and 13 lost efficacies against MGO-mediated AGE
formation.
1
7)
oxidation reactions accelerate AGE formation.
In this study, the effects of physiological concentra-
2
+
2+
2+
tion of metal ions, Cu , Zn , and Mg were investi-
gated on glucose-, ribose-, and MGO-mediated AGE
formation. Surprisingly, only glucose-mediated AGE
formation was significantly enhanced by metal ions,
whereas it was remarkably inhibited by DTPA having
high affinity for metal cations (p < 0.01). These results
suggested that metal ions were indispensable for the
AGE formation mediated by glucose. Although
ribose-mediated AGE formation was slightly enhanced
by metal ions (p < 0.05), MGO-mediated AGE
formation was not affected by these metal cations.
Aminoguanidine (AG) exhibited inhibitory activity
against glucose-, ribose-, and MGO-mediated AGE for-
mation in the presence of metal ions as shown in
Fig. 1. It is possible to explain why the effect of metal
ions was far weaker in ribose-mediated AGE formation
than in glucose-mediated AGE formation. A study
revealed that ribose is an effective glycation agent com-
1
8)
pared to glucose both in vitro and in vivo. Conforma-
tional change from helix to β-sheet occurs during
glycation of lysozyme, and rapid perturbation of the
1
9)
tertiary structure occurs in the presence of ribose. In
addition, divalent metal ions bind to 2,3-vicinal diol of
2
0,21)
ribose in aqueous solution,
whereas glucose che-
2
2)
lates such metal ions very poorly at neutral pH.
A
Catechol moiety was considered to act as the essen-
tial functional group for metal chelation in epicatechin
(1), epigallocatechin (2), epigallocatechin gallate (3),
gallic acid (4), procyanidin B2 (5), procyanidin C1 (6),
quercetin (10), rutin (11), or fisetin (13). These com-
pounds inhibited glucose-mediated and ribose-mediated
AGE formation, but compounds 1–6, 11, and 13 did
not have efficient inhibitory activities against
MGO-mediated AGE formation.
potent chelating effect of ribose might be a factor for
accelerating AGE formation by coordinating the
ribose–metal complex to arginine, lysine, asparagine,
glutamine, aspartic acid, or glutamic acid residue of
2
3,24)
protein.
Glucose, which has weak chelating effect
needs additional amounts of metal ions for acceleration
of AGE formation. Thus chelating effect of inhibitor
may be a limiting factor for glucose-mediated AGE for-
mation, whereas it would be only a minor factor for
ribose-mediated AGE formation.
Thus, it seemed that the inhibitory effect of quercetin
(10) against MGO-mediated AGE formation was

4
M.N.I. Bhuiyan et al.
H and C NMR data of isolated 3,4-dihydroxybenzoic
1
13
Table 1.
4
acid (15) in MeOH-d .
δ
δ
Position
C, type
H, (J in Hz)
HMBC
C=O
171.2, s
124.2, s
118.6, d
146.8, s
152.2, s
116.6, d
124.7, d
1
2
3
4
5
6
7.43, d, 1H (2.1)
3, 4, 6, C=O
6 6.80, d, 1H (8.1)
3, 4, 1
7.41, dd, 1H (8.1, 2.1)
2, 4, C=O
regulated by other mechanisms. Since morin (12) and
kaempferol (14) still retained their inhibitory activities
against MGO-mediated AGE formation, C-3′ hydroxyl
group seemed not to be essential for the inhibitory
activity of 10 against MGO-mediated AGE-formation.
Due to many evidences that increase in MGO levels in
serum can lead to development of diabetic complica-
Fig. 1. Effects of metal ions in glucose-BSA, ribose-BSA, and
MGO-BSA mediated AGE formation.
Notes: Diethylenetriaminepentaacetic acid (DTPA) and aminoguani-
dine (AG) were used as positive controls for removal of metal ions
and for inhibition of AGE formation, respectively. Bars represent
mean ± standard deviation of triplicate values. Paired t-test:
2
5–27)
tions,
many scavenging agents such as
aminoguanidine, tenilsetam, metformin, and pyridoxam-
2
8,29)
ine were introduced.
However, the side effects of
*
*p < 0.01, *p < 0.05, N.S. p = not significant).
these agents have prompted a search for less toxic
MGO-trapping agents. Li et al. independently reported
quercetin (10) as an efficient MGO-trapping agent.
1
0)
structures of MGO adducts of 10 or other flavonols. To
determine the structures of MGO-quercetin adducts
more precisely, we conducted the 1-L scale reaction of
10 with MGO in the presence of BSA and metal ions.
Contrary to expectation, the major product in this
reaction showed a pseudo-molecular ion at m/z 153
Identification of 3,4-dihydroxybenzoic acid (15)
derived from quercetin (10) in MGO-mediated AGE
formation
−
The structures of mono-MGO and di-MGO adducts
of quercetin (10) in MGO-mediated AGE formation
[M-H] in the negative-ion ESI-MS. High-resolution
electrospray ionization mass spectrometry (HR-ESl-
MS) indicated the molecular formula of C H O for
1
0)
have been proposed based on LC-MS/MS analyses.
However, there are yet no rigorous evidence for the
7
6 4
the major compound. UV-visible spectrum showed
Fig. 2. Inhibitory effects of polyphenols against glucose-BSA, ribose-BSA, and MGO-BSA mediated AGE formation.
Notes: (a) Glucose- or ribose-mediated AGE formation in the presence of each polyphenol (1–10). (b) MGO-mediated AGE formation in the
presence of each polyphenol (1–14). The concentration of each compound was 100 μM. Bars represent mean ± standard deviation of triplicate
values.

Action mechanism of quercetin as AGE inhibitor
5
three peaks at 210, 260, and 294 nm, respectively. The
structure of the major product was unambiguously
mixture of quercetin (10) without MGO in the presence
of BSA and metal ions. In this condition, the starting
material 10 was detected as a single peak in HPLC
shown as Condition A in Fig. 3(a) and (b). In second
experiment, a 1:50 mixture of 10 and MGO was incu-
bated in the presence of BSA and metal ions, and five
remarkable products were detected in HPLC indicated
as Condition B in Fig. 3(a) and (b). The products puri-
fied by HPLC were determined to be 3,4-dihydroxy-
benzoic acid (15), two mono-MGO adducts (16 and
17), di-MGO adduct (18), and the starting material 10
by ESI-MS spectral analyses. The third experiment was
performed in a 1:50 reaction mixture of quercetin (10)
and MGO in the presence of BSA and metal ions under
anoxic conditions. In this reaction, the major product
was assigned to be 18 and two minor products were
assigned to be 16 and 17 by HPLC as shown in
Condition C in Fig. 3(a) and (b). The structures were
confirmed by HR-ESI-MS of isolated samples.
1
determined to be 3,4-dihydroxybenzoic acid (15) by H
1
3
NMR and C NMR including 2D NMR analyses
Table 1), which was further confirmed by comparing it
(
with an authentic sample by HPLC.
It was suggested that oxidative degradation of 10
proceeded to give 15 by prolonged treatment with dis-
solved oxygen in the reaction mixture. If this hypothe-
sis was correct, the formation of 15 should be
suppressed by removing oxygen from the reaction mix-
ture in the assay system. To confirm this, we designed
three different experiments including anoxic conditions.
Productivities of 3,4-dihydroxybenzoic acid (15) and
MGO-quercetin adducts (16, 17, and 18) from 10 in
BSA-metal ion solution with or without MGO under
ambient or anoxic conditions
A model system under anoxic conditions provided a
sufficient amount of MGO-quercetin adducts as
expected (Fig. 3). The objective of the newly designed
experiments was mainly to confirm the impact of oxy-
gen. The first experiment was performed in a reaction
These results indicated that trapping of MGO in 10
proceeded without involvement of oxygen and resul-
tant di-MGO adduct (18) was degraded by oxygen to
give 15 under the ambient air level as indicated in
Fig. 4.
Fig. 3. Quantitative analysis of degradation product (15), MGO adducts (16, 17, 18) and quercetin (10).
Notes: (a) HPLC profiles of the samples originated from Condition A without MGO, Condition B with MGO, and Condition C with MGO under
an argon atmosphere. Column conditions: Mightysil (4.6 × 250 mm, 5 μm), 10% mobile phase A (H
2
O, 0.1% AcOH) and 90% mobile phase B
(
MeOH, 0.1% AcOH), 1.0 mL/min. Detection wavelength: 254 nm. (b) The relative yield of oxidative degradation product (15), di-MGO quercetin
adduct (18), mono-MGO quercetin adducts (16 and 17), and reduction rate of quercetin (10) in MGO-mediated AGE formation under Condition A,
Condition B or Condition C. The relative yield was calculated from each peak area value at 254 nm in HPLC.

6
M.N.I. Bhuiyan et al.
Fig. 4. Proposed mechanism for the fate of quercetin (10) in MGO-mediated AGE formation under ambient air conditions.
Notes: *Mono-MGO adduct (17) is shown as a representative. Trapping of two equivalent of MGO by quercetin (10) followed by trapping of
one equivalent of reactive oxygen species at C2 position of di-MGO adduct (18) to give 3,4-dihydroxy benzoic acid (15).
determined to be C H O by HR-ESI-MS. Fig. 5
1
8 14 9
indicates the possible structures of 16 and 17 based on
the molecular formula of each compound and NMR
assignment of di-MGO adduct (18) shown in Fig. 6(a).
The structure of mono-MGO adduct (17) was sup-
ported by its NMR using 0.5 mg of the sample and
NMR chemical shift calculations for hemiacetal or
hemiketal of mono-MGO adducts (Fig. 6(b)).
Fig. 5. Structure of mono-MGO adducts, 16 and 17.
The structures of mono-MGO adducts (16 and 17)
identified by HR-ESI-MS
Structural determination of di-MGO adduct (18) and
fate of quercetin (10) in MGO-mediated AGE
formation
The structural information of di-MGO adduct (18)
was obtained from NMR and ESI-MS in negative ion
Mono-MGO adducts (16 and 17) indicated the same
–
pseudo molecular ion at m/z 373 [M-H] in negative
ion mode. The molecular formulas of 16 and 17 were

Action mechanism of quercetin as AGE inhibitor
7
Fig. 6. Structure determination of di-MGO adduct (18) and mono-MGO (17).
Notes: (a) Observed NMR chemical shifts of 18 (I) and chemical shifts calculated for hemiacetal (II) or hemiketal (III). Hemiketal can be formed
10)
from di-MGO adduct of quercetin proposed by Li et al.
possible structures (II and III). ChemNMR was used for chemical shift calculations of
2.0.3.1216 (CambridgeSoft Corp, Cambridge, MA, USA).
(b) Observed NMR chemical shifts of 17 (I) and chemical shifts calculated for two
1H
13C
and
NMR in each structure on ChemDraw Ultra, ver.
1
mode, indicating a pseudo molecular ion at m/z 445
indicating a singlet at 7.7 ppm for 2′-H and a pair of
doublets at 7.5 and 6.86 ppm for 6′-H and 5′-H in B
ring, respectively, whereas no proton signal at C-6 and
C-8 in A ring was observed. HMQC of 18 indicated
that 4 singlet protons at 4.8–4.9 ppm assigned as hemi-
acetal protons correlated with the carbon signals at
–
[
M-H] . Molecular formula of 18 was determined to be
C H O by HR-ESI-MS. Although limited amount
2
1 18 11
of 18 existed as at least 8 distinguishable diastereomers
made it difficult to assign completely the NMR signals,
1
H NMR supported the proposed structure of 18

8
M.N.I. Bhuiyan et al.
around 101 ppm, whereas C–H correlations for 4 sin-
glet protons at around 5.0 ppm could not be detected
due to the lack of adequate S/N ratio, while 8 singlet
protons at 1.2–1.4 ppm assigned to methyl protons cor-
related with carbons at around 16 and 21.9 ppm. Mean-
while HMBC of 18 showed the long-range couplings
between the methyl protons at 1.2–1.4 ppm and car-
bons approximately at 106–112 ppm assigned to C6,
C8, and hemiacetal carbon.
cleavage by hydrolysis is unlikely in view of organic
chemistry. Cleavage of α-dicarbonyl compounds via
Baeyer–Villiger-like rearrangement was unambiguously
3
1)
evidenced in our previous work, thus hydrogen per-
oxide would attack the carbonyl group to afford acid
anhydride. Hydrolysis of acid anhydride would yield
3,4-dihydroxybenzoic acid (15) as indicated in Fig. 4.
In this article, we proposed three different efficacies
of quercetin (10) for inhibiting AGE formation. First,
10 functions as a metal chelating agent having catechol
Figure 6(a) indicates the comparison between NMR
assignment of di-MGO adduct (18) and NMR chemical
shifts calculated for newly proposed hemiacetal and for
hemiketal which can be formed from di-MGO adduct
3
2)
moiety and two other chelating sites, which suppress
glucose-mediated AGE formation. Excess amount of
glucose in human tissues is converted into 3-deoxy glu-
cosone (3DG) via Maillad reaction that leads to forma-
tion of such AGEs as glucosepane, versperlysine, and
1
0)
proposed by Li et al. The NMR data of 18 coincided
with the NMR chemical shifts for hemiacetal rather
than hemiketal. Aldehyde of MGO can be attacked first
by phenolic oxygen atom, which makes possible to
move the methyl ketone moiety closer to C6 or C8 in
1
7)
GLUCOLD.
Second, 10 traps two equivalents of
MGO, one of the major causes of accumulations of
AGEs in human tissues, and its scavenging effect
results in inhibition of MGO-mediated AGE formation.
Methylglyoxal (MGO) is produced via glycolysis, and
1
0. The optimal orientation effect could accelerate
intramolecular aromatic substitution at C6 or C8 to
form 18 as shown in Fig. 4. When quercetin (10) was
incubated with MGO at a ratio of 1:50 in the presence
of BSA and metal ions under anoxic conditions, the
accumulation of di-MGO adduct (18) was remarkably
enhanced in comparison with 3,4-dihydroxybenzoic
acid (15) or mono-MGO adducts (16 and 17). The
amount of unreacted quercetin (10) also dramatically
decreased in this experiment as shown in condition C
of Fig. 3. The structure of di-MGO adduct (18) and
overall oxidative degradation steps are illustrated in
Fig. 4.
1
7)
its concentration is elevated in diabetic patients.
Excess amount of MGO leads to formation of such
1
7)
AGEs as MOLD, MOLDIC, and argpyrimidine.
Third, di-MGO adduct (18) derived from 10 rapidly
consumes oxygen which accelerates formation of
AGEs. Oxygen contributes to oxidative degradation of
glucose with catalytic metals that leads to formation of
3
3)
pentosidine,
and oxidative degradation of lipid by
oxygen produces glyoxal (GO) which leads to the for-
mation of such advanced lipoxidation end products
(ALEs) or AGEs as GODIC, GOLA, GOLD, and
1
7,33)
CML.
Quercetin (10) has been recognized as one of the
most important ingredients in herbal medicines or sup-
plements, and the compound is found in a number of
plants such as onion, cocoa powder, cranberries, lin-
gonberries, kale, celery, broccoli, looseleaf lettuce, ripe
Conclusion
Structure–activity relationship (SAR) studies on 10
polyphenols (1–10) indicated that quercetin (10) was
the most potent inhibitor against AGE formation in glu-
cose-, ribose-, or MGO-mediated assay system. The
structural requirements of 10 for potent inhibition of
MGO-mediated AGE formation were determined using
additional four flavonols (11–14), indicating the impor-
tance of the 3,5,7,4′-tetrahydroxy groups and the C-4
keto group in 10 (Figs. 1 and 2). The reaction of 10
with MGO in the presence of BSA and metal ions
yielded four different types of products. The structures
of these products were determined to be hemiacetal
form of di-MGO adduct (18), two hemiacetal forms of
mono-MGO adducts (16 and 17), and 3,4-dihydroxy-
benzoic acid (15), respectively (Table 1, Figs. 3–6).
Under the conventional conditions, the oxidative degra-
dation of 10 was accelerated by MGO to form 3,4-di-
hydroxybenzoic acid (15). By contrast, predominant
accumulation of di-MGO adduct (18) was observed
under oxygen-free conditions, while degradation of 18
was suppressed (Fig. 3). The C-6 and C-8 positions of
the A ring in quercetin (10) were the major reactive
sites for trapping MGO to form both mono-MGO and
di-MGO adducts, and molecular oxygen oxidized di-
MGO adduct (18) to quinone methide by generating
hydrogen peroxide (Fig. 4). Inai et al. proposed a simi-
lar mechanism for oxidative degradation of quercetin
3
4)
tomatoes, and carrots. In addition, all such pharma-
3
5)
36,37)
41,42)
ceutical agents as aminoguanidine,
tenilsetam,
3
8)
39,40)
carnosine,
metformin,
and pyridoxamine,
developed for inhibiting AGE formation and preventing
the development of diabetic complications, have serious
1
2)
side effects. Thus, intervention using the food-borne
polyphenols like 10 is likely to be an effective strategy
to inhibit the formation of AGEs and prevent AGE-me-
diated processes linked to such age-related diseases as
diabetes complication, Alzheimer’s disease, and
3
1,42,43)
arteriosclerosis.
Author contributions
S.M and M.U designed and directed the project. S.M
performed experiments in Figs. 1–3 and Table 1.
M.N.I.B performed experiments in Table
1 and
Figs. 3–6. S.M, M.N.I.B, K.S, and M.U analyzed the
data. M.N.I.B and M.U wrote the article. All authors
have read and approved the final manuscript.
Disclosure statement
(
10) via quinone fornation, water addition, and hydrol-
3
0)
No potential conflict of interest was reported by the authors.
ysis.
However, the final step including C-C bond

Action mechanism of quercetin as AGE inhibitor
synthesized by sol–gel method.
011;60:41–47.
9
J
Sol–Gel Sci Technol.
References
2
[
1] Benavente-García O, Castillo J. Update on uses and properties
of citrus flavonoids: new findings in anticancer, cardiovascular,
and anti-inflammatory activity. J Agric Food Chem. 2008;56:
[23] Di Costanzo L, Flores LV, Christianson DW. Stereochemistry of
guanidine-metal interactions: Implications for L-arginine-metal
interactions in protein structure and function. Proteins. 2006;65:
637–642.
6185–6205.
[
[
[
[
[
[
[
[
2] Matsuura N, Aradate T, Sasaki C, et al. Screening system for
the Maillard reaction inhibitor from natural product extracts. J
Health Sci. 2002;48:520–526.
3] Matsuda H, Morikawa T, Ando S, et al. Structural requirements
of flavonoids for nitric oxide production inhibitory activity and
mechanism of action. Bioorg Med Chem. 2003;11:1995–2000.
4] Wu CH, Yen G. Inhibitory effect of naturally occurring flavo-
noids on the formation of advanced glycation endproducts. J
Agric Food Chem. 2005;53:3167–3173.
5] Sang S, Shao X, Bai N, et al. Tea polyphenol (−)-epigallocate-
chin-3-gallate: a new trapping agent of reactive dicarbonyl spe-
cies. Chem Res Toxicol. 2007;20:1862–1870.
6] Shao X, Bai N, He K, et al. Apple polyphenols, phloretin and
phloridzin: new trapping agents of reactive dicarbonyl species.
Chem Res Toxicol. 2008;21:2042–2050.
7] Lv L, Shao X, Chen H, et al. Genistein inhibits advanced glyca-
tion end product formation by trapping methylglyoxal. Chem
Res Toxicol. 2011;24:579–586.
8] Li D, Mitsuhashi S, Ubukata M. Protective effects of hesperidin
derivatives and their stereoisomers against advanced glycation
end-products formation. Pharm Biol. 2012;50:1531–1535.
9] Li D, Shigetomi K, Mitsuhashi S, et al. Maillard reaction inhibi-
tors produced by Paecilomyces sp. Biosci Biotechnol Biochem.
[24] Gold MH, Youngs HL, Sollewijn G, et al. In: Sigel A, Sigel H,
editors. Metal ions in biological systems: manganese and its role
in biological processes. Vol. 37, New York (NY): Marcel
Dekker, Inc; 2000. p. 559–586.
[25] Shamsi FA, Partal A, Sady C, et al. Immunological evidence for
methylglyoxal-derived modifications in vivo. Determination of
antigenic epitopes. J Biol Chem. 1998;273:6928–6936.
[26] Beisswenger P, Howell S, Nelson R, et al. α-Oxoaldehyde meta-
bolism and diabetic complications. Biochem Soc Trans. 2003;31:
1358–1363.
[27] Kilhovd BK, Juutilainen A, Lehto S, et al. Increased serum
levels of methylglyoxal-derived hydroimidazolone-AGE are
associated with increased cardiovascular disease mortality in
nondiabetic women. Atherosclerosis. 2009;205:590–594.
[28] Thomas M, Baynes J, Thorpe S, et al. The role of AGEs and
AGE inhibitors in diabetic cardiovascular disease. Curr Drug
Targets. 2005;6:453–474.
[29] Cervantes-Laurean D, Schramm DD, Jacobson EL, et al. Inhibi-
tion of advanced glycation end product formation on collagen by
rutin and its metabolites. J Nutr Biochem. 2006;17:531–540.
[30] Inai M, Miura Y, Honda S, et al. Metmyoglobin reduction by
polyphenols and mechanism of the conversion of metmyoglobin
to oxymyoglobin by quercetin. J Agric Food Chem. 2014;62:
893–901.
2013;77:2499–2501.
[
10] Li XL, Zheng T, Sang S, et al. Quercetin inhibits advanced
glycation end product formation by trapping methylglyoxal and
glyoxal. J Agric Food Chem. 2014;62:12152–12158.
[31] Nagamatsu R, Mitsuhashi S, Shigetomi K, et al. Cleavage of
α-dicarbonyl compounds by terpene hydroperoxide. Biosci
Biotechnol Biochem. 2012;76:1904–1908.
[
11] Matsuura N, Aradate T, Kurosaka C, et al. 2014. Potent protein
glycation inhibition of plantagoside in Plantago major seeds.
BioMed Res Int. 2014;2014. ID 208539. Available from http://
dx.doi.org/10.1155/2014/208539.
12] Shao X, Chen H, Zhu Y, et al. Essential structural requirements
and additive effects for flavonoids to scavenge methylglyoxal. J
Agric Food Chem. 2014;62:3202–3210.
[32] Kasprzak MM, Erxleben A, Ochocki J. Properties and
applications of flavonoid metal complexes. RSC Adv. 2015;5:
45853–45877.
[33] Peyroux J, Sternberg M. Advanced glycation endproducts
(AGEs): pharmacological inhibition in diabetes. Pathol Biol.
2006;54:405–419.
[34] Aguirre L, Arias N, Macarulla MT, et al. Beneficial effects of
quercetin on obesity and diabetes. Open Nutraceuticals J. 2011;
4:189–198.
[35] Thomas MC, Baynes JW, Thorpe SR, et al. The role of AGEs
and AGE inhibitors in diabetic cardiovascular disease. Curr Drug
Targets. 2005;6:453–474.
[36] Webster J, Urban C, Berbaum K, et al. The carbonyl scavengers
aminoguanidine and tenilsetam protect against the neurotoxic
effects of methylglyoxal. Neurotoxic Res. 2005;7:95–101.
[37] Price DL, Rhett PM, Thorpe SR, et al. Chelating activity of
advanced glycation end-product inhibitors. J Biol Chem. 2001;
276:48967–48972.
[38] Beisswenger P, Ruggiero-Lopez D. Metformin inhibition of
glycation processes. Diabetes Metab. 2003;29:6S95–6S103.
[39] Beisswenger PJ, Howell SK, Touchette AD, et al. Metformin
[
[
[
[
[
[
[
[
[
[
[
13] Yoon SR, Shim SM. Inhibitory effect of polyphenols in Houttuy-
nia cordata on advanced glycation end-products (AGEs) by trap-
ping methylglyoxal. LWT-Food Sci Technol. 2014;61:158–163.
14] Lo CY, Li S, Tan D, et al. Trapping reactions of reactive car-
bonyl species with tea polyphenols in simulated physiological
conditions. Mol Nutr Food Res. 2006;50:1118–1128.
15] Mira L, Fernadez MT, Santos M, et al. Interactions of flavonoids
with iron and copper ions: a mechanism for their antioxidant
activity. Free Radical Res. 2002;36:1199–1208.
16] Zhang W, Chen C, Shi H, et al. Curcumin is a biologically
active copper chelator with antitumor activity. Phytomedicine.
2016;23:1–8.
17] Vistoli G, De Maddis D, Cipak A, et al. Advanced glycoxidation
and lipoxidation end products (AGEs and ALEs): an overview of
their mechanisms of formation. Free Radical Res. 2013;47:3–27.
18] Han C, Lu Y, Wei Y, et al. D-ribose induces cellular protein
glycation and impairs mouse spatial cognition. PLoS One.
reduces systemic methylglyoxal levels in type
2 diabetes.
Diabetes. 1999;48:198–202.
[40] Nagaraj RH. Effect of pyridoxamine on chemical modification
of proteins by carbonyls in diabetic rats: characterization of a
major product from the reaction of pyridoxamine and methylgly-
oxal. Arch Biochem Biophys. 2002;402:110–119.
[41] Voziyan PA, Metz TO, Baynes JW, et al. A post-amadori inhibi-
tor pyridoxamine also inhibits chemical modification of proteins
by scavenging carbonyl intermediates of carbohydrate and lipid
degradation. J Biol Chem. 2002;277:3397–3403.
[42] Reddy VP, Zhu X, Perry G, et al. Oxidative stress in diabetes
and Alzheimer’s disease. J Alzheimers Dis. 2009;16:763–774.
[43] Jahan H, Choudhary MI. Glycation, carbonyl stress and AGEs
inhibitors: a patent review. Expert Opin Ther Pat. 2015;25:
1267–1284.
2011;6:e24623.
19] Ghosh S, Pandey NK, Roy AS, et al. Prolonged glycation of
hen egg white lysozyme generates non amyloidal structures.
PLoS One. 2013;8:e74336.
20] Luckey TD, Venugopal B. Metal toxicity in mammals. Vol. 1,
Physiologic and chemical basis for metal toxicity. New York
(
NY): Springer; 1977.
21] Symons MCR, Benbow JA, Pelmore H. Study of calcium ion bind-
ing to D-ribose in aqueous solutions using hydroxy-proton reso-
nance shifts. J Chem Soc Faraday Trans 1. 1982;78:3671–3677.
22] Ye Y, Liu Q, Wang J. Influence of saccharides chelating agent
2
on particle size and magnetic properties of Co Z hexaferrite