Oxidative conjugation of chlorogenic acid with glutathione: Structural characterization of addition products and a new nitrite-promoted pathway

Article abstract of DOI:10.1016/S0968-0896(03)00460-7

Chlorogenic acid (1), a cancer chemopreventive agent widely found in fruits, tea and coffee, undergoes efficient conjugation with glutathione (GSH), in the presence of horseradish peroxidase/H₂O₂ or tyrosinase at pH 7.4, to yield thr

Full text of DOI:10.1016/S0968-0896(03)00460-7

Bioorganic & Medicinal Chemistry 11 (2003) 4797–4805
Oxidative Conjugation of Chlorogenic Acid with Glutathione:
Structural Characterization of Addition Products and a New
Nitrite-Promoted Pathway
Lucia Panzella, Alessandra Napolitano and Marco d’Ischia*
Department of Organic Chemistry and Biochemistry, University of Naples ‘Federico II’, Via Cinthia 4, I-80126 Naples, Italy
Received 25 April 2003; accepted 10 July 2003
Abstract—Chlorogenic acid (1), a cancer chemopreventive agent widely found in fruits, tea and coffee, undergoes efficient con-
jugation with glutathione (GSH), in the presence of horseradish peroxidase/H2O2 or tyrosinase at pH 7.4, to yield three main
adducts that have been isolated and identified as 2-S-glutathionylchlorogenic acid (3), 2,5-di-S-glutathionylchlorogenic acid (4) and
2,5,6-tri-S-glutathionylchlorogenic acid (5) by extensive NMR analysis. The same pattern of products could be obtained by reaction
of 1 with GSH in the presence of nitrite ions in acetate buffer at pH 4. Mechanistic experiments suggested that oxidative conjuga-
tion reactions proceed by sequential nucleophilic attack of GSH on ortho-quinone intermediates. Overall, these results provide the
first complete spectral characterization of the adducts generated by biomimetic oxidation of 1 in the presence of GSH, and disclose
a new possible nitrite-mediated conjugation pathway of 1 with GSH at acidic pH of physiological relevance.
# 2003 Elsevier Ltd. All rights reserved.
Introduction
may serve as efficient nitrite scavengers and suppressors
of nitrosamine formation in the gastric com-
partment.2,13ꢀ15 Elucidation of the underlying chemistry
has underscored a marked reactivity of 1 toward nitrous
acid and reactive nitrogen species thereof, to give nitro-
chlorogenic acid 2 as the prevailing product.16,17
Epidemiological, biological and biochemical data
concur to support a beneficial role of chlorogenic acid
(5-caffeoyl-d-quinic acid, 1)1 and other dietary poly-
phenolic compounds in human health. Widely occuring
in various agricultural products at relatively elevated
levels, for example, 3.4–14 mg/100 g fresh weight in
potatoes, 12–31 mg/100 mL of apple juice, 89 mg/100 g
of dry tea shoots, and ca. 250 mg per cup of coffee,2,3
1
has been attributed anticarcinogenic,4ꢀ6 antimutagenic7
and antiinflammatory effects.8 It contributes sig-
nificantly to the total polyphenol intake and has been
suggested to play a role in the apparent association
between the regular consumption of polyphenol-rich
food and beverages and the prevention of inflammatory
and proliferative diseases. In particular, 1 has been
reported to exert inhibitory effects on carcinogenesis in
the large intestine, liver, and tongue,9ꢀ11 and a protec-
tive action on oxidative stress in vivo.12
Both antioxidant and antinitrosaminic properties of 1
stem from the oxidizable ortho-diphenolic functionality,
which can act as an H-atom donor toward reactive
oxygen species and other biological oxidants, as a
transition metal chelator preventing Fenton-type
processes,18 and as an efficient trap for electrophilic
nitrosating agents.16,17 It can also account for inhibition
of lipoxygenase activity,19 retinoic acid epoxidation20
and H2O2-induced erythrocyte haemolysis and lipid
peroxidation.21
Recent research has also emphasized the antinitrosating
properties of polyphenolic compounds such as 1, which
In spite of continuing advances, the oxidative metabo-
lism of 1 and its interactions with potential biological
*Corresponding author. Tel.: +39-081-674132; fax +39-081-674393;
e-mail: dischia@unina.it
0968-0896/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0968-0896(03)00460-7

4798
L. Panzella et al. / Bioorg. Med. Chem. 11 (2003) 4797–4805
targets in relation to cancer chemoprevention are still
not entirely elucidated. In a relevant study,22 it has been
reported that 1 reacts with cysteine in the presence of
apple polyphenol oxidase to give a main product iden-
tified as 2-S-cysteinylchlorogenic acid. More recently,23
it has been shown that 1, caffeic acid and dihydrocaffeic
acid form quinonoid and hydroxylated products when
oxidized by horseradish peroxidase (HRP)/H2O2 or
tyrosinase/O2. Mass spectrometric analyses of the
metabolites formed by HRP/H2O2 oxidation in the pre-
sence of glutathione (GSH), the most prevalent intra-
cellular thiol antioxidant and a potential anticarcinogen
deputed to mobilization of xenobiotic toxicants,
revealed formation of mono- and di-GSH conjugates
for all three compounds. However, the chemical struc-
tures and properties of these latter adducts were not
supported by detailed spectral analysis.
in 0.05 M phosphate buffer, pH 7.4, proceeded rapidly
to give a yellow solution displaying no chromato-
graphically isolable species under reverse phase HPLC
conditions. When the reaction was carried out in the
presence of excess GSH (200 mM), HPLC analysis of the
oxidation mixture revealed a well defined pattern of
products comprising a species eluting faster than 1 and
two more abundant components eluting at higher
retention times.
For preparative purposes the reaction was carried out
using the substrate at 2 mM concentration with 4 U/mL
of HRP and 4 mM H2O2 in the presence of 10 mM
GSH. Preparative HPLC allowed isolation of the more
retained products in pure form. One of these (32% iso-
lated yield) was formulated as 2-S-glutathionyl-
chlorogenic acid 324 based on the [M+H]+ peak at m/z
660 in the positive ion electrospray (ESI+)-MS spec-
trum and straightforward analysis of the 1H NMR
spectrum, showing two doublets in the aromatic region
(J=8.4 Hz) ascribed to the ortho-coupled H-5 and H-6
protons and the expected resonances for a GSH unit. The
E configuration at the double bond was maintained, as
apparent from the large coupling constant (J=16.0 Hz).
In agreement with previous data for 2-S-glutathio-
nylcaftaric acid,25 the vinyl a proton of 3 suffered a
marked downfield shift compared to 1. The integrity of
In this paper, we report the isolation and complete
spectral characterization of the adducts formed by
tyrosinase and HRP/H2O2 oxidation of 1 in the pre-
sence of GSH, including a hitherto unknown triadduct,
and describe for the first time the ability of nitrite ions
to promote the same oxidative conjugation processes
under mildly acidic conditions simulating those within
the gastric compartment.
1
the quinic acid unit was deduced by analysis of the H
and 13C NMR spectra, in which relevant resonances did
not differ appreciably from the chemical shift values of
1. Moreover, the chromophoric features of 3 were con-
sistent with those of catechol–thiol adducts.26
Results
HRP/H2O2 or tyrosinase-promoted oxidation of chloro-
genic acid (1) in the presence of GSH
Oxidation of 1 (50 mM) by HRP (1U/mL) and H2O2
The other product, showing a pseudomolecular peak at
m/z 965 in the ESI+-MS spectrum, was evidently a
(two molar equivalents) under the reported conditions23

L. Panzella et al. / Bioorg. Med. Chem. 11 (2003) 4797–4805
4799
diadduct (18% yield). The 1H NMR spectrum con-
sistently showed only a single proton resonance in the
aromatic region (d 7.38) appearing as a singlet, suggest-
ing either the 2,5- or the 2,6-disubstituted isomer. That
the former (4) was the correct structure was deduced
from extensive 2D homonuclear and 1H,13C hetero-
nuclear correlation experiments. As a most diagnostic
feature, the singlet at d 7.38 displayed a cross peak in
the 1H, 13C heteronuclear multiple bond correlation
spectrum with the vinyl a carbon resonance at d 144.2.
(plot B) molar equivalents of H2O2. Analysis of the
reaction mixtures was carried out 5 min after addition
of the oxidant. In most of the mixtures examined, mass
balance was higher than 80%. With 1molar equivalent
The third, more polar product was usually formed in
lower yields with respect to 3 and 4. However, after
several trials it was eventually obtained in satisfactory
amounts using higher concentrations of H2O2, for
example, 8 mM. Under such conditions, 1 was com-
pletely consumed and adducts 3 and 4 were formed in
negligible amounts. Chromatography of the reaction
mixture on a Sephadex G10 column, eluant water, yiel-
ded the product in practically pure form for spectral
analysis (90% yield). The [M+H]+ peak at m/z 1270
1
and the lack of signals for aromatic protons in the H
NMR spectrum, together with 13C NMR data, con-
curred to assign the product the structure of the novel
triadduct 5.
Selected NMR resonances of compounds 3, 4 and 5 and
assignments based on 2D correlation experiments are
listed in Table 1.
Figure 1. Substrate consumption and formation yields of adducts 3–5
as a function of molar equivalents of GSH in the HRP/H2O2 pro-
moted oxidation of 1. Reaction of 1 (50 mM) was run in 0.05 M phos-
phate buffer (pH 7.4), in the presence of HRP (1U/mL) and one (plot
A) or two (plot B) molar equivalents of H2O2, with GSH varying in
the range of 0–200 mM. The reaction mixture was analyzed by HPLC 5
min after addition of the oxidant. Percent concentration of 1 (black
bars), 3 (open bars), 4 (grey bars) and 5 (dashed bars) in the reaction
mixture. Shown are the mean values for three separate experiments
with SD ꢁ10%.
In another series of experiments the effect of various
parameters on the kinetics and product distribution in
the oxidation of 1 (50 mM) in the presence of GSH was
investigated.
Figure 1 reports substrate consumption and formation
yields of adducts 3–5 as a function of GSH equivalents
in the reaction of 1 with HRP and one (plot A) or two
Table 1. Selected NMR spectral data for adducts 3, 4 and 5a
3
4
5
1H (J, Hz)
13C
1H (J, Hz)
13C
1H (J, Hz)
13C
1b
2
—
6.27 (d, 16.0)
8.10 (d, 16.0)
—
—
170.0
117.8
145.2
130.9
121.2
147.4
147.7
118.2
121.2
36.7
—
6.42 (d, 16.0)
8.17 (d, 16.0)
—
—
—
—
—
7.38 (s)
3.06 (dd, 14.4, 8.8)
3.27 (dd, 14.4, 3.6)
3.23 (dd, 14.4, 8.4)
3.41 (dd, 14.4, 4.4)
—
4.20 (dd, 8.8, 3.6)
4.46 (dd, 8.4, 4.4)
—
169.6
118.6
144.2
130.8
121.2
147.5
147.5
122.9
124.1
37.1
—
6.38 (d, 16.0)
8.00 (d, 16.0)
169.1
126.4
144.5
138.1
120.9
148.3c
148.5c
127.1
129.9
37.4
3
10
—
—
—
—
—
—
20
30
—
—
40
50
6.88 (d, 8.4)
7.12 (d, 8.4)
3.05 (dd, 14.4, 8.8)
3.22 (dd, 14.4, 3.6)
—
60
CH2 b
3.00 (dd, 14.4, 9.6)
3.29 (dd, 14.4, 4.4)
3.28 (m)
CH2 b0
—
35.8
38.3
3.46 (dd, 13.6, 4.8)
3.09 (dd, 14.0, 9.2) 3.22 (m)
4.74 (m)
CH2 b00
CH a
—
4.22 (dd, 8.8, 3.6)
—
54.8
—
—
36.8
54.5
54.2
—
54.0d
54.4
CH a0
CH a00
—
—
4.24 (dd, 8.4, 4.8)
4.17 (dd, 9.2, 4.8)
—
54.1d
aSpectra run in D2O, chemical shift values given in ppm.
bNumbering as shown in structural formulas 3, 4 and 5.
cInterchangeable.
dInterchangeable.

4800
L. Panzella et al. / Bioorg. Med. Chem. 11 (2003) 4797–4805
of H2O2 (plot A), an increase in the number of molar
equivalents of GSH affected product distribution,
favouring the diadduct 4 at the expense of 3. Substrate
consumption, however, was not significantly affected by
the concentration of GSH in the reaction mixture. For-
mation of triadduct 5 became relevant only in the reac-
tions run with two molar equivalents of H2O2 (plot B):
in this case, consumption of 1 was complete in most of
the mixture analyzed and an increase in the number of
molar equivalents of GSH shifted the product distribu-
tion in favour of the more substituted adducts. Surpris-
ingly, with one molar equivalent of GSH the mass
balance was very poor. This could be explained con-
sidering that other reaction pathways prevailed when
GSH concentrations are too low for efficient coupling.
10% yield, in that order. With 1 at 2 mM concentration
and 5 molar equivalents of GSH the reaction was com-
plete after about 50 min to give 3 as the only detectable
adduct (76% yield). Consistently with previous
data,23,25 pure 3 or 4 were poor substrates of tyrosinase.
Nitrite-mediated reaction of chlorogenic acid (1) with
GSH under acidic conditions
In a recent paper,27 we demonstrated that nitrite ions
can mediate a remarkable decarboxylative conjugation
of caffeic acid with GSH under mildly acidic conditions
mimicking those within the stomach. As an extension of
that study, we found that GSH can likewise affect the
reaction of 1 with nitrite ions in acetate buffer at pH 4.
Comparative HPLC analysis of reaction mixtures in the
presence of even low GSH concentrations revealed sup-
pression of the 6-nitroderivative 2 and formation of a
main product which was identical with 3 by chromato-
graphic, mass spectrometric and NMR analysis (Table
2). Formation of 2 was observed only with high nitrite/
GSH molar ratios.
Addition of two molar equivalents of H2O2 in two por-
tions at 10-min intervals rather than as a bolus resulted
in comparable substrate consumption, but increased the
yields of the more substituted adducts 4 and 5 (data not
shown).
In separate experiments, 3 was allowed to react with
HRP/H2O2 in the presence of GSH. HPLC analysis
revealed quantitative conversion to 4. Moreover, reac-
tion of 4 with HRP/H2O2 in the presence of GSH was
found to yield 5 as the sole product.
To investigate in more detail the competition between
nitration and GSH adduct formation, the effect of
increasing nitrite concentration in the reaction of 2 mM
1 with 1mM GSH in acetate buffer, pH 4, was then
examined. The results in Figure 2 showed that forma-
tion of the nitroderivative 2 becomes significant only
when nitrite was ca. 10-fold with respect to GSH.
Using the glucose/glucose oxidase system as source of
H2O2, oxidation of 1 (50 mM) in the presence of 200 mM
GSH was complete in 1h leading to formation of 3, 4
and 5 in 8, 68 and 20% yields, in that order. As the
reaction proceeded further, for example at 2 h, triadduct
5 gradually accumulated up to 80% yield.
When the reaction was run at pH 4 with 1 at 2 mM
concentration in the presence of 4 and 5 molar equiva-
lents of nitrite and GSH, respectively, in addition to 3,
diadduct 4 and triadduct 5 were also formed in the
reaction mixture. Relative yields after 2 h were 33, 12
and 6% in that order. No other adduct was detectable
despite careful scrutiny of the reaction mixture. Analysis
of the reaction mixture at various times indicated that
all adducts were stable under the reaction conditions for
at least 16 h. An abundant species eluting under a less
retained peak was identified as S-nitrosoglutathione
(GSNO) by comparison with an authentic sample pre-
pared by a literature procedure.28 As confirmed in con-
trol experiments, in the early stages of the reaction GSH
was converted to GSNO which started to decompose
over about 2 h to yield GSH disulphide (GSSG). For-
mation of GSSG was observed also in the late stages of
the reaction of 1 with nitrite and GSH at pH 4.
When oxidation of 1 (50 mM) was carried out with tyr-
osinase in the presence of 200 mM GSH, substrate con-
sumption was about 95% after 5 min and 3 was
obtained in 72% yield, whereas after 3 h all of the con-
jugation products 3, 4 and 5 were formed in 27, 54 and
Table 2. Fomation of 2 and 3 in the reaction mixture of 1 with GSH
and NaNO2 at pH 4
Concentration (mM)
1
GSH
0.046
NaNO2
—
3a,b
2a,b
10.1
10.1
15
1
5
Figure 2. Consumption of 1 and product formation in the reaction
with GSH and nitrite ions as a function of nitrite concentration.
Reaction of 1 (2 mM) was run in the presence of GSH (1mM) and
varying amounts of NaNO2 (0–10 mM) in 0.05 M acetate buffer (pH
4) and the mixture was analyzed by HPLC at 3 h. Concentration of 1
(*), 2 (~), and 3 (&) in the reaction mixture. Shown are the
meanÆSD values for three separate experiments.
—
0.75
0.01—
5
7
aConcentration at 3-h reaction time.
bReported are the mean values for three separate experiments with SD
ꢁ10%.

L. Panzella et al. / Bioorg. Med. Chem. 11 (2003) 4797–4805
4801
To assess the possible role of GSNO in the nitrite-
mediated formation of conjugates, in separate experi-
ments GSNO was allowed to react with 1 at pH 4.
HPLC analysis, however, did not reveal appreciable
adduct formation. Similar reactions of 1 with nitrite in
the presence of GSSG led to 2 as the main product, with
no detectable GSH adducts.
variance with catecholamine quinones which react
mainly at the 5-position.33 A plausible explanation
would envisage the electronwithdrawing effect of the
unsaturated propenoate chain, which is lacking in those
ortho-quinones, such as dopaquinone and dopamine-
quinone, that react via the 5-position.33,34 Inspection of
canonical resonance structures for 1 quinone indicates
that the propenoate chain decreases electron density at
the 2-position with respect to the 5-position, thus sub-
verting the usual regiochemistry of S-conjugation on 4-
alkyl-ortho-quinones and rendering the 2-position more
susceptible to nucleophilic attack. To substantiate this
view, the total charge density and LUMO coefficients of
methyl caffeate quinone as a model system were pre-
liminarily investigated by semiempirical methods (AM1
and PM3). The results, summarized in Table 4, indi-
cated a more positive charge density on the 6-position
but a larger LUMO coefficient on the 2-position.
Significant formation of 3 was observed when 1 was
allowed to react at concentrations as low as 25 mM in
the presence of comparable amounts of GSH and nitrite
ions at pH 3, at 37 ꢂC, that is under conditions mimicking
those occurring in the gastric compartment29 (Table 3).
Diode array spectrophotometric analysis was used to
investigate the possible intermediacy of 1 quinone in the
nitrite-mediated conjugation of 1 with GSH. Reaction
of 1 (2 mM) with nitrite (1mM) at pH 4 in the presence
of 8 mM GSH led to the development of an inter-
mediate species exhibiting a chromophore centred at
400 nm, apparently identical with that ascribed to the
quinone of 1 or of caffeic acid,23,30,31 whose intensity
decreased when sodium borohydride was added. The
possible contribution of 2 [lmax (pH 4) 278, 340–360
(broad) nm] to the 400 nm chromophore was ruled out
by HPLC analysis of the mixture. The same chromo-
phore was obtained by oxidation of 1 with sodium
nitrite at pH 4, but in the absence of GSH, as well as
with periodate at pH 4 or 6.23 Under both conditions,
the chromophore disappeared by treatment with sodium
borohydride or GSH. HPLC analysis of the oxidation
mixture of 1 with acidic nitrite after addition of GSH
revealed formation of 3 and the 6-nitroderivative 2 in
comparable amounts.
Considering that the SH group is a soft nucleophile,37 it
can be speculated that the regiochemistry of the
nucleophilic attack is governed by frontier orbital
interactions, which consistently predicted the higher
reactivity for the 2-position.
The above discussion relied on a classical mechanism of
conjugation involving nucleophilic addition of GSH to
the quinone of 1 through the sulphydryl group. Such a
mechanism notoriously operates in the case of
tyrosinase, a copper enzyme which can catalyze two-
electron oxidation of catechols, and was previously
demonstrated for the HRP-promoted conjugation.23 In
the latter case, the quinone conceivably arises by
disproportionation of the semiquinone produced by
one-electron oxidation of 1. A detailed description of
the semiquinone radicals derived from 1 with enzyme
Discussion
Despite considerable interest in the oxidation chemistry
of 1, the mechanisms and mode of conjugation with
GSH have not yet been fully elucidated. Under at least
three different conditions of physiological relevance, we
have now shown that 1 is susceptible to sequential oxi-
dative conjugation with GSH up to the hitherto
unknown triadduct stage. Overall, these results expand
current knowledge of oxidative conjugation of caffeic
acid derivatives with GSH,23,25,32 but raise a number of
mechanistic issues.
Table 3. Formation of 3 in the reaction mixture of 1 with GSH and
NaNO2 at pH 3 and 37 ꢂC
Concentration (mM)
1
GSH
NaNO2
3a
25
500
500
25
25
500
25
25
500
4Æ0.7
16Æ1.8
13Æ1.2
aDetermined at 1-h reaction time. Reported are mean valuesÆSD for
experiments run in triplicate.
In a previous paper,23 it was speculated that GSH con-
jugation with 1 proceeded mainly via the 5-position of
the catechol ring, based on the known reactivity of 4-
substituted ortho-quinones with thiol compounds.33ꢀ35
However, caftaric acid was reported to react with GSH
to give exclusively the 2-S conjugate,25 as did 1 with
cysteine,22 whereas similar reaction of caffeic acid with
cysteine was shown to give also lower amounts of the 5-
S isomer.36 In line with these results was also the failure
to observe appreciable formation of the 5-S-isomer in
the oxidation reaction of 1 with GSH under all the
conditions examined in the present study. On this basis
it can be argued that the dominant mode of coupling of
caffeic acid-derived quinones is via the 2-position, at
Table 4. Total charge density (TCD) and LUMO coefficients of
methyl caffeate quinone
TCD
AM1PM3
LUMO
Atoma
AMP1M3
C-1
C-2
C-3
C-4
C-5
C-6
ꢀ0.017
ꢀ0.180
ꢀ0.023
ꢀ0.174
0.295
0.253
ꢀ0.183
ꢀ0.047
0.372
0.437
0.375
0.449
0.2010.254
0.279
0.205
ꢀ0.200
ꢀ0.079
0.246
0.312
0.272
0.233
0.327
0.273
aNumbering as indicated in ref. 24.

4802
L. Panzella et al. / Bioorg. Med. Chem. 11 (2003) 4797–4805
systems has been reported using the ESR spin stabili-
zation technique.38
Based on the present study, we can propose a similar
thiol–quinone interaction mechanism for the nitrite-
promoted conjugation, in which the role of nitrite
would be to generate the strong oxidant NO2 (Eo=0.99
V)39 by decomposition of nitrous acid, according to the
following equations:
NOꢀ2 +H+
HNO2+H+
2HNO2
N2O3
2NO+O2
NO2
ꢀ!
ꢀ
NO++H2O
N2O3+H2O
NO+NO2
2NO2
ꢀ!
ꢀ
ꢀ!
ꢀ
ꢀ!
ꢀ
!
Thus, one-electron oxidation of 1 by NO2 would gen-
erate the semiquinone which would suffer dis-
proportionation to give the ortho-quinone.
A schematic outline of the conjugation pathways of 1
with GSH described in the present study is shown in
Figure 3.
In this scheme, formation of adducts 3–5 is envisaged as
involving sequential addition of GSH to quinone inter-
mediates. These latter may be produced by HRP/H2O2
or acidic nitrite-promoted oxidations, or by a redox
cycling process brought about by 1 quinone.40
Figure 3. Oxidative conjugation pathways of 1 with GSH.
The demonstration that mildly acidic nitrite solutions
can promote the efficient coupling of 1 with GSH is, to
the best of our knowledge, unprecedented in the field of
catechol-thiol conjugation chemistry. Besides the
intrinsic chemical interest, the results of this study pro-
vide an improved background for studies aimed at clar-
ify the mechanisms through which 1 exerts its biological
effects. Dietary sources of GSH include raw meats, fresh
fruits and freshly cooked vegetables (ca. 40–150 mg/
kg).41 Since in the gastric compartment pH varies from
2.5 to 4.5 during digestion, conditions may exist for the
reported reaction to occur in vivo following high nitrite
intake, for example with cured/pickled meats or vege-
tables. Exposure of humans to excess nitrite from diet is
currently implicated as a potential contributory factor
in the etiology of cancers of the gastrointestinal tract. In
these circumstances, GSH-adduct formation, by pre-
venting formation of 2, would interfere with the nitrite-
scavenging properties of 1, at variance with the case of
caffeic acid.27 This would support the superior anti-
nitrosaminic and chemopreventive properties of caffeic
acid compared to 1.2,14,17 Moreover, GSH is found
widespread in phenol-rich plant foods, e.g., in grapes,
tomatoes and peaches, whereby conjugation reactions
may occur during oxidation and browning, modifying
both organoleptic properties and effects of phenols on
human health.
obtained by electrochemical oxidation of dopamine in
the presence of GSH35 as well as by HRP/H2O2-medi-
ated reaction of GSH with the flavonoid quercetin.42
This latter paper contains structural information for
ortho-quinone GSH adducts in a case close to that pre-
sented in this study. Clearly, more work is required
before the actual biological relevance of the reported
chemistry is assessed.
Experimental
Materials and methods
Chlorogenic acid (1), glutathione (GSH), d-glucose,
sodium nitrite and sodium borohydride were purchased
from Aldrich and used as obtained. Hydrogen peroxide
(30% solution) and sodium periodate were form Carlo
Erba. Glutathione disulphide (GSSG) was obtained
from Fluka. Horseradish peroxidase (HRP) (EC
1.11.1.7) type II, mushroom tyrosinase (EC 1.14.18.1)
and glucose oxidase (EC 1.1.3.4) type II were from
Sigma. S-nitrosoglutathione (GSNO) was prepared by
treatment of GSH with sodium nitrite in 2 M HCl.28 UV
spectra were performed on a diode array spectro-
photometer (Hewlett Packard model 8453E). 1H and
13C NMR spectra were recorded at 400.1and
100.6 MHz using a Bruker DRX-400MHz instrument
In addition to the discovery of the nitrite-mediated
conjugation of 1 with GSH, another major outcome of
this study was the first characterization of a GSH
triadduct with a caffeic acid derivative. Whereas cate-
chol–thiol diadducts have been reported in several
studies,32,36 the only triadducts so far known were those
1
fitted with a 5-mm H/broadband gradient probe with
inverse geometry. Heteronuclear multiple quantum
coherence, heteronuclear multiple bond correlation and
1H,1H correlation spectroscopy experiments were from
Bruker library. Chemical shifts are reported in d values

L. Panzella et al. / Bioorg. Med. Chem. 11 (2003) 4797–4805
4803
(ppm) downfield from TMS. D2O was used as the sol-
vent containing t-butanol as internal standard. Positive
ion electrospray (ESI +)-MS spectra were obtained in
1:1 v/v acetonitrile–H2O/1% acetic acid using a Micro-
massZQ Waters equipment. Main peaks are reported
with their relative intensities (percent values are in par-
entheses). Gel filtration was performed using Sephadex
G-10 (eluant water). HPLC analyses were performed
with a Gilson instrument equipped with an UV detector
set at 280 nm. Analyses were run on octadecylsilane
coated columns, 250ꢃ4.6 mm, 5 mm particle size
(Sphereclone, Phenomenex) using 0.1M formic acid–
methanol (80:20 v/v) at a 1.0 mL/min flow rate. For
preparative purposes a 250ꢃ22 mm, 10 mm particle size
(Econosil, Alltech) was used at a flow rate of 15.0 mL/
min (eluant: 0.1M formic acid–methanol 70:30 v/v).
Semiempirical AM1/PM3 calculations were carried out
with the Hyperchem 5.0 package produced by Hyper-
cube Inc. (Waterloo, Ontario, Canada) 1997.
(1U/mL) and H O2 (100 mM) added in two portions in
the presence of GSH (200 mM), and the reaction course
was followed by HPLC.
2
2-S-Glutathionylchlorogenic acid (3). UV: lmax (water)
1
251, 327 nm; H NMR (400 MHz, D2O) d (ppm): 2.05–
2.16 (5H, m), 2.23 (1H, m), 2.41 (2H, t, J=6.8 Hz), 3.05
(1H, dd, J=14.4, 8.8 Hz), 3.22 (1H, dd, J=14.4, 3.6
Hz), 3.68 (2H, s), 3.73 (1H, t, J=6.4 Hz), 3.90 (1H, m),
4.22 (1H, dd, J=8.8, 3.6 Hz), 4.27 (1H, m), 5.31 (1H,
m), 6.27 (1H, d, J=16.0 Hz), 6.88 (1H, d, J=8.4 Hz),
7.12 (1H, d, J=8.4 Hz), 8.10 (1H, d, J=16.0 Hz); 13C
NMR (100 MHz, D2O) d (ppm): 27.3 (CH2), 32.6
(CH2), 36.7 (CH2), 38.4 (CH2), 39.3 (CH2), 43.8 (CH2),
54.8 (CH), 55.3 (CH), 71.6 (CH), 72.4 (CH), 73.8 (CH),
77.5 (C), 117.8 (CH), 118.2 (CH), 121.2 (C), 121.2 (CH),
130.9 (C), 145.2 (CH), 147.4 (C), 147.7 (C), 170.0 (C),
173.0 (C), 175.1(C), 175.8 (C), 176.1 (C), 180.8 (C);
ESI+-MS: m/z 660 ([M+H]+, 100), 682 ([M+Na]+,
17), 698 ([M+K]+, 9).
Oxidation of chlorogenic acid (1) in the presence of
GSH
2,5 - Di - S - glutathionylchlorogenic acid (4). UV: lmax
(water) 268, 379 nm; 1H NMR (400 MHz, D2O) d
(ppm): 2.04–2.11 (7H, m), 2.26 (1H, m), 2.44 (4H, m),
3.06 (1H, dd, J=14.4, 8.8 Hz), 3.23 (1H, dd, J=14.4,
8.4 Hz), 3.27 (1H, dd, J=14.4, 3.6 Hz), 3.41 (1H, dd,
J=14.4, 4.4 Hz), 3.71–3.75 (6H, m), 3.92 (1H, dd,
J=9.2, 2.8 Hz), 4.20 (1H, dd, J=8.8, 3.6 Hz), 4.26 (1H,
d, J=2.8 Hz), 4.46 (1H, dd, J=8.4, 4.4 Hz), 5.33 (1H,
m), 6.42 (1H, d, J=16.0 Hz), 7.38 (1H, s), 8.17 (1H, d,
J=16.0 Hz); 13C NMR (100 MHz, D2O) d (ppm): 27.2
To a solution of 1 (50 mM) and GSH (200 mM) in
0.05 M phosphate buffer (pH 7.4), HRP (1pyrogallol
U/mL) was added followed by H2O2 in two portions up
to 100 mM concentration at 10-min intervals while the
mixture was taken under vigorous stirring at room
temperature. The reaction course was followed by
HPLC analysis.
In other experiments the reaction was run: (i) as above
but with GSH varying in the range of 0–3 molar
equivalents with respect to 1, adding H2O2 (100 mM
final concentration) in two portions at 10 min intervals
or as a bolus; (ii) as above with glucose (0.56 mM)/glu-
(2ꢃCH2), 32.6 (2ꢃCH2), 35.8 (CH2), 37.1(CH ), 38.3
2
(CH2), 39.4 (CH2), 43.8 (2ꢃCH2), 54.2 (CH), 54.5 (CH),
55.3 (2ꢃCH), 71.5 (CH), 72.5 (CH), 73.6 (CH), 77.4 (C),
118.6 (CH), 121.2 (C), 122.9 (C), 124.1 (CH), 130.8 (C),
144.2 (CH), 147.5 (2ꢃC), 169.6 (C), 173.0 (2ꢃC), 175.2
(2ꢃC), 175.8 (2ꢃC), 176.2 (2ꢃC), 180.9 (C); ESI+-MS:
m/z 483 ([M+2H]2+, 55), 965 ([M+H]+, 47), 987
([M+Na]+, 24).
cose oxidase (0.1U/mL) in place of H O2.
2
In those experiments using tyrosinase, the substrate at
50 mM or 2.0 mM concentration in 0.05 M phosphate
buffer, pH 7.4, was treated with the enzyme, 25 or 100
U/mL, respectively, in the presence of GSH 200 mM or
10 mM, respectively, and the mixture was taken under
vigorous stirring. The reaction course was followed by
HPLC analysis as above.
Isolation of 2,5,6-tri-S-glutathionylchlorogenic acid (5).
The reaction of 1 with HRP/H2O2 and GSH was carried
out as for the preparation of 3 and 4 using 50 mg of the
starting material and adding four portions of H2O2 (1
molar equivalent each) at 50-min intervals. At complete
consumption of 1, 3 and 4 (HPLC analysis), the reac-
tion mixture was worked up as above. The residue was
dissolved in water and purified by Sephadex G-10
chromatography to give 5 (tR 10.1 min, 161 mg, 90%
yield) as a pale yellow glassy oil.
Isolation of 2-S-glutathionylchlorogenic acid (3) and 2,5-
di-S-glutathionylchlorogenic Acid (4). For preparative
purposes, the reaction of 1 with HRP/H2O2 and GSH
was carried out using 200 mg of the starting material.
To a solution of 1 (2 mM) in 0.05 M phosphate buffer
(pH 7.4), GSH (10 mM) was added followed by HRP
(4 U/mL) and H2O2 in two portions up to 4 mM
concentration at 10 min intervals while the mixture
was taken under vigorous stirring at room tempera-
ture. After 1h, the reaction mixture was treated with
Na2S2O5 (20 mg), acidified with 4 M HCl to pH 3.0
and evaporated to dryness. The residue was dissolved
in water and purified by preparative HPLC to give 3
(tR 25.7 min, 120 mg, 32% yield) and 4 (tR 39.4 min,
98 mg, 18% yield) in pure form as pale yellow glassy
oils. Either 3 or 4 at 50 mM concentration were reac-
ted separately with tyrosinase (25 U/mL) or with HRP
2,5,6-Tri-S-glutathionylchlorogenic acid (5). UV: lmax
1
(water) 279 nm; H NMR (400 MHz, D2O) d (ppm):
2.08–2.23 (10H, m), 2.39–2.61 (6H, m), 3.00 (1H, dd,
J=14.4, 9.6 Hz), 3.09 (1H, dd, J=14.0, 9.2 Hz), 3.22
(1H, m), 3.28 (1H, m) 3.29 (1H, dd, J=14.4, 4.4 Hz),
3.46 (1H, dd, J=13.6, 4.8 Hz), 3.80 (3H, m), 3.82 (6H,
s), 3.91 (1H, m), 4.17 (1H, dd, J=9.2, 4.8 Hz), 4.24 (1H,
dd, J=8.4, 4.8 Hz), 4.29 (1H, m), 4.74 (1H, m), 5.37
(1H, m), 6.38 (1H, d, J=16.0 Hz), 8.00 (1H, d, J=16.0
Hz); 13C NMR (100 MHz, D2O) d (ppm): 27.3
(3ꢃCH2), 32.6 (3ꢃCH2), 36.8 (CH2), 37.4 (CH2), 38.3
(CH2), 39.1(CH 2), 39.5 (CH2), 44.1(3 ꢃCH2), 54.0

4804
L. Panzella et al. / Bioorg. Med. Chem. 11 (2003) 4797–4805
(CH), 54.1(CH), 54.4 (CH), 55.3 (3 ꢃCH), 71.6 (CH),
72.6 (CH), 73.7 (CH), 77.6 (C), 120.9 (C), 126.4 (CH),
127.1 (C), 129.9 (C), 138.1 (C), 144.5 (CH), 148.3 (C),
148.5 (C), 169.1 (C), 173.0 (C), 173.2 (C), 175.6(C),
176.1 (3ꢃC), 176.3 (3ꢃC), 176.5 (3ꢃC), 181.1 (C); ESI
+-MS: m/z 635 ([M+ 2H]2+, 78), 1270 ([M+H]+, 4).
Corsaro (Naples University) for mass spectra and help-
ful discussions. Thanks are also due to Miss Silvana
Corsani for excellent technical assistance.
References and Notes
Reaction of 1 with GSH and nitrite. A solution of 1 (200
mM) in methanol, was added to 0.05 M phosphate buf-
fer (pH 4.0) up to a final concentration of 2 mM, fol-
lowed by GSH (10 mM) and sodium nitrite (8 mM).
The mixture was taken under vigorous stirring at room
temperature and periodically analyzed by HPLC.
1. IUPAC nomenclature of cyclitols. Biochem. J. 1976, 153, 23.
2. Kono, Y.; Shibata, H.; Kodama, Y.; Sawa, Y. Biochem. J.
1995, 312, 947.
3. Clifford, M. N. J. Sci. Food Agric. 1999, 79, 362.
4. Yagasaki, K.; Miura, Y.; Okauchi, R.; Furuse, T. Cyto-
technology 2000, 33, 229.
5. Huang, M. T.; Smart, R. C.; Wong, C. Q.; Conney, A. H.
Cancer Res. 1988, 48, 5941.
6. Shimizu, M.; Yoshimi, N.; Yamada, Y.; Matsunaga, K.;
Kawabata, K.; Hara, A.; Morikawa, H.; Mori, H. J. Toxicol.
Sci. 1999, 24, 433.
7. Yamada, J.; Tomita, Y. Biosci. Biotechnol. Biochem. 1996,
60, 328.
8. Krakauer, T. Immunopharmacol. Immunotoxicol. 2002, 24,
113.
9. Mori, H.; Tanaka, T.; Shima, H.; Kuniyasu, T.; Takahashi,
M. Cancer Lett. 1986, 30, 49.
10. Tanaka, T.; Nishikawa, A.; Shima, H.; Sugie, S.; Shinoda,
T.; Yoshimi, N.; Iwata, H.; Mori, H. Basic Life Sci. 1990, 52,
429.
11. Tanaka, T.; Kojima, T.; Kawamori, T.; Wang, A.; Suzui,
M.; Okamoto, K.; Mori, H. Carcinogenesis 1993, 14, 1321.
12. Tsuchiya, T.; Suzuki, O.; Igarashi, K. Biosci. Biotechnol.
Biochem. 1996, 60, 765.
In other experiments the reaction of 1 was run: (i) as
above, but in the presence of GSNO (5 molar equiva-
lents) without addition of GSH and nitrite; (ii) using
GSSG in place of GSH; (iii) using GSH and nitrite
varying in the range 0.1–5 and 1–5 molar equivalents,
respectively, with the substrate at 1.0 mM concen-
tration; (iv) as in the general procedure but using 0.05 M
acetate buffer (pH 3.0) as the reaction medium taken at
37 ꢂC, with the substrate at 25 or 500 mM in the presence
of GSH and nitrite varying in the range 0.05–1molar
equivalents. In other experiments, GSH was reacted
with sodium nitrite under the conditions of the general
procedure but in the absence of 1.
Generation and identification of 1 quinone
13. Stich, H. F. Prev. Med. 1992, 21, 377.
14. Li, P.; Wang, H. Z.; Wang, X. Q.; Wu, Y. N. Biomed.
Environ. Sci. 1994, 7, 68.
15. Torres y Torres, J. L.; Rosazza, J. P. N. J. Agric. Food
Chem. 2001, 49, 1486.
16. Cotelle, P.; Vezin, H. Tetrahedron Lett. 2001, 42, 3303.
17. Napolitano, A.; d’Ischia, M. J. Org. Chem. 2002, 67, 803.
18. Iwahashi, H.; Ishii, T.; Sugata, R.; Kido, R. Arch. Bio-
chem. Biophys. 1990, 276, 242.
19. Koshihara, Y.; Neichi, T.; Murota, S.; Lao, A.; Fujimoto,
Y.; Tatsuno, T. Biochim. Biophys. Acta 1984, 792, 92.
20. Iwahashi, H.; Negoro, Y.; Ikeda, A.; Morishita, H.; Kido,
R. Biochem. J. 1986, 239, 641.
21. Ohnishi, M.; Morishita, H.; Iwahashi, H.; Toda, S.; Shir-
ataki, Y.; Kimura, M.; Kido, R. Phytochemistry 1994, 36, 579.
22. Richard, F. C.; Goupy, P. M.; Nicolas, J. J.; Lacombe, J.;
Pavia, A. A. J. Agric. Food Chem. 1991, 39, 841.
23. Moridani, M. Y.; Scobie, H.; Jamshidzadeh, A.; Salehi,
P.; O’Brien, P. J. Drug Metab. Dispos. 2001, 29, 1432.
24. The common catecholamine nomenclature system was
adopted which assigns numbers 3 and 4 to the OH-bearing
carbons.
25. Cheynier, V. F.; Trousdale, E. K.; Singleton, V. L.; Sal-
gues, M. J.; Wylde, R. J. Agric. Food Chem. 1986, 34, 217.
26. Bolton, J. L.; Acay, N. M.; Vukomanovic, V. Chem. Res.
Toxicol. 1994, 7, 443.
A solution of 1 (20 mM) in methanol was added to
0.05 M acetate buffer (pH 4) up to a 2 mM concen-
tration, followed by GSH (8 mM) and sodium nitrite (1
mM). The spectrum of the mixture was recorded within
2 s after addition of nitrite and then periodically over
10-min by a diode array spectrophotometer. Product
analysis was carried by HPLC. In other experiments the
reaction was run: (i) as above but with addition of
NaBH4 (2 mM) at 10 min reaction time; (ii) with GSH
(8 mM) or NaBH4 (2 mM) added 10 min after addition
of sodium nitrite; 5 min after addition of GSH, the
mixture was analyzed by HPLC. In control experiments
the absorbance at 400 nm of a solution of sodium nitrite
(1mM) or of 3 (2 mM) in the presence of equimolar
amounts of NaBH4 at pH 4 was measured. In other
experiments, oxidation of 1 by sodium periodate in
0.05M phosphate buffer (pH 6.0 or 4.0) was carried out
as described,23 with the substrate and the oxidant at 250
mM concentration, and the spectrum of the mixture was
recorded at the addition of the oxidant. When required,
NaBH4 (250 mM) or GSH (250 mM) was added to the
mixture 30 s after addition of the oxidant; 15 s after
addition of GSH, the reaction mixture was acidified to
pH 3.0 with 4 M HCl and analyzed by HPLC.
27. Panzella, L.; Napolitano, A.; d’Ischia, M. Bioorg. Med.
Chem. Lett. 2002, 12, 3547.
28. Hart, T. W. Tetrahedron Lett. 1985, 26, 2013.
29. Tricker, A. R.; Haubner, R.; Spiegelhalder, B.; Pre-
ussmann, R. Food Chem. Toxicol. 1988, 26, 861.
30. Kerry, N.; Rice-Evans, C. FEBS Lett. 1998, 437, 167.
31. Fulcrand, H.; Cheminat, A.; Brouillard, R.; Cheynier, V.
Phytochemistry 1994, 35, 499.
32. Salgues, M.; Cheynier, V.; Gunata, Z.; Wylde, R. J. Food
Sci. 1986, 51, 1191.
33. Prota, G. Fortsch. Chem. Org. Naturst. 1995, 64, 94.
Acknowledgements
This work was carried out in the frame of a research
program on antitumour agents sponsored by MURST
(PRIN 2001). We thank the ‘Centro Interdipartimentale
di Metodologie Chimico-Fisiche’ of Naples University
‘Federico II’ for NMR facilities and Prof. M. M.

L. Panzella et al. / Bioorg. Med. Chem. 11 (2003) 4797–4805
4805
34. Ito, S.; Palumbo, A.; Prota, G. Experientia 1985, 41, 960.
35. Zhang, F.; Dryhurst, G. J. Electroanal. Chem. 1995, 398,
117.
36. Cilliers, J. J. L.; Singleton, V. L. J. Agric. Food Chem.
39. Koppenol, W. H.; Moreno, J. J.; Pryor, W. A.; Ischir-
opoulos, H.; Beckman, J. S. Chem. Res. Toxicol. 1992, 5, 834.
40. Cheynier, V. F.; Van Hulst, M. W. J. J. Agric. Food Chem.
1988, 36, 1 0.
1990, 38, 1789.
37. Pearson, R. G.; Songstand, J. J. Am. Chem. Soc. 1967, 89,
41. Wierzbicka, G. T.; Hagen, T. M.; Jones, D. P. J. Food
Compos. Anal. 1989, 2, 327.
1827.
38. Grace, S. C.; Yamasaki, H.; Pryor, W. A. Basic Life Sci.
1999, 66, 435.
42. Awad, H. M.; Boersma, M. G.; Boeren, S.; van Bladeren,
P. J.; Vervoort, J.; Rietjens, I. M. C. M. Chem. Res. Toxicol.
2002, 15, 343.