A synthetic approach to the generation of quercetin sulfates and the detection of quercetin 3′-O-sulfate as a urinary metabolite in the rat

Article abstract of DOI:10.1016/j.bmc.2005.07.021

To study the biological effects of quercetin, authentic products of quercetin metabolism are required as standards. The synthesis of quercetin sulfate standards is thus described. Quercetin was reacted with a 10-fold molar excess of sulfur trioxide-N-trie

Full text of DOI:10.1016/j.bmc.2005.07.021

Bioorganic & Medicinal Chemistry 13 (2005) 6727–6731
A synthetic approach to the generation of quercetin sulfates
0
q
and the detection of quercetin 3 -O-sulfate as a urinary
metabolite in the rat
a,
Donald J. L. Jones, Rebekah Jukes-Jones, Richard D. Verschoyle,
b
a
*
a
Peter B. Farmer and Andreas Gescher
a
a
Cancer Biomarkers and Prevention Group, Departments of Biochemistry and Cancer Studies & Molecular Medicine,
University of Leicester, Leicester LE1 7RH, UK
b
MRC Toxicology Unit, University of Leicester, Leicester LE1 9NH, UK
Received 18 May 2005; revised 14 July 2005; accepted 20 July 2005
Available online 18 August 2005
Abstract—To study the biological effects of quercetin, authentic products of quercetin metabolism are required as standards. The
synthesis of quercetin sulfate standards is thus described. Quercetin was reacted with a 10-fold molar excess of sulfur trioxide-
N-triethylamine, and the products were analyzed by HPLC and mass spectrometry. Four monosulfates and three disulfates were
1
identified, and structural inferences were drawn by H NMR spectrometry of HPLC peak isolates. Analysis of the urine of rats that
had received quercetin (1.9 g/kg po) yielded a single peak, which by comparison with the products of the reaction between quercetin
0
and sulfur trioxide-N-triethylamine was identified as quercetin 3 -O-sulfate.
ꢀ
2005 Elsevier Ltd. All rights reserved.
1
. Introduction
macological deactivation step. Alternatively, metabo-
lites may contribute to the pharmacological or, indeed,
toxicological efficacy of the parent polyphenol. Intrigu-
ingly, it has recently been shown that the methylated
quercetin conjugates isorhamnetin and tamarixetin were
much more potent inhibitors of the activity of the en-
zyme cyclooxygenase-2 (COX-2) than their metabolic
Naturally occurring flavonoids and isoflavonoids in the
diet are thought to exert beneficial effects on human
health. Elucidation of the mechanisms underlying
these effects is currently the focus of much research.
Epidemiological evidence links diets rich in quercetin
1
6
(for structure, see Fig. 1) with decreased incidence of
2
cardiovascular and neoplastic diseases.
precursor. This example illustrates the principle that
–5
to rationalize advice as to the potential benefit of con-
sumption of flavonoids as diet constituents or food addi-
tives, it is important to clarify the role that their
metabolites may play in the pharmacology of the parent
substance. Such studies require the availability of
authentic standards. Asynthetic approach to furnish
quercetin glucuronides led to products, in which four
of the five hydroxyl groups were substituted with glucu-
There is accumulating evidence to suggest that polyphe-
nolic phytochemicals including flavonoids are rapidly
metabolized in the human or animal organism. Glucu-
ronidation appears to be the most prevalent conjugation
pathway for quercetin in rats and humans. Sulfation can
occur to form mono- or bis-conjugates or combine with
other conjugation pathways to form sulfo-glucuronides
8
ronyl moieties. In general, phenols not only undergo
6
–8
and sulfate conjugates of methylated quercetin. Most
often, the metabolism of polyphenols constitutes a phar-
avid glucuronidation but also metabolic conjugation
with activated sulfate to generate O-sulfates. The rela-
tive prevalence of the different conjugates will depend
on the concentration of quercetin, activity of the meta-
bolic enzymes and cell type. Sulfation of quercetin has
previously been attempted using a liver extract with
Keywords: Sulfate synthesis; Quercetin; NMR; Mass spectrometry.
q
This work was supported by two UK Medical Research Council
programme grants.
Corresponding author. Tel.: +44 0116 223 1827; fax: +44 0116 223
840; e-mail: djlj1@le.ac.uk
9
activated sulfate and cofactors. This synthesis allowed
the generation of one quercetin monosulfate that was
*
1
0
968-0896/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2005.07.021

6728
D. J. L. Jones et al. / Bioorg. Med. Chem. 13 (2005) 6727–6731
OH
3
'
9
OH
2
'
2
.0
.5
4'
8
B
HO
O
2
5 '
1
'
7
6 '
A
C
3
6
6
1
4
OH
5
OH
O
AU
1
0
0
.0
.5
.0
8
Quercetin
7
4
3
5
2
1
0
10
20
30
40
Retention time (min)
Figure 1. HPLC chromatogram of the products of the reaction of quercetin with Esulfur trioxide-N-triethylamine. On the basis of mass
spectrometric evidence (Table 1) peaks 2–4 are quercetin disulfates and peaks 5, 6, 7 and 8 monosulfates. Peaks 5 and 9 co-eluted with authentic
quercetin 3-O-sulfate and quercetin, respectively. Also shown is the structure of quercetin. For details of chromatography, see Section 4.
0
not characterized. Day et al. synthesized quercetin 3 -O-
7
sulfate specifically using a chemical synthesis method.
sulfates. Preliminary studies using a 2-, 4- and 10-fold
excess of sulfur trioxide-N-triethylamine complex found
the 10-fold excess to be optimal, and was used in
subsequent studies.
One of the initial objectives was to generate a number of
sulfate analogues that would allow comparison in a bio-
logical assay. The advantage of a chemical synthetic ap-
proach is that the potential yield of product could be
even bigger since with biological approaches the enzymes
and cofactor concentrations could be a limiting factor.
The relative yield of each of the peaks using a 10-fold ex-
cess of sulfur trioxide-N-triethylamine complex is given
in Table 1. The yield varies from 1 to 16% for the mono-
sulfates and 0.8–3.8% for the disulfates. The synthesized
sulfates were at least 90% pure (as determined by
HPLC), with the major contaminant being the parent
molecule, quercetin. The purchased sulfate was >98%
pure. Whilst a detailed study with regard to stability
was not undertaken, the dried compound, stored at
The aim of the present study was to synthesize authentic
quercetin sulfates and to explore in a preliminary fash-
ion their occurrence in biomatrices of rats that have
received quercetin.
+
4 ꢁC, when reconstituted showed no decomposition
for a month. Material eluting with each peak was col-
lected individually, and the eluant was evaporated to
dryness.
2
. Results and discussion
2
.1. Synthesis
2
.2. NMR and mass spectrometric data
Quercetin was reacted with a 10-fold molar excess of sul-
fur trioxide-N-triethylamine complex, and an extract of
the mixture was analyzed by HPLC. The resultant chro-
matogram (Fig. 1) shows the presence of quercetin (peak
Isolated peak material was subjected to MS and NMR
analyses, the latter after reconstitution in DMSO-d6.
Table 1 shows mass spectral inferences with regard to
peak identity.
9
) and eight peaks that were surmised to be quercetin
Table 1. Molecular ions of species contained in HPLC peaks isolated from a reaction mixture of quercetin with sulfur trioxide-N-triethylamine
complex
Fraction number
Rt (min)
Molecular weight
Inference
Yield (%)
1
2
3
4
5
6
7
8
9
4.5
8.3
n.a.
462
462
462
382
382
382
382
302
—
Quercetin disulfate
Quercetin disulfate
Quercetin disulfate
Quercetin monosulfate
Quercetin monosulfate
Quercetin monosulfate
Quercetin monosulfate
Quercetin
0.8
2.6
9.0
9.3
3.8
1.3
10.7
12.3
14.5
15.4
19.1
16.3
8.3
10.9
—
Yield of individual sulfate is included in the table.
n.a., the molecular weight could not be assigned.

D. J. L. Jones et al. / Bioorg. Med. Chem. 13 (2005) 6727–6731
6729
The materials derived from peaks with numbers 5, 6, 7,
8
spectra. Figure 2A and B show spectra for quercetin
(peak 9 in Fig. 1) and purchased authentic quercetin
3-O-sulfate, respectively. Figures 2C and D depict H
1
and 9 in Figure 1 afforded interpretable H NMR
1
NMR spectra of material contained in peaks 5 and 6,
respectively, and Figure 2E the spectrum of a mixture
of substances eluting with peaks assigned numbers 7
and 8 in Figure 1 and Table 1. The spectrum of quer-
cetin is characterized by coupling of protons 6-H and
8-H on ring A (Fig. 2A) that appear as a pair of
meta-coupled doublets at 6.2 and 6.4 ppm, respectively,
A
5
'-H
2
'-H
6
-H
8
-H
0
with a coupling constant of J = 2.1 Hz. The 5 -H on
0
6
'-H
ring B (6.9 ppm) is split by 6 -H and, vice versa, the
0
0
proton 6 -H is split by 5 H. This ortho-coupling results
0
in a large J value of 8.5 Hz. The 6 -H is split by further
0
7
.5
7.0
6.5
0
8
.0
long range coupling to 2 -H. Consequently, the 6 -H
appears as
a
doublet of doublets (7.65 ppm,
B
0
J = 8.5 Hz, J = 2.13 Hz), and the 2 -H signal appears
as a fine doublet (7.75 ppm, J = 2.13 Hz). The spectrum
of HPLC peak 5 (Fig. 2C) is identical to that of authen-
tic quercetin 3-O-sulfate (Fig. 2B) purchased from
Extrasynthese (Genay, France). In comparison to the
H NMR spectrum of unconjugated quercetin, the
spectrum of this molecule is characterized by an upfield
shift of the 2 -H signal. This upfield shift is intriguing,
as one might rather expect conjugation at the 3-posi-
1
0
8
.0
7
.5
7.0
6.5
0
C
tion to affect the resonance frequency of 6 -H. An anal-
ogous observation was made in the analytical
comparison of quercetin and quercetin 3-O-glucuro-
0
nide, where the resonance frequency of 2 -H was shifted
1
0
upfield by glucuronidation at the 3-position. It is
likely that the quercetin molecule could allow rotation
0
around the C-2 to C-1 bond. This rotation is analo-
0
gous to the C5-C1 bond rotation in fluoronitropyri-
methamine that exists as two slowly interconverting
1
1
rotamers. NMR evidence for the existence of two
rotamers of quercetin could not be generated, although
X-ray crystallography data suggest that ring B can be
8
.0
7.5
7.0
6.5
D
1
2
situated at 0ꢁ and 180ꢁ to the residual structure.
The major difference between the spectra of quercetin
and HPLC peak 6 is a change in chemical shift of both
the 6- and 8-protons (Fig. 2D). The material eluting
with HPLC peak 6 might be quercetin 7-O-sulfate,
based on the similar magnitude of the effect of sulfate
substitution on the shifts of 6-H and 8-H. It proved
impossible to completely resolve the materials con-
tained in the HPLC peaks 7 and 8 by semi-preparative
HPLC. The NMR spectrum of the mixture allowed dis-
crimination between the two components by exploiting
the difference in integration (Fig. 2E). The chemical
8
.0
7.5
7.0
6.5
E
7,8
7,8
8
8
0
shift of the 2 -H in the spectrum of the material eluting
with 7 (Fig. 2E) displays a considerable downward shift
in comparison with the equivalent proton in quercetin,
7
8
7
7
0
compatible with sulfate substitution at the 3 -position.
The spectrum of the material eluting with 8 (Fig. 2E)
0
is marked by a large downfield shift of the 5 -H (d
8
.0
7.5
7.0
6.5
0.5 ppm) in comparison to quercetin, suggesting that
it is quercetin 4 -O-sulfate.
0
Chemical shift (ppm)
1
Figure 2. H NMR spectra of quercetin (A), quercetin 3-O-sulfate (B)
and of peak isolates obtained by HPLC separation of products of the
reaction of quercetin with sulfur trioxide-N-triethylamine as assigned
in Figure 1 and Table 1, as follows: peak 5 (C), peak 6 (D) and a
mixture of substances eluting with peaks 7 and 8 (E). For details of
2
.3. Biological data
Urine from rats that had received quercetin (1900 mg/kg
po) was analyzed by LC–MS (Fig. 3). Figure 3A shows
the urine sample of a rat. There is a single peak, consis-
tent in retention time with material that eluted with peak
1
reaction and H NMR spectrometry, see Section 4, for analysis of
spectra see Section 2.

6730
D. J. L. Jones et al. / Bioorg. Med. Chem. 13 (2005) 6727–6731
0
1
00
A
cetin 3 -O-sulfate is the most prominent metabolite in
human plasma after consumption of a quercetin-con-
7
6
taining diet or after iv infusion of pure quercetin.
Many polyphenols are currently being subjected to
investigation as potential clinical therapeutic agents.
The metabolism of polyphenols often follows a similar
path to quercetin, with sulfate, glucuronides and methyl-
ated derivatives being generated. It should be possible to
use this method for the generation of authentic stan-
dards for a whole range of polyphenols and also to
understand both mechanistic and metabolic attributes
of these interesting molecules.
0
7
B
1
00
6
8
5
0
0
5
10
15
20
25
In conclusion, the work described here helps one with
the understanding of the reactivity of quercetin hydroxy
moieties vis- a` -vis sulfating agents under artificial chem-
ical and biological conditions. Mechanistic studies of
bioactive polyphenols exemplified by quercetin tend to
concentrate on the parent molecules alone. Many of
these polyphenolic phytochemicals undergo avid meta-
bolic conjugation, and one cannot exclude the possibili-
ty that such conjugates harbour biological activity.
Therefore, the inclusion of such conjugates in pharma-
cological profiling seems propitious.
Retention time (min)
Figure 3. LC–MS ion chromatograms (m/z 381) of extracts of (A)
urine collected over 24 h from a rat that had received quercetin (1.9 g/
kg by gavage), and (B) products of the reaction of quercetin to sulfur
trioxide-N-triethylamine. Peak numbers are identical to those in Figure
1 and Table 1. On the basis of mass spectrometric evidence (Table 1),
peaks 5, 6, 7 and 8 are monosulfates. Peak 7 co-eluted with authentic
0
quercetin 3 -O-sulfate and is seen in the urine of rats treated with
quercetin. For details of LC–MS, see Section 4.
7
in the HPLC analysis of the products of the reaction of
quercetin with sulfur trioxide-N-triethylamine. Figure
B shows an extracted LC–MS ion chromatogram
m/z 381) of the extract from the synthesis. Shown are
4. Experimental
4.1. Chemicals
3
(
four peaks which from left to right are the 3-O-, 7-O-,
0
0
3
that the urinary metabolite was quercetin 3 -O-sulfate.
-O- and 4 -O-sulfates of quercetin. This result suggests
0
Quercetin 3-O-sulfate was obtained from Extrasynthese
(Genay, France), while all other chemicals and solvents
were from either Sigma (Poole, UK) or Fisher (Lough-
borough, UK). For administration in rats in vivo, quer-
cetin was dissolved in DMSO.
3. Conclusions
Polyphenolic phytochemicals are good substrates of
metabolic conjugation reactions with activated glucu-
ronic acid or sulfate. The identification of such conju-
gates often poses an analytical challenge because of
the multitude of hydroxy moieties contained in such
molecules available for reaction with the conjugating
species, furnishing potentially isomeric mono- and
poly-conjugates. Quercetin is a suitable example that
harbours five OH groups. Whilst quercetin glucuronides
4.2. Reaction of quercetin with sulfur trioxide-N-
triethylamine
To remove any compound-associated water, dry pyri-
dine was added to quercetin (0.5 g, 1.7 mmol) until dis-
solved. The mixture was rotor evaporated and the
procedure repeated twice. The dried quercetin was then
dissolved in dioxane (50 mL), and sulfur trioxide-N-tri-
ethylamine complex (3.1 g, 16.9 mmol) was added under
argon to avoid contact with air. The reaction vessel was
placed in a water bath (40 ꢁC) for 90 min, after which
products of quercetin sulfation precipitated out and
stuck to the glass. After incubation, the dioxane was
decanted and the conjugates were redissolved in metha-
nol. Aliquots of this solution were diluted in water and
methanol, and used for HPLC analysis.
8
,10
have been the subject of investigation,
there is a
fspaucity of information on the chemistry and biochem-
istry of quercetin sulfates. The results outlined above de-
scribe a facile synthetic route to four mono- and three
di-sulfates of quercetin. Spectral elucidation suggests
monosulfation of quercetin to occur in positions 3, 7,
0
0
3
and 4 . The data did not allow the positional alloca-
tion of the sulfate groups with regard to the disulfates,
but—in analogy to the preferred loci of monosulf-
ation—it is likely that the hydroxy moieties in positions
4.3. Analysis of rat urine
0
0
3
, 7, 3 and 4 , but not the 5 hydroxy, are involved.
Four F344 rats (150 g, obtained from Harlan, UK)
received quercetin (1.9 g/kg) by gavage. Four control
rats received DMSO only. Rats were kept in metabolism
cages for 24 h post-administration and urine was
collected. Aliquots of urine were mixed with twice the
volume of DMSO/methanol (1:4, v/v). The mixture
was vortexed and centrifuged (17,000g, 15 min). The
The results provide for the first time evidence for the for-
0
mation of quercetin 3 -O-sulfate as the major product of
0
conjugation of quercetin with 3-phospho-adenosine 5 -
phosphosulfate in the rat. This finding is closely akin
to that in humans, as previous work suggests that quer-

D. J. L. Jones et al. / Bioorg. Med. Chem. 13 (2005) 6727–6731
6731
1
supernatant was diluted with water (1:1) and injected
onto the HPLC column (injection volume 50 lL).
Proton ( H) spectra were generated with a Bruker ARX
250 MHz spectrometer, using deuterated dimethylsulf-
oxide as the solvent.
4
.4. Analysis of quercetin sulfates by HPLC
The products of the chemical reaction of quercetin with
sulfur trioxide-N-triethylamine or extracts of urine were
separated on a BDS-Hypersil C18 column (250 · 4.6 mm
I.D., 5 lm particle size, Phenomenex, Macclesfield,
UK). The instrument used was a Varian Prostar system
Acknowledgments
We thank Professor Gordon Roberts, Leicester Univer-
sity, for help with the interpretation of the NMR analy-
ses and Dr. Robert Britton for proof-reading the
manuscript.
(
(
Walton-on-Thames, UK), comprising a UV detector
Model 310), solvent delivery system (Model 230) and
an autosampler (Model 410) with a 100 lL injection
loop. The gradient elution programme commenced with
References and notes
7
5% ammonium acetate (0.1 M, pH 5.15) in methanol,
which was followed by a linear decrease over 10 min
to 55% ammonium acetate in methanol, and then to
1. Hollman, P. C. H.; Hertog, M. G. L.; Katan, M. B.
Biochem. Soc. Trans. 1996, 24, 785–789.
2
3
4
. Hertog, M. G. L.; Feskens, E. J. M.; Hollman, P. C. H.;
Katan, M. B.; Kromhout, D. Lancet 1993, 342, 1007–
4
2
5% ammonium acetate in methanol over a further
0 min, after which the mobile phase remained isocratic
1
. Hertog, M. G. L.; Feskens, E. J. M.; Hollman, P. C. H.;
011.
for a further 5 min. The instrument was operated with a
flow rate of 1.0 mL/min and detection by UV spectro-
photometry (375 nm).
Katan, M. B.; Kromhout, D. Nutr. Cancer 1994, 22, 175–
1
84.
. Hertog, M. G. L.; Kromhout, D.; Aravanis, C.; Black-
burn, H.; Buzina, R.; Fidanza, F.; Giampaoli, S.; Jansen,
A.; Menotti, A.; Nedeljkovic, S.; Pekkarinen, M.; Simic,
B. S.; Toshima, H.; Feskens, E. J. M.; Hollman, P. C. H.;
Katan, M. B. Arch. Intern. Med. 1995, 155, 381–386.
. Keli, S. O.; Hertog, M. G. L.; Feskens, E. J. M.;
Kromhout, D. Arch. Intern. Med. 1996, 156, 637–642.
. Jones, D. J. L.; Lamb, J. H.; Verschoyle, R. D.; Howells,
L. M.; Butterworth, M.; Lim, C. K.; Ferry, D.; Farmer, P.
B.; Gescher, A. J. Br. J. Cancer 2004, 91, 1213–1219.
. Day, A. J.; Mellon, F.; Barron, D.; Sarrazin, G.; Morgan,
M. R. A.; Williamson, G. Free Radic. Res. 2001, 35, 941–
A second system was used to fraction-collect the querce-
tin sulfates from HPLC eluant. A Gilson semi-prepara-
tive system (Gilson Inc., Middleton, WI, USA)
comprising two pumps and a UV detector with a Rheo-
dyne 7125 manual injector connected to a 500 ll loop
was used. Fractions were collected when they eluted
from a semi-preparative column (250 · 21.2 mm I.D.,
5
6
5
lm particle size, Phenomenex, Macclesfield). The
chromatographic conditions were as described above,
except that the flow rate was 10 mL/min.
7
9
8. Day, A. J.; Bao, Y. P.; Morgan, M. R. A.; Williamson, G.
52.
4.5. Analysis of quercetin sulfates by mass spectrometry
and H NMR spectrometry
1
Free Radic. Biol. Med. 2000, 29, 1234–1243.
. Justino, G. C.; Santos, M. R.; Canario, S.; Borges, C.;
9
Florencio, M. H.; Mira, L. Arch. Biochem. Biophys. 2004,
4
LC–MS was carried out using a Micromass Platform
mass spectrometer coupled with a Hewlett Packard Ser-
ies 1100 HPLC system and a UV detector absorbing at
32, 109–121.
0. Moon, J. H.; Tsushida, T.; Nakahara, K.; Terao, J. Free
Radic. Biol. Med. 2001, 30, 1274–1285.
1
1
3
75 nm. The eluant was split 1:7 post column so that
1. Tendler, S. J. B.; Griffin, R. J.; Birdsall, B.; Stevens, M. F.
G.; Roberts, G. C. K.; Feeney, J. FEBS Lett. 1988, 240,
approximately 142 lL/min of eluant entered the mass
spectrometer. Source temperature was maintained at
2
01–204.
110 ꢁC. Cone voltage was 42 V, while the capillary volt-
age was maintained at 3.82 kV.
12. Jin, G. Z.; Yamagata, Y.; Tomita, K. Acta Crystallogr.
Sect. C Crystal Struct. Commun. 1990, 46, 310–313.