Hydroxycinnamic acids as DNA-cleaving agents in the presence of Cu II ions: Mechanism, structure-activity relationship, and biological implications

Article abstract of DOI:10.1002/chem.200901627

The effectiveness of hydroxycinnamic acids (HCAs), that is, caffeic acid (CaA), chlorogenic acid (ChA), sinapic acid (SA), ferulic acid (FA), 3hydroxycinnamic acid (3-HCA), and 4hydroxycinnamic acid (4-HCA), as pBR322 plasmid DNA-cleaving agents in the pr

Full text of DOI:10.1002/chem.200901627

FULL PAPER
DOI: 10.1002/chem.200901627
Hydroxycinnamic Acids as DNA-Cleaving Agents in the Presence of CuII
Ions: Mechanism, Structure–Activity Relationship, and Biological
Implications
Gui-Juan Fan, Xiao-Ling Jin, Yi-Ping Qian, Qi Wang, Ru-Ting Yang, Fang Dai,*
Jiang-Jiang Tang, Ya-Jing Shang, Li-Xia Cheng, Jie Yang, and Bo Zhou*[a]
Abstract: The effectiveness of hydroxy-
cinnamic acids (HCAs), that is, caffeic
acid (CaA), chlorogenic acid (ChA), si-
napic acid (SA), ferulic acid (FA), 3-
hydroxycinnamic acid (3-HCA), and 4-
hydroxycinnamic acid (4-HCA), as
pBR322 plasmid DNA-cleaving agents
in the presence of CuII ions was investi-
gated. Compounds bearing o-hydroxy
or 3,5-dimethoxy groups on phenolic
rings (CaA, SA, and ChA) were re-
markably more effective at causing
DNA damage than the compounds
bearing no such groups; furthermore,
CaA was the most active among the
HCAs examined. The involvement of
reactive oxygen species (ROS) and CuI
ions in the DNA damage was affirmed
by the inhibition of the DNA breakage
by using specific scavengers of ROS
and a CuI chelator. The interaction be-
tween CaA and CuII ions and the influ-
ence of ethylenediaminetetraacetic
acid (EDTA), the solvent, and pH
value on the interaction were also stud-
ied to help elucidate the detailed
prooxidant mechanism by using UV/
Vis spectroscopic analysis. On the basis
of these observations, it is proposed
that it is the CaA phenolate anion, in-
stead of the parent molecule, that che-
lates with the CuII ion as a bidentate
ligand, hence facilitating the intramo-
lecular electron transfer to form the
corresponding CaA semiquinone radi-
cal intermediate. The latter undergoes
a second electron transfer with oxygen
to form the corresponding o-quinone
and a superoxide, which play a pivotal
role in the DNA damage. The interme-
diacy of the semiquinone radical was
supported by isolation of its dimer
from the CuII-mediated oxidation prod-
ucts. Intriguingly, CaA was also the
most cytotoxic compound among the
HCAs toward human promyelocytic
leukemia (HL-60) cell proliferation.
Addition of exogenous CuII ions result-
ed in an effect dichotomy on cell via-
bility depending on the concentration
of CaA; that is, low concentrations of
CaA enhanced the cell viability and,
conversely, high concentrations of CaA
almost completely inhibited the cell
proliferation. On the other hand, when
superoxide dismutase was added
before, the two stimulation effects of
exogenous CuII ions were significantly
ameliorated, thus clearly indicating
that the oxidative-stress level regulates
cell proliferation and death. These
findings provide direct evidence for the
antioxidant/prooxidant mechanism of
cancer chemoprevention.
Keywords: caffeic acid · cancer ·
chemoprevention · DNA damage ·
reaction mechanisms
oxygen species
·
reactive
Introduction
Chemoprevention by using naturally occurring or synthetic
antioxidants has recently emerged as a viable alternative
strategy for preventing carcinogenesis.[1] Growing evidence
has revealed that oxidative DNA damage induced by reac-
tive oxygen species (ROS) contributes to human tumori-
[a] G.-J. Fan, Dr. X.-L. Jin, Y.-P. Qian, Q. Wang, R.-T. Yang, Dr. F. Dai,
J.-J. Tang, Y.-J. Shang, L.-X. Cheng, J. Yang, Prof. Dr. B. Zhou
State Key Laboratory of Applied Organic Chemistry
Lanzhou University
AHCTUNGTRENNUNG
genesis[2] and that regular consumption of certain fruits and
Lanzhou, Gansu 730000 (China)
Fax : (+86)931-8915557
E-mail: bozhou@lzu.edu.cn
vegetables containing substantial amounts of various poly-
phenolics with antioxidative properties can decrease the risk
of suffering from cancer.[3] From this evidence, the idea that
antioxidants in these foods that are effective cancer chemo-
daifang@lzu.edu.cn
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200901627.
Chem. Eur. J. 2009, 15, 12889 – 12899
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
12889

preventive agents has developed.[3] Generally speaking, anti-
oxidants can counteract ROS production and inhibit ROS-
induced oxidative DNA damage, hence decreasing the risk
of cancer.[4] However, every antioxidant is in fact a redox
agent and thus might become a prooxidant to induce DNA
damage under special conditions. Antioxidant vitamins, such
as vitamins E[5] and C[6] and other polyphenolic antioxidants,
including quercetin,[7] curcumin,[8] tea catechins,[9] salsoli-
nol,[10] resveratrol,[11] and phenolic acids,[12] have been re-
ported to act as DNA-cleaving agents in the presence of
cupric ions by producing ROS. Over the past few years, in-
creasing experimental evidence has indicated that antioxi-
dant-mediated production of ROS (prooxidant action) may
be responsible for the induction of apoptosis of cancer cells
and cancer chemopreventive.[13] Recently, Levine and co-
workers demonstrated that ascorbate (a well-known antioxi-
dant) at pharmacological concentrations is a prooxidant,
thus generating hydrogen peroxide-dependent cytotoxicity
toward a variety of cancer cells in vitro without adversely af-
fecting normal cells.[13b,c] Therefore, it is desirable to know
how an antioxidant molecule can switch from an antioxidant
to a prooxidant and from an inhibitor of DNA damage to
an inducer of DNA damage as well as its implication in
cancer chemoprevention.
the biological implications of the copper-dependent prooxi-
dative action of polyphenolic antioxidants. Therefore, we
were motivated to see whether the correlation between the
prooxidative activity on DNA damage and cytotoxicity
against cancer cells exists in other polyphenolic antioxidants
and whether the addition of cupric ions can result in the en-
hancement of the cytotoxicity of polyphenolic antioxidants.
Phenolic acids, especially hydroxycinnamic acid deriva-
tives, are widely distributed in plants and are present in con-
siderable amounts in fruits, vegetables, and beverages in the
human diet.[22] The daily uptake of caffeic acid (3,4-dihy-
droxycinnamic acid) has been estimated to be 206 mg in
subjects drinking coffee.[23] The antioxidative properties of
hydroxycinnamic acids have been extensively studied,[24] but
their prooxidative properties, especially the detailed mecha-
nism, have not been well understood. Recently, we conduct-
ed a preliminary investigation into the prooxidative action
of hydroxycinnamic acids,[12] but the detailed prooxidative
mechanism and biological implications have not been well
exploited. Therefore, in the present study hydroxycinnamic
acids (HCAs), including caffeic acid (CaA), chlorogenic
acid (ChA), sinapic acid (SA), ferulic acid (FA), 3-hydroxy-
cinnamic acid (3-HCA), and 4-hydroxycinnamic acid (4-
HCA; Scheme 1) were chosen to probe their abilities to act
as DNA-cleaving agents in the presence of CuII ions. The in-
teraction between HCAs and CuII ions and the oxidation
products of CaA and SA obtained in the presence of CuII
ions were also studied to help elucidate the prooxidative
mechanism. To verify the biological implications of the
prooxidative action, the antiproliferative effects of HCAs on
HL-60 cells in the absence and presence of exogenous CuII
ions were assessed with a trypan-blue dye-exclusion test.
Copper is an important metal ion present in chromatin
and is closely associated with DNA bases, in particular gua-
nine.[14] It is also one of the most redox-active metal ions
present in cells.[15] Several studies have shown that serum
and tumor copper levels are significantly elevated in various
types of cancers, such as breast cancer,[16] lung cancer,[17] and
leukemia.[18] For example, it was shown that copper concen-
trations in the serum and cells of leukemic patients
(328 mgmLꢀ1 and 52 mg1010 cellsꢀ1 in the serum and cells, re-
spectively) were significantly higher than in healthy donors
(114 mgmLꢀ1 and 15 mg1010 cellsꢀ1 in the serum and cells, re-
spectively).[18a] Thus, compared with normal cells, preneo-
plastic and neoplastic cells might be more sensitive to
copper-related redox reactions that switch an antioxidant to
a prooxidant to generate ROS, thus resulting in DNA
damage. DNA damage induced by antioxidants in the pres-
ence of copper might be an important pathway through
which preneoplastic and neoplastic cells can be killed while
normal cells survive.[19] This behavior could be one reason
for the ability of antioxidants to differentiate between
normal and abnormal tumor cells. Indeed, many antioxi-
dants are nontoxic to normal cells.[20] Resveratrol, a well-
known antioxidant and cancer chemopreventive agent found
in a wide variety of dietary sources including grapes, plums,
and peanuts, has been shown to induce apoptosis in human
leukemia (HL-60) cells but not in normal peripheral blood
lymphocytes.[20b] In our ongoing research on bioanti-
Scheme 1. Molecular structures of hydroxycinnamic acids (HCAs).
ACHTUNGTRENNUNG
oxidants,[4a,11f,21] we recently found an interesting correlation
between antioxidant activity, prooxidant activity, cytotoxici-
ty, and apoptosis-inducing activity of resveratrol and its ana-
logues.[21a] However, little is known about the detailed
prooxidative mechanism of polyphenolic antioxidants in the
presence of copper from a chemical point of view and about
Results
Strand breakage of plasmid pBR322 DNA induced by
HCAs and CuII ions and the involvement of ROS and CuI
ions: The single-strand breakage of supercoiled plasmid
12890
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 12889 – 12899

Cytotoxicity of Hydroxycinnamic Acids in the Presence of CuII Ions
FULL PAPER
DNA with a relatively high electrophoretic mobility leads to
the formation of an open circular conformation with a de-
creased electrophoretic mobility in agarose, whereas the for-
mation of linear DNA is indicative of a double-strand
breakage and has a mobility intermediate between that of
the supercoiled and open circular conformation.[25] We ex-
amined the effect of HCAs in combination with CuII ions on
the plasmid DNA breakage by using agarose gel electropho-
resis (Figure 1A). Neither CuII ions (0.2 mm) nor HCAs
and fragmental forms (lane 8, Figure 1A). ChA completely
converted the supercoiled DNA into a relaxed circular form
in the presence of the CuII ions and FA showed a similar but
somewhat weaker effect (lanes 6 and 10, respectively, Fig-
ure 1A). On the other hand, 3-HCA and 4-HCA were com-
pletely ineffective even in the presence of CuII ions (lanes
12 and 13, respectively, Figure 1A). The order of DNA-
cleaving effectiveness is CaA>SA>ChA>FA >4-HCA
and 3-HCA. It is worth noting that the compounds bearing
o-hydroxy or 3,5-dimethoxy groups on the phenolic rings
(i.e., CaA, SA, and ChA) are most active followed by a
compound bearing o-methoxyhydroxy groups (i.e., FA),
whereas the monohydroxy-substituted substrates (4-HCA
and 3-HCA) are inactive.
To analyze the role of ROS in the HCAs/CuII-dependent
DNA breakage, we used specific scavengers of activated
oxygen and a CuI chelator to define the nature of the reac-
tive species. Glutathione (GSH), a ROS scavenger, and
bathocuproinedisulfonic acid disodium salt (BCDS), a spe-
cific CuI chelator, both provided protection against DNA
strand breakage induced by CaA in the presence of CuII
ions (Figure 1C). GSH and BCDS at concentrations of 0.5
and 0.4 mm, respectively, completely inhibited the CaA/CuII-
mediated stand breakage. Catalase (CAT; 50 mgmLꢀ1) effec-
tively protected plasmid DNA from DNA strand breakage
induced by CaA (0.1 mm) and CuII ions (0.1 mm), thus clear-
ly indicating the involvement of hydrogen peroxide in the
process (Figure 1D). These findings suggest that both ROS
and the CuII/CuI redox cycle are critical to the DNA
damage.
Figure 1. A) Effect of HCAs on pBR322 DNA strand breakage in the
presence of CuII ions. Supercoiled pBR322 DNA (125 ng) was incubated
with 0.2 mm CuII and 0.1 mm HCAs at 378C for 1 h in 10 mm PBS
(pH 7.4). Lanes: 1: control; 2: CuII alone; 3: CaA alone; 4: CaA and
CuII; 5: ChA alone; 6: ChA and CuII; 7: SA alone; 8: SA and CuII; 9: FA
alone; 10: FA and CuII; 11: 4-HCA alone; 12: 4-HCA and CuII; 13: 3-
HCA and CuII. B) Effect of CaA on pBR322 DNA strand breakage in
the presence of CuII ions. The conditions were the same as for (A),
except that the total concentration of CaA plus CuII ions was maintained
constant (0.3 mm). Lanes: 1: control; 2: 0.225 mm CaA and 0.075 mm
CuII; 3: 0.2 mm CaA and 0.1 mm CuII; 4: 0.15 mm CaA and 0.15 mm CuII;
5: 0.1 mm CaA and 0.2 mm CuII; 6: 0.075 mm CaA and 0.225 mm CuII; 7:
0.06 mm CaA and 0.24 mm CuII; 8: 0.24 mm CuII alone. C) Effect of GSH
and BCDS on CaA and CuII-mediated DNA strand breakage. The exper-
imental conditions were the same as for (A): 1: control; 2: 0.1 mm CaA
plus 0.2 mm CuII; 3–5: 0.25, 0.5, and 1 mm GSH, respectively; 7–9: 0.05,
0.1, 0.2, and 0.4 mm BCDS, respectively. D) Effect of catalase on CaA
and CuII-mediated DNA strand breakage. The experimental conditions
were the same as for (A). Lanes: 1: control; 2: 0.1 mm CaA and 0.1 mm
CuII; 3–7: 12.5, 25, 50, 75, and 100 mgmLꢀ1 CAT, respectively.
UV/Vis spectral changes of HCAs in the presence of CuII
ions and influence of EDTA, solvent, and pH value on the
interaction of CaA and CuII ions: To clarify the mechanism
of the DNA damage, the UV/Vis absorption changes of
HCAs upon the addition of CuII ions were examined. Fig-
ure 2A was obtained when CuII ions were added to CaA in
phosphate-buffered saline (PBS; pH 7.4). The rapid disap-
pearance of the absorption bands of CaA centered at l=
287 and 311 nm was accompanied by the appearance of
bathochromic-shifted peaks at l=304 and 349 nm and the
hypsochromic-shifted peak at l=259 nm, which is character-
istic for the formation of a chelate complex of a CaA pheno-
late anion with a CuII ion (see below). The intensity of the
former two peaks (l=304 and 349 nm) decreased with an
increase in time, and the absorption bands in the range l=
259–270 nm and at about l=400 nm were enhanced due to
the formation of CaA o-quinone (see below). Two isosbestic
points at l=288 and 402 nm suggested a direct transition
from one form (CaA phenoxide/CuII chelate) to another
one (CaA o-quinone). A similar bathochromic shift was also
observed in the case of ChA, but its decay was remarkably
slower than that of CaA (spectrum not shown). A compari-
son of Figure 2A with Figure 2B indicates clearly that the
decay of SA is significantly faster than the decay of CaA in
the presence of CuII ions. The addition of CuII ions to SA re-
sulted in the decrease and gradual redshift of the maximal
absorption at l=306 nm accompanied by the increased ab-
(0.1 mm) alone caused detectable DNA damage (lanes 2, 3,
5, 7, 9, and 11, Figure 1A). However, CaA (0.1 mm) could
work cooperatively with CuII ions (0.2 mm) to produce a
smear of fragments, which is indicative of extensive DNA
damage (lane 4, Figure 1A). When the total concentration
of CaA and CuII ions was maintained at a constant (0.3 mm)
and the CaA/CuII molar ratio was varied (Figure 1B), ratios
of 1:1 and 1:2 were the most active in inducing DNA
damage. Therefore, concentrations of 0.1 mm HCAs and 0.2
or 0.1 mm CuII ions were chosen for the following experi-
ments. SA and CuII ions resulted in the conversion of the
substrate DNA almost completely into open circular, linear,
Chem. Eur. J. 2009, 15, 12889 – 12899
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12891

F. Dai, B. Zhou et al.
Figure 2. Absorption spectral changes of HCAs (50 mm) in the absence
(dashed line) and in the presence (solid line) of 100 mm CuII ions in PBS
buffer (pH 7.4) in air. A) CaA (interval=15 min), the dotted line shows
the effect of 300 mm EDTA on the interaction of CaA and CuII ions for
120 min. B) SA (interval=5 min), the square line shows the UV/Vis spec-
trum of the SA dimer (50 mm). The arrows show the time-related absorb-
ance changes.
Figure 3. A) The absorption spectral changes of CaA. Lines: 1: CaA
(50 mm); 2: CaA (50 mm) and CuII (100 mm) for 60 min in air; 3: the same
as (2) but with EDTA (300 mm) in air; 4: the same as (2) but in argon; 5:
the same as (4) but with EDTA (300 mm). B) The absorption spectrum of
CaA (dashed line) and the effect of EDTA (300 mm) on the absorption
spectrum of the CaA/CuII system (solid line) at different time intervals
(interval=15 min). All the spectra were recorded in 10 mm PBS (pH 7.4)
at room temperature. The arrows show the time-related absorbance
changes.
sorption in the range l=400–550 nm, and the appearance of
three isosbestic points at l=246, 259, and 333 nm, respec-
tively. This type of absorption between l=400 and 550 nm
could be due to the formation of dimerization products.[26]
At this point, we identified the oxidative product (furofuran
bislactone) of SA in the presence of CuII ions (see below)
and recorded its UV/Vis spectrum to substantiate the forma-
tion of a dimer. The formation of a dimer corresponds to
the observed UV/Vis spectral changes of SA (square line,
Figure 2B). There was no clear change of absorption spec-
trum for FA, 4-HCA, and 3-HCA in the presence of CuII
ions. The decay rate of HCA or the HCA/CuII chelate com-
plex follows the sequence of SA>CaA>ChA>FA, 4-
HCA, and 3-HCA, which is similar to the activity sequence
of the DNA damage mentioned above.
CuII ions as a bidentate ligand, hence facilitating intramolec-
ular electron transfer to form o-quinone (see below). The
formation of o-quinone became clear by the addition of
EDTA at different reaction times (Figure 3B). Two new
bands (lmax =252 and 392 nm) observed in the range l=
232–270 nm and at around l=400 nm can be assigned to
CaA o-quinone (see below; Figure 3B). The spectral
changes showed two isosbestic points at l=270 and 347 nm,
thus suggesting that a simple equilibrium between two spe-
cies (CaA and CaA o-quinone) was operating. The partici-
pation of oxygen in the reaction of CaA with CuII ions was
studied by mixing CaA with CuII ions in an inert atmos-
phere. Although the CuII chelate still formed, the bands at
l=304 and 349 nm did not decrease after the CuII ions had
undergone a reaction with CaA for 60 minutes in the inert
atmosphere (line 4, Figure 3A). When EDTA was added at
this point, the redshifted bands diminished further and the
absorption bands of CaA recovered almost to their initial
intensities (line 5, Figure 3A). Therefore, CuII ions could
not induce the reaction of CaA in the absence of O2.
The formation of the CaA phenolate anion/CuII chelate
complex was confirmed by reaction with ethylenediamine-
ACHTUNGTRENNUNGtetraacetic acid (EDTA), a well-known chelating agent for
metal ions (Figure 3A). EDTA was added after the CuII
ions underwent a reaction with CaA for 60 minutes. Upon
the addition of 300 mm EDTA, the redshifted bands (l=304
and 349 nm) returned back to their initial position with a de-
crease in absorbance (line 3, Figure 3A). This outcome indi-
cates unambiguously that CaA phenoxide can chelate with
In solvents that support ionization, such as water and al-
cohols, phenol (ArOH) may equilibrate with the corre-
12892
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 12889 – 12899

Cytotoxicity of Hydroxycinnamic Acids in the Presence of CuII Ions
FULL PAPER
sponding phenolate anion (ArOꢀ) [Eq. (1)], which is a much
stronger electron donor relative to the parent ArOH. Ethyl
acetate has a much lower dielectric constant than ethanol
(e=6.02 versus 24.30, respectively)[27] and hence is less able
to support ionization of the substrate. To rationalize the
mechanism and actual electron donor, the effect of the sol-
vent (ethanol and ethyl acetate) on the decay rate of CaA
was examined (see Figures S1A and S1B in the Supporting
Information for the results). In ethanol, the absorbance of
CaA at lmax =325 nm was indeed slightly redshifted to lmax
=
329 nm on the addition of CuII ions, but the decay of the
latter was significantly slower than in PBS (pH 7.4; see Fig-
ure S1A in the Supporting Information). However, in ethyl
acetate, no redshifted peak was observed and the CuII ions
could not induce the reaction of CaA (see Figure S1B in the
Supporting Information). This outcome clearly suggests that
the actual electron donor is the CaA phenolate anion
(ArOꢀ) instead of the parent molecule (ArOH).
ArOH Ð ArOꢀ þ Hþ
ð1Þ
It has been reported that the oxidation rate of phenol in-
creases remarkably with the increasing anionic character of
phenol in alkaline media.[28] Therefore, the influence of the
pH value on the decay rate of CaA in the presence of CuII
ions and the spectral change of CaA was investigated. In an
NH4Cl/NH4OH buffer solution (pH 10.0), the maximum ab-
sorption of the CaA phenolate anion at lmax =344 nm was
slightly redshifted to lmax =349 nm on the addition of CuII
ions (Figure 4A). This latter peak is the same as the absorp-
tion observed in PBS (pH 7.4), thus indicating unambiguous-
ly that it is the CaA phenolate anion, instead of the parent
molecule, that chelates with the CuII ions in PBS (pH 7.4).
This spectral evolution in the NH4Cl/NH4OH buffer solution
(pH 10.0) was characterized by the decrease in the absorp-
tion bands at l=255 and 349 nm, the enhancement of the
absorption bands in the range l=267–297 nm and above l=
400 nm, and the presence of three isosbestic points at l=
267, 297, and 391 nm. Notably, the decay of the CaA pheno-
late anion/CuII chelate complex was much faster in the
NH4Cl/NH4OH buffer solution (pH 10.0) than that in PBS
(pH 7.4). On the other hand, no redshifted peak was ob-
served in a NaAc/HAc buffer solution (pH 6.0) and the
decay of CaA is very slow (Figure 4B), thus further confirm-
ing the obligatory role of the phenolate anion in the reac-
tion process.
Figure 4. Absorption spectral changes of CaA (50 mm) in the absence
(dashed line) and in the presence (solid line) of 100 mm CuII ions in
A) NH4Cl/NH4OH buffer solution (pH 10.0) and B) NaAc/HAc buffer
solution (pH 6.0) in air (interval=15 min). The arrows show the time-re-
lated absorbance changes.
Figure 5. Oxidative products of A) CaA and B) SA in the presence of
CuII ions.
mation). The CaA dimer has been isolated previously from
the cultured mushroom Inonotus sp. K-1410[29] and obtained
by oxidizing CaA with NaIO4 in water.[30]
Oxidative products of CaA and SA in the presence of CuII
ions in a mixed solvent: To further verify the prooxidative
mechanism, we subsequently isolated and identified the oxi-
dative products of CaA and SA in the presence of CuII ions
in a mixed solvent (water/acetonitrile=2:1 (v/v)) at room
temperature. The major products (Figure 5) were the dimers
(furofuran bislactone) 2,6-bis(3’,4’-dihydroxyphenyl)-3,7-
Effect of HCAs on HL-60 cell proliferation in the absence
and presence of exogenous CuII ions: The antiproliferative
effect of HCAs on HL-60 cells (seeded density=2ꢁ
105 cellsmLꢀ1) was assessed by using a trypan-blue dye-ex-
clusion test[31] (the results are exemplified in Figure 6 and
summarized in Table 1). CaA exhibited dose-dependent in-
hibitory effects on HL-60 cell proliferation (Figure 6), and
the cell viability was decreased to 67, 50, and 35% after
dioxabicycloACHTUNGTRENNUNG[3.3.0]octane-4,8-dione and 2,6-bis(4’-hydroxy-
3’,5’-dimethoxyphenyl)-3,7-dioxabicycloACTHNUTRGNE[NUG 3.3.0]octane-4,8-
dione as characterized with HRMS (ESI) and 1H and
13C NMR spectroscopic analysis (see the Supporting Infor-
Chem. Eur. J. 2009, 15, 12889 – 12899
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12893

F. Dai, B. Zhou et al.
Figure 6. Inhibitory effects of CaA on HL-60 cell viability (seeded densi-
ty=2ꢁ105 cellsmLꢀ1). Values are the mean ꢁS.D. for three independent
experiments.
Figure 7. Inhibitory effects of CaA in the a) absence and b) presence of
250 mm exogenous CuII ions on HL-60 cell viability. Line (c) is the same
as (b) but with SOD (0.5 mgmLꢀ1). Cells (seeded density=5ꢁ105
cellsmLꢀ1) were treated continuously with CaA for 24 h. Values are the
mean ꢁS.D. for three independent experiments.
Table 1. Cytotoxicity against HL-60 cells of HCAs in vitro.[a]
HCAs
IC50 [mM]
150 mm CaA almost completely inhibited the cell prolifera-
tion (line b, Figure 7). On the other hand, when superoxide
dismutase (SOD; 0.5 mgmLꢀ1, a noncytotoxic concentra-
tion) was added before biphasic stimulation effects of exoge-
nous CuII ions (pro-proliferative and antiproliferative effects
in the presence of low and high concentrations of CaA, re-
spectively) were significantly ameliorated (line c, Figure 7),
thus suggesting that superoxide produced by CaA in the
presence of exogenous CuII ions might play a pivotal role in
the stimulation effects and the level of this superoxide mod-
ulates cell proliferation and death.[19a,b,c] Figure 8 shows the
CaA
ChA
SA
14.9ꢁ0.3
18.8ꢁ0.8
>250
>250
>250
>250
1.4ꢁ0.4
FA
4-HCA
3-HCA
VP-16
[a] Cytotoxicity is expressed as IC50, that is, the concentration of the com-
pound that causes 50% inhibition of the cell viability. Cells (seeded den-
sity=2ꢁ105 cellsmLꢀ1) were treated continuously with HCAs for 48 h.
The data are expressed as the mean ꢁS.D. for three determinations.
treatment with 10, 15, and 20 mm CaA, respectively, for 48 h.
The IC50 values for all of these compounds are listed in
Table 1. Interestingly, the molecules bearing o-dihydroxy
groups (CaA and ChA) exhibited remarkably higher cyto-
toxicity than other analogues, and CaA was the most reac-
tive among the HCAs.
Furthermore, to investigate whether CuII ions can act syn-
ergistically with HCAs against cancer-cell proliferation, the
antiproliferative effect of HCAs in the presence of exoge-
nous CuII ions on HL-60 cells was also measured (Figures 7
and 8). In contrast to the experiments in the absence of
exogenous CuII ions, the relatively high cell-seeded density
(5ꢁ105 cellsmLꢀ1) and short exposure time (24 h) were
chosen for the following experiments to investigate clearly
the effect of exogenous CuII ions on the cytotoxicity of
HCAs against HL-60 cells. At this cell-seeded density and
exposure time, the presence of CaA increased the cell via-
bility slightly from 50 to 300 mm (line a, Figure 7). A concen-
tration of 250 mm exogenous CuII ions was selected because
CuII ions alone do not influence the cell proliferation. The
addition of exogenous CuII ions enhanced the cell viability
in the presence of a low concentration of CaA (50 mm).
However, the cell viability was sharply decreased with an in-
crease in concentration of CaA (line b, Figure 7). At a con-
centration of CaA that approached 150 mm, the cell viability
was decreased to 30%, and at concentrations higher than
Figure 8. Inhibitory effects of HCAs (250 mm) and exogenous CuII
(250 mm) on HL-60 cell viability. Cells (seeded density=5ꢁ105 cellsmLꢀ1
)
were treated continuously with HCAs for 24 h. Values are the mean
ꢁS.D. for three independent experiments.
synergistic effect against cancer-cell proliferation of HCAs
with exogenous CuII ions (250 and 250 mm, respectively),
thus indicating clearly that the synergism is much more pro-
nounced in the case of CaA and ChA than in the other
cases. Specifically, the cell viability in the presence of CaA
and CuII ions (250 and 250 mm, respectively) was 10%,
which is much lower than those produced by CaA and CuII
12894
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 12889 – 12899

Cytotoxicity of Hydroxycinnamic Acids in the Presence of CuII Ions
FULL PAPER
ions (60 and 99%, respectively) when they were used indi-
vidually under the same experimental conditions. Again, the
molecules bearing o-dihydroxy groups exhibited remarkably
higher activities than those bearing no such groups.
pH value on the spectral changes of CaA in the presence of
CuII ions, that it is the ArOꢀ ion instead of the parent mole-
cule (ArOH) that chelates with the CuII ions. The ArOꢀ/
CuII complex can undergo intramolecular electron transfer
to form the corresponding semiquinone radical and CuI ion.
It has been pointed out that a protonated phenolic group is
not a particularly good ligand for metal cations but once de-
protonated an oxygen center is generated that possesses a
high charge density, the so-called “hard” ligand.[35] Although
the pKa2 value of CaA is 8.48,[36] in the presence of CuII ions,
the proton is displaced at much lower pH values, for exam-
ple, pH 5.0–8.0.[35] Therefore, CaA should dissociate to form
a phenoxide, which chelates CuII ions as a bidentate ligand
and undergoes intramolecular electron transfer to form an
o-hydroxyphenoxyl radical (semiquinone radical). The radi-
cal intermediate was also proved by the formation of the
dimer (furofuran bislactone) (Figure 5) of CaA in the pres-
ence of CuII ions. The dimer must be formed by the 8,8’-cou-
pling of the two radicals followed by the intramolecular nu-
cleophilic addition of the carboxylate groups onto the qui-
none methide moieties (Scheme 2).
The higher activity of the compounds bearing o-dihydrox-
yl groups in the DNA damage can be understood because
the oxidation intermediate, the o-hydroxyphenoxyl radical,
is more stable due to the intramolecular hydrogen-bonding
interaction, as evidenced recently from both spectrophoto-
metric measurements[37] and theoretical calculations.[38] The
acidity dissociation constant of the o-hydroxyphenoxyl radi-
cal is much lower (pKa1 =4.1)[39] than that of catechol
(pKa1 =9.25).[40] Thus, the o-hydroxyphenoxyl radical should
dissociate and form o-semiquinone radical anions. The o-
semiquinone radical anion will be more easily further oxi-
dized to form the final product o-quinone (Scheme 2). Two
new peaks that appeared at l=252 and 392 nm in the case
of CaA demonstrates the formation of CaA o-quinone. This
outcome is in accordance with the previous observation of
the absorbances of CaA o-quinone at l=248 and 400 nm,
which appeared during the chemical and enzymatic oxida-
tion of CaA by o-chloranil[41] and polyphenol oxidase,[42]
Discussion
The notion of cancer chemoprevention through antioxidant
intervention arises from the fact that fruits and vegetables
contain antioxidants that are linked to low cancer rates in
those who consume them. In contrast, a recent meta-analysis
of the clinical data from 68 randomized human trials with
232606 participants revealed that an increased risk of mor-
tality was associated with the regular use of certain putative
antioxidants, such as vitamins A and E and b-carotene.[32] In
several large-scale interventions with disease or death as the
endpoint, supplementation with b-carotene resulted in no
effect or an increase in cancer incidence.[33] These mixed re-
sults have made people reconsider the role of antioxi-
dants.[34] As a matter fact, in addition to the counteraction
of ROS production, antioxidants induce a multitude of ef-
fects on many other cellular functions including cell signal-
ing, apoptosis, production of phase-II enzymes, and so forth.
Recently, it has been proposed in some comprehensive re-
views that the prooxidant action of plant-derived phenolics
rather than their antioxidant action may play an important
role in their cancer-chemopreventive properties, as ROS
produced by prooxidant action can mediate apoptotic DNA
fragmentation.[19] In the present study, HCAs with different
structural features were selected as the representative phe-
nolic antioxidants to investigate their prooxidant effect on
DNA damage in the presence of CuII ions and the related
mechanisms and biological implications. We not only ex-
plored the detailed prooxidative mechanism for HCAs in
the presence of cupric ions from a chemical point of view
but also provided direct evidence for prooxidant cancer-che-
mopreventive action.
Mechanism and structure–activity relationship: HCAs,
which are generally effective antioxidants, can switch to
being prooxidants in the presence of CuII ions to induce
DNA damage, except for monohydroxy-substituted HCAs
(4-HCA and 3-HCA; Figure 1A). It is clearly seen that the
compounds bearing o-dihydroxy and o-dimethoxyhydroxy
groups (i.e., CaA, SA, and ChA) are most active in inducing
plasmid pBR322 DNA strand breakage in the presence of
CuII ions.
The high activity of CaA and ChA is obviously due to
their o-dihydroxy groups, which can lose a proton to pro-
duce the corresponding phenolate anion (ArOꢀ) followed
by chelation with CuII ions to form ArOꢀ/CuII complexes
(see Figures 2–4 and Figure S1 in the Supporting Informa-
tion). Although the formation of metal complexes of poly-
phenols bearing o-dihydroxy groups has been well docu-
mented,[10] the present work provides unambiguous evi-
dence, by means of the effects of EDTA, the solvent, and
ACHTUNGTRENNUNG
in the absence of oxygen (line 4, Figure 3B) demonstrates
unambiguously the involvement of O2 in the process. The
CaA/CuII-mediated DNA damage was ameliorated by CAT,
GSH, and BCDS, thus suggesting the involvement of ROS
and CuI ions in the process. Therefore, a possible mecha-
nism of DNA damage induced by CaA in the presence of
CuII ions can be clarified as shown in Scheme 2. The initial
electron-transfer oxidation of CaA by CuII ions generates
the corresponding semiquinone radical, which undergoes a
ꢀC
second electron transfer with O2 to form o-quinone and O2
ꢀC
I
radicals. The O2 radicals react with Cu ions to give hydro-
gen peroxide, which is readily converted by a Fenton-type
reaction into hydroxyl radicals to induce the oxidative DNA
damage. In addition, CuI ions can be oxidized back to CuII
ions by oxygen. This CuII/CuI redox cycle renders CaA be
catalytically oxidized.
Chem. Eur. J. 2009, 15, 12889 – 12899
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12895

F. Dai, B. Zhou et al.
Scheme 2. Proposed mechanism for CaA/CuII-mediated DNA damage.
It is also noticeable that SA bearing o-dimethoxyhydroxy
groups exhibits the fastest reaction rate with CuII ions
among the examined HCAs (Figure 2B), which is in line
with the previous observation in the electron-transfer reac-
lowest value. It has been well documented that complexa-
tion of metal ions such as Cu, Fe, and Mn with phenolic
compounds results in the changes in their redox poten-
tials;[45] for example, the reduction potential of a catechol/
CuII complex is about E=0.277 V (versus NHE), whereas it
is E=0.442 V (versus NHE) for catechol.[45]
C
tion of HCAs with 2,2-diphenyl-1-picrylhydrazyl (DPPH ),
with SA as the most reactive.[24c] The possible reason for this
behavior is related to the structure of the ArOꢀ(SA)/CuII
complex, the formation of which was also elucidated by the
gradual redshift of the maximal absorption of SA (Fig-
ure 2B). In the complex, there are two electron-donating
(ED) methoxy groups in the ortho position of the ArOꢀ ion,
which could further enhance the electron density of the
ArOꢀ ion and hence result in the fastest intramolecular elec-
tron-transfer rate. By comparing the DNA-cleaving effec-
tiveness of the three compounds (CaA, ChA, and SA), we
found the active sequence CaA>SA>ChA and can con-
clude that in addition to the electron-transfer reaction rate
(i.e., the formation rate of ROS), the formation of o-qui-
none is the other important factor to influence the DNA-
cleaving effectiveness. It has been well documented that o-
quinones are involved in DNA damage by forming covalent
adducts with DNA.[10,43] Although SA exhibited the fastest
electron-transfer reaction rate, it could not form o-quinone
in its reaction with CuII ions. Therefore, SA was the second
most effective at DNA cleavage.
Biological implications: Intriguingly, the compounds bearing
o-dihydroxy groups (i.e., CaA and ChA) exhibited remark-
AHCTUNGTREGaNNNU bly higher cytotoxicity toward HL-60 cells than those com-
pounds bearing no such groups (Table 1). Notably, these
compounds also exhibited enhanced activity in inducing
DNA damage. A similar structure–activity relationship was
also observed in resveratrol and its analogues.[21a] In contrast
to ChA, SA had a relatively high activity in inducing DNA
damage, but exhibited low cytotoxicity toward HL-60 cells.
This behavior highlights the involvement of o-quinone in
the cytotoxicity, in line with reports of o-quinone products
from the oxidation of catecholic estrogen[43a] and dopa-
AHCTUNGTRENNUNG
mine[43b] that are responsible for the observed apoptotic ef-
fects of these chemicals in mutagenic and neuroblastoma
cells.
In the absence of exogenous CuII ions, CaA alone exhibit-
ed cytotoxicity toward HL-60 cells (Figure 6), thus implying
the existence of endogenous CuII ions. Recently, Hadi and
co-workers proposed a mechanism for the cytotoxic action
of plant polyphenols against cancer cells that involves mobi-
lization of endogenous copper and the consequent pro-
The oxidative potential was reported to be Epa =0.212,
0.261, 0.314, 0.430, and 0.583 V (versus Ag/AgCl) for CaA,
ChA, SA, FA, and 4-HCA respectively.[44] It should be
noted that the ranking of the oxidation potentials does not
corroborate with the fact that SA was the most reactive spe-
cies in the electron-transfer reaction. The reason for this in-
consistency could be the formation of the ArOꢀ/CuII com-
plex, which could shift the oxidation potential of SA to the
AHCTUNGTRENNUNG
oxidant action.[11c,d,19e] It was also reported that exogenous
copper could enhance cytotoxic and apoptosis-inducing ac-
tivities of phenolic antioxidants against human leukemia
and breast cancer cells.[46] In the present study, the incorpo-
ration exogenous CuII ions into the cells at a concentration
12896
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 12889 – 12899

Cytotoxicity of Hydroxycinnamic Acids in the Presence of CuII Ions
FULL PAPER
nology. HL-60 cell lines were originally obtained from the Shanghai Insti-
tute of Biochemistry and Cell Biology (Chinese Academy of Sciences).
All the other chemicals were of the highest quality available.
of 250 mm demonstrated that the biphasic pattern of cell via-
bility depended on the concentration of CaA (Figure 7). A
concentration of 50 mm CaA enhanced cell viability and,
conversely, CaA at high concentrations (i.e., >150 mm)
almost completely inhibited cell proliferation. Furthermore,
the addition of SOD significantly inhibited the two stimula-
tion effects of exogenous CuII ions, thus clearly indicating
that the ROS level modulates cell proliferation and death
and the two stimulation effects of exogenous CuII ions come
from weak and strong oxidative stress (prooxidant action),
respectively. It is now well recognized that ROS might func-
tion as a doubled-edged sword.[19a,b,c] A moderate increase in
ROS may promote cell proliferation and survival; however,
when the increase in ROS reaches a certain level (the toxic
threshold), it may trigger cell death.[19a,b,c] In contrast to
normal cells, cancer cells exhibit increased intrinsic ROS
stress and copper levels.[19a,b,c,47] Therefore, a further increase
in ROS stress in cancer cells by using an exogenous ROS-
generating agent (prooxidant) is likely to cause elevation of
ROS above the threshold level, thus leading to death of
these cells.[19a,b,c] However, the same concentration of the
exogenous ROS-generating agent (prooxidant) in normal
cells could not induce the ROS level to reach the toxic
threshold, and thus the normal cells survive. As a whole, in
the case of antioxidant-based cancer chemoprevention,
strong oxidative stress acts as a “good” stress and may play
a crucial role in the cytotoxicity of antioxidants against
cancer cells (Scheme 2). Additionally, it is worth noting that
CaA and ChA clearly exhibit a synergistic effect with CuII
ions against HL-60 cell proliferation (Figure 8), thus further
highlighting the importance of o-dihydroxy groups in phe-
nolic antioxidants in the cancer chemoprevention.
Assay for oxidative DNA strand breakage: The induction of DNA strand
breakage by HCAs was assessed by measuring the conversion of the su-
percoiled pBR322 plasmid DNA into open circular and linear forms or
into further fragmentation by gel electrophoresis.[25] pBR322 DNA
(125 ng) was incubated with HCAs and/or CuII ions in PBS at pH 7.4 and
378C for 1 h in 1.5-mL microcentrifuge tubes. The total volume was
20 mL, that is, DNA (5 mL), HCAs (5 mL), CuII (5 mL), and PBS (5 mL). In
the inhibition experiments, specific scavengers of ROS and the CuI chela-
tor were preincubated before addition of the CuII ions. After incubation,
the samples (10 mL) were mixed with gel loading buffer (2 mL; 0.25%
bromophenol blue and 30% (w/v) glycerol) and immediately loaded in a
1% agarose gel containing 40 mm tris(hydroxymethyl)aminomethane
(Tris), 20 mm sodium acetate, and 2 mm EDTA and subjected electropho-
resis in a horizontal slab gel apparatus in Tris/acetate/EDTA gel buffer
for 1 h. The gels were stained with 0.5 mgmLꢀ1 ethidium bromide for 1 h
followed by destaining in water for 0.5 h, and photographed under UV
light.
UV/Vis spectral measurements: The UV/Vis spectra were measured at
room temperature with a Hitachi 557 spectrophotometer. PBS containing
50 mm HCAs was kept at room temperature, and the spectral tracing was
started by the addition of 100 mm CuSO4. All the spectra were run against
blanks containing the buffer and the metal ions. The spectra were record-
ed every appointed time after addition of 100 mm CuII ions.
Oxidative product analysis of CaA and SA in the presence of CuII ions:
CaA (500 mg, 2.8 mmol) or SA (672 mg, 2.8 mmol) was dissolved in ace-
tonitrile (100 mL) and an aqueous solution of CuSO4·5H2O (2.8 mm,
200 mL) was added. The reaction mixture was stirred for 7 h at room
temperature and evaporated to dryness at 408C under reduced pressure.
The residue was redissolved in water and extracted with ethyl acetate
(3ꢁ150 mL). The combined organic layers were washed with brine, dried
over Na2SO4, and taken to dryness to give a residue. The residues of
CaA and SA were purified by column chromatography on silica gel and
eluted with chloroform/acetone/acetic acid (4:1:0.5, v/v/v) and petroleum
ether/acetone (4:1, v/v), respectively. The CaA dimer (170 mg) and SA
dimer (520 mg) were obtained as
a white and yellow powders,
AHCTUNGTREGUNrNN espectively.
CaA dimer: 1H NMR (400 MHz, (CD3)2CO): d=3.99 (brs, 2H; H-1 and
H-5), 5.72 (brs, 2H; H-2 and H-6), 6.81 (dd, J=8.0, 2.0 Hz, 2H; H-6’and
H-6’’), 6.92 (d, J=2 Hz, 2H; H-2ꢂ and H-2’’), 6.87 ppm (d, J=8.0 Hz,
2H; H-5ꢂ and H-5’’); 13C NMR (100 MHz, (CD3)2CO): d=49.3 (C-1, C-
5), 83.2 (C-2, C-6), 113.9 (C-6’, C-6’’), 116.5 (C-5’, C-5’’), 118.4 (C-2’, C-
2’’), 131.3 (C-1’, C-1’’), 146.5 (C-4’, C-4’’), 146.8 (C-3’, C-3’’), 176.1 ppm
(C-4, C-8); HRMS (ESI): m/z: calcd for C18H14O8 +NH4+: 376.1027 [M+
NH4]+; found: 376.1029, error=0.5 ppm.
Conclusions
In conclusion, our data substantiate the important role of
HCAs as DNA-cleaving agents in the presence of CuII ions.
The observation that compounds bearing o-dihydroxy moi-
eties exhibit remarkably high activities in DNA damage and
an antiproliferative effect on HL-60 cells in the absence and
presence of exogenous CuII ions, provides useful information
for the identification and development of more potent
cancer-chemopreventive agents. The detailed prooxidative
mechanism also provides the necessary groundwork for un-
derstanding antioxidant-based cancer chemoprevention.
1
SA dimer: H NMR (400 MHz, (CD3)2CO): d=7.45 (s, 2H; OH), 6.74 (s,
4H; H-2’, H-6’, H-2’’, and H-6’’), 5.76 (s, 2H; H-2 and H-6), 4.13 (s, 2H;
H-1 and H-5), 3.84 ppm (s, 12H; H-3’, H-5’, H-3’’ and H-5’’-OCH3);
13C NMR (100 MHz, (CD3)2CO): d=176.2 (C-4, C-8), 149.2 (C-3’, C-5’,
C-3’’ and C-5’’), 137.6 (C-4’, C-4’’), 130.0 (C-1’, C-1’’), 104.4 (C-2’, C-6’, C-
2’’ and C-6’’), 83.5 (C-2, C-6), 56.9 (C-3’, C-5’, C-3’’, C-5’’-OCH3),
49.2 ppm (C-1, C-5); ESI-MS: m/z: 446.9 [M+H]+.
Assessment of cell viability: The numbers of viable cells were counted
using a trypan-blue dye-exclusion test.[31] Trypan-blue dye-exclusion test
is based on the principle that live cells possess intact cell membranes that
exclude certain dyes, such as trypan blue, whereas dead cells do not. The
viable cell number was counted. In the experiment, HL-60 cells were
seeded at a density of 2ꢁ105 cellsmLꢀ1 in 24-well multiwell plates and
treated with various concentrations of HCAs. After incubation in a hu-
midified CO2 (5%) incubator at 378C for 48 h, cells were harvested and
stained with trypan blue and observed under a microscope for cell-
number counting. For the cell-viability assay of HCAs in the presence of
cupric ions, 5ꢁ 105 cells were plated onto 24-well multiwell plates and
treated with various concentrations of CaA and cupric ions (250 mm) for
24 h. For the effect of SOD on the cell viability of CaA in the present of
Experimental Section
Materials: Caffeic acid (CaA, Acros), chlorogenic acid (ChA, Aldrich),
sinapic acid (SA, Acros), ferulic acid (FA, Aldrich), 3-hydroxycinnamic
acid (3-HCA, Aldrich), 4-hydroxycinnamic acid (4-HCA, Fluka), batho-
cuproinedisulfonic acid disodium salt (BCDS, Fluka), etoposide (VP-16,
Sigma), trypan blue (Sigma), ethidium bromide (Sigma), catalase (CAT,
3000 Umg proteinꢀ1, Sigma), and pBR322 DNA (Sigma) were purchased
in the highest purity available and used as received. Superoxide dismu-
tase (SOD, 6500 Umg proteinꢀ1) was purchased from Beyotime Biotech-
Chem. Eur. J. 2009, 15, 12889 – 12899
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12897

F. Dai, B. Zhou et al.
cupric ions, SOD (0.5 mgmLꢀ1) was incubated for 5 h before addition of
CaA and cupric ions.
Arpaci, A. Ozet, B. Doner, T. Karayilanoglu, A. Sayar, O. Berk,
Biol. Trace Elem. Res. 1994, 40, 31–37.
[17] a) M. Diez, M. Arroyo, F. J. Cerdan, M. Munoz, M. A. Martin, J. L.
Balibrea, Oncology 1989, 46, 230–234; b) A. Scanni, L. Licciardello,
M. Trovato, M. Tomirot II, M. Biraghi, Tumori 1977, 63, 175–180.
[18] a) U. Carpentieri, J. Myers, L. Thorpe, C. W. Daeschner III, M. E.
Haggard, Cancer Res. 1986, 46, 981–984; b) X. L. Zuo, J. M. Chen,
X. Zhou, X. Z. Li, G. Y. Mei, Biol. Trace Elem. Res. 2006, 114, 41–
54.
[19] a) D. Trachootham, J. Alexandre, P. Huang, Nat. Rev. Drug Discov-
ery 2009, 8, 579–591; b) A. Gupte, R. J. Mumper, Cancer Treat. Rev.
2009, 35, 32–46; c) N. Hail, Jr., M. Cortes, E. N. Drake, J. E. Spall-
holz, Free Radical Biol. Med. 2008, 45, 97–110; d) J. Antosiewicz,
W. Ziolkowski, S. Kar, A. A. Powolny, S. V. Singh, Planta Med.
2008, 74, 1570–1579; e) S. M. Hadi, S. H. Bhat, A. S. Azmi, S. Hanif,
U. Shamim, M. F. Ullah, Semin. Cancer Biol. 2007, 17, 370–376;
f) S. Nair, W. Li, A. N. Kong, Acta Pharmacol. Sin. 2007, 28, 459–
472; g) H. Pelicano, D. Carney, P. Huang, Drug Resist. Updates 2004,
7, 97–110; h) Q. Kong, J. A. Beel, K. Q. Lillehei, Med. Hypotheses
2000, 55, 29–35.
Acknowledgements
This work was supported by the National Natural Science Foundation of
China (grant nos. 20972063 and 20621091), the 111 Project, and Program
for New Century Excellent Talents in University (NCET-06–0906)
[1] a) M. B. Sporn, N. Suh, Nat. Rev. Cancer 2002, 2, 537–543; b) M. B.
Sporn, K. T. Liby, Nat. Clin. Pract. Oncol. 2005, 2, 518–525; c) Che-
moprevention Working Group to the American Association for
Cancer Research: Cancer Res. 1999, 59, 4743–4758.
[2] a) M. S. Cooke, M. D. Evans, M. Dizdaroglu, J. Lunec, FASEB J.
2003, 17, 1195–1214; b) S. Perwez Hussain, L. J. Hofseth, C. C.
Harris, Nat. Rev. Cancer 2003, 3, 276–285; c) M. D. Evans, M. Diz-
daroglu, M. S. Cooke, Mutat. Res. 2004, 567, 1–61.
[20] a) Z. P. Chen, J. B. Schell, C. T. Ho, K. Y. Chen, Cancer Lett. 1998,
129, 173–179; b) M. V. Clꢃment, J. L. Hirpara, S. W. Chawdhury, S.
Pervaiz, Blood 1998, 92, 996–1002.
[3] a) Y. J. Surh, Nat. Rev. Cancer 2003, 3, 768–780; b) A. R. Collins,
Eur. J. Cancer 2005, 41, 1923–1930; c) M. H. Pan, G. Ghai, C. T. Ho,
Mol. Nutr. Food Res. 2008, 52, 43–52.
[21] a) Y. P. Qian, Y. J. Cai, G. J. Fan, Q. Y. Wei, J. Yang, L. F. Zheng,
X. Z. Li, J. G. Fang, B. Zhou, J. Med. Chem. 2009, 52, 1963–1974;
b) Y. J. Shang, Y. P. Qian, X. D. Liu, F. Dai, X. L. Shang, W. Q. Jia,
Q. Liu, J. G. Fang, B. Zhou, J. Org. Chem. 2009, 74, 5025–5031; c) B.
Zhou, Q. Miao, L. Yang, Z. L. Liu, Chem. Eur. J. 2005, 11, 680–691;
d) W. F. Chen, S. L. Deng, B. Zhou, L. Yang, Z. L. Liu, Free Radical
Biol. Med. 2006, 40, 526–535; e) B. Zhou, L. M. Wu, L. Yang, Z. L.
Liu, Free Radical Biol. Med. 2005, 38, 78–84.
[4] a) B. Zhou, Z. L. Liu, Pure Appl. Chem. 2005, 77, 1887–1903;
b) D. E. Brash, P. A. Havre, Proc. Natl. Acad. Sci. USA 2002, 99,
13969–13971; c) C. A. Rice-Evans, A. T. Diplock, Free Radical Biol.
Med. 1993, 15, 77–96.
[5] N. Yamashita, M. Murata, S. Inoue, M. J. BurKitt, L. Miline, S. Ka-
wanishi, Chem. Res. Toxicol. 1998, 11, 855–862.
[6] I. D. Podmore, H. R. Griffiths, K. E. Herbert, N. Mistry, P. Mistry, J.
Lunec, Nature 1998, 392, 559.
[7] N. Yamashita, H. Tanemura, S. Kawanishi, Mutat. Res. 1999, 425,
107–115.
[8] a) H. Ahsan, S. M. Hadi, Cancer Lett. 1998, 124, 23–30; b) M. Yosh-
ino, M. Haneda, M. Naruse, H. H. Htay, R. Tsubochi, S. L. Qiao,
W. H. Li, K. Murakami, T. Yokochi, In Vitro Toxicol. 2004, 18, 783–
789.
[9] F. Hayakawa, T. Kimura, T. Maeda, M. Fujita, H. Sohmiya, M. Fujii,
T. Ando, Biochim. Biophys. Acta 1997, 1336, 123–131.
[10] Y. J. Jung, Y. J. Surh, Free Radical Biol. Med. 2001, 30, 1407–1417.
[11] a) K. Fukuhara, N. Miyata, Bioorg. Med. Chem. Lett. 1998, 8, 3187–
3192; b) A. Ahmad, S. F. Asad, S. Singh, S. M. Hadi, Cancer Lett.
2000, 154, 29–37; c) A. S. Azmi, S. H. Bhat, S. M. Hadi, FEBS Lett.
2005, 579, 3131–3135; d) A. S. Azmi, S. H. Bhat, S. Hanif, S. M.
Hadi, FEBS Lett. 2006, 580, 533–538; e) K. Fukuhara, M. Nagaka-
wa, I. Nakanishi, K. Ohkubo, K. Imai, S. Urano, S. Fukuzumi, T.
Ozawa, N. Ikota, M. Mochizuki, N. Miyata, H. Okuda, Bioorg. Med.
Chem. 2006, 14, 1437–1443; f) L. F. Zheng, Q. Y. Wei, Y. J. Cai, J. G.
Fang, B. Zhou, L. Yang, Z. L. Liu, Free Radical Biol. Med. 2006, 41,
1807–1816.
[22] R. J. Robbins, J. Agric. Food Chem. 2003, 51, 2866–2887.
[23] J. Radtke, J. Linseisen, G. Wolfram, Zeitschrift fꢀr Ernahrungswis-
senschaft 1998, 37, 190–197.
[24] a) J. C. Cheng, F. Dai, B. Zhou, L. Yang, Z. L. Liu, Food Chem.
2007, 104, 132–139; b) M. Roche, C. Dufour, N. Mora, O. Dangles,
Org. Biomol. Chem. 2005, 3, 423–430; c) M. C. Foti, C. Daquino, C.
Ceraci, J. Org. Chem. 2004, 69, 2309–2314; d) D. Taubert, T. Brei-
tenbach, A. Lazar, P. Censarek, S. Harlfinger, R. Berkels, W. Klaus,
R. Roesen, Free Radical Biol. Med. 2003, 35, 1599–1607; e) M. Nar-
dini, M. D’Aquino, G. Tomassi, V. Gentili, M. Di Felice, C. Scaccini,
Free Radical Biol. Med. 1995, 19, 541–552; f) J. Laranjinha, O.
Vieira, V. Madeira, L. Almeida, Arch. Biochem. Biophys. 1995, 323,
373–381.
[25] A. Rahman, F. Fazel, J. Greensill, K. Ainley, J. H. Parish, S. M.
Hadi, Mol. Cell. Pharmacol. 1992, 111, 3–9.
[26] a) P. Hapiot, A. Neudeck, J. Pinson, H. Fulcrand, P. Neta, C. Rolan-
do, J. Electroanal. Chem. 1996, 405, 169–176; b) R. Petrucci, P. As-
tolfi, L. Greci, O. Firuzi, L. Saso, G. A. Marrosu, Electrochim. Acta
2007, 52, 2461–2470.
[27] G. Litwinienko, K. U. Ingold, J. Org. Chem. 2004, 69, 5888–5896.
[28] a) K. Mukai, W. Oka, K. Watanabe, Y. Egawa, S. I. Nakaoka, J.
Phys. Chem. A 1997, 101, 3746–3753; b) K. Mukai, S. Mitani, K.
Ohara, S. I. Nagaoka, Free Radical Biol. Med. 2005, 38, 1243–1256.
[29] Y. Kumada, H. Naganawa, H. Ilnuma, M. Matsuzaki, T. Takeuchi,
H. Umezawa, J. Antibiot. 1976, 29, 882–889.
[30] H. Tazaki, D. Taguchi, T. Hayashida, K. Nabeta, Biosci. Biotechnol.
Biochem. 2001, 65, 2613–2621.
[31] S. I. Schlager, A. C. Adams, Methods Enzymol. 1983, 93, 233–245.
[32] G. Bjelakovic, D. Nikolova, L. L. Gluud, R. G. Simonetti, C. Gluud,
JAMA J. Am. Med. Assoc. 2007, 297, 842–857.
[12] L. F. Zheng, F. Dai, B. Zhou, L. Yang, Z. L. Liu, Food Chem. Toxi-
col. 2008, 46, 149–156.
[13] a) T. Nakazato, K. Ito, Y. Ikeda, M. Kizaki, Clin. Cancer Res. 2005,
11, 6040–6049; b) Q. Chen, M. G. Espey, A. Y. Sun, J. H. Lee, M. C.
Krishna, E. Shacter, P. L. Choyke, C. Pooput, K. L. Krik, G. R.
Buettner, M. Levine, Proc. Natl. Acad. Sci. USA 2007, 104, 8749–
8754; c) Q. Chen, M. G. Espey, A. Y. Sun, C. Pooput, K. L. Krik,
M. C. Krishna, D. B. Khosh, J. Drisko, M. Levine, Proc. Natl. Acad.
Sci. USA 2008, 105, 11105–11109; d) M. H. Pan, M. C. Hsieh, J. M.
Kuo, C. S. Lai, H. Wu, S. Sang, C. T. Ho, Mol. Nutr. Food Res. 2008,
52, 527–537; e) P. Javvadi, A. T. Segan, S. W. Tuttle, C. Koumenis,
Mol. Pharmacol. 2008, 73, 1491–1501.
[14] K. Agarwal, A. Sharma, G. Talukder, Chem.-Biol. Interact. 1989, 69,
1–16.
[15] Y. Yoshida, S. Furuta, E. Niki, Biochim. Biophys. Acta 1993, 1210,
81–88.
[16] a) A. E. Schwartz, G. W. Leddicotte, R. W. Fink, E. W. Friedman,
Surgery 1974, 76, 325–329; b) P. M. Santoliquido, H. W. Southwick,
J. H. Olwin, Surg. Gynecol. Obstet. 1976, 142, 65–69; c) I. Yucel, F.
[33] a-Tocopherol b-carotene cancer prevention study group: New Engl.
J. Med. 1994, 330, 1029–1035.
[34] a) J. M. Gutteridge, B. Halliwell, Ann. N. Y. Acad. Sci. 2000, 899,
136–147; b) A. Azzi, K. J. A. Davies, F. Kelly, FEBS Lett. 2004, 558,
3–6; c) B. Halliwell, J. Rafter, A. Jenner, Am. J. Clin. Nutr. 2005, 81,
268S-278S; d) R. P. Howes, Ann. N. Y. Acad. Sci. 2006, 1067, 22–26.
[35] B. C. Hider, Z. D. Liu, H. H. Khodr, Methods Enzymol. 2001, 335,
190–203.
12898
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 12889 – 12899

Cytotoxicity of Hydroxycinnamic Acids in the Presence of CuII Ions
FULL PAPER
[36] F. A. M. Silva, F. Borges, C. Guimar¼es, J. L. F. C. Lima, C. Matos, S.
Reis, J. Agric. Food Chem. 2000, 48, 2122–2126.
[37] M. Foti, G. Ruberto, J. Agric. Food Chem. 2001, 49, 342–348.
[38] J. S. Wright, E. R. Johnson, G. A. Dilabio, J. Am. Chem. Soc. 2001,
123, 1173–1183.
[39] N. Schweigert, R. Hunziker, B. J. Escher, R. I. L. Eggen, Environ.
Toxicol. Chem. 2001, 20, 239–247.
[40] K. B. Patel, R. L. Wilson, J. Chem. Soc. Faraday Trans. 1973, 69,
814–825.
[41] F. Kader, M. Irmouli, N. Zitouni, J. P. Nicolas, M. Metche, J. Agric.
Food Chem. 1999, 47, 4625–4630.
[42] V. Cheynier, M. Moutounet, J. Agric. Food Chem. 1992, 40, 2038–
2044.
5390–5395; b) M. E. Haque, M. Asanuma, Y. Higashi, I. Miyazaki,
K. I. Tanaka, N. Ogawa, Biochim. Biophys. Acta 2003, 1619, 39–52.
[44] H. Hotta, S. Nagano, M. Ueda, Y. Tsujino, J. Koyama, T. Osakai,
Biochim. Biophys. Acta 2002, 1572, 123–132.
[45] H. S. Mahal, S. Kapoor, A. K. Satpati, T. Mukherjee, J. Phys. Chem.
B 2005, 109, 24197–24202.
[46] a) K. Satoh, T. Kadofuku, H. Sakagami, Anticancer Res. 1997, 17,
2487–2490; b) H. N. Yu, J. J. Yin, S. R. Shen, J. Agric. Food Chem.
2004, 52, 462–466; c) A. Gupte, R. J. Mumper, Free Radical Biol.
Med. 2007, 43, 1271–1278.
[47] a) M. F. Renschler, Eur. J. Cancer 2004, 40, 1934–1940; b) S. S. Leo-
nard, G. K. Harris, X. Shi, Free Radical Biol. Med. 2004, 37, 1921–
1942.
[43] a) A. M. Samuni, E. Y. Chuang, M. C. Krishna, W. Stein, W. De-
Graff, A. Russo, J. B. Mitchell, Proc. Natl. Acad. Sci. USA 2003, 100,
Received: June 15, 2009
Published online: October 21, 2009
Chem. Eur. J. 2009, 15, 12889 – 12899
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
12899