Oligomeric oxidation products of the flavonoid quercetin

Article abstract of DOI:10.1007/s10600-008-9092-1

Some of the main oxidation products of quercetin were shown to be compounds formed by oligomerization of the starting flavonoid. Conditions for the preparative synthesis of these compounds were developed. Their structures were established using HPLC-MS an

Full text of DOI:10.1007/s10600-008-9092-1

Chemistry of Natural Compounds, Vol. 44, No. 4, 2008
OLIGOMERIC OXIDATION PRODUCTS OF
THE FLAVONOID QUERCETIN
E. M. Chervyakovsky,1,2 D. A. Bolibrukh,1
D. L. Kurovskii,1 A. A. Gilep,1 T. M. Vlasova,2
V. P. Kurchenko,2 and S. A. Usanov1*
UDC 547.814.5+577.151.03
Some of the main oxidation products of quercetin were shown to be compounds formed by oligomerization
of the starting flavonoid. Conditions for the preparative synthesis of these compounds were developed. Their
structures were established using HPLC—MS and NMRmethods. Quercetin oligomers in the natural sample,
Allium cepa
outer leaves of modified runners of
L., were found using chromatographic procedures. The use
of quercetin oligomers as indicators of its oxidation was proposed.
Key words:
quercetin, oxidation products,
, redox reactions, NMR.
Allium cepa
We have previously shown that oxidation of quercetin
produces primarily oligomeric products without any
in vitro
significantaccumulation ofhydroxybenzoicacidderivatives[1]. The question ofthe existence ofoligomeric quercetin oxidation
products remains open. Herein we describe possible formation pathways and structural properties of the two main
in vivo
oligomeric oxidation products of the flavonoid. The possibility of using data on the physical chemical properties of these
compounds to search for them in biological samples is discussed.
It was shown previously that quercetin was chemically modified upon reaction with certain hemoproteins of the and
b
types [1]. As a result, several oxidation products, the principal ones of which were product B, a dimer of the starting
c
compound, and product C, the structure of which could not be established, were formed. Use of hemoproteins as oxidants did
not produce product C in quantities required for structural studies. The main obstacle was the limited solubility of quercetin
in the aqueous solutions that were used to carry out the oxidation. Addition of an organic solvent to the reaction mixture could
have increased the solubility of quercetin. However, this would have made it impossible to use proteinaceous oxidants because
of their possible denaturation. We used K [Fe(CN) ] as an alternate oxidant that is not destroyed in organic media. The redox
3
6
potential of K [Fe(CN) ] is high and positive, like that of cytochrome [2, 3].
c
3
6
HO
3'
OH
HO
7
8
5'
8a
4
OH
HO
6
O
2
1'
8
7
OH
7
O
5
8a
4
5'
4a
8
6
5
3
O
HO
HO
1'
2
8a
4
5'
O
6
5
O
O
O
4a
1'
2
HO
3
OH
O
O
3'
2
4a
i
8a
8
3
OH
HO
O
3
O
O
2
3'
7
4
8a
8
6
3
HO
HO
O
ii
O
4a
OH
ii
7
4
O
HO
6
5
4a
OH
O
OH
i
5
OH
B
iii
C
1) Institute of Bioorganic Chemistry, National Academy of Sciences, Belarus, Republic of Belarus, Minsk, 220141,
ul. Akad. Kuprevicha, 5/2, e-mail: usanov@iboch.bas-net.by; 2) Belorusian State University, Republic of Belarus, Minsk,
220050, pr. Nezavisimosti, 4. Translated from Khimiya Prirodnykh Soedinenii, No. 4, pp. 344-347, July-August, 2008.
Original article submitted November 1, 2007.
0009-3130/08/4404-0427 ©2008 Springer Science+Business Media, Inc.
427

TABLE 1. PMR Spectra for Dimer B and Trimer C (δ, J/Hz)
Subunits
H atom
i
ii
iii
C
C
B
C
B
H-6 6.00 (1H, 8d, J = 2.1)
6.03 (d, J = 2.1)
6.27 (1H, 8d, J = 2.1)
6.28 (d, J = 2.1) and
6.29 (d, J = 2.1)
6.61 (d, J = 2.1) and
6.63 (d, J = 2.1)
6.04 (0.5H, 3d, J = 2.1)
6.06 (0.5H, 2d, J = 2.1)
6.13 (0.3H, 2d, J = 2.1)
6.15 (0.2H, 2d, J = 2.1)
6.155 (0.25H, d, J = 2.1)
6.165 (0.25H, d, J = 2.1)
7.02 (0.5H, 4dd, J = 0.3; 8.6)
7.15 (0.5H, 2d, J = 8.6)
H-8 6.03 (0.2H, 2d, J = 2.1)
6.05 (0.8H, 6d, J = 2.1)
6.08 (d, J = 2.1)
6.82 (d, J = 8.4)
6.59 (1H, m)
′
H-5 6.76 (0.5H, 4d, J = 8.5)
7.17 (0.2H, 2d, J = 0.3; 8.7)
7.18 (0.3H, 2d, J = 0.3; 8.7)
7.31 (0.5H, m)
7.17 (d, J = 8.7) and
7.29 (d, J = 8.7)
6.77 (0.5H, 2d, J = 8.5)
′
H-6 7.09 (1H, 4dd, J = 2.4; 8.5) 7.17 (dd, J = 2.4; 8.4) and 7.98 (0.7H, 3dd, J = 2.1; 8.7)
7.98 (dd, J = 2.2; 8.7) 7.46 (1H, m)
8.01 (dd, J = 2.2; 8.7)
7.19 (dd, J = 2.4; 8.4)
7.35 (d, J = 2.4) and
7.36 (d, J = 2.4)
8.00 (0.3H, m)
′
H-2 7.29 (1H, 2d, J = 2.4)
7.915 (0.3H, 2dd, J = 0.3; 2.1)
7.925 (0.3H, 2d, J = 2.1)
8.04 (0.2H, 2dd, J = 0.3; 2.1)
8.06 (0.2H, 2d, J = 2.1)
7.91 (d, J = 2.2) and
8.05 (d, J = 2.2)
7.43 (0.5H, m)
7.56 (0.1H, 2d, J = 0.3; 2.4)
7.58 (0.3H, 2dd, J = 0.3; 2.4)
7.59 (0.1H, 2d, J = 0.3; 2.4)
HPLC studies ofthe reaction products showed that the qualitative compositions ofthe reaction mixtures obtained using
cytochrome c and K [Fe(CN) ] were identical. However, the quantitative compositions were slightly different. The quercetin
3
6
content varied in the range of 30%; product B, 35%; product C, 17%. Use of acetone as the solvent for preparative purposes
could increase the quercetin content in the reaction mixture and; therefore, the overall yield of the oxidation products.
Chromatographic separation produced highly pure products B and C. Product B was the conjugate of two quercetin
molecules joined through a dioxane ring [1, 4-10]. Product C was most probably a trimer of quercetin, as was demonstrated
earlier using HPLC—MS [1]. An attempt was made to analyze spectra of the trimeric adduct of this flavonoid [4]. However,
some ambiguity remained in the assignments of certain resonances. Use of acetone-d as solvent enabled all resonances in the
6
spectrum to be assigned. Chemical shifts and spin—spin coupling constants (SSCC) in the PMR spectrum of fragments (i) and
(ii) of the trimeric adduct (Table 1) turned out to be close to those for compound B [1, 4, 7]. This suggests that the principal
fragments of B and C had similar chemical structures and allowed the trimeric structure to be proposed for C. Use ofacetone-d
6
as solvent made it possible to observe resonances for hydroxyls of compound C, for which a detailed spectral analysis was
also carried out using 2D NMR spectroscopy methods (HSQC, HMBC, COSY, NOESY) and all resonances in the PMR and
13
C NMR spectra were assigned.
Table 2 shows clearly that the described correlations for resonances of H atoms in the PMR spectra between chemical
shifts of pyrone and phenol parts of subunits (i), (ii), and (iii) of product C and pyrone and phenol parts of subunits (i) and (ii)
13
of product B remained valid. Differences in chemical shifts of resonances in the C NMR spectrum of subunits (i) and (ii) of
the dimeric and trimeric adducts were less than 0.2 ppm for most C atoms; for subunit (iii), 0.3 ppm.
The presence of resonances for C atoms that did not correlate with the overall pattern could be explained by the effect
of the through-space (more than four bonds) chemical environment. For example, the differences of chemical shifts for C1′ and
C2′ of the phenol part of subunit (iii) of the trimer relative to the phenol part of subunit (ii) of the dimer were 2.95and
1.00 ppm, respectively. This was due to the different substitutions on C3′ and C4′. For the trimer, C3′-C4′ are fused to subunit
(i); for the dimer, C3′ and C4′ have hydroxyls. Furthermore, an Overhauser effect between protons on C-8 of the corresponding
pyrone part of the molecule and protons on C-2′ and C-6′ of the phenol part was observed using the 2D NMR NOESY method.
13
According to high-resolution C NMR and PMR spectra for the trimer, it was concluded that product C is a
chromatographically inseparable mixture of eight diastereomers.
All resonances corresponding to the stereomers have not yet been assigned because of the large number of them.
428

TABLE 2. 13C NMR Spectra for Dimer B and Trimer C
Subunits
C
C atom
i
ii
iii
C
C
B
B
2
3
4
4a
5
6
101.38
91.49
188.95
100.80
165.04
97.95
101.05
91.5
189.05
100.85
165.1
98.0
145.33
137.65
176.78
104.18
162.2
145.4
137.6
176.7
104.2
162.3
99.3
101.02
91.29
188.71
100.71
165.18
98.28
99.23
7
8
169.25
97.28
169.2
97.3
165.33
94.64
165.2
94.6
169.49
97.38
8a
1′
2′
3′
4′
5′
6′
160.68
126.15
116.49
145.43
147.66
115.39
121.02
160.82
126.25
116.5
145.4
147.6
115.4
121.1
157.84
127.05
117.62
141.35
142.98
118.35
123.6
157.8
126.8
117.6
141.6
143.2
118.3
123.5
160.4
129.75
118.6
141.1
143.15
117.92
123.65
______
Average values are given for mixtures of diastereomers of B and C.
Because quercetin is readily oxidized, it can be assumed that oligomeric oxidation products will be encountered in
living organisms in addition to the starting flavonoid. Allium cepa L. is a natural sample that contains significant quantities
of quercetin in various chemical forms. The flavonoid is localized in the leaves of a modified runner (bulbs). It is observed as
the aglycon in outer dry leaves; as glycosides, in inner moist ones. In view of this information, we selected outer bulb leaves
(skin) to search for quercetin derivatives. Phenolic compounds were extracted by EtOAc. HPLC—MS analysis of the extract
showed the presence of compounds similar to products B and C (retention times 9.6 and 11.7 min) that were formed upon
oxidation of quercetin by cytochrome c and K [Fe(CN) ]. Mass spectra of the first compound gave two ions at 601 (43,
3
6
[M - 1]−) and 299 (100) m/z; of the second, 901 (49, [M - 1]−) and 299 (100) m/z. Addition of highly pure products B and C
to the onion skin extract as an internal chromatographic standard did not produce new peaks. These facts indirectly indicate
that onion skin contains oligomeric quercetin derivatives, the structures of which agree fully with those published for products
B and C [4]. However, there are reports in which other oligomeric forms of quercetin are described [11, 12].
We tend to believe that the natural oligomers of the flavonoid contains a dioxane fusion as indicated by the detailed
NMR spectral analyses obtained in our and previous work [4]. Enzymes of the oxidoreductase class, namely peroxidase, are
most probablyinvolved in the conversion of quercetin to oligomeric adducts in onion skin. Onion skin has been shown to have
peroxidase activity [13, 14]. Herein we used a system containing peroxidase from horseradish root, H O , and quercetin in
2
2
phosphate buffer as a model of peroxidase oxidation. Peroxidase oxidation of quercetin formed a rather large amount of
oxidation products compared with systems containing cytochrome c or K [Fe(CN) ]. Using the aforementioned methods,
3
6
compounds with properties identical to those of products B and C could be observed. This suggested that the formation of
oligomeric oxidation products of quercetin in onion skin involves enzymatic systems. The reaction mechanism presumably
included oxidation to form a C-3′–C-4′ semiquinone of quercetin with subsequent Diels—Alder heterocyclization with an
activated C-2–C-3 double bond of the pyrone part of the other quercetin molecule. Apparently the cyclization process occurred
randomly because mixtures of all possible isomers were formed for the dimer and trimer. Only steric factors and hindered
rotation of large fragments of quercetin determined the predominance of one product or another. Proteins that could influence
the selectivity of the process were not involved in the reaction at this stage. It has been demonstrated here and previously that
oligomeric quercetin adducts are formed in reactions with a variety of oxidants [1, 5, 7, 8]. They have been analyzed by
chromatographic and spectral methods. These facts suggest that these compounds can be used as indicators of the biological
oxidation of the flavonoid regardless of the pathway taken. The extent of the oxidative processes occurring in living systems,
429

in particular plants, can be estimated using them. The relationship of these processes to changes in environmental factors of
the plant habitat can be determined.
EXPERIMENTAL
Analytical Methods. HPLC, mass spectrometry, and NMR conditions have been published [1].
Oxidation of Quercetin in Reactions with Various Electron Acceptors. Aqueous phosphate buffer (100 mL,
0.05 M, pH 7.4) containing quercetin (3.02 mg, 0.01 mmol) was stirred vigorouslyat 25°C and treated over 3 min with oxidants
a) ferricytochrome c (125 mg, 0.01 mmol); b) K [Fe(CN) ] (3.29 mg, 0.01 mmol); or c) horseradish peroxidase solution (0.01
3
6
mL, 0.01 µmol) and H O solution (0.01 mL, 0.1 mmol). The mixture was stirred for 30 min. The course of the reaction was
2
2
monitored by TLC. The reaction mixture was evaporated to one half the volume and treated with EtOAc (30 mL). The organic
fraction was separated. The aqueous layer was extracted with EtOAc (3 × 20 mL). The combined organic fractions were dried
over Na SO . The precipitate was filtered off. The filtrate was evaporated. The solid was analyzed by HPLC—MS.
2
4
Preparative Synthesis and Production of Pure Oligomeric Quercetin Oxidation Products. Aqueous acetone
phosphate buffer (100 mL, 0.05 M, 40/60 by vol., pH 7.4) containing quercetin (604 mg, 0.2 mmol) was stirred vigorously at
25°C and treated over 3 min with K [Fe(CN) ] (658 mg, 0.2 mmol). The mixture was stirred for 30 min. Completion of the
3
6
reaction was monitored byTLC. The reaction mixture was evaporated to one half the volume, extracted with EtOAc, and dried
over Na SO . The solid was chromatographed over a column of Sephadex LH-20 [1] to produce (in order of elution) unreacted
2
4
quercetin (181 mg, 30%) and oligomeric products B (52.1 mg, 9%) and C (43.3 mg, 7%).
1,3,11a-Trihydroxy-9-(3,5,7-trihydroxy-4H-1-benzopyran-4-on-2-yl)-5a-(3,4-dihydroxyphenyl)-5,6,11-hexahydro-
5,6,11-trioxanaphthacen-12-one (B). UV spectrum (EtOH, λ , nm): 270, 303, 365 (log ε 4.31, 4.36, 4.33). Mass spectrum
max
(EI, 4.5 kV, m/z, I , %): 601 (43) [M - 1]−, 299 (100).
rel
1,3,11a-Trihydroxy-9-(3,5,7-trihydroxy-4H-1-benzopyran-4-on-2-yl)-5a-[1,3,11a-trihydroxy-5a-(3,4-
dihydroxyphenyl)-5,6,11-hexahydro-5,6,11-trioxanaphthacen-12-on-9-yl]-5,6,11-hexahydro-5,6,11-trioxanaphthacen-12-
one (C). UV spectrum (EtOH, λ , nm): 302, 363 (log ε 4.61, 4.39). Mass spectrum (EI, 4.5 kV, m/z, I , %): 901 (49)
max
rel
[M - 1]−, 299 (100).
Extraction of Phenolic Compounds from Outer Skin of Allium cepa. Outer light-brown onion skin was separated
manually from the bulb and ground in a Waring 1L Blender. The ground skin (4 g) was placed in a flask and treated with
EtOAc (50 mL). The resulting mixture was incubated at 25°C for 12 h with constant stirring. The extract was clarified by
filtration and evaporated to dryness. The solid was dissolved in EtOAc or EtOH.
REFERENCES
1.
E. M. Cherviakovsky, D. A. Bolibrukh, A. V. Baranovsky, T. M. Vlasova, V. P. Kurchenko, A. A. Gilep, and
S. A. Usanov, Biochem. Biophys. Res. Commun., 342, 2, 459 (2006).
C. J. Wallace and A. E. Proudfoot, Biochem. J., 245, 3, 773 (1987).
N. Ohno and M. A. Cusanovich, Biophys. J., 36, 3, 589 (1981).
T. N. Ly, C. Hazama, M. Shimoyamada, H. Ando, K. Kato, and R. Yamauchi, J. Agric. Food Chem., 53, 21, 8183
(2005).
2.
3.
4.
5.
6.
7.
8.
9.
V. Krishnamachari, L. H. Levine, and P. W. Pare, J. Agric. Food Chem., 50, 15, 4357 (2002).
V. Krishnamachari, L. H. Levine, C. Zhou, and P. W. Pare, Chem. Res. Toxicol., 17, 6, 795 (2004).
Y. Hirose, T. Fujita, and M. Nakayama, Chem. Lett., 775 (1999).
A. Gulsen, D. P. Makris, and P. Kefalas, Food Res. Int., 40, 1, 7 (2007).
M. Furusawa, H. Tsuchiya, M. Nagayama, T. Tanaka, K. Nakaya, and M. Iinuma, J. Health Sci., 49, 6, 475
(2003).
10.
M. Furusawa, H. Tsuchiya, M. Nagayama, T. Tanaka, M. Oyama, T. Ito, M. Iinuma, and H. Takeuchi, J. Health
Sci., 52, 5, 578 (2006).
430

11.
12.
P. Schreier and E. Miller, Food Chem., 18, 301 (1985).
F. A. Ramos, Y. Takaishi, M. Shirotori, Y. Kawaguchi, K. Tsuchiya, H. Shibata, T. Higuti, T. Tadokoro, and
M. Takeuchi, J. Agric. Food Chem., 54, 10, 3551 (2006).
13.
14.
D. Stajner, N. Milic, J. Canadanovic-Brunet, A. Kapor, M. Stajner, and B. M. Popovic, Phytother. Res., 20, 7, 581
(2006).
U. Takahama and S. Hirota, Plant Cell Physiol., 41, 9, 1021 (2000).
431