First synthesis, characterization, and evidence for the presence of hydroxycinnamic acid sulfate and glucuronide conjugates in human biological fluids as a result of coffee consumption

Article abstract of DOI:10.1039/c0ob00137f

A systematic investigation of the human metabolism of hydroxycinnamic acid conjugates was carried out. A set of 24 potential human metabolites of coffee polyphenols has been chemically prepared, and used as analytical standards for unequivocal identifications. These included glucuronide conjugates and sulfate esters of caffeic, ferulic, isoferulic, m-coumaric and p-coumaric acids as well as their dihydro derivatives. A particular focus has been made on caffeic and 3,4-dihydroxyphenylpropionic acid derivatives, especially the sulfate conjugates, for which regioselective preparation was particularly challenging, and have so far never been identified as human metabolites. Ten out of the 24 synthesized conjugates have been identified in human plasma and/or urine after coffee consumption. A number of these conjugates were synthesized, characterized and detected as hydroxycinnamic acid metabolites for the first time. This was the case of dihydroisoferulic acid 3′-O-glucuronide, caffeic acid 3′-sulfate, as well as the sulfate and glucuronide derivatives of 3,4-dihydroxyphenylpropionic acid.

Full text of DOI:10.1039/c0ob00137f

View Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
First synthesis, characterization, and evidence for the presence of
hydroxycinnamic acid sulfate and glucuronide conjugates in human biological
fluids as a result of coffee consumption†
Rene´ Fumeaux,a Candice Menozzi-Smarrito,a Angelique Stalmach,b Caroline Munari,a Karin Kraehenbuehl,a
Heike Steiling,a Alan Crozier,b Gary Williamsona and Denis Barron*a
Received 17th May 2010, Accepted 1st July 2010
DOI: 10.1039/c0ob00137f
A systematic investigation of the human metabolism of hydroxycinnamic acid conjugates was carried
out. A set of 24 potential human metabolites of coffee polyphenols has been chemically prepared, and
used as analytical standards for unequivocal identifications. These included glucuronide conjugates and
sulfate esters of caffeic, ferulic, isoferulic, m-coumaric and p-coumaric acids as well as their dihydro
derivatives. A particular focus has been made on caffeic and 3,4-dihydroxyphenylpropionic acid
derivatives, especially the sulfate conjugates, for which regioselective preparation was particularly
challenging, and have so far never been identified as human metabolites. Ten out of the 24 synthesized
conjugates have been identified in human plasma and/or urine after coffee consumption. A number of
these conjugates were synthesized, characterized and detected as hydroxycinnamic acid metabolites for
the first time. This was the case of dihydroisoferulic acid 3¢-O-glucuronide, caffeic acid 3¢-sulfate, as well
as the sulfate and glucuronide derivatives of 3,4-dihydroxyphenylpropionic acid.
is the small intestine. Since the presence of an esterase activity
Introduction
has been detected in the human small intestine mucosa,12,13 this
With more than one billion drinkers over the world, coffee is
results in the limited release of a first pool of free hydroxycinnamic
one of the most popular beverages. Coffee consumption may be
acids. The second site of hydrolysis of chlorogenic acids is the
associated with several health benefits,1,2 some of them related to
the antioxidant activity of its phenolic constituents.1,3 Coffee is
one of the major sources of chlorogenic acids.4,5 The chlorogenic
acid compositions of commercial roasted6 and brewed7 coffees
have been reported. On the other hand, green coffee beans
contain the 3 isomers (3-, 4-, and 5-) of O-caffeoyl quinic acids,
5-O-feruloylquinic acid, and a number of di-O-caffeoylquinic
acids (3,5-, 3,4-, and 4,5-).8 Recently, our knowledge of the
phytochemical composition of green coffee has been considerably
increased by the use of LC-MSn, leading to identification of com-
plex mixtures of mono- and di-O-p-coumaroyl, caffeoyl, feruloyl,
and/or dimethoxycinnamoyl quinic acid esters.9,10 However, it is
well known that the polyphenols contained in food are extensively
metabolized by the human body. Thus, the real entities that the
target cells or tissues will be exposed to after consumption are
likely to be the metabolites, rather than the native forms present
in plants. These metabolites could significantly contribute to the
biological activity.11 Indeed, chlorogenic acids are transformed
during digestion and the first site of their human metabolism
colon, catalyzed by the colonic microflora.14,15 The released caffeic
acid is subsequently dehydroxylated into m-coumaric acid by the
microflora.16 Alternatively, caffeic acid may be reduced to 3,4-
dihydroxyphenylpropionic acid (dihydrocaffeic acid), and subse-
quently dehydroxylated into 3-hydroxyphenylpropionic acid.17 All
the metabolites released in the digestive tract can be absorbed and
further metabolized in the enterocyte and the liver by phase-II
enzymes into their sulfated and/or glucuronidated conjugates.
The absorption and the human metabolism of phenolic acids
from coffee have not been studied in detail. Indeed, caffeic
acid was found in plasma,18 and caffeic acid and metabolites
resulting from methylation and/or reduction were present in urine.
These metabolites included ferulic, isoferulic, dihydroferulic, and
dihydrocaffeic acids.19,20 p-Coumaric acid was also identified in
the urine of most subjects.20 Since all these compounds have been
identified after treatment of plasma or urine by b-glucuronidase
and sulfatase, no identification of the original conjugates was
carried out, and these were only suspected to be glucuronidated
and/or sulfated derivatives. In contrast, Caco-221 and HepG222
cells have been clearly shown to produce hydroxycinnamic acid
glucuronides and sulfates, and the conjugates identified by LC-
MS. However, the technique was obviously not able to assign
the structure of caffeic acid derivatives, where two different
positions of conjugation are possible on the catechol ring. On the
other hand, ferulic acid is metabolized in rats into two different
monoglucuronides, one presumably conjugated at the phenolic
hydroxyl of ferulic acid, and a second one presumably conjugated
at the carboxylic acid function of ferulic acid.23 Again, no
distinction could be made by LC-MS between these two isomers.
However, the exact knowledge of the position of conjugation is of
aNestle´ Research Center, PO Box 44, CH-1000, Lausanne 26, Switzerland.
E-mail: Denis.Barron@rdls.nestle.com; Fax: + 41 21 785 8554; Tel: + 41
21 785 9497
bPlant Products and Human Nutrition Group, Graham Kerr Building,
Division of Developmental Medicine, Faculty of Medicine, University of
Glasgow, Glasgow, UK G12 8QQ
† Electronic supplementary information (ESI) available: Detailed synthetic
strategies, procedures and characterization for compounds 1 to 16, and 19
to 22 can be found in ESI part A. Purity data (HPLC control, HRMS,
and/or quantification of sulfate esters) and NMR data (1H and 13C) of
the entire set of standard conjugates 1–24 can be found in ESI part B. See
DOI: 10.1039/c0ob00137f
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 5199–5211 | 5199
©

View Online
Fig. 1 Structures of the most probable circulating human hydroxycinnamic acid glucuronide 1–12, and sulfate 13–24 conjugates.
the utmost important since this parameter is known to have an
impact on the biological activity.11,24
procedure involving glucuronidation (using glucuronic acid donor
25, see Scheme 1) of their ethyl ester precursors. The latter com-
pounds have been prepared by the Horner–Wadsworth–Emmons
reaction of the corresponding substituted hydroxybenzaldehydes
with ethyl (triphenylphosphoranyliden)acetate. Catalytic hydro-
genation of hydroxycinnamic acid glucuronides 1–4 yielded the
corresponding dihydro derivatives 13–16. Caffeic acid 3¢- and 4¢-O-
glucuronides (5 and 6) have been regioselectively obtained for the
first time by glucuronidation of ethyl 4¢-acetyl- and ethyl 3¢-acetyl
caffeic acids, respectively. On the other hand, the sulfate conjugates
7–10 and 19–22 have been prepared by direct sulfation (with a
sulfur trioxide-trimethylamine complex, in aqueous Na2CO3) of
the corresponding hydroxycinnamic acids or their reduced forms.
In the main body of this paper, only the synthetic strategy for
the four caffeic- and dihydrocaffeic acid sulfates (11, 12, 23, and
24), and of the two dihydrocaffeic acid glucuronides (17 and 18) is
presented. These are newly characterized compounds, the presence
of which has never been evidenced in human biological fluids.
However the detailed synthetic strategies and preparations of the
above mentioned compounds 1 to 10, 13 to 16, and 19 to 22 are
displayed in the ESI of the paper (Part A; Schemes S1 to S3).†
For the preparation of dihydrocaffeic acid 3¢-O-b-D-
glucuronide, we used the benzyl group as protector (Scheme 1).
The main advantage was to combine, in a single step, the benzyl
deprotection and the reduction of the 2,3-double bond. Acetyla-
tion of 3¢,4¢-dihydroxybenzaldehyde 26 by acetyl chloride in the
presence of DBU as a base yielded the desired compound 27 in
88% yield. The regioselective deacetylation of 27 using thiophenol
and imidazole gave access to 3¢-acetoxy-4¢-hydroxybenzaldehyde
28 in 77% yield.
Furthermore, in the absence of chromatographic standards
of known purity, all the previous identifications only remained
qualitative and no quantification was made. All these examples
clearly demonstrate the limitations of the LC-MS technology
as a unique analytical tool. Beside the fact that we still do not
have an exact picture of the nature of the human metabolites of
coffee phenolics, only the synthesis of well characterized standard
conjugates would allow the investigation of their contribution to
the overall health effects of coffee, as well as their distribution in
circulating fluids and tissues. Very few reports have appeared on
the synthesis of animal and human polyphenol metabolites, and
the subject has been reviewed by us.25 The methods of synthesis
of glucuronic acid conjugates have been improved tremendously
due to the introduction of trichloroacetimidates as glycosyl
donors, but in contrast, the synthesis of sulfate conjugates is still
difficult. Even more problematic is the purification of phenolic
sulfate conjugates, which remain frequently contaminated by
some inorganic material, thus precluding determination of their
exact purity. In the present paper, we report on the efficient
synthesis of a set of 24 sulfated or glucuronidated conjugates of
hydroxycinnamic and hydroxyphenylpropionic acids, as potential
human metabolites of coffee phenolics. Furthermore, the purity
of all conjugates including sulfates was determined. This allowed
us to identify the circulating human phenolic acid metabolites
resulting from coffee consumption.
Results and discussion
Considering i) the phenolic composition of green and brewed
coffee, and ii) the previous metabolic studies related to coffee
consumption, the metabolites that may potentially be present in
human circulating fluids could be glucuronidated and sulfated
conjugates of hydroxycinnamic and dihydro hydroxycinnamic
(hydroxyphenylpropionic) acids. The structures of these potential
metabolites 1–24 are displayed in Fig. 1. Hydroxycinnamic acid
mono-O-glucuronides 1–4 have been synthesized by an improved
Compound 28 was benzylated under Mitsunobu conditions
to give aldehyde 29 (97%), which was submitted to a Horner–
Wadsworth–Emmons reaction, yielding the ethyl cinnamate 30
(89%). Deacetylation of 30 with sodium methoxide in dry
methanol and dry CH2Cl2 yielded compound 31 (97%). The
latter compound was glucuronidated at position 3¢, using 25
as glucuronic acid donor, to give the protected glucuronide 32
(76%). The correct aromatic substitution of compound 32 was
5200 | Org. Biomol. Chem., 2010, 8, 5199–5211
This journal is
The Royal Society of Chemistry 2010
©

View Online
Scheme 1 The preparation of dihydrocaffeic acid 3¢-O-b-D-glucuronide 17. a) Acetyl chloride, DBU, CH2Cl2, 88%; b) Thiophenol, imidazole, NMP, 77%;
c) Benzyl alcohol, triphenylphosphine, DEAD, 97%; d) Ethyl (triphenylphosphoranyliden)acetate, CH2Cl2, 89%; e) NaOMe, dry MeOH–dry CH2Cl2,
97%; f) 2,3,4-Triacetyl-D-methyl glucuronopyranosyl-(N-phenyl)-2,2,2-trifluoroacetimidate 25, BF3 etherate, CH2Cl2, 76%; g) NaOH, MeOH–H2O, then
AmberliteꢀR IR-120, 73%; h) H2, 5% Pd/C, MeOH–H2O, 96%.
evidenced by the presence, on its long distance HMBC spectrum,
of a correlation between the benzyl protons H7b at 5.11 ppm and
C4¢ at 151.42 ppm, as well as a correlation between the anomeric
proton H1a at 5.09 ppm and C3¢ at 146.12 ppm. Similar H7b/C4¢
correlations, indicative of a 4¢-benzylation, were detected on the
HMBC spectra of compounds 29–31. Saponification of compound
32 furnished 4¢-O-benzylcaffeic acid 3¢-O-b-D-glucuronide 33
(73%), which was hydrogenated to dihydrocaffeic acid 3¢-O-b-D-
glucuronide 17 in 96% yield.
again, was easily eliminated by extraction of the products by
ethyl acetate at alkaline pH. Final purification of 37 was carried
out by VLC on silica and RP-18, with an overall yield of 42%
from 26. Neither on the basis of the melting points, nor on the
1
analysis of the H NMR data, a clear distinction could be made
in the past between 3¢-benzyloxy-4¢-hydroxybenzaldehyde and
its 4¢-benzyloxy-3¢-hydroxy isomer.27 In compound 37, however,
3¢-O-benzylation was clearly evidenced by the presence of the
long distance H7b/C3¢ correlation on its HMBC spectrum. The
remainder of Scheme 2 followed similar procedure as above,
i.e. i) reaction of aldehyde 37 to cinnamic acid 38 (56%), ii)
glucuronidation of 38 with activated sugar 25 to give the protected
glucuronide 39 (73%), and iii) saponification of 39 to benzylated
glucuronide 40 (78%). Again, hydrogenation of 40 led to both the
removal of the 3¢-O-benzyl protective group, and to the reduction
of the 2,3-double bond, affording dihydrocaffeic acid 4¢-O-b-D-
glucuronide 18 in quantitative yield.
For the preparation of hydroxycinnamic acid sulfates, again
access to regioselectively protected precursors was necessary.
Furthermore, the protecting group had to be removed under
conditions compatible with the stability of the sulfate ester
bond.
The use of alkyl chlorosulfates in the synthesis of sulfate
monoesters has recently been reported,28 and 2,2,2-trichloroethyl
chlorosulfate has been involved in the preparation of a sulfated
phenolic alkaloid.29 Among the main advantages of this approach,
the sulfation reaction was compatible with the presence of a
wide range of protecting groups. Thus, 3¢-O-benzylcaffeate 38
was sulfated with neopentyl chlorosulfate,28,30 to afford compound
41 in 86% yield. Reduction of 41 yielded compound 42 (90%).
Elimination of the neopentyl (sodium azide), and ethyl ester (aq.
NaOH) substituents yielded dihydrocaffeic acid 4¢-sulfate 24 in
34% overall yield from 42.
The preparation of dihydrocaffeic acid 4¢-O-b-D-glucuronide
18 required access to a 3¢-O-benzylated precursor (Scheme 2).
Our strategy was based on our previous observation that
methoxymethylation of quercetin only affected the most acidic
phenolic hydroxyl groups when N-ethyldiisopropylamine was used
as a base.26 Taking that into consideration, methoxymethylation
of protocatechualdehyde 26 was expected to mainly affect the
4¢-hydroxyl group (conjugated to the carbonyl). We found it
more convenient to combine the three steps a–c of Scheme 2
into a one pot protocol. By manipulating and optimizing the
extraction procedures, the side products of each reaction could
be eliminated. In fact, reaction of 26 with chloromethyl methyl
ether in presence of N-ethyldiisopropylamine, yielded 3¢-hydroxy-
4¢-methoxymethoxybenzaldehyde 34. Other compounds present in
the reaction medium were unreacted 3¢,4¢-dihydroxybenzaldehyde
26, and 3¢,4¢-bis(methoxymethoxy)benzaldehyde 35. Extraction of
the products of the reaction in dichloromethane under alkaline
conditions led to the elimination of unreacted protocatechualde-
hyde 26. The subsequent benzylation reaction (step b) was
carried out on the mixture of 34 plus 35. This resulted in the
formation of 3¢-benzyloxy-4¢-methoxymethoxybenzaldehyde 36,
while of course 35 was not affected by this step. Hydrolysis of
the MOM groups in step c led to a mixture of 3¢-benzyloxy-
4¢-hydroxybenzaldehyde 37 and protocatechualdehyde 26 which,
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 5199–5211 | 5201
©

View Online
Scheme 3 The preparations of caffeic acid 3¢-sulfate 11, and dihy-
drocaffeic acid 3¢-sulfate 23. a) Cl–SO3–C5H11, DBU, CH2Cl2, 91%; b)
Cyclohexene, 10% Pd/C, DMF, 94%; c) Ethyl (triphenylphosphoranyli-
den)acetate, CH2Cl2–THF, 83%; d) NaN3, DMF; e) 1 N aq. NaOH
then AmberliteꢀR CG-50, d + e: 43%; f) Ethyl (triphenylphosphoranyli-
den)acetate, CH2Cl2–THF, 86%; g) H2, 5% Pd/C, MeOH–H2O, 100%; h)
NaN3, DMF; i) 1 N aq. NaOH then AmberliteꢀR CG-50, h + i: 70%.
Scheme 2 The preparations of dihydrocaffeic acid 4¢-O-b-D-glucuronide
18, and of dihydrocaffeic acid 4¢-sulfate 24. a) Chloromethyl methyl ether,
iPr2NEt, THF; b) Benzyl bromide, K2CO3, MeCN; c) 37% HCl, MeOH;
a + b + c: 42% d) Ethyl (triphenylphosphoranyliden)acetate, CH2Cl2–THF,
56%; e) Compound 25, BF3 etherate, CH2Cl2, 73%; f) NaOH, MeOH/H2O,
then AmberliteꢀR IR-120, 78%; g) H2, 5% Pd/C, MeOH–H2O, 100%; h)
Cl–SO3–C5H11, Et3N, DMAP, THF, 86%; i) H2, 5% Pd/C, MeOH–H2O,
90%; j) NaN3, DMF; k) 1 N aq. NaOH then AmberliteꢀR CG-50, j + k:
34%.
In the preparation of caffeic acid 4¢-sulfate, we submitted 3¢,4¢-
diacetoxycinnamic acid ethyl ester 49 (prepared in quantitative
yield by the reaction of 3¢,4¢-diacetoxybenzaldehyde 27 with
ethyl (triphenylphosphoranyliden)acetate) to deacetylation using
thiophenol as acetyl acceptor (Scheme 4). This step turned out
not to be regioselective, and a 1/1 mixture of mono deacetylated
products 50 and 51 was obtained (overall mono deacetylation
yield: 54%). However, during sulfation of the mixture of 50 and
51 with neopentyl chlorosulfate, partial acetate hydrolysis and/or
rearrangement took place, and the 4¢-neopentyl sulfate ester 52
As the precursor of caffeic acid 3¢-sulfate 11, we used 4-
benzyloxy-3-hydroxybenzaldehyde 43.31,32 The latter compound
was sulfated with neopentyl chlorosulfate, to give the neopentyl
sulfate ester 44 in 91% yield (Scheme 3). Attempts to debenzylate
compound 44 using classical hydrogenation conditions (H2 + 5%
Pd/C), led to the reduction of the aldehyde functional group
as well. However, under milder catalytic transfer hydrogenation
conditions (cyclohexene + 10% Pd/C), only debenzylation took
place, leaving both the aldehyde and the neopentyl sulfate
ester groups intact. This yielded aldehyde 45 in excellent 94%
yield.
The latter compound was reacted with ethyl (triphenylphospho-
ranyliden)acetate, yielding the protected sulfated ethyl caffeate 46
in 83% yield. Elimination of the neopentyl (sodium azide), and
the ethyl ester (aq. NaOH) groups, gave the desired caffeic acid
3¢-sulfate 11 in 43% yield from 46, and in the form of a 9/1 mixture
of (E)- and (Z)-isomers.
Alternatively, aldehyde 44 was directly submitted to the Horner–
Wadsworth–Emmons reaction, to give protected sulfated ethyl
caffeate 47 in 86% yield (Scheme 3). The reduction of compound
47 directly afforded the debenzylated ethyl dihydrocaffeate 48
(quantitative yield), which was submitted to the usual deprotection
sequence (NaN3, followed by 1 N aq. NaOH), to give the desired
dihydrocaffeic acid 3¢-sulfate 23 in 70% yield from 48.
Scheme 4 The preparation of caffeic acid 4¢-sulfate 12. a) Ethyl (triph-
enylphosphoranyliden)acetate, CH2Cl2, 100%; b) Thiophenol, imidazole,
NMP, 54% of 50 + 51; c) Cl–SO3–C5H11, Et3N, DMAP, THF, 80% of 52;
d) NaN3, DMF; e) 1 N aq. NaOH then AmberliteꢀR CG-50, d + e: 67%.
5202 | Org. Biomol. Chem., 2010, 8, 5199–5211
This journal is
The Royal Society of Chemistry 2010
©

View Online
Table 1 13C NMR sulfation shiftsa for sulfate conjugates 7–12, and 19–24
were expressed in mmol of sulfate conjugate per mg of sample.
Depending on the conjugate, the quantification varied from 1.57
to 2.81 mmol/mg. The quantification enabled us to calculate the
yields of the reactions: Hydroxycinnamic acid sulfates, 43–92%
yield; dihydro hydroxycinnamic acid sulfates, 36–40%.
Compound
C-1¢
C-2¢
C-3¢
C-4¢
C-5¢
C-6¢
7
-3.7
0.5
-3.5
0.9
-1.2
-7.4
-5.7
-1
-4.8
1.1
1.1
-5.2
-0.3
-5.3
-6.4
-0.5
0.8
-5.2
-0.2
-5.3
-7.3
-2.7
-4.6
3.8
-2.5
4
4.4
-5
-4.6
0.5
1.1
-3.9
1.5
-3.7
-3.2
2.7
0.8
-4.1
0.8
8
9
4.4
-4.8
-0.4
-1.6
-7.0
-5.3
0.9
-5.6
-0.2
-2.8
-7.0
Previous studies on the metabolism of hydroxycinnamic acid
derivatives have been carried out on rats. Furthermore, in a number
of cases, the evidence for the presence of glucuronidated and/or
sulfated metabolites was only indirect. For example both the
glucuronide and sulfate conjugates of m-coumaric, dihydro-m-
coumaric, ferulic, and dihydroferulic acids have been shown to be
present in rat urine after administration of g-oryzanol.34 Ferulic
acid sulfate, glucuronide and sulfoglucuronide have been shown
to be metabolites of ferulic acid sugar esters in rats.35,36 During the
sole human trial previously performed, ferulic acid glucuronide or
sulfate has been found in human urine after consumption of a pine
bark extract.37 However, in all these studies, no identification of the
real circulating metabolites was made since the data was collected
after enzymatic deconjugation using b-glucuronidase and/or aryl
sulfatase.
A set of other identifications has been carried out online by chro-
matography and/or mass spectrometry techniques. When only one
position of conjugation is available, the LC-MS identification of
the metabolite was straightforward. In fact p-coumaric acid sulfate
has been identified as an urinary metabolite of nonylphenol,38 to-
gether with p-coumaric acid glucuronide, and dihydro-p-coumaric
acid sulfate and glucuronide. Similarly, administration of ferulic or
isoferulic acids led to the detection of their respective glucuronides
in rat urine.39 In contrast the identification of positional isomers
was problematic. For example caffeic acid sulfates have been
identified in rat urine by HPLC-MSn after administration of caffeic
acid,39 but the position of sulfation was not determined. Similarly
the two caffeic acid glucuronides have been identified in rat urine
and plasma after dosing of caffeic acid. In the absence of proper
standards, the identification of each isomer was only based on
an unconvincing comparison of their relative chromatographic
behavior with that of ferulic and isoferulic acids.39
The phenolic acid conjugates present in human plasma and in
human urine collected 0–24 h after drinking coffee containing
412 mmol of chlorogenic acids were investigated. The results of
peak co-elution with our 24 chromatographic standards were
combined with LC-MS-MS data (Table 2). All the conjugates
that could be detected in human plasma and urine after coffee
consumption were caffeic acid derivatives, or compounds resulting
from the methylation (ferulic and isoferulic), the reduction (dihy-
drocaffeic), or both methylation and reduction (dihydroferulic and
dihydroisoferulic) of caffeic acid.
Four conjugates were present in human plasma, i.e. caffeic-
and dihydrocaffeic acid 3¢-sulfates (11 and 23), ferulic- and
dihydroferulic acid 4¢-sulfate (9 and 11). Ten conjugates were
present in human urine, i.e. caffeic acid 3¢- and 4¢-sulfates (11 and
12), dihydrocaffeic 3¢-O-glucuronide and 3¢-sulfate (17 and 23),
ferulic acid 4¢-sulfate 9, isoferulic acid 3¢-sulfate 10, dihydroferulic
acid 4¢-O-glucuronide and 4¢-sulfate (15 and 21), isoferulic acid-
and dihydroisoferulic acid 3¢-O-glucuronides (4 and 16). No p-
coumaric acid derivative was detected in plasma or urine, probably
because of the low p-coumaroyl quinic acid content of coffee. On
the other hand, dehydroxylation (by microflora in the colon) of
caffeic acid to m-coumaric acid was a minor pathway since no
10
11
12
19
20
21
22
23
24
-2.6
-2.7
6.9
4.3
-4.8
4.0
-2.9
-6.6
3.9
4.6
-3.6
-5.3
3.9
-2.9
3.9
3.8
-9.3
-4
-6.9
4.2
-0.7
-8.3
a Sulfation shift = d non sulfated compound – d corresponding sulfated
conjugate.
was the major product of the reaction (80%), contaminated with
a small amount of the 3¢,4¢-neopentyl disulfate ester 53.
The latter contaminant was eliminated by MPLC on silica and
was not further characterized. Finally treatment of pure 52 by
sodium azide (removal of the neopentyl group), followed by 1 N
aq. NaOH (hydrolysis of the ethyl ester), yielded caffeic acid 4¢-
sulfate 12 in 67% yield from 52, again as a 9/1 mixture of (E)- and
(Z)-isomers.
The identification of all the conjugates was based on their
complete 1D and 2D NMR data, as well as HRMS analysis. In
addition the structure of all the sulfated conjugates was confirmed
on the basis of the carbon shifts, induced by sulfation, of the 13
C
NMR spectra. In fact, when comparing the spectra of phenolic
compounds with their sulfated analogues, the carbon bearing the
sulfate group is shifted upfield, while in contrast the carbons ortho
and para to this position are moved downfield.33 The sulfation
shifts measured for the complete panel of sulfate esters prepared
in the course of this work are summarized in Table 1.
All the shifts were in accordance with the expected structures.
The chromatographic purity of all the conjugates was determined
by HPLC-UV detection. All conjugates were at least 94% pure
(See Table S1 of ESI part B).† Sulfated conjugates, however,
represented a special case. First the sulfate ester linkage may
partially hydrolyze on storage, releasing the free hydroxycinnamic
acid. On the other hand, the last step of the synthesis involved
exposure to either sodium carbonate or sodium hydroxide and
consequently the final sulfated conjugates were possibly contami-
nated with non UV-absorbing inorganic material. This inorganic
salt was difficult to totally eliminate, even after purification by gel
filtration and/or reversed phase chromatography. Furthermore
we found out that the complete elimination of the inorganic
material tended to increase the risk of degradation by hydrolysis
of the sulfate ester bond. Thus, the presence of inorganic material
played a stabilization role on the sulfate conjugates. HPLC analysis
demonstrated that no free hydroxycinnamic acids were present
in our samples (Figures S7–S12 and S19–S24 of ESI part B).†
On the other hand examination of the NMR spectra of the
sulfate conjugates demonstrated that they were also free of other
organic contaminants (Figures S31–S36 and S43–S48 of ESI
part B).† Thus, the exact amount of sulfate conjugate present
in the samples was deduced after acid hydrolysis of an aliquot,
followed by the quantification of the released aglycone. The results
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 5199–5211 | 5203
©

View Online
Table 2 HPLC-MSn identification of chlorogenic acid derivatives and metabolites detected in the plasma and urine of human volunteers 0–24 h following
the ingestion of 200 mL of instant coffee beverage
HPLC Rt (min)a
[M - H]- (m/z)b
Metabolite
Location
10.3
11.1
11.7
12.1
12.8
13.9
14.5
16.9
17.7
19.6
261
357
259
275
259
371
273
273
371
369
Dihydrocaffeic acid 3¢-sulfate 23
Dihydrocaffeic acid 3¢-O-glucuronide 17
Caffeic acid 4¢-sulfate 12
Plasma, Urine
Urine
Urine
Plasma, Urine
Plasma, Urine
Urine
Plasma, Urine
Urine
Dihydroferulic acid 4¢-sulfate 21
Caffeic acid 3¢-sulfate 11
Dihydroferulic acid 4¢-O-glucuronide 15
Ferulic acid 4¢-sulfate 9
Isoferulic acid 3¢-sulfate 10
Dihydroisoferulic acid 3¢-O-glucuronide 16
Isoferulic acid 3¢-O-glucuronide 4
Urine
Urine
a Rt, retention time. b [M - H]-, negatively charged molecular ion.
m-coumaric derivatives were detected. Furthermore, in plasma, all
phenolic acid conjugates were found in the sulfated form, although
some glucuronides (4 out of 11 total) were found in urine. This
implies that glucuronides are produced in lower quantities and
that they are more rapidly cleared from the blood circulation.
Neither dihydroisoferulic acid 3¢-glucuronide, nor dihydrocaf-
feic acid sulfates or dihydrocaffeic acid glucuronides have been
previously identified. In the two latter cases, only one of the
two possible isomers was present in human plasma and/or
urine. Co-elution with our analytical standards unambiguously
identified these compounds as dihydrocaffeic acid 3¢-sulfate, and
dihydrocaffeic acid 3¢-glucuronide, respectively. Caffeic acid 3¢-
sulfate is also a novel compound and its 4¢-sulfate analogue is now
detected for the first time as human metabolites. Furthermore the
exact order of elution of the two isomers was determined with
accuracy. Finally both dihydroferulic acid 4¢-sulfate and isoferulic
acid 3¢-sulfate were novel human metabolites as well.
Sulfur trioxide-trimethylamine complex was purchased from
Lancaster. BF3-etherate, chloromethyl methyl ether, tributyltin
methoxide, and triphenylphosphine were obtained from Aldrich.
2,3,4-Tri-O-acetyl-D-methyl
glucuronopyranosyl-(N-phenyl)-
2,2,2-trifluoroacetimidate 25 was prepared by reaction of
N-phenyl-2,2,2-trifluoroacetimidoyl chloride with methyl 2,3,4-
tri-O-acetyl-D-glucuronate. The former compound was prepared
in one step,40 by heating a mixture of trifluoroacetic acid and
aniline in carbon tetrachloride in the presence of triphenyl-
phosphine and triethylamine. The latter compound was prepared
in two steps by established procedures from D-(+)-glucuronic
acid g-lactone. This yielded first methyl 1,2,3,4-tetra-O-acetyl-
b-D-glucuronate,41 and secondly methyl 2,3,4-tri-O-acetyl-D-
glucuronate after deacetylation with tributyltin methoxide.42
4-Benzyloxy-3-hydroxybenzaldehyde,32 and neopentyl chloro-
sulfate,28,30 were prepared according to published procedures.
General methods
Conclusion
Thin layer chromatography (TLC) was carried out on Silica 60
F254 (Merck), RP-18 F254s and RP-18 WF254s (Merck), DIOL F254s
(Merck) or CN F254s plates. The detection was carried out by UV
254 nm or after spraying with Pancaldi reagent. Vacuum liquid
chromatography (VLC) was performed on a 500 mL sintered glass
filter (100 ¥ 80 mm, P2 porosity) filled with 170 g of silica gel 60
(40–63 mM) Merck, and using a gradient of EtOAc in hexane.
Alternatively, the filter was filled with an octadecyl-bonded phase
(C18) for flash chromatography (40 mM average particle diameter,
Merck), using a gradient of MeOH in water as solvent. Column
chromatography (CC) was performed on a 22 ¥ 400 mm column,
filled with 100 g of silica gel 60 (63–200 mM) Merck; or on a 22 ¥
This is the first time that such an extended panel of potential
conjugates is produced and used as analytical standards for
unequivocal identifications. Out of the 24 synthesized potential
conjugates, 4 have never identified before in any animal or human
circulating fluid, and are now evidenced for the first time to
be hydroxycinnamic acid metabolites. One the other hand, 7
metabolites are novel conjugates in humans.
Experimental Section
Source of chemicals and materials
R
30 mm column filled with 50 g of Sephadexꢀ LH 20 (Amersham
GF 92 glass fiber prefilters were purchased from Schleicher &
Schuell (Dassel, Germany). Acetyl chloride, benzyl alcohol,
diethyl azodicarboxylate (DEAD), 3,4-dihydroxybenzaldehyde,
dry dichloromethane, dry THF, dry methanol, N-ethyldiiso-
propylamine, ethyl (triphenylphosphoranyliden)acetate, D-
(+)-glucuronic acid g-lactone, trifluoroacetic acid, 3-(3,4-
dihydroxyphenyl)propionic acid, dry DMF, sodium azide, Pd/C
5 and 10%, cyclohexene, Amberliteꢀ IR 120, Amberliteꢀ CG
50, sodium carbonate, potassium carbonate, 1,8-diazabicyclo-
[5,4,0]undec-7-ene (DBU), triethylamine (TEA), tetrabutyl-
ammonium bromide and benzyl bromide were from Fluka.
Bioscience), using MeOH–water 1/1 as solvent (the flow rate was
10 mL/10 min and fractions of 10 mL were collected); or on a 22 ¥
30 mm column filled with 20 g of Lichropep RP-18 Merck, using
water (100 mL), and water–MeOH 9/1 (100 mL), successively.
Medium pressure liquid chromatography (MPLC) was carried out
on a Bu¨chi apparatus comprising a 688 chromatographic pump, a
R
684 fraction collector, and equipped with a C 690 Sepacoreꢀ glass
R
R
column (dimensions: 36 ¥ 460 mm, volume 470 mL) plus a pre-
column (volume 11 mL). The column plus the pre-column were
filled with 222 g of silica gel 60 (40–63 mm), Merck. For elution, a
gradient of EtOAc in hexane was used. Alternatively MPLC was
5204 | Org. Biomol. Chem., 2010, 8, 5199–5211
This journal is
The Royal Society of Chemistry 2010
©

View Online
performed on a Biotage HPFC SP1 system, using normal silica
gel or RP-18 manufactured columns (Biotage). The solvents were
mixtures of hexane in EtOAc for normal phase and MeOH in water
for reverse phase. The gradients were calculated by the software,
on the basis of the TLC behavior. A multiwavelength detector
and a fraction collector completed the equipment. High pressure
liquid chromatography (HPLC) was carried out on a Agilent 1100
tower including a binary pump, an autosampler, a column oven,
a UV detector (DAD 1040A), and a CC 250/4Nucleosil 100-
5C18 column (Macherey Nagel). The flow rate was 1 mL/min,
with an injection volume of 5 mL. The column was held at 40 ◦C
and the UV detector set to 210, 254, and 280 nm. The following
solvents were used. Method A: solvent A: 0.1% TFA in water;
solvent B: MeCN; gradient from 10 to 100% B over 20 min;
5 min 100% B; from 0 to 90% A over 1 min; 4 min 90% A.
Method B: solvent A: 0.1% TFA in water; solvent B: methanol–
water (1/1); gradient from 20 to 50% B over 15 min; from 50 to
90% B over 10 min; 10 min 90% B. 1H NMR (360.13 MHz) and
13C NMR (90.56 MHz): The spectra were recorder on a Bruker
DPX-360 spectrometer equipped with a broadband multinuclear
z-gradient probehead. The chemical shifts (in ppm) were expressed
with respect to tetramethylsilane (TMS) as an internal reference,
and using CDCl3 as solvent. The attribution of the signals was
based on the complete interpretation of direct and long dis-
tance heteronuclear correlations. Liquid Chromatography-High-
Resolution Mass Spectrometry (ESI-) was carried out on an
Agilent-1200 series LC System including a binary pump SL, an
autosampler, a diode-array detector and a thermostatted column
compartment with an Agilent Eclipse XDB (44.6 ¥ 150 mm, 5 mm)
coupled with an Agilent 6210 time-of-flight mass spectrometer.
resulting suspension was stirred at rt for 4 h. The medium was
diluted with 40 mL of H2O, cooled to 0 ◦C (ice bath) and
R
neutralized (to pH 4.0–4.5) by addition of 4–6 g of Amberliteꢀ
IR-120. The resin was eliminated by filtration and washed with
3 ¥ 50 mL of H2O. The filtrate plus the three washings were mixed,
concentrated under reduced pressure to a volume of about 5 mL
(or lyophilized and the residue dissolved in 5 mL of H2O), and
directly purified by MPLC on RP-18 using a gradient of MeOH
in H2O as solvent.
General procedure D: reduction of hydroxycinnamic acids, or
hydroxycinnamic acid glucuronides to 3-(hydroxyphenyl)propionic
acids, or 3-(hydroxyphenyl)propionic acid glucuronides
To 100 mg of hydroxycinnamic acid or hydroxycinnamic acid
glucuronide, dissolved in 5 mL of water plus 5 mL of MeOH,
10 mg of palladium (5%) on activated carbon were added. The
suspension was submitted to hydrogenation for 2–4 h (until no
hydrogen consumption). After filtration on GF 92 (Schleicher-
Schuell, 421030) and evaporation of the solvent, the product
was dried overnight under vacuum (0.01 mbar). If necessary,
the residue was purified by MPLC on RP-18 using a gradient
of MeOH in 1% aqueous AcOH as solvent.
General procedure E: synthesis of protected caffeic and
3-(3,4-dihydroxyphenyl)propionic acid 3¢- and 4¢-sulfates
A solution of 1 part of the starting material, 2 parts of DBU,
and 2 parts of neopentyl chlorosulfate in dry dichloromethane
(10 mL/mmol), or a solution of 1 part of the starting material,
2 parts of DMAP, 2.5 parts of TEA, and 2.5 parts of neopentyl
chlorosulfate in dry THF (10 mL/mmol) was stirred at room
temperature for 24 h under nitrogen atmosphere. The medium was
directly purified by MPLC (Biotage) on silica, using a gradient of
EtOAc in hexane as solvent.
General procedure A: synthesis and purification of ethyl
hydroxycinnamic acids.
In a 250 mL Erlen flask, one part of the desired hydroxybenzalde-
hyde was dissolved in dry CH2Cl2, or in a mixture of dry THF +
dry CH2Cl2. One part of ethyl (triphenylphosphoranyliden)acetate
was added and the reaction was allowed to take place at rt under
stirring for 4 h. The medium was directly purified by VLC on silica,
using a gradient of EtOAc in hexane as solvent.
General procedure F: deprotection of protected caffeic and
3-(3,4-dihydroxyphenyl)propionic acid 3¢- and 4¢-sulfates
A solution of 1 part of the protected acid, and 2 parts of sodium
azide in DMF (3 mL/mmol of acid) was stirred at 60 ◦C for 24 h.
The DMF was removed under high vacuum (1 mbar/40 ◦C) and
methanol (2 mL/mmol of starting material) and 1.0 M NaOH
(4 mL/mmol of starting material) were added to the residue. The
solution was stirred at room temperature for 4 h. After dilution
with water (5 to 10 times) the solution was brought to pH 6 to
General procedure B: synthesis and purification of protected
hydroxycinnamic acid O-b-D-glucuronides
One part of hydroxycinnamic acid ethyl ester and 2 parts of
2,3,4-triacetyl-D-methyl glucuronopyranosyl-(N-phenyl)-2,2,2-
trifluoroacetimidate 25 were dissolved in 10 mL of dry CH2Cl2.
0.30 parts of BF3 etherate were added and the mixture was stirred
at rt for 15 h. The medium was diluted with 50 mL of CH2Cl2,
washed with 3 ¥ 50 mL of H2O and concentrated under reduced
pressure. The liquid residue was purified by MPLC on silica using a
gradient of EtOAc in hexane as solvent. The fractions containing
the product were further purified by MPLC on RP-18, using a
gradient of MeOH in 50% aqueous MeOH as solvent.
R
6.5 with Ambertiteꢀ CG50. The resin was eliminated by filtration
through a P zero sintered glass filter. The filtrate was filtered again
though a GF 92 pre-filter and lyophilized. The filtrate was purified
by MPLC (Biotage) on RP-18 using a gradient of MeOH in
water as solvent. After evaporation of the methanol under reduced
pressure, the water solution was freeze-dried.
General procedure G: quantification of the amount of sulfate
conjugate in the isolated products
General procedure C: deprotection of the protected glucuronides
and purification of hydroxycinnamic acid O-b-D-glucuronides
Before acid hydrolysis, the sulfate conjugates were analyzed
by HPLC to check for the absence of free acids in the sam-
ple. 5–8 mg of hydroxycinnamic acid sulfate were hydrolyzed
To solid protected hydroxycinnamic acid O-b-D-glucuronides
were added 10 mL of 1 N aq. NaOH and 10 mL of MeOH. The
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 5199–5211 | 5205
©

View Online
to the corresponding acids in 1 mL of water, plus 1 mL
of 6 M aq. HCl, for 1 h at room temperature in Pyrex
tubes (SVL18/10 ml) under stirring (shaker). After evaporation
(3H/50 ◦C/0.1 mbar/Jouan RC10.22/vacuum concentrator/
centrifugal evaporator), the residue was taken in water–THF 1/1,
and transferred into a 50 mL volumetric flask. A 20 mL aliquot of
the solution was analyzed by HPLC and the free hydroxycinnamic
acid quantified. A similar procedure was followed for the hydrol-
ysis of 3-(hydroxyphenyl)propionic acid sulfates, except that due
to the lower UV response of this class of compounds, samples
of 10 mg were hydrolyzed. Calibration curves (4 points) were
obtained by submitting 2, 5, 8, and 12 mg of hydroxycinnamic
acid, or 5, 10, 15, or 20 mg of 3-(hydroxyphenyl)propionic acid to
the same procedure. The results were expressed in mmol of sulfate
conjugate/mg of sample.
NMR (90 MHz, CDCl3) d 20.9 (3¢-OAc), 117.6 (C5¢), 124.2 (C2¢),
129.8 (C1¢ + C6¢), 138.7 (C3¢), 153.5 (C4¢), 169.2 (3¢-OAc), 190.6
(C1). H NMR (300 MHz, Acetone-d6) d 2.30 (s, 3H, 3¢-OAc),
7.15 (d, 1H, J = 8.3 Hz, H5¢), 7.63 (d, 1H, J = 2.0 Hz, H2¢), 7.72
(dd, 1H, J = 8.3 and 2.0 Hz, H6¢), 9.86 (s, 1H, H1). 13C NMR
(75 MHz, Acetone-d6) d 20.7 (3¢-OAc), 118.0 (C5¢), 125.0 (C2¢),
130.2 (C6¢), 130.6 (C1¢), 140.0 (C3¢), 155.6 (C4¢), 169.3 (3¢-OAc),
190.7 (C1).
1
3¢-Acetoxy-4¢-benzyloxybenzaldehyde (29)
1.7088 g (9.485 mmol) of 3¢-acetoxy-4¢-hydroxybenzaldehyde 28
were dissolved in 20 mL of dry THF. 3.23 g (12.31 mmol) of triph-
enylphosphine and 1.5 mL (14.5 mmol) of benzyl alcohol were
added. The medium was cooled to 0 ◦C (ice bath), equilibrated
under argon, and 6.6 mL (15.16 mmol) of a solution of 40% of
diethyl azodicarboxylate (DEAD) in toluene were added in 15 min
dropwise under stirring. The medium was allowed to warm up
slowly to room temperature, and left to react at rt for 17 h. The
medium was diluted with 100 mL of CH2Cl2 and the extract was
washed with 100 mL of H2O. The CH2Cl2 extract was purified
by VLC using a gradient of EtOAc in hexane as solvent, to give
2.5 g (9.25 mmol, 97%) of compound 29. TLC Rf = 0.37 (Silica
3¢,4¢-Diacetoxybenzaldehyde (27)
In a 250 mL 3-neck round-bottom flask, equipped with a
condenser and a dropping funnel, 4 g (28.96 mmol) of 3¢,4¢-
dihydroxybenzaldehyde 26 were suspended in 40 mL of CH2Cl2.
Acetyl chloride (6 mL, 84.46 mmol) was added dropwise under
stirring to this suspension. On the other hand, 13 mL (87.01 mmol)
of 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) were diluted with
20 mL of CH2Cl2 and introduced into the dropping funnel. The
round-bottom flask was cooled to 0 ◦C (ice bath) and the apparatus
was equilibrated under argon. The DBU solution was added to the
3,4-dihydroxybenzaldehyde suspension dropwise under stirring at
0 ◦C under inert atmosphere. After complete addition of the DBU,
total dissolution of the solid was observed. The medium was
removed from the ice bath and allowed to react at rt under stirring
for 14 h. The CH2Cl2 medium was washed with 100 mL of 1 N aq.
HCl and with 4 ¥ 100 mL of H2O, and evaporated under reduced
pressure. The residue was dissolved in 5 mL of EtOAc and 5 mL
of hexane were added. This solution was purified by VLC using
a gradient of EtOAc in hexane as solvent. This yielded 5.6844 g
(25.58 mmol, 88%) of compound 27. TLC Rf = 0.46 (Silica 60
F254, Hexane–EtOAc 6/4). The 1H-, and 13C NMR spectral data
of 27 were in accordance with published values.43–45
1
60 F254, Hexane–EtOAc 4/1). H NMR (300 MHz, CDCl3) d
2.29 (s, 3H, 3¢-OAc), 5.18 (s, 2H, H7b), 7.10 (d, 1H, J = 8.5 Hz,
H5¢), 7.34–7.39 (m, 5H, H2b + H3b + H4b + H5b + H6b), 7.61
(d, 1H, J = 2.0 Hz, H2¢), 7.71 (dd, 1H, J = 8.5 and 2.1 Hz, H6¢),
9.85 (s, 1H, H1). 13C NMR (75 MHz, CDCl3) d 20.5 (3¢-OAc),
70.8 (C7b), 113.3 (C5¢), 123.4 (C2¢), 127.1 (C2b/6b), 128.3 (C4b),
128.7 (C3b/5b), 130.1 (C6¢), 130.1 (C1¢), 135.6 (C1b), 140.6 (C3¢),
155.4 (C4¢), 168.7 (3¢-OAc), 190.1 (C1). The NMR data was in
accordance with previously published information.32
3¢-Acetoxy-4¢-benzyloxycinnamic acid ethyl ester (30)
Procedure A was carried out using 2.00 g (7.4 mmol) of 3¢-acetoxy-
4¢-benzyloxybenzaldehyde 29 in 10 mL of dry CH2Cl2, and 2.6 g
(7.47 mmol) of ethyl (triphenylphosphoranyliden)acetate. This
afforded, after purification, 2.2441 g (6.59 mmol; 89%) of novel
1
compound 30. H NMR (300 MHz, CDCl3) d 1.32 (t, 3H, J =
7.1 Hz, H2¢¢), 2.29 (s, 3H, 3¢-OAc), 4.25 (q, 2H, J = 7.1 Hz, H1¢¢),
5.12 (s, 2H, H7b), 6.29 (d, 1H, J = 16.0 Hz, H2), 6.98 (d, 1H, J =
8.5 Hz, H5¢), 7.26 (d, 1H, J = 2.6 Hz, H2¢), 7.32 (dd, 1H, J = 8.5
and 2.2 Hz, H6¢), 7.31–7.38 (m, 5H, H2b + H3b + H4b + H5b +
H6b), 7.59 (d, 1H, J = 16.0 Hz, H3). 13C NMR (75 MHz, CDCl3)
d 14.3 (C2¢), 20.6 (3¢-OAc), 60.4 (C1¢¢), 70.7 (C7b), 113.8 (C5¢),
117.0 (C2), 122.0 (C2¢), 127.1 (C2b/6b), 127.5 (C6¢), 127.9 (C1¢),
128.1 (C4b), 128.6 (C3b/5b), 136.2 (C1b), 140.4 (C3¢), 143.3 (C3),
151.9 (C4¢), 167.0 (C1), 168.9 (3¢-OAc).
3¢-Acetoxy-4¢-hydroxybenzaldehyde (28)
3.00 g (13.56 mmol) of 3¢,4¢-diacetoxybenzaldehyde 27 were
dissolved in 30 mL of THF. The solution was cooled to 0 ◦C
in an ice bath and 350 mg (5.1307 mmol) of imidazole, followed
by 1.65 mL (16.2 mmol) of thiophenol, were added. The reaction
medium was removed from the ice bath and allowed to react at rt
for 1.5 h under stirring. 100 mL of CH2Cl2 were added to the
medium. The organic solution was washed with 100 mL of 1
N aq. HCl, and with 3 ¥ 100 mL of H2O. The CH2Cl2 extract
was concentrated to constant volume under reduced pressure. The
liquid residue was purified by MPLC on silica using a gradient of
EtOAc in hexane as solvent. This afforded 1.8763 g (10.41 mmol,
77%) of compound 28. The 1H and 13C NMR data for 3¢-acetoxy-
4¢-hydroxybenzaldehyde in methanol-d4 has been published.46 The
NMR data in CDCl3 and acetone-d6 are presented here for the
first time. TLC Rf = 0.21 (Silica 60 F254, Hexane–EtOAc 6/4).1H
NMR (360 MHz, CDCl3) d 2.36 (s, 3H, 3¢-OAc), 7.07 (d, 1H,
J = 9.0 Hz, H5¢), 7.66 (m, 2H, H2¢+H6¢), 9.82 (s, 1H, H1). 13C
Ethyl 4¢-Benzyloxy-3¢-hydroxycinnamate (31)
In a 100 mL Erlen flask, 401.4 mg (1.18 mmol) of 3¢-acetoxy-4¢-
benzyloxycinnamic acid ethyl ester 30 were dissolved in 2 mL of
dry CH2Cl2. 95.23 mg (1.76 mmol) of sodium methoxide, dissolved
in 15 mL of dry MeOH were added, and the mixture was stirred
at rt for 30 min. The medium was acidified by addition of 10 mL
of 1 N aq. HCl, diluted with 50 mL of H2O, and extracted with
50 mL of CH2Cl2. The CH2Cl2 extract was washed with 3 ¥ 10 mL
of H2O, and evaporated under reduced pressure, to give 344.1 mg
5206 | Org. Biomol. Chem., 2010, 8, 5199–5211
This journal is
The Royal Society of Chemistry 2010
©

View Online
(1.15 mmol; 97%) of compound 31. TLC Rf = 0.49 (Silica 60 F254,
Hexane–EtOAc 7/3). 1H NMR (300 MHz, CDCl3) d 1.33 (t, 3H,
J = 7.2 Hz, H2¢¢), 4.25 (q, 2H, J = 7.2 Hz, H1¢¢), 5.14 (s, 2H, H7b),
5.74 (s, 1H, 3¢-OH), 6.28 (d, 1H, J = 15.9 Hz, H2), 6.90 (d, 1H,
J = 8.3 Hz, H5¢), 7.00 (dd, 1H, J = 8.3 and 2.0 Hz, H6¢), 7.15 (d,
1H, J = 2.0 Hz, H2¢), 7.36–7.42 (m, 5H, H2b + H3b + H4b +
H5b + H6b), 7.58 (d, 1H, J = 15.9 Hz, H3). 13C NMR (75 MHz,
CDCl3) d 14.3 (C2¢¢), 60.4 (C1¢¢), 71.1 (C7b), 111.9 (C5¢), 113.3
(C2¢), 116.5 (C2), 121.6 (C6¢), 127.8 (C2b/6b), 128.4 (C1¢), 128.6
(C4b), 128.8 (C3b/5b), 135.8 (C1b), 144.3 (C3), 146.1 (C3¢), 147.6
(C4¢), 167.3 (C1). The 1H NMR data was in accordance with the
previously published one.47
Dihydrocaffeic acid 3¢-O-b-D-glucuronide (17)
145 mg 0.325 mmol) of (E)-4¢-O-benzylcaffeic acid 3¢-O-b-D-
glucuronide 33 was dissolved in 20 mL of 50% aq. MeOH. 22 mg
of 5% Pd/C were added and the compound was hydrogenated
at rt and 1 atmosphere for 4 h under stirring. The solid was
removed by filtration on a GF 92 glass fibre prefilter and the
filter was washed with 2 ¥ 5 mL of 50% aq. MeOH. The filtrate
plus the two washings were mixed, concentrated under reduced
pressure (removal of MeOH) and the resulting aqueous phase
was lyophilized, to give 111.69 mg (0.312 mmol; 96%) of the novel
compound dihydrocaffeic acid 3¢-O-b-D-glucuronide 17. TLC Rf =
0.29 (Silica 60 F254, BuOH–AcOH–water 4/1/1). HPLC Rt = 5.46
(method A; 99% purity at 280 nm). 1H NMR (360 MHz, DMSO-
d6) d 2.44 (t, 2H, J = 7.5, H2), 2.69 (t, 2H, J = 7.6, H3), 3.27–3.42
(m, 3H, H2a + H3a + H4a), 3.83 (d, J = 9.5 Hz, 1H, H5a), 4.83
(d, 1H, J = 7.5 Hz, H1a), 6.71 (brs, 2H, H6¢), 6.90 (brs, 1H, H5¢).
13C NMR (90 MHz, DMSO-d6) d 29.9 (C3), 35.6 (C2), 71.4 (C4a),
73.1 (C2a), 75.3 (C3a + C5a), 101.8 (C1a), 115.9 (C2¢), 116.8 (C5¢),
122.6 (C6¢), 131.7 (C1¢), 144.7 (C3¢), 145.2 (C4¢), 170.3 (C6a) 173.8
(C1). LC-HRMS [M - H]- calc for C15H17O10: 357.0821; found:
357.0816.
2-Propenoic acid, 3-[4-benzyloxy-3-[(2,3,4-tri-O-acetyl-6-methyl-
b-D-glucopyranurosyl)oxy]phenyl]-, ethyl ester (32)
Using procedure B, 211.3 mg (0.71 mmol) of 4¢-benzyloxy-3¢-
hydroxycinnamic acid ethyl ester 31 were reacted with 0.716 g
(1.42 mmol) of 2,3,4-triacetyl-D-methyl glucuronopyranosyl-
(N-phenyl)-2,2,2-trifluoroacetimidate 25,48 and with 30 mL
(0.24 mmol) of BF3 etherate. This yielded, after purification by
MPLC on silica and RP-18, 332.1 mg (0.54 mmol; 76%) of novel
compound 32. TLC Rf = 0.10 (Silica 60 F254, Hexane–EtOAc
7/3). 1H NMR (360 MHz, CDCl3) d 1.33 (t, 3H, J = 7.1 Hz, H2¢¢),
1.78 (s, 3H, CH3–CO), 2.03 (s, 3H, CH3–CO), 2.04 (s, 3H, CH3–
CO), 3.78 (s, 3H, H7a), 4.11 (d, 1H, J = 9.3 Hz, H5a), 4.25 (q, 2H,
J = 7.1 Hz, H1¢¢), 5.09 (m, 1H, H1a), 5.11 (s, 2H, H7b), 5.30–5.39
(m, 3H, H2a + H3a + H4a), 6.29 (d, 1H, J = 15.9 Hz, H2), 6.94
(d, 1H, J = 8.5 Hz, H5¢), 7.21 (dd, 1H, J = 8.5 and 2.1 Hz, H6¢),
7.31–7.41 (m, 5H, H2b + H3b + H4b + H5b + H6b), 7.42 (d,
1H, J = 2.1 Hz, H2¢), 7.58 (d, 1H, J = 16.0 Hz, H3). 13C NMR
(90 MHz, CDCl3) d 14.3 (C2¢¢), 20.3 (CH3–CO), 20.5 (CH3–CO),
20.6 (CH3–CO), 53.0 (C7a), 60.4 (C1¢), 69.1 (C4a), 70.9 (C7b +
C2a), 72.0 (C3a), 72.5 (C5a), 100.4 (C1a), 114.2 (C5¢), 116.8 (C2),
119.6 (C2¢), 125.6 (C6¢), 127.4 (C2b/6b), 128.0 (C1¢), 128.2 (C4b),
128.7 (C3b/5b), 136.2 (C1b), 143.7 (C3), 146.1 (C3¢), 151.4 (C4¢),
166.8 (C6a), 167.1 (C1), 169.3 (2 ¥ CH3–CO), 170.1 (CH3–CO).
3¢-Benzyloxy-4¢-hydroxybenzaldehyde (37) from
3¢,4¢-dihydroxybenzaldehyde (26) (3 steps)
Step 1: 3 g (21.72 mmol) of 3¢,4¢-dihydroxybenzaldehyde 26 were
dissolved in 10 mL of THF. The solution was cooled in an ice bath.
4.5 mL (26.06 mmol) of N-ethyl diisopropylamine were added
dropwise with stirring. Finally 2 mL (26.06 mmol) of chloromethyl
methyl ether were added dropwise under stirring. The medium was
removed from the ice bath and allowed to react at rt under stirring
for 4 h. The medium was diluted with 50 mL of 10% aq. KHCO3
and extracted with 2 ¥ 50 mL of CH2Cl2. The CH2Cl2 extract was
washed with 50 mL of 1 N aq. HCl, and 3 ¥ 50 mL of H2O. It
was evaporated under reduced pressure, to give 3.7928 g of crude
residue. TLC Rf = 0.51 (Silica 60 F254, Hexane–EtOAc 6/4).
Step 2: the 3.7928
g of crude 3¢-hydroxy-4¢-methoxy-
methoxybenzaldehyde 34 were dissolved in 50 mL of MeCN.
3 g (21.72 mmol) of anhydrous K2CO3, 0.71 g (2.2 mmol) of
tetrabutylammonium bromide, and 2.6 mL (21.72 mmol) of benzyl
bromide were added, successively at rt under stirring. The medium
was allowed to react at rt under stirring for 1 h. The medium was
diluted with 100 mL of H2O and extracted with 2 ¥ 100 mL of
EtOAc. The EtOAc extract was washed with 3 ¥ 200 mL of H2O
and concentrated under reduced pressure, to give 5.33 g of crude
product. TLC Rf = 0.09 (RP-18 F254s, MeOH–water 6/4).
(E)-4¢-O-Benzylcaffeic acid 3¢-O-b-D-glucuronide (33)
Following procedure C, 314.9 mg (0.512 mmol) of 2-
propenoic acid, 4-[4-benzyloxy-3-[(2,3,4-tri-O-acetyl-6-methyl-b-
D-glucopyranurosyl)oxy]phenyl]-, ethyl ester 32 were treated with
10 mL of 1 N aq. NaOH. After acidification (6.3 g of wet
R
Amberliteꢀ IR-120) and purification by MPLC on RP-18, 167 mg
of the novel (E)-4¢-O-benzylcaffeic acid 3¢-O-b-D-glucuronide 33
were isolated (yield: 154 mg, 0.373 mmol, 73%). HPLC Rt = 30.31
(method B). 1H NMR (360 MHz, DMSO-d6) d 3.31–3.45 (m, 3H,
H2a + H3a + H4a), 3.98 (d, 1H, J = 9.6 Hz, H5a), 5.20 (s, 2H,
H7b), 5.28 (d, 1H, J = 7.5 Hz, H1a), 6.42 (d, 1H, J = 15.9 Hz, H2),
7.06 (d, 1H, J = 8.5 Hz, H5¢), 7.24 (dd, 1H, J = 8.5 and 1.8 Hz,
H6¢), 7.31 (m, 1H, H4b), 7.38 (m, 2H, H3b + H5b), 7.46–7.49 (m,
3H, H2¢ + H2b+ H6b), 7.50 (d, 1H, J = 15.9 Hz, H3). 13C NMR
(90 MHz, DMSO-d6) d 69.7 (C7b), 71.2 (C4a), 72.8 (C2a), 75.3
(C5a), 76.0 (C3a), 99.5 (C1a), 114.5 (C5¢), 114.7 (C2¢), 117.0 (C2),
123.6 (C6¢), 127.3 (C1¢), 127.3 (C2b + C6b), 127.6 (C4b), 128.3
(C3b + C5b), 136.8 (C1b), 143.7 (C3), 146.5 (C3¢), 149.8 (C4¢),
167.7 (C1), 170.1 (C6a).
Step 3: the 5.33
g of crude 3¢-benzyloxy-4¢-methoxy-
methoxybenzaldehyde 36 were dissolved in 20 mL of MeOH and
2 mL of 37% HCl (fuming) were added dropwise under stirring.
The medium was refluxed (oil bath 80 ◦C) under stirring for 1 h.
The medium was diluted with 50 mL of 10% aq. KHCO3 and
extracted with 2 ¥ 50 mL of EtOAc. The EtOAc extract was washed
with 50 mL of 1 N aq. HCl, and 3 ¥ 50 mL of H2O. It was purified
by VLC on silica, followed by VLC on RP-18, using gradients
of EtOAc in hexane, and of MeOH in 20% aqueous MeOH
as solvents, respectively. TLC Rf = 0.40 (RP-18 F254s, MeOH–
water 7/3). Final yield in 3¢-benzyloxy-4¢-hydroxybenzaldehyde
37:49,50 2.0898 g (9.16 mmol; 42% from 26). 1H NMR (300 MHz,
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 5199–5211 | 5207
©

View Online
1
Acetone-d6) d 5.25 (s, 2H, H7b), 7.05 (d, 1H, J = 8.1 Hz, H5¢),
7.34–7.42 (m, 3H, H3b + H4b + H5b), 7.47 (dd, 1H, J = 8.1 and
1.8 Hz, H6¢), 7.53 (m, 2H, H2b/6b), 7.56 (d, 1H, J = 1.8 Hz, H2¢),
9.81 (s, 1H, H1). 13C NMR (75 MHz, Acetone-d6) d 71.4 (C7b),
112.6 (C2¢), 116.4 (C5¢), 127.1 (C6¢), 128.9 (C4b + C2b/6b), 129.3
(C3b/5b), 130.6 (C1¢), 137.5 (C1a), 148.0 (C3¢), 153.8 (C4¢), 191.0
(C1).
HPLC Rt = 27.85 (method B). H NMR (360 MHz, DMSO-d6)
d 3.30–3.45 (m, 3H, H2a + H3a + H4a), 3.89 (d, 1H, J = 9.6 Hz,
H5a), 5.17 (d, 1H, J = 7.3 Hz, H1a), 5.20 (s, 2H, H7b), 6.43 (d, 1H,
J = 16.0 Hz, H2), 7.12 (d, 1H, J = 8.6 Hz, H5¢), 7.23 (dd, 1H, J =
8.5 and 1.6 Hz, H6¢), 7.29–7.51 (m, 7H, H3 + H2¢ + H2b+ H3b +
H4b + H5b + H6b). 13C NMR (90 MHz, DMSO-d6) d 69.9 (C7b),
71.2 (C4a), 72.8 (C2a), 75.3 (C5a), 75.9 (C3a), 99.7 (C1a), 113.7
(C2¢), 115.8 (C5¢), 117.4 (C2), 122.5 (C6¢), 127.5 (C2b + C6b),
127.6 (C4b), 128.3 (C3b + C5b), 128.5 (C1¢), 137.0 (C1b), 143.6
(C3), 148.1 (C3¢), 148.3 (C4¢), 167.7 (C1), 170.1 (C6a).
3¢-Benzyloxy-4¢-hydroxycinnamic acid ethyl ester (38)
Procedure A was carried out using 1.9 g (8.32 mmol) of 3¢-
benzyloxy-4¢-hydroxybenzaldehyde 37 in 10 mL of dry THF +
10 mL of dry CH2Cl2, and 2.96 g (8.5 mmol) of ethyl (triph-
enylphosphoranyliden)acetate. This afforded, after purification by
VLC on silica, 2.34 g of product. The compound was further
purified by CC on silica using hexane–EtOAc 75/25 as solvent,
to give 1.39 g (4.66 mmol; 56%) of compound 38.51 TLC Rf =
Dihydrocaffeic acid 4¢-O-b-D-glucuronide (18)
Following procedure D, 300 mg (0.65 mmol) of compound 40
were hydrogenated for 5 h. This yielded after filtration and freeze-
drying 240 mg (0.67 mmol; 100%) of the novel compound 18.
1
HPLC Rt = 5.23 (method A; 98% purity at 280 nm). H NMR
1
0.28 (Silica 60 F254, Hexane–EtOAc 2/1). H NMR (300 MHz,
(360 MHz, DMSO-d6) d 2.46 (t, 2H, J = 7.6, H2), 2.69 (t, 2H,
J = 7.6, H3), 3.27–3.42 (m, 3H, H2a + H3a + H4a), 3.83 (d, 1H,
J = 9.6 Hz, H5a), 4.78 (d, 1H, J = 7.4 Hz, H1a), 6.59 (dd, 1H,
J = 8.3 and 2.0 Hz, H6¢), 6.68 (d, 1H, J = 7.4 Hz, H2¢), 6.90 (d,
1H, J = 8.3 Hz, H5¢). 13C NMR (90 MHz, DMSO-d6) d 29.7 (C3),
35.3 (C2), 71.3 (C4a), 73.0 (C2a), 75.1 (C5a), 75.3 (C3a), 102.1
(C1a), 115.9 (C2¢), 116.7 (C5¢), 118.8 (C6¢), 135.9 (C1¢), 143.1
(C4¢), 146.7 (C3¢), 170.0 (C6a) 173.7 (C1). LC-HRMS [M - H]-
calc for C15H17O10: 357.0821; found: 357.0824.
CDCl3) d 1.32 (t, 3H, J = 7.1 Hz, H2¢¢), 4.25 (q, 2H, J = 7.1 Hz,
H1¢¢), 5.13 (s, 2H, H7b), 5.92 (brs, 1H, 4¢-OH), 6.26 (d, 1H, J =
15.9 Hz, H2), 6.94 (d, 1H, J = 8.2 Hz, H5¢), 7.09 (m, 2H, H2¢ +
H6¢), 7.37–7.43 (m, 5H, H2b + H3b + H4b + H5b + H6b), 7.59 (d,
1H, J = 15.9 Hz, H3). 13C NMR (75 MHz, CDCl3) d 14.4 (C2¢¢),
60.4 (C1¢¢), 71.2 (C7b), 111.0 (C2¢), 115.0 (C5¢), 115.7 (C2), 123.2
(C6¢), 127.0 (C1¢), 127.8 (C2b/6b), 128.6 (C4b), 128.8 (C3b/5b),
135.8 (C1b), 144.6 (C3), 146.0 (C3¢), 148.1 (C4¢), 167.3 (C1).
2-Propenoic acid, 3-[3-benzyloxy-4-[(2,3,4-tri-O-acetyl-6-methyl-
b-D-glucopyranurosyl)oxy]phenyl]-, ethyl ester (39)
(E)-Ethyl 3-(3-(benzyloxy)-4-
(pentyloxysulfonyloxy)phenyl)acrylate (41)
Using procedure B, 400 mg (1.34 mmol) of 3-benzyloxy-4-
hydroxycinnamic acid ethyl ester 38 were reacted with 1.5 g
(2.96 mmol) of 2,3,4-triacetyl-D-methyl glucuronopyranosyl-
(N-phenyl)-2,2,2-trifluoroacetimidate 25,48 and with 43 mL
(0.43 mmol) of BF3 etherate. This yielded, after purification by
MPLC on silica and RP-18, 617 mg (0.982 mmol; 73%) of novel
compound 39. TLC Rf = 0.40 (Silica 60 F254, Hexane–EtOAc
Following procedure E, 300 mg (1.0 mmol) of compound 38 were
dissolved in 10 mL of THF. 220 ml (1.50 mmol) of TEA, 122 mg of
DMAP (1.00 mmol), and 300 mg of neopentylchlorosulfate28,30
(1.5 mmol) were added successively under stirring at rt and
nitrogen atmosphere. After removal of the solvent under reduced
pressure, the oily colored residue was purified by MPLC (Biotage
25 M) using a gradient of EtOAc in hexane as solvent. Yield:
401 mg (0.86 mmol; 86%) of novel compound 41. TLC Rf =
1
1/1). H NMR (360 MHz, CDCl3) d 1.33 (t, 3H, J = 7.1 Hz,
1
H2¢¢), 1.80 (s, 3H, CH3–CO), 2.02 (s, 3H, CH3–CO), 2.04 (s, 3H,
CH3–CO), 3.73 (s, 3H, H7a), 4.11 (m, 1H, H5a), 4.26 (q, 2H, J =
7.1 Hz, H1¢¢), 5.09–5.15 (m, 3H, H7b + H1a), 5.31–5.36 (m, 3H,
H2a + H3a + H4a), 6.30 (d, 1H, J = 15.9 Hz, H2), 7.10–7.45 (m,
8H, H2b + H3b + H4b + H5b + H6b + H2¢ + H5¢ + H6¢), 7.59 (d,
1H, J = 16.0 Hz, H3). 13C NMR (90 MHz, CDCl3) d 14.3 (C2¢¢),
20.3 (CH3–CO), 20.5 (CH3–CO), 20.6 (CH3–CO), 53.0 (C7a), 60.5
(C1¢), 69.1 (C4a), 70.9 (C7b), 71.1 (C2a), 71.8 (C3a), 72.6 (C5a),
99.9 (C1a), 113.7 (C2¢), 117.7 (C2), 120.1 (C5¢), 122.1 (C6¢), 127.1
(C2b/6b), 128.0 (C1¢), 128.2 (C4b), 128.7 (C3b/5b), 136.2 (C1b),
143.8 (C3), 146.1 (C3¢), 151.4 (C4¢), 166.8 (C6a), 167.1 (C1), 169.3
(2 ¥ CH3–CO), 170.1 (CH3–CO).
0.61 (Silica 60 F254, Hexane–EtOAc 2/1). H NMR (360 MHz,
CDCl3) d 0.85 (s, 9H, H1c), 1.34 (t, 3H, J = 7.13 Hz, H2¢¢), 4.03 (s,
2H, H3c), 4.27 (q, 2H, J = 7.13 Hz, H1¢¢), 5.14 (s, 2H, H7b), 6.38
(d, 1H, J = 16.0 Hz, H2), 7.14 (dd, 1H, J = 8.4 and 1.9 Hz, H6¢),
7.19 (d, 1H, J = 1.9 Hz, H2¢), 7.30–7.50 (m, 6H, H2b + H3b +
H4b + H5b + H6b + H5¢), 7.61 (d, 1H, J = 16.0 Hz, H2),. 13C
NMR (90 MHz, CDCl3) d 14.6 (C2¢¢), 26.4 (C1c), 32.0 (C2c), 61.0
(C1¢¢), 71.4 (C7b), 84.0 (C3c), 113.6 (C2¢), 119.9 (C2), 121.5 (C6¢),
123.8 (C5¢), 128.1 (C2b/6b), 128.8 (C4b), 129.1 (C3b/5b), 134.9
(C1¢), 135.9 (C1b), 140.6 (C4¢), 143.4 (C3), 151.1 (C3¢), 166.9 (C1).
Ethyl 3-(3-hydroxy-4-(pentyloxysulfonyloxy)phenyl)propanoate
(42)
(E)-3¢-O-Benzylcaffeic acid 4¢-O-b-D-glucuronide (40)
Following procedure D, 400 mg (0.87 mmol) of compound 41 were
hydrogenated for 4 h. This yielded after filtration, evaporation and
purification by MPLC (Biotage 25 M) using a gradient of EtOAc
in hexane as solvent, 277 mg (0.77 mmol; 90%) of novel compound
42. TLC Rf = 0.65 (Silica 60 F254, Hexane–EtOAc 2/1). 1H NMR
(360 MHz, CDCl3) d 0.99 (s, 9 H, H1c), 1.24 (t, 3H, J = 7.13 Hz,
H2¢¢), 2.59 (t, 2H, J = 8.12 Hz, H2), 2.88 (t, 2H, J = 7.54 Hz, H3),
4.13 (q, 2H, J = 7.17 Hz, H1¢¢), 4.16 (s, 2H, H3c), 6.71 (dd, 1H,
Following procedure C, 617 mg (0.98 mmol) of 2-
propenoic acid, 3-[3-benzyloxy-4-[(2,3,4-tri-O-acetyl-6-methyl-b-
D-glucopyranurosyl)oxy]phenyl]-, ethyl ester 39 were treated with
10 mL of 1 N aq. NaOH. After acidification (17 g of wet
R
Amberliteꢀ IR-120) and purification by MPLC on RP-18, 352 mg
of the novel compound (E)-3¢-O-benzylcaffeic acid 4¢-O-b-D-
glucuronide 40 were isolated (yield: 352 mg, 0.7643 mmol, 78%).
5208 | Org. Biomol. Chem., 2010, 8, 5199–5211
This journal is
The Royal Society of Chemistry 2010
©

View Online
J = 8.36 and 2.16 Hz, H6¢), 6.90(d, 1H, J = 2.12 Hz, H2¢), 7.20 (d,
1H, J = 8.32 Hz, H5¢). 13C NMR (90 MHz, CDCl3) d 14.3 (C2¢¢),
26.0 (C1c), 30.5 (C3), 31.9 (C2c), 35.6 (C2), 60.5 (C1¢¢), 83.7 (C3c),
118.0 (C2¢), 120.0 (C6¢), 122.4 (C5¢), 136.4 (C4¢), 141.2 (C1¢), 148.4
(C3¢), 172.7 (C1).
(E)-Ethyl 3-(4-hydroxy-3-
(neopentyloxysulfonyloxy)phenyl)acrylate (46)
Procedure A was carried out using 390 mg (1.35 mmol) of 5-
formyl-2-hydroxyphenyl neopentyl sulfate 45 in 10 mL of dry
CH2Cl2, and 0.5 g (1.43 mmol) of ethyl (triphenylphosphoranyli-
den)acetate. This afforded, after purification, 400 mg (1.12 mmol;
83%) of novel compound 46. 1H NMR (360 MHz, CDCl3) d 1.03
(s, 9H, H1c), 1.35 (t, 3H, J = 7.13 Hz, H2¢¢), 4.17 (s, 2H, H3c),
4.28 (q, 2H, J = 7.13 Hz, H1¢¢), 6.34 (d, 1H, J = 16.0 Hz, H2), 7.09
(d, 1H, J = 8.5 Hz, H5¢), 7.42 (dd, 1H, J = 8.5 and 1.8 Hz, H6¢),
Dihydrocaffeic acid 4¢-sulfate (24)
Following procedure F, 277 mg (0.77 mmol) of compound 42 in
2 mL of dry DMF were reacted with 71 mg (1.09 mmol) of sodium
azide and hydrolyzed in methanol and NaOH 1.0 M. This yielded,
after purification by CC on Biotage RP-18, 165 mg (Quantification
of the sulfate ester content following procedure G: 1.55 mmol/mg;
0.26 mmol; 34%) of novel compound 24. HPLC Rt = 5.70 (method
A). 1H NMR (360 MHz, DMSO-d6) d 2.14 (m, 2H, H2), 2.59 (m,
2H, H3), 6.23 (dd, 1H, J = 8.0 and 2.0 Hz, H6¢), 6.50 (d, 1H, J =
2.0 Hz, H2¢), 6.83 (d, 1H, J = 8.0 Hz, H5¢). 13C NMR (90 MHz,
DMSO-d6) d 32.3 (C3), 40.1 (C2), 114.4 (C6¢), 118.3 (C2¢), 122.4
(C5¢), 139.4 (C4¢), 139.9 (C1¢), 154.2 (C3¢), 177.2 (C1). LC-HRMS
[M - H]- calc for C9H9O7S: 261.0069; found: 261.0108.
7.49 (d, 1H, J = 1.7 Hz, H2¢), 7.60 (d, 1H, J = 16.0 Hz, H3). 13
NMR (90 MHz, CDCl3) d 14.3 (C2¢¢), 25.8 (C1c), 32.0 (C2c), 60.6
(C1¢¢), 84.8 (C3c), 117.7 (*C5¢), 118.7 (*C2¢), 121.9 (*C2), 128.1
(C1¢), 128.6 (C6¢), 137.8 (C3¢), 142.7 (C3), 149.4 (C4¢), 166.9 (C1).
C
(E)- and (Z)-Caffeic acid 3¢-sulfates (11a) and (11b)
Following procedure F, 400 mg (1.12 mmol) of compound 46 in
4 mL of dry DMF were reacted with 150 mg (2.2 mmol) of sodium
azide and hydrolyzed in methanol and NaOH 1.0 M. This yielded,
after purification by CC on Biotage RP-18, 210 mg (Quantification
of the sulfate ester content following procedure G: 2.30 mmol/mg;
0.48 mmol; 43%) of compound 11 as a 9/1 mixture of (E)-11a and
(Z)-11b forms. HPLC Rt = 5.50 (method A). 1H NMR (360 MHz,
DMSO-d6) d 5.75 (d, 0.1H, J = 12.8 Hz, H2 Z form), 5.96 (d,
0.1H, J = 12.8 Hz, H3 Z form), 6.16 (d, 0.9H, J = 15.8 Hz, H2
E-form), 6.71 (d, 0.1 H, J = 8.4 Hz, H5¢ Z form), 6.78 (d, 0.9H,
J = 8.3 Hz, H5¢ E-form), 7.06 (d, 0.9H, J = 15.8 Hz, H3 E-form),
7.08 (dd, 0.9H, J = 8.3 and 1.9 Hz, H6¢ E-form), 7.32 (d, 0.9H,
J = 1.9 Hz, H2¢ E-form), 7.36 (dd, 0.1H, J = 8.3 and 1.9 Hz,
H6¢ Z form), 7.49 (d, 0.1H, J = 1.9 Hz, H2¢ Z form). 13C NMR
(90 MHz, DMSO-d6) d 117.2 (C5¢), 120.9 (C2¢), 124.2 (C6¢), 125.3
(C2), 126.8 (C1¢), 136.6 (C3), 140.8 (C3¢), 150.7 (C4¢), 171.7 (C1).
LC-HRMS [M - H]- calc for C9H7O7S: 258.9912; found: 258.9960.
NMR data of compound 11 in D2O has been previously reported,
but only as a inseparable mixture with its 4¢-O-sulfate isomer.52
Thus compound 11 is now fully characterized for the first time.
2-(Benzyloxy)-5-formylphenyl neopentyl sulfate (44)
Following procedure E, 1.14 g (5 mmol) of 4-(benzyloxy)-3-
hydroxybenzaldehyde 43,31,32 were dissolved in 25 mL of CH2Cl2.
1.14 g (7.5 mmol; 1.12 mL) of DBU and 1.52 g (7.5 mmol) of
neopentylchlorosulfate28,30 were added successively under stirring
at RT and nitrogen atmosphere. After removal of the solvent under
reduced pressure, the oily colored residue was purified by MPLC
(Biotage 40 M) using a gradient of EtOAc in hexane as solvent.
Yield: 1.71 g (4.5 mmol; 91%) of novel compound 44. TLC Rf =
1
0.36 (Silica 60 F254, Hexane–EtOAc 2/1). H NMR (360 MHz,
CDCl3) d 0.89 (s, 9H, H1c), 4.06 (s, 2H, H3c), 5.23 (s, 2H, H7b),
7.18 (d, 1H, J = 8.55 Hz, H5¢), 7.38–7.48 (m, 5H, H2b + H3b +
H4b + H5b + H6b), 7.82 (dd, 1H, J = 8.55 and 2.0 Hz, H6¢), 7.91
(d, 1H, J = 2.19 Hz, H2¢). 13C NMR (90 MHz, CDCl3) d 25.8
(C1c), 31.8 (C2c), 71.3 (C7b), 83.9 (C3c), 114.0 (C5¢), 123.9 (C2¢),
127.7 (C2b/6b), 128.7 (C4b), 128.8 (C3b/5b), 130.0 (C1¢), 130.6
(C6¢), 135.0 (C1b), 139.5 (C3¢), 155.6 (C4¢), 189.5 (C1).
(E)-Ethyl 3-(4-(benzyloxy)-3-
(pentyloxysulfonyloxy)phenyl)acrylate (47)
5-Formyl-2-hydroxyphenyl neopentyl sulfate (45)
Procedure A was carried out using 590 mg (1.56 mmol) of 2-
(benzyloxy)-5-formylphenyl neopentyl sulfate 44 in 10 mL of dry
CH2Cl2, and 0.6 g (1.72 mmol) of ethyl (triphenylphosphoranyli-
den)acetate. This afforded, after purification, 600 mg (1.34 mmol;
86%) of novel compound 47. 1H NMR (360 MHz, CDCl3) d 0.87
(s, 9H, H1c), 1.33 (t, 3H, J = 7.12 Hz, H2¢¢), 4.04 (s, 2H, H3c),
4.26 (q, 2H, J = 7.13 Hz, H1¢¢), 5.15 (s, 2H, H7b), 6.33 (d, 1H, J =
15.99 Hz, H2), 7.05 (d, 1H, J = 8.57 Hz, H5¢), 7.35–7.70 (m, 8H,
H2b + H3b + H4b + H5b + H6b + H2¢+ H6¢ + H3).
To avoid the reduction of the aldehyde group, procedure D was
modified as follows. 1.08 g (2.85 mmol) of 44 was added to
a suspension of 1.5 g (1.42 mmol) of Pd–C (10%, standard,
unreduced, wet) and 4.6 mL of cyclohexene in 100 mL of
DMF. The suspension was heated 3 h at 70 ◦C. After filtration
on GF 92 and evaporation of the DMF, the resulting residue was
dissolved in EtOAc and washed with water. The organic layer was
dried over sodium sulfate, filtered and concentrated under reduced
pressure. Purification by MPLC (Biotage 40 M) using a gradient
of EtOAc in hexane as solvent yielded 770 mg (2.67 mmol, 94%)
of novel compound 45. TLC Rf = 0.61 (Silica 60 F254, Hexane–
Ethyl 3-(4-hydroxy-3-(pentyloxysulfonyloxy)phenyl)propanoate
(48)
1
EtOAc 1/1). H NMR (360 MHz, CDCl3) d 1.02 (s, 9H, H1c),
4.19 (s, 2H, H3c), 7.20 (d, J = 8.43 Hz, 1H, H5¢), 7.77 (dd, 1H, J =
8.41 and 1.78 Hz, H6¢), 7.87 (d, 1H, J = 1.7 Hz, H2¢), 9.86 (s, 1H,
H1). 13C NMR (90 MHz, CDCl3) d 25.8 (C1c), 32.0 (C2c), 84.7
(C3c), 118.6 (*C5¢), 123.7 (*C2¢), 130.0 (C1¢), 130.9 (C6¢), 138.1
(C3¢), 153.7 (C4¢), 189.8 (C1).
Following procedure D, 300 mg (0.67 mmol) of compound 47 were
hydrogenated for 4 h. This yielded, after filtration, evaporation and
drying under high vacuum, 240 mg (0.67 mmol; 100%) of novel
compound 48. TLC Rf = 0.49 (Silica 60 F254, Hexane–EtOAc
1
1/1). H NMR (360 MHz, CDCl3) d 1.01 (s, 9H, H1c), 1.23 (t,
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 5199–5211 | 5209
©

View Online
3H, J = 7.13 Hz, H2¢¢), 2.58 (t, 2H, J = 7.67 Hz, H2), 2.89 (t, 2H,
J = 7.89 Hz, H3), 4.12 (q, 2H, J = 7.13 Hz, H1¢¢), 4.13 (s, 2H, H3c),
6.97 (d, 1H, J = 8.36 Hz, H5¢), 7.06 (dd, 1H, J = 8.32 and 2.12 Hz,
H6¢), 7.13 (d, 1H, J = 2.12 Hz, H2¢). 13C NMR (90 MHz, CDCl3)
d 14.2 (C2¢¢), 25.8 (C1c), 29.9 (*C3), 32.0 (C2c), 35.8 (*C2), 60.6
(C1¢¢), 84.5 (C3c), 118.4 (*C5¢), 122.1 (*C2¢), 128.6 (C6¢), 133.9
(C1¢), 137.4 (*C3¢), 145.9 (*C4¢), 172.6 (C1).
neopentylchlorosulfate28,30 were added successively under stirring
at RT and nitrogen atmosphere. After removal of the solvent under
reduced pressure, the oily colored residue was purified by MPLC
(Biotage 25 M) using a gradient of EtOAc in hexane as solvent.
Yield: 168 mg (0.42 mmol; 80%) of novel compound 52. TLC Rf =
1
0.55 (Silica 60 F254, Hexane–EtOAc 2/1). H NMR (360 MHz,
CDCl3) d 1.01 (s, 9H, H1c), 1.34 (t, 3H, J = 7.13 Hz, H2¢¢), 2.37
(s, 3H, COCH3), 4.11 (s, 2H, H3c), 4.27 (q, 2H, J = 7.13 Hz, H1¢¢),
6.40 (d, 1H, J = 16.0 Hz, H2), 7.38 (d, 1H, J = 2.1 Hz, H2¢), 7.42
(dd, 1H, J = 8.50 and 2.1 Hz, H6¢), 7.47 (d, 1H, J = 8.5 Hz, H5¢).
13C NMR (90 MHz, CDCl3) d 14.60 (C2¢¢), 21.0 (COCH3), 26.2
(C1c), 32.3 (C2c), 61.1 (C1¢¢), 84.4 (C3c), 120.7 (C2), 122.8 (C5¢),
123.7 (C2¢), 126.9 (C6¢), 134.8 (C1¢), 142.3 (C3), 142.7 (*C3¢), 143.0
(*C4¢), 166.7 (C1), 168.4 (COCH3).
Dihydrocaffeic acid 3¢-sulfate (23)
Following procedure F, 240 mg (0.67 mmol) of compound 48 in
2 mL of dry DMF were reacted with 100 mg (1.5 mmol) of sodium
azide and hydrolyzed in methanol and NaOH 1.0 M. This yielded,
after purification by CC on Biotage RP-18, 270 mg (Quantification
of the sulfate ester content following procedure G: 1.75 mmol/mg;
0.47 mmol; 70%) of novel compound 23. HPLC Rt = 4.82 (method
A). 1H NMR (360 MHz, DMSO-d6) d 2.20 (m, 2H, H2), 2.63 (m,
2H, H3), 6.67 (d, J = 8.18 Hz, 1H, H5¢), 6.77 (dd, 1H, J = 8.18 and
2.12 Hz, H6¢), 6.99 (d, 1H, J = 2.12 Hz, H2¢). 13C NMR (90 MHz,
DMSO-d6) d 32.2 (C3), 40.8 (C2), 118.1 (C5¢), 122.8 (C2¢), 125.6
(C6¢), 132.3 (C1¢), 141.0 (C3¢), 149.9 (C4¢), 179.1 (C1). LC-HRMS
[M - H]- calc for C9H9O7S: 261.0069; found: 261.0115.
(E)- and (Z)-Caffeic acid 4¢-sulfate (12a) and (12b)
Following procedure F, 168 mg (0.42 mmol) of compound 52 in
1 mL of dry DMF were reacted with 41 mg (0.65 mmol) of sodium
azide and hydrolyzed in methanol and NaOH 1.0 M. This yielded,
after purification by CC on Biotage RP-18, 148 mg (Quantification
of the sulfate ester content following procedure G: 1.91 mmol/mg;
0.28 mmol; 67%) of compound 12 as a 9/1 mixture of (E)-12a and
(Z)-12b forms. HPLC Rt = 5.29 (method A). 1H NMR (360 MHz,
DMSO-d6) 5.81 (d, 0.1H, J = 12.9 Hz, H2 Z-form), 5.91 (d, 0.1H,
J = 12.9 Hz, H3 Z-form), 6.24 (d, 0.9H, J = 15.8 Hz, H2 E-form),
6.87 (dd, 0.9H, J = 8.4 and 2.1 Hz, H6¢ E-form), 6.95 (d, 0.9H, J =
2.1 Hz, H2¢ E-form), 7.06 (d, 0.9H, J = 15.8 Hz, H3 E-form), 7.12
(d, 0.9H, J = 8.3 Hz, H5¢ E-form). 13C NMR (90 MHz, DMSO-d6)
115.1 (C2¢), 118.3 (C6¢), 122.7 (C5¢), 127.8 (C2), 133.0 (C1¢), 136.1
(C3), 141.1 (C4¢), 149.1 (C3¢), 170.9 (C1). 1H and 13C NMR data
of the (E)-isomer 12a were in accordance with published values
in DMSO-d6.54 LC-HRMS [M - H]- calc for C9H7O7S: 258.9912;
found: 258.9955.
3¢,4¢-Diacetoxycinnamic acid ethyl ester (49)
Procedure A was carried out using 2.7154 g (12.22 mmol) of 3¢,4¢-
diacetoxybenzaldehyde 27 in 20 mL of dry CH2Cl2, and 4.33 g
(12.42 mmol) of ethyl (triphenylphosphoranyliden)acetate. This
afforded, after purification, 3.9 g of compound 49.53 1H NMR
(360 MHz, CDCl3) d 1.33 (t, 3H, J = 7.1 Hz, H2¢¢), 2.30 (s, 3H,
OAc), 2.31 (s, 3H, OAc), 4.26 (q, 2H, J = 7.1 Hz, H1¢¢), 6.38 (d,
1H, J = 16.0 Hz, H2), 7.22 (d, 1H, J = 8.4 Hz, H5¢), 7.36 (d, 1H,
J = 2.0 Hz, H2¢), 7.40 (dd, 1H, J = 8.4 and 2.0 Hz, H6¢), 7.62 (d,
1H, J = 16.0 Hz, H3). 13C NMR (90 MHz, CDCl3) d 14.3 (C2¢¢),
20.6 (OAc), 20.7 (OAc), 60.7 (C1¢¢), 119.5 (C2), 122.7 (C2¢), 123.9
(C5¢), 126.4 (C6¢), 133.4 (C1¢), 142.4 (C3¢), 142.6 (C3), 143.4 (C4¢),
166.6 (C1), 168.0 (OAc), 168.1 (OAc).
Human intervention study design
Trial Design. The study and protocol were reviewed by Glasgow
Royal Infirmary NHS Research Ethics Committee. Eleven healthy
human volunteers followed a low polyphenol diet for 48 h prior
the beginning of the study. On the night preceding the trial, the
subjects stopped consuming any food and came to the trial unit
at 8.30 am the next day in a fasted state and consumed 200 mL of
an instant coffee containing 412 mmol of chlorogenic acids, and
remained on a low polyphenol diet for 24 h. Plasma and urine were
collected over the ensuing 24 h period and aliquots were stored at
-80 ◦C prior to analysis.
4¢-Acetoxy-3¢-hydroxycinnamic acid ethyl ester (50) and
3¢-acetoxy-4¢-hydroxycinnamic acid ethyl ester (51)
1.5289 g (5.23 mmol) of 3¢,4¢-diacetoxycinnamic acid ethyl ester
49 were dissolved in 5 mL of 1-methyl-2-pyrrolidone (NMP).
The solution was cooled to 0 ◦C in an ice bath and 135 mg
(1.98 mmol) of imidazole, followed by 0.7 mL (6.85 mmol) of
thiophenol were added. The reaction was allowed to take place
at rt under stirring for 4 h. The reaction medium was directly
purified by MPLC on silica using a gradient of EtOAc in hexane
as solvent, affording 707.5 mg (2.83 mmol, 54%) of a mixture
of 4¢-acetoxy-3¢-hydroxycinnamic acids ethyl ester 50, and of 3¢-
acetoxy-4¢-hydroxycinnamic acids ethyl ester 51 (in a ratio of about
1/1, according to NMR analysis). TLC Rf = 0.53 (Silica 60 F254,
Hexane–EtOAc 3/2).
Plasma and urine analyses. Plasma samples were extracted using
a method developed by Day et al.,55 and along with urine samples
were analysed by HPLC-PDA-MSn using a Surveyor HPLC with
a PDA detector and a LCQ Duo ion trap mass spectrometer
fitted with an electrospray interface (Thermo Finnigan, San Jose,
◦
USA). Separations were performed at 40 C using a SYNERGI
4 mm POLAR-RP 250 ¥ 4.6 mm i.d. reverse phase column
(Phenomenex, Macclesfield, UK). Injections were carried out with
an autosampler maintained at 4 ◦C. The mobile phase comprising
a 40 min, 5–16% gradient of acetonitrile in 0.5% aqueous acetic
acid was pumped at a flow rate of 1 mL/min. The column eluate
initially passed through the PDA detector and was then split,
(E)-Ethyl 3-(3-(acetoxy)-4-(pentyloxysulfonyloxy)phenyl)acrylate
(52)
Following procedure E, 125 mg (0.5 mmol) of the mixture of 50
and 51 were dissolved in 5 mL of THF. 150 mL (1.00 mmol) of
TEA, 122 mg (1.00 mmol) of DMAP and 203 mg (1.0 mmol) of
5210 | Org. Biomol. Chem., 2010, 8, 5199–5211
This journal is
The Royal Society of Chemistry 2010
©

View Online
with 0.3 mL/min-1 directed to the mass spectrometer operating
in full scan negative ionisation mode (100–1000 m/z). Analyses
were carried out using full scan, data dependent MS2 scanning
from m/z 100 to 1000. The tuning of the mass spectrometer
was optimised by infusing a standard of ferulic acid, dissolved
in the initial HPLC mobile phase, into the sou◦rce at a flow rate of
0.3 mL/min. Capillary temperature was 300 C, sheath gas and
auxiliary gas were 80 and 60 units/min-1 respectively, and source
voltage was 3.0 kV, and collision energy set at 35%.
25 D. Barron, Recent Advances in the Chemical Synthesis and Biological
Activity of Phenolic Metabolites, in Recent Advances in Polyphenol
Research., ed. F. Daayf and V. Lattanzio, Wiley-Blackwell, Chichester,
UK, 2008, pp. 317–358.
26 J. B. Daskiewicz, F. Depeint, L. Viornery, C. Bayet, G. Comte-Sarrazin,
G. Comte, J. M. Gee, I. T. Johnson, K. Ndjoko, K. Hostettmann and
D. Barron, J. Med. Chem., 2005, 48, 2790–2804.
27 B. L. Allwood, F. H. Kohnke, A. M. Z. Slawin, J. F. Stoddart and D. J.
Williams, J. Chem. Soc., Chem. Commun., 1985, 311–314.
28 L. S. Simpson and T. S. Widlanski, J. Am. Chem. Soc., 2006, 128,
1605–1610.
29 A. Fu¨rstner, M. M. Domostoj and B. Scheiper, J. Am. Chem. Soc.,
2006, 128, 8087–8094.
30 E. Buncel and J. P. Millington, Can. J. Chem., 1965, 43, 556–564.
31 G. L. Plourde and R. R. Spaetzel, Molecules, 2002, 7, 697–705.
32 P. Li, C. D. Evans, Y. Wu, B. Cao, E. Hamel and M. M. Joullie, J. Am.
Chem. Soc., 2008, 130, 2351–2364.
33 D. Barron, L. Varin, R. K. Ibrahim, J. B. Harborne and C. A. Williams,
Phytochemistry, 1988, 27, 2375–2395.
Notes and references
1 P. Viola, J. Appl. Cosmetol., 2005, 23, 129–137.
2 L. Taylor and J. Antonio, Coffee As a Functional Beverage, in Hand-
book of Nutraceuticals and Functional Foods, ed. R. E. C. Wildman,
CRC Press, Boca Raton, Florida, 2007, pp. 453–465.
3 J. S. Bonita, M. Mandarano, D. Shuta and J. Vinson, Pharmacol. Res.,
2007, 55, 187–198.
4 M. N. Clifford, J. Sci. Food Agric., 1999, 79, 362–372.
5 A. Farah and C. M. Donangelo, Brazilian Journal of Plant Physiology,
2006, 18, 23–36.
6 I. Lopez-Galilea, M. P. Pena and C. Cid, Eur. Food Res. Technol., 2008,
227, 141–149.
7 K. Fujioka and T. Shibamoto, Food Chem., 2008, 106, 217–221.
8 K. Iwai, N. Kishimoto, Y. Kakino, K. Mochida and T. Fujita, J. Agric.
Food Chem., 2004, 52, 4893–4898.
9 M. N. Clifford, S. Knight, B. Surucu and N. Kuhnert, J. Agric. Food
Chem., 2006, 54, 1957–1969.
10 M. N. Clifford, S. Marks, S. Knight and N. Kuhnert, J. Agric. Food
Chem., 2006, 54, 4095–4101.
11 G. Williamson, D. Barron, K. Shimoi and J. Terao, Free Radical Res.,
2005, 39, 457–469.
12 M. F. Andreasen, P. A. Kroon, G. Williamson and M. T. Garcia-
Conesa, J. Agric. Food Chem., 2001, 49, 5679–5684.
13 M. F. Andreasen, P. A. Kroon, G. Williamson and M. T. Garcia-
Conesa, Free Radical Biol. Med., 2001, 31, 304–314.
14 G. W. Plumb, M. T. Garcia-Conesa, P. A. Kroon, M. Rhodes, S. Ridley
and G. Williamson, J. Sci. Food Agric., 1999, 79, 390–392.
15 D. Couteau, A. L. McCartney, G. R. Gibson, G. Williamson and C. B.
Faulds, J. Appl. Microbiol., 2001, 90, 873–881.
16 A. N. Booth and R. T. Williams, Biochem. J., 1963, 88, 66P–67P.
17 A. R. Rechner, M. A. Smith, G. Kuhnle, G. R. Gibson, E. S. Debnam,
S. K. Srai, K. P. Moore and C. A. Rice-Evans, Free Radical Biol. Med.,
2004, 36, 212–225.
18 M. Nardini, E. Cirillo, F. Natella and C. Scaccini, J. Agric. Food Chem.,
2002, 50, 5735–5741.
19 A. R. Rechner, J. P. E. Spencer, G. Kuhnle, U. Hahn and C. A. Rice-
Evans, Free Radical Biol. Med., 2001, 30, 1213–1222.
20 M. Monteiro, A. Farah, D. Perrone, L. C. Trugo and C. Donangelo,
J. Nutr., 2007, 137, 2196–2201.
34 S. Fujiwara, S. Sakurai, I. Sugimoto and N. Awata, Chem. Pharm. Bull.,
1983, 31, 645–652.
35 Z. Zhao, Y. Egashira and H. Sanada, Journal of Nutrition, 2003, 133,
1355–1361.
36 L. Rondini, M. N. Peyrat-Maillard, A. Marsset-Baglieri and C. Berset,
J. Agric. Food Chem., 2002, 50, 3037–3041.
37 F. Virgili, G. Pagana, L. Bourne, G. Rimbach, F. Natella, C. Rice-Evans
and L. Packer, Free Radical Biol. Med., 2000, 28, 1249–1256.
38 D. Zalko, R. Costagliola, C. Dorio, E. Rathahao and J. P. Cravedi,
Drug Metab. Dispos., 2003, 31, 168–178.
39 Z. c. Zhang, M. Xu, S. f. Sun, X. Qiao, B. r. Wang, J. Han and D. a.
Guo, J. Chromatogr. B., 2008, 871, 7–14.
40 K. Tamura, H. Mizukami, K. Maeda, H. Watanabe and K. Uneyama,
J. Org. Chem., 1993, 58, 32–35.
41 G. N. Bollenback, J. W. Long, D. G. Benjamin and J. A. Lindquist,
J. Am. Chem. Soc., 1955, 77, 3310–3315.
42 A. Nudelman, J. Herzig, H. E. Gottlieb, E. Keinan and J. Sterling,
Carbohydr. Res., 1987, 162, 145–152.
43 L. Corda, A. M. Fadda, A. Maccioni, A. M. Maccioni and G. Podda,
J. Heterocycl. Chem., 1988, 25, 311–314.
44 P. Manini, A. Pezzella, L. Panzella, A. Napolitano and M. d¢Ischia,
Tetrahedron, 2005, 61, 4075–4080.
45 R. Mo¨rck and K. P. Kringstad, Holzforschung, 1985, 39, 109–
119.
46 J. G. Buchanan, D. G. Hill, R. H. Wightman, I. K. Boddy and B. D.
Hewitt, Tetrahedron, 1995, 51, 6033–6050.
47 E. A. Couladouros, I. C. Soufli, V. I. Moutsos and R. K. Chadha,
Chem.–Eur. J., 1998, 4, 33–43.
48 P. Wang, Z. Zhang and B. Yu, J. Org. Chem., 2005, 70, 8884–
8889.
49 D. Planchenault, R. Dhal and J. P. Robin, Tetrahedron, 1995, 51, 1395–
1404.
50 F. Santangelo and C. Casagrande, Synth. Commun., 1996, 26, 2863–
2873.
51 F. Wang, L. Yang, K. Huang, X. Li, X. Hao, J. Stockigt and Y. Zhao,
Nat. Prod. Res. Part A, 2007, 21, 196–202.
52 J. S. Todd, R. C. Zimmerman, P. Crews and R. S. Alberte, Phytochem-
istry, 1993, 34, 401–404.
53 K. Herrmann, Pharmazie, 1953, 8, 303.
54 T. Nakazawa and K. Ohsawa, J. Nat. Prod., 1998, 61, 993–996.
55 A. J. Day, F. Mellon, D. Barron, G. Sarrazin, M. R. A. Morgan and G.
Williamson, Free Radical Res., 2001, 35, 941–952.
21 S. M. Kern, R. N. Bennett, P. W. Needs, F. A. Mellon, P. A. Kroon and
M. T. Garcia-Conesa, J. Agric. Food Chem., 2003, 51, 7884–7891.
22 R. Mateos, L. Goya and L. Bravo, J. Agric. Food Chem., 2006, 54,
8724–8732.
23 J. L. Zhang, G. D. Zhang and T. H. Zhou, J. Asian Nat. Prod. Res.,
2005, 7, 49–58.
24 A. J. Day, Y. Bao, M. R. A. Morgan and G. Williamson, Free Radical
Biol. Med., 2000, 29, 1234–1243.
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 5199–5211 | 5211
©