Mechanism of toxicity of esters of caffeic and dihydrocaffeic acids

Article abstract of DOI:10.1016/S0968-0896(00)00238-8

Ten esters each of caffeic acid and dihydrocaffeic acid have recently been synthesized. Cytotoxicity evaluations of these esters versus L1210 leukemia and MCF-7 breast cancer cells in culture have led to the delineation of substantially different QSAR for each series. The L1210 QSAR for dihydrocaffeic acid esters resembles the QSAR obtained for simple phenols and estrogenic phenols. However, the QSAR pertaining to the caffeic acid esters differs considerably from its sister QSAR. This difference may be attributed to the presence of the olefinic linkage in the side chain. The octyl ester of caffeic acid is nearly ten times as toxic to the leukemia cells than the widely studied phenethyl ester, CAPE.

Full text of DOI:10.1016/S0968-0896(00)00238-8

Bioorganic & Medicinal Chemistry 9 (2001) 199±209
Mechanism of Toxicity of Esters of Ca€eic and Dihydroca€eic
Acids
Beth Etzenhouser, Corwin Hansch, Sanjay Kapur and C. Dias Selassie*
Chemistry Department, Pomona College, 645 N. College Ave, Claremont, CA 91711, USA
Received 25 July 2000; accepted 28 August 2000
AbstractÐTen esters each of ca€eic acid and dihydroca€eic acid have recently been synthesized. Cytotoxicity evaluations of these
esters versus L1210 leukemia and MCF-7 breast cancer cells in culture have led to the delineation of substantially di€erent QSAR
for each series. The L1210 QSAR for dihydroca€eic acid esters resembles the QSAR obtained for simple phenols and estrogenic
phenols. However, the QSAR pertaining to the ca€eic acid esters di€ers considerably from its sister QSAR. This di€erence may be
attributed to the presence of the ole®nic linkage in the side chain. The octyl ester of ca€eic acid is nearly ten times as toxic to the
leukemia cells than the widely studied phenethyl ester, CAPE. # 2000 Elsevier Science Ltd. All rights reserved.
Introduction
Michael reaction acceptors, diphenols and phenols to
generate reactive intermediates by redox cycling has
been linked to their ability to modulate gene expression
and to inhibit nuclear transcription regulators.15
Ca€eic acid phenethyl ester (CAPE) has been identi®ed
as one of the major components of honeybee propolis.1
This phenolic compound elicits several interesting and
varied biological responses; it wields antimicrobial,
antiin¯ammatory, antineoplastic and antioxidant
activity.2 5 It also inhibits the development of azoxy-
methane-induced tumors in the colon of rats and phorbol
induced tumors in mouse skin.6,7 CAPE is also a potent
inhibitor of ornithine decarboxylase, cyclooxygenase,
lipoxygenases and HIV-1 integrase.8 11
In recent years a great deal of attention has focused on
the role and mechanism of action of polyphenolic com-
pounds as inhibitors of oxidative processes.16 However,
little attention has focused on simple phenols, some of
which have been shown to be carcinogenic17 and estro-
genic.18 The dearth of data pertaining to the e€ects of
various substituted phenols on mammalian cytotoxicity
led to our examination of the growth inhibitory e€ects
of a series of simple and complex X-phenols on a
rapidly growing murine cell line and the development of
appropriate quantitative structure±activity relationships
(QSAR).19
The ability of CAPE to alter oxidative processes is well
documented.12 It can inhibit the formation of intra-
cellular H2O2 and oxidized bases in DNA of 12-O-tet-
radecanoylphorbol-13-acetate (TPA) treated human
cells.13 CAPE has also demonstrated an ability to
induce apoptosis in certain transformed cells but not in
parental normal cells.14 This was one of the ®rst indica-
tions that CAPE could act via a redox mechanism as a
pro-oxidant since its induced apoptosis could be
reduced by treatment with the antioxidant N-acetyl-l-
cysteine. In addition, the reduced glutathione inhibitor,
buthionine sulfoximine, induced apoptosis after CAPE
treatment in nontumorigenic cells indicating that these
cells have an enhanced oxidant defense system in com-
parison to the tumor cells. Recently, the ability of
Chemical reactivity and biological activity have been
well correlated by Brown's variation of the Hammett
electronic constant, sigma-plus (s+). The sigma plus
parameter has found great success in delineating radical
reactions involving phenols.20 Thus this variable was
utilized in the study of the cytotoxic pro®les of X-Phe-
nols versus L1210 murine cells in culture. Eqs. 1 and 2
were formulated for electron-releasing and electron-
withdrawing substituents respectively.19
‡
log 1=C ˆ 1:58ꢀÆ0:26†s ‡ 0:21ꢀÆ0:06†logP ‡ 3:10ꢀÆ0:24†
*Corresponding author. Tel.: +1-909-621-8446; fax: +1-909-607-
7726; e-mail: cselassie@pomona.edu
n ˆ 23; r2 ˆ 0:898; s ˆ 0:191; q2 ˆ 0:868
ꢀ1†
0968-0896/01/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(00)00238-8

200
B. Etzenhouser et al. / Bioorg. Med. Chem. 9 (2001) 199±209
and para substituted phenols, but do not appear to
make hydrophobic contact with a receptor.
log 1=C ˆ 0:62ꢀÆ0:16†logP ‡ 2:35ꢀÆ0:31†
n ˆ 15; r2 ˆ 0:845; s ˆ 0:232; q2 ˆ 0:800
ꢀ2†
Eqs. 1, 3, and 4 with their marked dependence on s+ or
BDE suggest that homolytic cleavage of the O-H bond
is critical to the development of cytotoxicity. The small
coecient with the hydrophobic term suggests that
minimal binding on the surface of a speci®c receptor
also mediates this toxicity. Note that the estrogenic
phenols like Bisphenol A, b-estradiol, nonylphenol and
estriol are well predicted by these models. Preliminary
results with a few antioxidants such as quercetin indi-
cate that they ®t this model well; they are cytotoxic
albeit at a lower level. In order to gain further under-
standing of this ambivalent behavior of phenolic com-
pounds we have synthesized and evaluated 20 ester
analogues of ca€eic acid and dihydroca€eic acid (Fig. 1).
In these equations C is the molar concentration of X-
phenol that induces 50% inhibition of growth after 48 h.
Sigma-plus (s+) represents the through-resonance elec-
tronic attributes of the substituents while logP is the
octanol±water partition coecient of the phenol. The
number of data points in the study is represented by n,
the correlation coecient by r, the standard deviation
by s, and the cross-validated r2 by q2. Eq. 1 correlates a
heterogeneous set of substituted phenols including
complex ones such as bisphenol A, diethylstilbesterol,
estradiol, equilin and equilenin. A larger data set of
electron-releasing X-phenols, which include ortho sub-
stituted phenols, led to the development of Eqs. 3 and
4.21
Results
‡
log 1=C ˆ 1:35ꢀÆ0:15†ꢀ ‡ 0:18ꢀÆ0:04†logP ‡ 3:31ꢀÆ0:11†
n ˆ 51; r2 ˆ 0:895; s ˆ 0:227; q2 ˆ 0:882
ꢀ3†
From the results in Table 1, we have derived Eq. 6 for
the inhibition of growth of L1210 cells by esters of caf-
feic acid.
log 1=C ˆ 0:19ꢀÆ0:02†BDE ‡ 0:21ꢀÆ0:03†logP ‡ 3:11ꢀÆ0:10†
n ˆ 52; r2 ˆ 0:920; s ˆ 0:202; q2 ˆ 0:909
ꢀ4†
log 1=C ˆ 0:46ꢀÆ0:12†ClogP ‡ 3:84 Æ 0:37
n ˆ 9; r2 ˆ 0:915; s ˆ 0:165; q2 ˆ 0:881
ꢀ6†
In QSAR 4, BDE represents the calculated homolytic
bond dissociation energy for each phenol. In the analy-
sis of only ortho substituted congeners, logP was set to
zero when it became clear that hydrophobic contact was
not important in the interactions. Cytotoxicities were
then modelled by Eq. 5.21
outliers : R ˆ C6H13
In this equation, C represents the concentration of the
analogue that induces 50% inhibition of growth (IC50)
while ClogP is the calculated partition coecient of
each compound. In this equation, the absence of an
electronic term such as s+ or BDE is to be expected
since the catecholic moiety is present and constant in all
the analogues. The coecient with the hydrophobic
term is larger than what is normally seen with the
phenolic compounds; in fact it is twice as large as that in
Eqs. 3 and 4. One data point was not included in the
analysis: the hexyl ester of ca€eic acid, which is sig-
ni®cantly less active than predicted. The reduced solu-
bility of this analogue may contribute to its reduced
cytotoxicity. Although the t-Butyl ester is not as well
log 1=C ˆ 0:17ꢀÆ0:03†BDE ‡ 3:18ꢀÆ0:16†
n ˆ 14; r2 ˆ 0:936; s ˆ 0:191; q2 ˆ 0:915
ꢀ5†
Note that the slope and intercept of QSAR 5 are essen-
tially the same as QSAR 4. The addition of a steric term
for ortho substituted phenols did not improve Eq. 5.
Thus, ortho substitued phenols appear to undergo radi-
calization via a similar mechanism that prevails for meta
Table 1. Biological and physicochemical constants used to derive
QSAR 6 for the inhibition of growth of L1210 cells by ca€eic acid
esters
No. Substituent Obsd log 1/C Calcd log 1/C Á log 1/C ClogP
1a
1b
1c
1d
1e
1f
1g
1h
1i
CH3
C2H5
C4H9
C6H13
C8H17
CHMe2
CH2C6H5
(CH2)2C6H5
(CH2)3C6H5
CMe3
4.46
4.62
4.99
5.20
6.10
4.70
5.01
5.20
5.74
5.20
4.39
4.63
5.12
5.60
6.08
4.77
5.20
5.35
5.52
4.96
0.07
0.01
0.13
0.40
0.02
0.07
0.19
0.15
0.22
0.24
1.20
1.73
2.79
3.85
4.90
2.04
2.97
3.30
3.68
2.44
a
1j
aNot used to derive Eq. 6.
Figure 1. Structures of ca€eic acid analogues.

B. Etzenhouser et al. / Bioorg. Med. Chem. 9 (2001) 199±209
201
predicted by this model due to its added bulk and sub-
stantial branching at the a-carbon, its activity still falls
within an acceptable range.
of the ca€eic acid esters in QSAR 6 but similar to that
seen in the case of the phenols.
The growth inhibitory pro®les of these two sets of esters
versus MCF-7 (ER+) cells were also determined and
the following QSAR models were developed. Eqs. 8 and
9 describe the QSAR for the inhibition of growth of
MCF-7 cells by ca€eic esters. See Table 3.
Equation 7 is derived from data in Table 2, enumerating
the cytotoxicities of esters of dihydroca€eic acids.
log 1=C ˆ 0:15ꢀÆ0:05†ClogP ‡ 0:14ꢀÆ0:13†ES ‡ 4:62ꢀÆ0:25†
log 1=C ˆ 0:37ꢀÆ0:07†ClogP ‡ 2:64ꢀÆ0:20†
ꢀ7†
n ˆ 9; r2 ˆ 0:906; s ˆ 0:067; q2 ˆ 0:793
n ˆ 9; r2 ˆ 0:956; s ˆ 0:075; q2 ˆ 0:931
ꢀ8†
ꢀ9†
outlier : R ˆ CH2C6H5
log 1=C ˆ 0:64ꢀÆ0:30†ClogP 0:06ꢀÆ0:05†ClogP2
‡ 2:34ꢀÆ0:42†
In QSAR 7, ClogP once again represents the calculated
partition coecient of each analogue while ES repre-
sents Taft's steric parameter for each alkyl group of the
alkoxy moiety of each ester. The low coecient with the
ES term indicates that the cytotoxicity is not markedly
attenuated by bulky esters. The positive coecient with
the ES term indicates that cytotoxicity is enhanced by a
decrease in size of the substituents. However, the low
magnitude of this coecient and its relatively large
con®dence interval suggests that size plays a minor role
in inhibiting cell growth. In this analysis the benzylic
ester was a statistical outlier. A reason for its slightly
anomalous behavior is not apparent. The coecient
with the hydrophobic term in QSAR 7 is less than that
n ˆ 10; r2 ˆ 0:958; s ˆ 0:081; q2 ˆ 0:882;
ClogPo ˆ 5:75
ꢀ9†
The linear model explains 95% of the variance in the
data. Although the parabolic model (Eq. 9) is a mar-
ginal improvement it is able to establish an ideal
hydrophobicity for optimal inhibitory activity. The
optimum hydrophobicity 5.75 as delineated by ClogP is
high and suggests that hydrophobic ca€eic esters excel
at inhibiting cell growth in this breast cancer line. It is
dicult to ascertain whether this toxicity is due to
increased permeability or enhanced interaction with a
molecular target.
Table 2. Biological and physicochemical constants used to derive
QSAR 7 for the inhibition of growth of L1210 cells by dihydroca€eic
acid esters
Eq. 10 was obtained for the inhibition of growth of
MCF-7 cells by esters of dihydroca€eic acid. See Table 4.
No. Substituent Obsd log Calcd log Á log ClogP
1/C
ES
1/C
1/C
log 1=C ˆ 0:36ꢀÆ0:12†B1 ‡ 3:40ꢀÆ0:20†
n ˆ 9; r2 ˆ 0:885; s ˆ 0:050; q2 ˆ 0:367
ꢀ10†
2a
2b
2c
2d
2e
2f
2g
2h
2i
Me
C2H5
C4H9
C6H13
C8H17
CHMe2
CH2C6H5
(CH2)2C6H5
(CH2)3C6H5
CMe3
4.62
4.63
4.70
5.06
5.05
4.72
4.66
4.92
4.98
4.62
4.60
4.67
4.79
4.96
5.12
4.70
4.84
4.89
4.95
4.61
0.02
0.04
0.09
0.10
0.07
0.02
0.18
0.03
0.03
0.01
1.02
1.54
2.60
3.66
4.72
2.07
2.79
3.11
3.49
2.47
1.24
1.31
1.63
1.54
1.57
1.71
1.51
1.47
1.46
2.78
B1 represents Verloop's width parameter; its importance
in equation 10 suggests that shape is an important con-
straint in the interaction between these dihydroca€eic
acid esters and a potential receptor. One point was
poorly ®t by this equation: the n-octyl ester. Perhaps the
ease of ¯exibility of the octyl chain combined with its
ability to fold on itself is not adequately represented by
a
2j
aNot used to derive Eq. 7.
Table 4. Biological and physiocochemical constants used to derive
QSAR 10 for the inhibition of growth of MCF-7 cells by
dihydroca€eic acid esters
Table 3. Biological and physicochemical constants used to derive
QSAR 8 for the inhibition of growth of MCF-7 cells by ca€eic acid
esters
a
No.
Substituent
Obsd log
1/C
Calcd log
1/C
Á log
1/C
B1
No.
Substituent
Obsd log
1/C
Calcd log
1/C
Á log
1/C
ClogPa
2a
2b
2c
2d
2e
2f
2g
2h
2i
CH3
C2H5
CH2(CH2)2CH3
CH2(CH2)4CH3
CH2(CH2)6CH3
CH(CH3)2
CH2C6H5
CH2CH2C6H5
CH2(CH2)2C6H5
C(CH3)3
3.94
3.98
3.96
3.98
4.22
4.16
3.92
3.88
3.89
4.30
3.94
3.94
3.94
3.94
3.94
4.08
3.94
3.94
3.94
4.33
0.00
0.04
0.02
0.04
0.28
0.08
0.02
0.06
0.05
0.03
1.52
1.52
1.52
1.52
1.52
1.90
1.52
1.52
1.52
2.60
1a
1b
1c
1d
1e
1f
1g
1h
1i
CH3
C2H5
CH2(CH2)2CH3
CH2(CH2)4CH3
CH2(CH2)6CH3
CH(CH3)2
CH2C6H5
(CH2)2C6H5
(CH2)3C6H5
C(CH3)3
3.02
3.30
3.60
4.08
4.10
3.40
3.80
3.70
3.98
3.59
3.07
3.27
3.65
4.04
4.43
3.38
3.72
3.84
3.98
3.53
0.05
0.03
0.05
0.04
0.33
0.02
0.08
0.14
0.00
0.06
1.20
1.73
2.79
3.85
4.90
2.04
2.97
3.30
3.68
2.44
b
a
2j
1j
aVerloop's parameter.
bNot included in the derivation of QSAR 10.
aNot used in the derivation of QSAR 8.

202
B. Etzenhouser et al. / Bioorg. Med. Chem. 9 (2001) 199±209
Verloop's B1 descriptor. Hence the large discrepancy
between the observed and predicted value. Overall, the
lack of diversity in biological activity of these analogues
is striking. The average activity of all ten analogues falls
within the following range: 4.02Æ0.15. This suggests
that the binding receptor in the case of these esters must
have shape constraints that are non-restrictive.
to be derivatives of cinnamic acid which have been
implicated as antioxidants. A recent review shows that
there has been considerable interest in the radical
scavenging activity of cinnamic acids.22 The following
equation illustrates the mechanism of radical-mediated
oxidation of substituted cinnamic acids.23
Oxidation of X-C6H4CH=CHCOOH in 30% acetic
acid by quinolinium dichromate
The overall results versus MCF-7 cells closely parallel
those obtained versus L1210 cells. QSAR 8 demon-
strates a strong dependence on hydrophobicity akin to
QSAR 6. The coecients with the hydrophobic term are
very similar: 0.46 (L1210) versus 0.37 (MCF-7) for the
ca€eic acid esters. In the case of the dihydroca€eic acid
esters, steric e€ects once again come into play although
they are more pronounced in the case of the MCF-7
cells, which lack a hydrophobic component. Both
QSAR 7 and QSAR 10 contain steric terms such as ES
and B1 and their coecients vary slightly (0.14 for
L1210 and 0.35 for MCF-7) in magnitude but di€er in
e€ect. QSAR 7 also shows a marginal dependence on
hydrophobicity. Nevertheless, the emphasis on steric
parameterization in both cells is of great interest and
suggests the involvement of a receptor with speci®c
spatial constraints. QSAR 7 suggests that small sub-
stituents enhance cytotoxicity while QSAR 10 indicates
that an increase in size enhances toxicity. This apparent
di€erence could be attributed to di€erences in cell type
and/or variations in receptor steric requirements. Over-
all, it appears that rapidly dividing cells such as L1210
are more susceptible to the cytotoxic e€ects of both the
ca€eic acid and dihydroca€eic acid esters.
‡
log k2 ˆ 0:66ꢀÆ0:16†ꢀ ‡ 0:95ꢀÆ0:08†
n ˆ 7; r2 ˆ 0:959; s ˆ 0:084; q2 ˆ 0:897
ꢀ11†
QSAR 12 deals with inhibition of lipid peroxidation by
substituted cinnamic acids. There is a strong depen-
dence on the electron-releasing capabilities of the sub-
stituents as delineated by the Hammett s+ term.
ED50 for reducing accumulation of malonaldehyde in
mice by X-cinnamic acids24
‡
log 1=C ˆ 0:64ꢀÆ0:25†ꢀ ‡ 0:41ꢀÆ0:25†B53 ‡ 2:41ꢀÆ0:48†
n ˆ 10; r2 ˆ 0:867; s ˆ 0:219; q2 ˆ 0:798;
ꢀ12†
outlier : 3; 4-di-OMe
The 3-OCH2COOH and 4-OCH2COOH derivatives
were omitted from the analysis for lack of adequate
sigma-plus values. Once again, a pronounced depen-
dence on s+ is observed.
The striking similarity in dependence on s+ for the
radical oxidation by dichromate in Eq. 11, and the pre-
vention of the oxidation of polyunsaturated fatty acids
(PUFAs) to malonaldehyde (QSAR 12) indicates that
both are radical mediated processes in which the p elec-
trons of the double bond are involved. Thus, ca€eic
acids have two possible reaction centers-the phenolic
OH or p electrons of the side chain. The results in
QSAR 6 tend to discount the involvement of the cate-
cholic OH group, because of the unusual strong depen-
dence on ClogP and the subsequent marked deviation of
QSAR 6 from QSARs 3 and 4.
Discussion
Eqs. 1, 3, and 4 clearly establish the importance of
electronic e€ects in the toxicity of phenols to fast grow-
ing leukemia cells. Since catechol is well ®t by these
equations, one can presume that ca€eic and dihydro-
ca€eic esters would act via the same mechanism. The
alkoxy groups of these esters will have essentially little
or no e€ect on the electron density of the catechol ring.
Thus, we are left to consider two of the three main
structural features that generally a€ect biological
response: hydrophobicity and steric bulk.
One can also look for similarity with styrenes and their
associated biological activities. In QSAR 13, C is the
concentration that induces an elevation of serum ala-
nine transaminase, a measure of hepatotoxicity.25 Thus
the more electron releasing the substituents on the styr-
enes, the greater the resulting hepatoxicity.
In examining QSAR 6, we ®nd that the data are well
correlated by hydrophobicity alone, while QSAR 7
yields a much di€erent model. The most important term
in QSAR 7 is hydrophobicity as represented by ClogP.
Using the hydrophobic parameter alone, r2=0.800 with
slope of 0.15, while using ES alone, r2=0.067. Thus
80% of the variance in the data in QSAR 7 can be
explained by hydrophobicity alone while 10% of the
variance in the data can be attributed to steric e€ects.
Except for the small contribution by ES the coecient
with ClogP in QSAR 7 is similar to that in QSAR 3 and
4.
‡
log 1=C ˆ 0:46ꢀÆ0:26†s ‡ 3:22ꢀÆ0:18†
n ˆ 6; r2 ˆ 0:862; s ˆ 0:118; q2 ˆ 0:738
ꢀ13†
outlier : X ˆ H
However, QSAR 6 suggests an entirely di€erent
mechanism. Unfortunately there is no electronic refer-
ence point, but ca€eic acid analogues can be considered
It is also conceivable that both reaction centers may be
involved in cytotoxicity. A comparison of growth inhi-
bitory values (cytotoxic potencies) in Tables 1 and 2,

B. Etzenhouser et al. / Bioorg. Med. Chem. 9 (2001) 199±209
203
reveal that the compounds with relatively low hydro-
phobicities have parallel activity in the two series. With
increased hydrophobicity, the ca€eic esters are more
potent, which suggests that oxidation of the ole®nic
linkage may be enhanced in a more lipophilic milieu.
See Figure 2.
human glioblastoma cells GBM-18. From their data, it
can be ascertained that the log 1/C of CAPE in the
melanoma line would fall between 4.45 to 5.05, while
the value in the astrocytoma line would be closer to
4.45. They deduced that HO-1 cells were more suscep-
tible to CAPE induced growth suppression than GBM-
18 cells. Our results for CAPE versus a rapidly growing
cell line versus slow growing breast cancer cells are in
agreement. The log 1/C values for L1210 and MCF-7
are 5.20 and 3.70, respectively. This suggests that CAPE
and other ca€eic acid esters should show promise versus
aggressively growing tumor cells such as melanomas.1
Extensive studies by Fukaya et al.26,27 on various ca€eic
acid derivatives have revealed that they are excellent
inhibitors of arachidonate 5-lipoxygenase (5-LO), 12-
lipoxygenase (12-LO) and prostaglandin (PG) synthase.
The n-butyl ester showed potent inhibitory activity
towards 5-LO as well as selectivity versus this enzyme.
Various ca€eic acid derivatives such as ca€eic acid
amides, 1-(3,4-dihydroxyphenyl)-1-alken-3-ones and 1-
(3,4-dihydroxylphenyl)-1-alkenes were also investigated
as 5-LO inhibitors. It was determined that the catecholic
moiety and a hydrophobic alkyl side chain were critical
for strong binding to 5-LO. Rigorous QSAR studies of
an extensive set of catechol derivatives by Fujita et al.27
delineated the structural requirements for maximal
binding to 5-LO. These include an optimal hydro-
phobicity of 4.3±4.7, high electron density of the cate-
chol moiety and a lack of thickness of the lipophilic ab
side chain close to the catechol benzene ring. There are
some similarities between our QSAR for the ca€eic acid
esters (QSAR 7 and 9) and the ones elaborated by
Fujita et al. There is a strong possibility that the inhibi-
tion of cell growth by the ca€eic acid esters results from
an interrupted synthesis of leukotrienes mediated by
inhibition of 5-LO.
Conclusion
A comparison of the cytotoxicities of ca€eic and
dihydroca€eic acid esters versus L1210 cells reveals that
there are subtle di€erences in potency between similar
analogues. The ole®nic linkage in ca€eic esters appears
to play a critical role in overall toxicity. Its susceptibility
to peroxidation may be the underlying reason for its
toxicity. Studies with similar analogues suggests that
these esters may be inhibitors of 5-LO. N-octylca€eate
is a more potent inhibitor of cell growth versus L1210
and MCF-7 cells than CAPE. In the dihydroca€eic acid
ester series, n-octyldihydroca€eate maintains its cyto-
toxic superiority over the corresponding hydrogenated
CAPE analogue.
Experimental
General
In undertaking this research, we hoped to learn about
the mechanism of action of the ca€eic esters as well as
to ®nd a compound more potent than CAPE. Com-
pound 1e, the n-octyl-ca€eate, is almost ten times as
potent as CAPE versus L1210 cells in culture and three
times more active than CAPE in MCF-7 cells.
All reagents and solvents were used as received, unless
otherwise indicated. Ca€eic acid (3,4-dihydroxy-
cinnamic acid, 97%), dihydroca€eic acid (3-(3,4-dihy-
droxyphenyl)-propanoic acid, 98%), and tert-Butyl
bromoacetate (98%) were purchased from Aldrich
Chemical Co. (Milwaukee, WI). Phenethyl ca€eate
Guarini et al.,28 examined the cytotoxic potential of
CAPE versus a human melanoma cell line HO-1 and
Figure 2. Cytotoxicity versus hydrophobicity of ca€eic and dihydroca€eic esters.

204
B. Etzenhouser et al. / Bioorg. Med. Chem. 9 (2001) 199±209
(Phenethyl-3,4-dihydroxycinnamate, CAPE) used for
cell culture testing was purchased from LKT Labora-
tories, Inc. (St. Paul, MN). tert-Butoxycarbonylmethyl
triphenylphosphonium chloride (98+%) was purchased
from Lancaster Chemical Company. Diethyl ether and
hexanes (or petroleum ether) used for recrystallization
were dried over CaH2 and distilled before use. Column
chromatography was carried out using 70±230 mesh
silica gel (Aldrich). Silica gel IB TLC plates (Baker,
precoated, plastic-backed) were used for thin layer
chromatography. Melting points were determined on a
Mel-Temp melting point apparatus with digital ther-
mometer and are uncorrected. Elemental analyses (C,
H) were conducted by Desert Analytics (Tucson, AZ).
The acceptable range was Æ0.4% of calculated values.
Mass spectral analyses were performed by the UCR
Mass Spectrometry Facility (University of California,
Riverside). Proton (1H) and 13C FT-NMR spectra were
obtained using either a General Electric QE-300
(300 MHz) or a Bruker Spectrospin 400 MHz spectro-
meter. NMR samples were prepared using deuterated
acetone (acetone-d6) or deuterated chloroform (CDCl3)
purchased from Aldrich Chemical Co. Either the pro-
tonated solvent impurity or TMS was used as the spec-
tral reference peak. FT-IR spectra were obtained using
a Perkin Elmer 1600 Series spectrophotometer. FT-IR
samples were prepared as pressed KBr wafers or as thin
®lms on single or between two NaCl plates.
petroleum ether) was also carried out and repeated until
the desired degree of purity was achieved. Pure (based
on melting point, TLC, and proton NMR data) methyl
ester was dried in an Aberhalden drying pistol (re¯uxing
methanol, about 6 h) and submitted for elemental ana-
lysis and cell culture testing. Yield (pure) data re¯ects
the amount of recovered product deemed pure enough
for analytical testing. The ®nal product was recovered
as a white to very pale yellow, crystalline solid. Yield
(pure): 0.26 g (14%). Melting point 161±163 ꢀC. 1H
NMR (acetone-d6) d 7.52 (d, J=16 Hz, 1H), 7.15 (d,
J=2 Hz, 1H), 7.04, 7.01 (dd, J=2.0, 2.1, 8.2 Hz, 1H),
6.85 (d, J=8.2 Hz, 1H), 6.27 (d, J=15.9 Hz, 1H), 3.70
(s, 3H); 13C NMR (acetone-d6) d 167.9, 148.6, 146.1,
145.7, 127.5, 122.5, 116.3, 115.2, 115.1, 115.1, 51.5. FT-
IR (thin ®lm, single NaCl plate) cm 1: 3495, 3312, 1684,
1635, 1602. Anal. calcd (C10H10O4): 61.86 %C,
5.20 %H; Found: 61.77 %C, 5.22 %H.
Synthesis of ethyl-3,4-dihydroxycinnamate (ethyl ca€e-
ate, 1b). Similar to the procedure for 1a, except ethyl
alcohol was used. The ®nal product was recovered as a
light-tan solid. Percent yield is based on 3.6 g
(19.9 mmol) of ca€eic acid starting material. Yield
(pure): 0.56 g (13%). Melting point: 149±151 ꢀC. 1H
NMR (acetone-d6) d 7.52 (d, J=15.9 Hz, 1H), 7.15 (d,
J=1.8 Hz, 1H), 7.04, 7.01 (dd, J=1.8, 1.9 Hz; 8.1 Hz,
1H), 6.85 (d, J=8.2 Hz, 1H), 6.26 (d, J=15.9 Hz, 1H),
4.17 (quartet, 2H), 1.26 (t, 3H); 13C NMR (acetone-d6) d
167.5, 148.5, 146.1, 145.5, 127.5, 122.4, 116.3, 115.6,
115.1, 60.5, 14.6. FT-IR (KBr) cm 1: 3442, 3249, 1670,
1598. Anal. calcd (C11H12O4): 63.46 %C, 5.82 %H;
Found: 63.32 %C, 5.99 %H.
Method 1: Fisher esteri®cation of ca€eic acid29
Synthesis of methyl-3,4-dihydroxycinnamate (methyl
ca€eate, 1a). Ca€eic acid (1.8 g, 9.9 mmol), p-toluene-
sulfonic acid (PTSA, 0.10 g, 0.5 mmol), methyl alcohol
(50 mL), and benzene (50 mL), were placed in a 250 mL
3-neck round bottom ¯ask equipped with a magnetic
stirring bar and a Dean±Stark trap. The mixture was
stirred, heated to re¯ux, and monitored by thin layer
chromatography (diethyl ether eluent; versus ca€eic
acid). As heating continued, the Dean±Stark trap was
periodically emptied. The level of solvent in the reaction
¯ask decreased gradually, so more methyl alcohol and
benzene were added to maintain an azeotropic mixture.
More PTSA was added (approx. 0.1 to 0.2 g) to the
reaction mixture at 8±16 h intervals. The reaction was
considered complete by TLC after about 48 h of re¯ux-
ing. The reaction mixture was poured into a separatory
funnel containing approximately 200 mL of distilled
water. Diethyl ether (50±75 mL) was added and the
mixture was shaken to extract the organic components.
The organic layer was washed with distilled water (2±
3Â, 50±75 mL), 5% NaHCO3 (2±3Â, 50±75 mL), and
saturated aqueous NaCl (1±2Â,50±75 mL). The result-
ing yellow ether solution was dried over anhydrous
MgSO4, ®ltered, and concentrated on a rotary eva-
porator to leave a light to dark yellow solid residue
(0.9 g, 47% crude yield). The crude solid was further
puri®ed by treating a diethyl ether solution of the solid
with decolorizing carbon and ®ltering the recovered
solution through a silica gel column (2Â, diethyl ether
eluent). For some uses, puri®cation to this extent was
sucient. For cell culture and elemental analysis sam-
ples, recrystallization from diethyl ether and hexanes (or
Synthesis of n-butyl-3,4-dihydroxycinnamate (n-butyl
ca€eate, 1c). Similar to the procedure for 1a, except n-
butyl alcohol was used. Excess n-butyl alcohol and
benzene were removed from the crude reaction mixture
by steam distillation (insitu, atmospheric pressure).
Subsequent work up and puri®cation methods were
similar to those for 1a. The ®nal product was recovered
as a light-tan, crystalline solid. Percent yield is based on
3.6 g (19.9 mmol) of ca€eic acid starting material. Yield
(pure): 1.1 g (23%). Melting point 110±111 ꢀC. 1H NMR
(acetone-d6=2.04 ppm) d 8.33 (broad singlet, 2H), 7.53
(d, J=15.9 Hz, 1H), 7.16 (s, 1H), 7.03 (d, 8.1 Hz, 1H),
6.86 (d, J=8.1 Hz, 1H), 6.28 (d, J=15.9 Hz, 1H), 4.13
(t, 2H), 1.63 (quintet, 2H), 1.40 (sextet, 2H), 0.92 (t,
3H); 13C NMR (acetone-d6=29.8 ppm) d 167.5, 148.6,
146.2, 145.5, 127.5, 122.4, 116.3, 115.6, 115.1, 64.3, 31.5,
1
19.8, 14.0. FT-IR (thin ®lm, single NaCl plate) cm
3486, 3342, 2952, 2868, 1683, 1638, 1602, 1536. Anal.
calcd (C13H16O4): 66.09 %C, 6.84 %H; Found:
66.07 %C, 6.90 %H.
Synthesis of n-hexyl-3,4-dihydroxycinnamate (n-hexyl
ca€eate, 1d). Similar to the procedure for 1c, except n-
hexyl alcohol (30 mL/3.6 g of ca€eic acid) was used. The
®nal product was recovered as a white to very pale-
yellow, ®nely crystalline solid. Percent yield is based on
3.6 g (19.9 mmol) of ca€eic acid starting mꢀaterial. Yield
(pure): 0.33 g (6%). Melting point 128±129 C. 1H NMR
(acetone-d6) d 8.33 (broad singlet, 2H), 7.53 (d, J=15.9,

B. Etzenhouser et al. / Bioorg. Med. Chem. 9 (2001) 199±209
205
1H), 7.15 (d, J=2.0 Hz, 1H), 7.04, 7.02 (dd, J=2.1,
8.2 Hz, 1H), 6.86 (d, J=8.2 Hz, 1H), 6.27 (d,
J=15.9 Hz, 1H), 4.13 (t, J=6.7 Hz, 2H), 1.66 (quintet,
2H), 1.34 (mult., 6H), 0.88 (t, J=6.8, 6.7; 3H); 13C
NMR (acetone-d6=29.8 ppm) d 67.5, 148.6, 146.2,
145.5, 127.5, 122.4, 116.3, 115.6, 115.1, 64.6, 32.1, 29.4,
26.3, 23.1, 14.2. FT-IR (thin ®lm, single NaCl plate)
cm 3491, 3335, 2854, 1686, 1639, 1602, 1535. Anal.
calcd (C15H20O4): 68.16 %C, 7.64%H; Found: 68.27
%C, 7.69%H.
roleum ether; second round: 4:1 CHCl3/ethyl acetate),
and Kugelrohr distillation (80±85 ꢀC, 2 h) to remove
any volatile impurities. A pale, champagne colored oil
was recovered. Percent yield is based the sum of the
theoretical yields of the combined reaction products
(11 mmol). Yield (pure): 0.88 g (38%). 1H NMR
(CDCl3=7.26 ppm) d 6.75 (d, J=8.1 Hz, 1H), 6.69 (d,
J=1.5 Hz, 1H), 6.59, 6.57 (dd, J=1.4, 8.1 Hz, 1H), 6.22
(broad singlet, 2H), 4.13 (quartet, 2H), 2.81 (t, J=7.6,
7.7 Hz; 2H), 2.58 (t, J=7.7, 7.6 Hz; 2H), 1.23 (t, 3H);
13C NMR (CDCl3=77.0 ppm) d 174.3, 143.7, 142.1,
132.8, 120.2, 115.4 (2 overlapped peaks), 60.9, 36.0,
30.1, 13.9. FT-IR (neat, thin ®lm between two NaCl
1
Synthesis of isopropyl-3,4-dihydroxycinnamate (isopro-
pyl ca€eate, 1f). Similar to the procedure for 1a, except
isopropyl alcohol was used. The reaction appeared
complete by TLC after about six days of re¯uxing. work
up and puri®cation methods were similar to those for
1a. The ®nal product was recovered as an o€-white,
crystalline solid. Percent yield is based on 1.8 g
(9.9 mmol) of ca€eic acid starting material. Yield (pure)
0.50 g (23%). Melting point 145±147 ꢀC. 1H NMR
1
plates) cm 3368, 2981, 1706, 1606, 1519. Exact Mass
Data (Electron Impact, 20 eV): For (C11H14O4): calcd
M+: 210.089209; Found M+: 210.089015.
Synthesis of 3-(3,4-dihydroxyphenyl)propanoic acid, n-
Butyl ester (n-Butyl dihydroca€eate, 2c). Similar to the
procedure for 1c, except dihydroca€eic acid (1 g,
5.5 mmol) was used and 0.3±0.5 g of PTSA were used
initially, with no more added later. Oily products from
two separate reactions (1 g, 5.5 mmol starting material
each) were recovered and combined for puri®cation.
Puri®cation to generate analytical samples consisted of
two rounds of column chromatography (®rst round:
100% CHCl3, followed by 2:1 diethyl ether/petroleum
ether; second round: 4:1 CHCl3/ethyl acetate). Kugel-
rohr distillation (80±85 ꢀC, 2 h) to remove volatile
impurities followed. A pale-orange oil was recovered.
Yield (pure): 0.45 grams (17%). 1H NMR
(CDCl3=7.26 ppm) d 6.75 (d, J=8.1 Hz, 1H), 6.69 (d,
J=1.9 Hz, 1H), 6.60, 6.58 (dd, J=1.9, 8.1 Hz, 1H), 6.15
(s, 1H), 5.99 (s, 1H), 4.07 (t, 2H), 2.82 (t, J=7.5, 7.8 Hz;
2H), 2.59 (t, J=7.8, 7.5 Hz, 2H), 1.58 (quintet, 2H), 1.33
(sextet, 2H), 0.91 (t, 3H); 13C NMR (CDCl3=77.0 ppm)
d 174.4, 143.6, 142.1, 132.7, 120.1, 115.4 (two over-
lapping peaks), 64.7, 36.0, 30.3, 30.0, 18.8, 13.4. FT-IR
(acetone-d6)
d 8.4 (broad singlet, 2H), 7.51 (d,
J=15.9 Hz, 1H), 7.15 (d, J=1.8 Hz, 1H), 7.03, 7.01 (dd,
J=1.7, 8.1 Hz, 1H), 6.85 (d, J=8.1 Hz, 1H), 6.23 (d,
J=15.9 Hz, 1H), 5.04 (septet, 1H), 1.24 (d, 6H); 13C
NMR (acetone-d6) d 167.0, 148.5, 146.1, 145.3, 127.5,
1
122.4, 116.2, 116.1, 115.0, 67.7, 22.1. FT-IR (KBr) cm
3466, 1679, 1633, 1601, 1534. Anal. calcd (C12H14O4):
64.85 %C, 6.36%H; Found: 64.81 %C, 6.28 %H.
Synthesis of 3-(3,4-dihydroxyphenyl)propanoic acid,
methyl ester (methyl dihydroca€eate, 2a). Similar to the
procedure for 1a, except dihydroca€eic acid (1 g,
5.5 mmol) and 25 mL each of methyl alcohol and ben-
zene were used. During work up, the aqueous layer was
extracted two additional times with 50 mL of diethyl
ether. Ether extracts were combined, and work up con-
tinued essentially as described for 1a. After column
chromatography and recrystallization, pure 2a was
recovered as a white to very-pale yellow, crystalline
solid. Yield (pure): 0.28 g (26%). Melting point 73±
1
(neat, thin ®lm between two NaCl plates) cm 3390,
2960, 2935, 1707, 1607, 1519. Anal. calcd (C13H18O4):
65.53 %C, 7.63 %H; Found: 65.15 %C, 7.71 %H. Exact
Mass Data (Electron Impact, 20 eV): For (C13H18O4):
calcd M+: 238.120509; Found M+: 238.120999.
75 ꢀC. 1H NMR (CDCl3=7.26 ppm)
d 6.77 (d,
J=8.1 Hz, 1H), 6.71 (d, J=2.0 Hz, 1H), 6.62, 6.60 (dd,
J=2.0, 8.1 Hz, 1H), 5.73 broad singlet, 2H), 3.67 (s,
3H), 2.83 (t, J=7.5, 7.8 Hz, 2H), 2.59 (t, J=7.8, 7.5 Hz,
2H); 13C NMR (acetone-d6=29.8 ppm) d 173.6, 145.6,
144.0, 133.2, 120.2, 116.1, 115.9, 51.5, 36.4, 30.8. FT-IR
Synthesis of 3-(3,4-dihydroxyphenyl)propanoic acid, n-
octyl ester (n-octyl dihydroca€eate, 2e). Similar to the
procedure for 2c except n-octyl alcohol (20 mL) was
used. Excess n-octyl alcohol and benzene were removed
by azeotropic steam distillation. Crystalline product
recovered after workup was recrystallized one time
(diethyl ether/hexanes) to yield product pure enough for
analytical samples. Compound 2e was recovered as a
champagne colored, crystalline solid. Percent yield is
based on 5.5 mmol of dihydroca€eic acid starting mate-
rial. Yield (pure): 0.89 g (55%). Melting point 56±58 ꢀC.
1H NMR (CDCl3=7.26 ppm) d 6.76 (d, J=8.1 Hz, 1H),
6.71 (d, J=2.0 Hz, 1H), 6.63, 6.61 (dd, J=2.0, 8.1 Hz,
1H), 5.53 (s, 1H, OH), 5.37 (s, 1H, OH), 4.06 (t, 2H),
2.83 (t, J=7.5, 7.8 Hz, 2H), 2.58 (t, J=7.9, 7.5 Hz, 2H),
1.59 (quintet, 2H), 1.32±1.23 (broad mult., 10H), 0.88 (t,
3H); 13C NMR (CDCl3=77.0 ppm) d 174.5, 143.7,
142.2, 132.9, 120.3, 115.3 (two overlapping peaks), 65.2,
36.2, 31.7, 30.2, 29.1, 29.09, 28.4, 25.8, 22.6, 14.0. FT-IR
1
(KBr) cm 3485, 3312, 2947, 1712, 1608, 1517. Anal.
calcd (C10H12O4): 61.22 %C, 6.18%H; Found: 61.50
%C, 6.13 %H.
Synthesis of 3-(3,4-dihydroxyphenyl)propanoic acid,
ethyl ester (ethyl dihydroca€eate, 2b). Similar to the
procedure for 1a, except dihydroca€eic acid (1 g,
5.5 mmol), 50 mL of ethyl alcohol, and 25 mL of ben-
zene were used for the Fisher esteri®cation reaction.
Workup was similar to that described for 2a. An oily
product was recovered. The products of two separate
reactions (5.5 mmol of dihydroca€eic acid starting
material each) were puri®ed in combination. Puri®ca-
tion to obtain analytical samples involved decoloriza-
tion of a diethyl ether solution of the combined reaction
products, two rounds of column chromatography (®rst
round: 100% CHCl3 followed by 2:1 diethyl ether/pet-

206
B. Etzenhouser et al. / Bioorg. Med. Chem. 9 (2001) 199±209
1
(KBr) cm 3509, 3390, 2926, 2849, 1710, 1608, 1537.
Anal. calcd (C17H26O4): 69.35 %C, 8.92%H; Found:
69.35 %C, 9.16 %H.
1.68 (quintet, 2H), 1.34 (mult., 10H), 0.89 (t, 3H); 13C
NMR (acetone-d6=29.8 ppm) d 167.4, 148.6, 146.2,
145.5, 127.8, 127.5, 122.4, 116.2, 115.6, 115.0, 64.6, 32.5,
1
29.92, 29.89, 29.5, 26.6, 23.2, 14.3. FT-IR (KBr) cm
Synthesis of 3-(3,4-dihydroxyphenyl)propanoic acid, iso-
propyl ester (isopropyl dihydroca€eate, 2f). Procedure
was similar to that described for 2a, except that isopropyl
alcohol was used. The products of two separate reac-
tions (5.5 mmol starting material each) were combined
for puri®cation by column chromatography (2:1 diethyl
ether/petroleum ether). The recovered oil was subjected
to Kugelrohr distillation (80±85 ꢀC, 2 h) to remove
volatile impurities. The ®nal product was recovered as a
pale brown oil. Percent yield is based on the theoretical
3489, 3316, 2919, 2854, 1684, 1604, 1533. Anal. calcd
(C17H24O4): 69.83 %C, 8.29%H; Found 69.78 %C, 8.16
%H.
Synthesis of benzyl-3,4-dihydroxycinnamate (benzyl
ca€eate, 1g). The procedure is similar to that described
for 1e, except that benzyl bromide was used. Puri®ca-
tion involved an initial decolorization (carbon) and
round of chromatography (100% diethyl ether) fol-
lowed by three separate recrystallizations (diethyl ether/
hexanes). The recrystallization product was chromato-
graphed again (2:1 diethyl ether/petroleum ether) and
recrystallized a ®nal time (diethyl ether/petroleum
ether). Pure 1g was recovered as a ¯u€y, white to pale-
yellow solid. Percent yield is based on 5.5 mmol of caf-
feic acid starting material. Yield (pure): 0.05 g (3%).
Melting point 151±154 ꢀC. 1H NMR (acetone-d6,
TMS=0.00 ppm) d 8.49 (broad s, 2H), 7.59 (d, 1H,
J=15.9 Hz), 7.47±7.30 (m, 5H), 7.18 (d, J=2.0 Hz,
1H,), 7.07, 7.05 (dd, J=2.1, 8.2 Hz, 1H), 6.87 (d,
J=8.2 Hz, 1H), 6.34 (d, J=16.0 Hz, 1H), 5.22 (s, 2H);
13C NMR (acetone-d6=29.8 ppm) d 167.3, 148.8, 146.1,
146.05, 137.7, 129.2 (two equivalent carbon atoms),
128.9 (two equivalent carbon atoms), 128.7, 127.5,
yields
of
the
combined
reaction
products
(2Â5.5 mmol=11 mmol theoretical yield). Yield (pure):
1
1.2 g (48%). H NMR (CDCl3=7.26 ppm) d 6.75 (d,
J=8.1 Hz, 1H), 6.69 (d, J=1.6 Hz, 1H), 6.60, 6.58 (dd,
J=1.7, 8.1 Hz, 1H), 6.23 (s, 1H, OH), 6.08 (s, 1H, OH),
5.00 (septet, J=6.2 Hz, 1H), 2.81 (t, J=7.5, 7.7 Hz, 2H),
2.56 (t, J=7.7, 7.5 Hz, 2H), 1.21 (d, J=6.2 Hz, 6H); 13C
NMR (CDCl3=77.0 ppm) d 173.9, 143.7, 142.2, 132.7,
120.2, 115.4, 115.3, 68.4, 36.4, 30.2, 21.6. FT-IR (neat,
1
thin ®lm between two NaCl plates) cm 3380, 2981,
2935, 1703, 1606, 1520. Exact Mass Data (Electron
Impact, 20 eV): M+ for (C12H16O4): calcd 224.104859;
Found: 224.104339.
1
122.6, 116.3, 115.3, 115.1, 66.2. FT-IR (KBr) cm
3465, 1688, 1636, 1601, 1533. Anal. calcd (C16H14O4):
71.10 %C, 5.23 %H; Found 71.23 %C, 5.10 %H.
Method 2: Alkylation of the ca€eic acid carboxylate
salt30
Synthesis of n-octyl-3,4-dihydroxycinnamate (n-octyl
ca€eate, 1e). Ca€eic acid (1.0 g, 5.5 mmol) was dis-
solved in 40±50 mL of N,N-Dimethylformamide
(DMF). A dark-brown solution resulted. 1.14 mL of
aqueous NaOH (2.8 g NaOH in 10 mL of H2O) was
added slowly. The mixture was stirred for 2.5 h after
which, a precipitate was observed. A solution of 1-bro-
mooctane (3.4 mL, 24 mmol) in DMF was added drop-
wise. The mixture was allowed to stir for 7 days. Work
up consisted of pouring the reaction mixture into ice
water (1200 mL, distilled), extracting the aqueous mix-
ture with diethyl ether (until extract had very little
color), and washing the ether extract successively with
1 N HCl (3Â150 mL), H2O (3Â150 mL), and brine
(2Â75 mL). The extract was dried over anhydrous
MgSO4, ®ltered, and concentrated on a rotary eva-
porator. When oily residues were obtained after con-
centration, trituration with petroleum ether or hexanes
yielded a solid product. Puri®cation proceeded similarly
to that described for compound 1a, except additional
column chromatography (2:1 diethyl ether/hexanes) and
recrystallization (diethyl ether/hexanes) steps were
necessary. Pure n-octyl ca€eate was recovered as a can-
ary yellow, crystalline solid. Percent yield is based on
the sum of the theoretical yields of two individual reac-
tions (5.5 mmol+11 mmol=16.5 mmol total theoretical
yield) that were combined for late stage puri®cation.
Synthesis of phenethyl-3,4-dihydroxycinnamate (phen-
ethyl ca€eate, CAPE, 1h). This compound was syn-
thesized to serve as the hydrogenation substrate for
formation of the saturated side chain analogue. The
procedure was similar to that for 1e, except 2.0 g
(11 mmol) of ca€eic acid and 150 mL of DMF (reagent
grade) were used. 2.3 mL of aq NaOH (2.7 g NaOH in
10 mL H2O) was added. This mixture was allowed to
stir for 2.5 h. A solution of 5.6 mL (41 mmol) 1-bromo-
2-phenylethane in 10 mL of DMF was added dropwise
to the stirring mixture. Stirring was continued for 7
days. Workup was essentially the same as done for 1e
except the reaction mixture was initially poured into
1600 mL of distilled water. Puri®cation consisted of
treating a diethyl ether solution of the crude compound
with decolorizing carbon, followed by ®ltration through
a short, small diameter silica gel column (Et2O eluant).
Compound 1h was recovered as a tan to o€-white solid
after initial puri®cation and judged pure enough to use
for hydrogenation reactions. Yield: 1.45 g (46%), melting
point 116±123 ꢀC. 1H NMR (acetone-d6, TMS=
0.00 ppm) d 8.39 (broad s, 2H), 7.56 (d, J=15.9 Hz,
1H), 7.30 (d, 3H), 7.22 (t, 2H), 7.18 (d, J=1.9 Hz, 1H),
7.05, 7.03 (dd J=1.9, 8.2 Hz, 1H), 6.89 (d, J=8.2 Hz,
1H), 6.29 (d, J=15.9 Hz, 1H), 4.36 (t, J=7.0 Hz, 2H),
2.99 (t, J=7.0 Hz, 2H); 13C NMR (acetone-
d6=29.8 ppm) d 167.3, 148.6, 146.1, 145.7, 139.1, 129.7
(two equivalent carbon atoms), 129.1 (two equivalent
carbon atoms), 127.5, 127.1, 122.5, 116.3, 115.4, 115.1,
ꢀ
1
Yield (pure): 0.32 g (7%), melting point 111±112 C. H
NMR (acetone-d6, TMS=0.00 ppm) 8.39, (broad sing-
let, 2H), 7.54 (d, J=15.9 Hz, 1H), 7.17 (d, J=2.0 Hz,
1H), 7.06, 7.04 (dd, J=2.0 Hz, 8.1 Hz, 1H), 6.87 (d,
J=8.2 Hz, 1H), 6.29 (d, J=15.9 Hz, 1H), 4.14 (t, 2H),
1
65.3, 35.7. FT-IR (KBr) cm 3480, 3327, 1684, 1636,
1602.

B. Etzenhouser et al. / Bioorg. Med. Chem. 9 (2001) 199±209
207
Synthesis of 3-phenylpropyl-3,4-dihydroxycinnamate (3-
phenylpropyl ca€eate, 1i). Procedure was similar to that
for 1e except that 4.0 g (22 mmol) of ca€eic acid were
used. Other reagents were scaled up accordingly. Only a
portion of the initially puri®ed (decolorization (carbon),
column chromatography (100% diethyl ether)) reaction
product was carried through a ®nal puri®cation process
(column chromatography (1:1 diethyl ether/hexanes),
second decolorization, recrystallization (diethyl ether/
hexanes). The ®nal product was recovered as a white to
very pale yellow, ®nely crystalline solid. Yield (pure):
0.63 g (10%, based on 22 mmol of ca€eic acid starting
material). More was available in a less pure form.
Melting point 123±124 ꢀC. 1H NMR (acetone-d6,
TMS=0.00 ppm) d 8.39 (broad s, 2H), 7.55 (d,
J=15.9 Hz, 1H) 7.31±7.19 (m, 5H) 7.17 (d, J=2.1 Hz,
1H), 7.07, 7.05 (dd, J=2.1, 8.2 Hz, 1H), 6.88 (d,
J=8.1 Hz, 1H), 6.31 (d, J=16.0 Hz, 1H), 4.16 (t, 2H),
2.75 (t, 2H), 2.01 (quintet, 2H); 13C NMR (acetone-
d6=29.8 ppm) d 167.4, 148.6, 146.1, 145.6, 142.3, 129.1,
127.5, 126.6, 122.5, 116.3, 115.6, 115.1, 64.0, 32.7, 31.1.
NaCl plate) cm 1: 3394, 3033, 2953, 1710, 1606, 1519.
Anal. calcd (C16H16O4) 70.57 %C, 5.94 %H; Found:
70.71 %C, 5.86%H. Exact Mass Data (Electron
Impact, 20 eV): M+ for (C16H16O4): Calcd: 272.104859;
Found: 272.104524.
Method 3: Hydrogenation of the unsaturated analogue29
Synthesis of 3-(3,4-dihydroxyphenyl)propanoic acid, n-
Hexyl ester (n-hexyl dihydroca€eate, 2d). One gram
(3.8 mmol) of n-hexyl ca€eate (1d) was combined with
0.5 g of 10% Pd/C and absolute ethanol (approx.
125 mL). Using a Parr hydrogenator, the mixture was
placed under a hydrogen atmosphere (40 psi) and shaken
for about 3 h. The mixture was ®ltered through Celite,
concentrated, and analyzed by proton NMR, which
revealed that conversion was complete. The residue was
taken up in acetone and ®ltered with air pressure
through a short silica gel column to remove traces of
catalyst. Upon rotary evaporation, the product solidi®ed
to become a soft green-grey residue. The solid was dis-
solved and the solution treated with decolorizing car-
bon. Concentration, followed by trituration with
hexanes yielded a yellow solid. An initial recrystalliza-
tion (diethyl ether/hexanes) was followed by an additional
decolorization step (carbon) and column chromatography
(1:1 diethyl ether/hexanes). A ®nal recrystallization
(diethyl ether/hexanes) resulted in product pure enough
for analytical samples. n-Hexyl dihydroca€eate (2d) was
recovered as a white crystalline solid. Yield (pure) 0.23 g
(23%). Melting point 67±68 ꢀC. 1H NMR (CDCl3=
7.26 ppm) d 6.76 (d, J=8.1 Hz, 1H), 6.70 (d, J=1.9 Hz,
1H), 6.62, 6.60 (dd, J=1.9, 8.1 Hz, 1H), 4.06 (t, 2H),
2.83 (t, J=7.5, 7.8 Hz, 2H), 2.58 (t, J=7.9, 7.5 Hz, 2H),
1.59 (quartet, 2H), 1.34±1.26 (multiplet, 6H), 0.88 (t,
2H); 13C NMR (CDCl3=77.0 ppm) d 174.4, 143.7,
142.2, 132.9, 120.4, 115.3 (two overlapping peaks), 65.2,
1
FT-IR (KBr) cm 3488, 1684, 1637, 1602. Anal. calcd
(C18H18O4): 72.47 %C, 6.09 %H; Found 72.48 %C,
6.06 %H.
Synthesis of 3-(3,4-dihydroxyphenyl)propanoic acid, ben-
zyl ester (benzyl dihydroca€eate, 2g). Similar to the pro-
cedure for 1e, except that 0.8 g (4.4 mmol) of
dihydroca€eic acid, 0.91 mL of aq NaOH (approx. 22%
w/w) and benzyl bromide (1.05 mL, 8.8 mmol) were
used. A smaller excess of benzyl bromide was used to
facilitate easier removal of unreacted reagent, while still
driving the reaction forward. After about 2.5 days of
stirring, TLC (diethyl ether eluant; versus benzyl bro-
mide and dihydroca€eic acid) indicated the reaction to
be reasonably complete. Workup was similar to that for
1e. The oily reaction product was chromatographed
(100% CHCl3 followed by 100% diethyl ether) and
fractions containing a faster moving component (likely
benzyl bromide) were discarded. After a second round
of column chromatography (2:1 diethyl ether/petroleum
ether), the product was combined with the product of a
second reaction. The combination was chromato-
graphed again (4:1 CHCl3/ethyl acetate; gravity elu-
tion). TLC of the combined fractions still showed the
presence of a slightly faster moving ¯uorescent com-
ponent. Nevertheless, the oily, orange product was
subjected to Kugelrohr distillation (80±85 ꢀC, 2 h) to
remove any volatile impurities and samples were pre-
pared for cell culture analysis, mass spectrometry, and
elemental analysis. The percent yield is based on the
sum of the theoretical yields of the combined reaction
products (4.4 mmol+5.5 mmol=9.9 mmol). Yield (pure)
0.91g (34%). 1H NMR (CDCl3=7.26 ppm) d 7.4±7.3
(mult., 5H), 6.74 (d, J=8.1 Hz, 1H), 6.65 (d, J=1.9 Hz,
1H), 6.60, 6.58 (dd, J=1.9, 8.1 Hz, 1H), 5.57 (s, 1H,
OH), 5.48 (s, 1H, OH), 5.11 (s, 2H), 2.85 (t, J=7.5,
7.7 Hz, 2H), 2.64 (t, J=7.7, 7.5 Hz, 2H); 13C NMR
(CDCl3=77.0 ppm) d 173.9, 143.6, 142.2, 135.4, 132.8,
128.5, 128.2 (two equivalent carbon atoms), 128.1 (two
equivalent carbon atoms), 120.4, 115.4 (two overlapping
signals), 66.6, 36.0, 30.1. FT-IR (neat, thin ®lm on one
1
36.2, 31.3, 30.3, 28.4, 25.5, 22.4,13.9. FT-IR (KBr) cm
3507, 2955, 1708, 1608, 1540. Anal. calcd (C15H22O4):
67.64 %C, 8.34 %H; Found: 68.03 %C, 8.71 %H.
Compound 2d was also successfully prepared by Fisher
esteri®cation of dihydroca€eic acid with n-hexyl
alcohol.
Synthesis of 3-(3,4-dihydroxyphenyl)propanoic acid, t-
butyl ester (t-butyl dihydroca€eate, 2j). t-Butyl ca€eate
(1j) (0.72 g, 3 mmol) was combined with 10% Pd/C
catalyst (0.39 g) and absolute ethanol (125 mL). Parr
hydrogenation (40 psi, 6 h) followed. The mixture was
®ltered through Celite and solvent removed. The oily
residue was chromatographed twice (®rst round: 3:1
diethyl ether/petroleum ether; second round: 4:1 CHCl3/
ethyl acetate), followed by Kugelrohr distillation (80±
85 ꢀC, 2 h) to remove volatile impurities. A pale orange
1
oil was recovered. Yield (pure): 0.19 g (26%). H NMR
(CDCl3=7.26 ppm) d 6.75 (d, J=8.1 Hz, 1H), 6.69 (d,
J=1.9 Hz, 1H), 6.61, 6.59 (dd, J=1.9, 8.1 Hz, 1H), 5.90
(s, 1H), 5.70 (s, 1H), 2.78 (t, J=7.5, 7.8 Hz, 2H), 2.50 (t,
J=7.9, 7.5 Hz, 2H), 1.42 (s, 9H); 13C NMR
(CDCl3=77.0 ppm) d 173.7, 143.7, 142.2, 132.9, 120.2,
115.4, 115.2, 81.1, 37.3, 30.4, 27.9. FT-IR (neat, thin

208
B. Etzenhouser et al. / Bioorg. Med. Chem. 9 (2001) 199±209
1
®lm between two NaCl plates) cm 3390, 2978, 2932,
1699, 1606, 1520. Anal. calcd (C13H18O4) 65.53 %C,
7.63 %H; Found: 65.28 %C, 7.69 %H. Exact Mass
Data (Electron Impact, 20 eV): M+ for (C13H18O4):
Calcd: 238.120509; Found: 238.119568.
Method 4: Wittig Reaction31a
Synthesis of tert-butoxycarbonylmethyl triphenylphos-
phonium bromide (3).31a,b,c Triphenylphosphine (6.0 g,
23 mmol) was dissolved in benzene or toluene (10±
20 mL). tert-Butyl bromoacetate (3.8 mL, 26 mmol) was
added to the solution. The mixture was swirled well and
almost immediately became cloudy. White precipitate
formed within 10 min. The mixture was allowed to stand
in the refrigerator for at least 24 h before collecting the
precipitate. The solid was ®ltered, washed with benzene
or toluene, then with hexanes, and allowed to dry over
vacuum. The product was not puri®ed further but used
directly for Wittig reactions. The product was stored in
the refrigerator. Yield: 10.2 g (97%), melting point 169±
173 ꢀC, dec. 1H NMR (CDCl3=7.26 ppm) 8.00±7.60
(mult. 15H), 5.32 (d, J=14 Hz, 2H), 1.19 (s, 9H). FT-IR
Synthesis of 3-(3,4-dihydroxyphenyl)propanoic acid,
phenethyl ester (phenethyl dihydroca€eate, 2h). Similar
to the procedure for 2d, except 2.0 g (7 mmol) of phe-
nethyl ca€eate, 0.7 g of 10% Pd/C catalyst, and 100 mL
of absolute ethanol were combined. Parr hydrogenation
(40 psi, 4 h) followed. After ®ltering the ethanol mix-
ture through Celite, the solvent was removed and the
residue taken up in acetone. This mixture was ®ltered
with air pressure through silica gel to remove traces of
catalyst. The oily reaction product was chromato-
graphed two times (®rst round: 2:1 diethyl ether/petro-
leum ether; second round: 100% CHCl3, followed by
100% diethyl ether or 2:1 diethyl ether/petroleum
ether). Upon standing, the combined fractions solidi-
®ed. Recrystallization (diethyl ether/petroleum ether)
led to recovery of a white to pale champagne colored,
crystalline solid. Yield (pure): 0.5 g (25%), melting
point 73±74 ꢀC. 1H NMR (acetone-d6, TMS=0.00 ppm)
d 7.74 (broad singlet, 2H), 7.31±7.19 (multiplet, 5H),
6.72 (d (up®eld half of doublet overlapped by down
®eld half of 6.70 ppm doublet), J=8.1 Hz, 1H), 6.70 (d,
J=1.9 Hz, 1H), 6.53, 6.51 (dd, J=2.0, 8.0 Hz, 1H),
4.24 (t, J=7.0 Hz, 2H), 2.90 (t, J=7.0 Hz, 2H), 2.74 (t,
J=7.5, 7.7 Hz, 2H), 2.52 (t,J=7.8, 7.3 Hz, 2H); 13C
NMR (CDCl3=77.0 ppm) d 174.1, 143.6, 142.1, 137.5,
132.9, 128.8, 128.5, 126.5, 120.4, 115.3 (two over-
lapping peaks), 65.3, 36.1, 34.8, 30.1. FT-IR (KBr)
1
(KBr) cm 2876, 1721, 1586, 1436, 1273, 1138, 1112,
757.
Synthesis of t-butyl-3,4-dihydroxycinnamate (t-butyl
ca€eate, 1j).31a Although the above reference utilizes
sonication as the main energy source for the Wittig
reaction, preliminary reactions during this study showed
that re¯uxing was indeed more convenient, if not just as
e€ective. 3,4-Dihydroxybenzaldehyde (1.0 g, 7.2 mmol)
was dissolved in distilled 1,4-dioxane (20 mL). 4.5 grams
of Wittig reagent 3 (9.8 mmol) was dissolved in reagent
grade CHCl3 (20 mL). The solutions were added to a
¯ask containing 2.10 g (21 mmol) of KHCO3 and a stir
bar. A re¯ux condenser was attached, and the mixture
was heated to re¯ux temperature. The resulting mixture
was ®ltered and concentrated on a rotary evaporator.
Column chromatography (2:1 ethyl acetate/petroleum
ether) and recrystallization (diethyl ether/petroleum
ether) led to recovery of a champagne colored, crystal-
lineꢀsolid. Yield (pure): 0.30 g (18%), melting point 130±
1
cm 3330, 3027, 2958, 1707, 1604, 1524. Anal. calcd
(C17H18O4): 71.31 %C, 6.35 %H; Found: 71.41 %C,
6.52 %H.
Synthesis of 3-(3,4-dihydroxyphenyl)propanoic acid, 3-
phenylpropyl ester (3-phenylpropyl dihydroca€eate, 2i).
One-half gram (1.7 mmol) of 3-phenylpropyl ca€eate
(1i) was combined with 0.3 g of 10% Pd/C and 120 mL
of absolute ethanol. Parr hydrogenation (40 psi, 4.5 h)
followed. The mixture was ®ltered two times (through
Celite as an ethanol solution, then through silica gel as
an acetone solution) to remove excess catalyst, then
chromatographed (2:1 diethyl ether/petroleum ether).
Crystals evolved from the combined fractions after
adding more petroleum ether and cooling. A second
recrystallization (diethyl ether/petroleum ether) resulted
in a white to pale-yellow crystalline product. Yield
1
134 C. H NMR (acetone-d6, TMS=0.00 ppm) d 8.38
(broad singlet, 2H), 7.46 (d, J=15.9 Hz, 1H), 7.14 (d,
J=2.1 Hz, 1H), 7.03, 7.01 (dd, J=2.0, 8.2 Hz, 1H), 6.86
(d, J=8.2 Hz, 1H), 6.20 (d, J=15.8 Hz, 1H), 1.50 (s,
9H); 13C NMR (acetone-d6=29.8 ppm) d 166.8, 148.3,
146.1, 144.6, 127.6, 122.2, 117.5, 116.2, 115.0, 80.1, 28.3.
1
FT-IR (KBr) cm 3488, 3061, 2976, 1680, 1606, 1531.
Anal. calcd (C13H16O4) 66.09 %C, 6.84 %H; Found
66.05 %C, 7.04 %H.
Cytotoxicity assays
(pure): 0.11 g (22%), melting point 69±71 ꢀC. H NMR
L1210 cells. The IC50 values in the L1210 cell line were
determined according to previously published proto-
cols.19 The IC50 is de®ned as the concentration of a
compound that inhibits cell growth by 50%.
1
(acetone-d6=2.04 ppm) d 7.74 (broad singlet, 2H), 7.28±
7.14 (multiplet, 5H), 6.709 (d, J=8.0 Hz, 1H), 6.71 (d,
J=2 Hz, 1H), 6.56, 6.54 (dd, J=2.0, 8.0 Hz, 1H), 4.02
(t, 2H), 2.76 (t, J=7.5, 7.7 Hz, 2H), 2.64 (t, 2H), 2.54 (t,
J=7.7, 7.5 Hz, 2H), 1.89 (quintet, 2H); 13C NMR
(CDCl3=77.0 ppm) d 174.3, 143.7, 142.1, 140.9, 132.9,
128.3 (two overlapping peaks representing two sets of
equivalent carbon atoms), 125.9, 120.4, 115.3 (two
overlapping peaks) 64.2, 36.1, 31.9, 30.2, 29.9. FT-IR
MCF-7 cells. Human Breast cancer cells were main-
tained in asynchronous logarithmic growth at 37 ꢀC in
Phenol red free Iscoves modi®ed Dulbeccos medium
with l-glutamine supplemented with 10% (v/v) FBS.
The population doubling time was 24±36 h. Every 48 h,
the old media was replaced by the fresh media. All stock
solutions and dilutions were made in unsupplemented
Dulbeccos medium.
1
(KBr): cm 3496, 3284, 2937, 1701, 1611. Anal. calcd
(C18H20O4) 71.98 %C, 6.73 %H; Found: 72.26 %C,
6.71 %H.

B. Etzenhouser et al. / Bioorg. Med. Chem. 9 (2001) 199±209
209
Cell cultures were seeded at 2±5Â104 cells/mL in dupli-
cate for each inhibitor concentration in a 24-well
microtiter plate (0.9 mL/well). The test compounds
(0.1 mL) were then added to the cell cultures in 1:10
dilution in order to achieve the desired concentration.
Each inhibitor was tested at a minimum of 8 con-
centrations. After 4 days of continuous drug exposure
the cells were counted by using the CyQUANT GR
assay kit from Molecular Probes. For this purpose, the
media was removed from the plates which were then
frozen at 80 ꢀC for a minimum of 1 h. The cells were
thawed at 37 ꢀC and 200 uL of CyQUANT GR dye/cell
lysis bu€er added to each well. The plates were incu-
bated for 5 min at 37 ꢀC and ¯uorescence was measured
using Cyto¯uor II multiwell ¯uorescence plate reader.
The excitation maximum was 485 nM and the emission
maximum was 530 nM.
Grunberger, D.; Sacks, P. G.; Tanabe, T.; Dannenberg, A.J.
Cancer Res. 1999, 59, 2347.
10. Sud'ina, G. F.; Mirzoeva, O. K.; Puskareva, M. A.; Kor-
shunova, G. A.; Sumbatyan, N. V.; Varfolomeev, S. D. FEBS
Lett. 1993, 329, 21.
11. Fesen, M. R.; Pommier, Y.; Leteurtre, F.; Hiroguchi, S.;
Yung, J.; Kohn, K. W. Biochem. Pharmacol. 1994, 48, 595.
12. Bhimani, R.; Troll, W.; Grunberger, D.; Frenkel, K.
Cancer Res. 1993, 53, 4528.
13. Huang, M. T.; Ma, W.; Yen, P.; Xie, J. A.; Frenkel, K.;
Grunberger, D.; Conney, A. Carcinogenesis (Lond) 1996, 17,
761.
14. Chiao, C. M.; Carothers, A.; Grunberger, D.; Solomon,
G.; Preston, G. A.; Barrett, J. C. Cancer Res. 1995, 55, 3576.
15. Jaiswal, A. Biochem. Pharmacol. 1994, 48, 439.
16. Van Acker, S.; van der Berg, D.; Tromp, M.; Gri€oen, D.;
van Bennekom, W.; van der Vijg, W.; Bast, A. Free Radical
Biol. Med. 1996, 20, 331.
17. Asadawa, E. M.; Hagiwara, A.; Takahoshi, S.; Ito, N. Int.
J. Cancer 1994, 56, 146.
From the data, a dose-response curve was drawn and
the IC50 determined. Physicochemical constants were
then utilized to formulate a quantitative structure
activity relationship (QSAR) for these compounds.32
Calculated logP values were obtained using ClogP pro-
gram.33 (Biobyte Corp. version 4.1).
18. Soto, A. M.; Justicia, H.; Wray, J. W.; Sonnenschein, C.
Environ. Health Persp. 1991, 92, 167.
19. Selassie, C. D.; DeSoyza, T. V.; Rosario, M.; Gao, H.;
Hansch, C. Chem-Bio. Interactions 1998, 113, 175.
20. Hansch, C.; Gao, H. Chem. Rev. 1997, 97, 2995.
21. Selassie, C. D.; Shusterman, A. J.; Kapur, S.; Verma, R.
P.; Zhang, L.; Hansch, C. J. Chem. Soc., Perkin Trans. 2 1999,
2729.
22. Simonyan, A. V. Khim. Farm. Zh. 1993, 27, 92 (English
Ed).
Acknowledgement
23. Aruma, K.; Manikyamba, P. Indian J. Chem. 1995, 34A,
822.
24. Simonyan, A. V.; Bogdashev, N. A.; Tukhovskaya, N. A.;
Sysoeva, T. N. Isv. Vyssh. Uched. Sev.-kav. Regiona. Ser. Est.
Vyssh. Nauki 1993, 3±4, 150.
This research was supported by a grant from R. J.
Reynolds for fundamental studies in toxicology and
NIEHS grant R01 ES 07595.
25. Yamamoto, K.; Kato, S.; Mizutani, T.; Irie, Y. Res.
Commun. Pharm. Toxicol. 1996, 1, 211.
References and Notes
26. Sugiura, M.; Naito, Y.; Yamaura, Y.; Fukaya, C.;
Yokoyama, K. Chem. Pharm. Bull. 1989, 37, 1039.
27. Naito, Y.; Sugiura, M.; Yamaura, Y.; Fukaya, C.;
Yokoyama, K.; Nakagawa, Y.; Ikeda, T.; Senda, M.; Fujita,
T. Chem. Pharm. Bull. 1991, 39, 1736.
28. Guarini, L.; Su, Z. Z.; Zucker, S.; Lin, J.; Grunberger, D.;
Fisher, P. B. Cell. Mol. Biol. 1992, 38, 513.
29. Burke Jr., T. R.; Fesen, M. R.; Mazumder, A.; Wang, J.;
Carothers, A. M.; Grunberger, D.; Driscoll, J.; Kohn, K.;
Pommier, Y. J. Med. Chem. 1995, 38, 4171.
30. Hashimoto, T.; Tori, M.; Asakawa, Y.; Wollenweber, E.
Z. Naturforsch. 1988, 43c, 470.
31. (a) Bankova, V. S. J. Nat. Prod. 1990, 53, 821. (b) Con-
sidine, W. J. J. Org. Chem. 1962, 27, 647. (c) Denney, D. B.;
Ross, S. T. J. Org. Chem. 1962, 27, 998.
1. Grunberger, D.; Banerjee, R.; Eisinger, K.; Oltz, E. M.;
Efros, L.; Caldwell, M.; Estevez, V.; Nakanishi, K. Experi-
entia 1988, 44, 230.
2. Challis, B. C.; Bartlett, C. D. Nature 1975, 254, 532.
3. Koshihara, Y.; Neichi, T.; Murota, S.; Lao, A.; Fujimoto,
Y.; Tatsuno, T. Biochim. Biophys. Acta 1984, 792, 92.
4. Tanaka, T.; Kojima, T.; Kawamori, T.; Yoshimi, N.; Mori,
M. Cancer Res. 1993, 53, 2775.
5. Chimi, M.; Cillard, J.; Cillard, P. J. Am. Oil. Chem. Soc.
1991, 68, 307.
6. Rao, C. V.; Dhimant, D.; Simi, B.; Kulkarni, N.; Amin, S.;
Reddy, B. S. Cancer Res. 1993, 53, 4182.
7. Rao, C. V.; Dhimant, D.; Rivenson, A.; Simi, B.; Amin, S.;
Reddy, B. Cancer Res. 1995, 55, 2310.
8. Rao, C. V.; Desai, D.; Kaul, B.; Amin, S.; Reddy, B. S.
Chem. Biol. Int 1992, 84, 277.
9. Michaluart, P.; Masferrer, J. L.; Carothers, A. M.; Sub-
baramaiah, K.; Zweifel, B. S.; Koboldt, C.; Mestre, J. R.;
32. Hansch, C.; Leo, A. In Exploring QSAR. Fundamentals
and Applications in Chemistry and Biology; American Chemi-
cal Society: Washington D.C., 1995; pp 1±121.
33. Leo, A. J. Chem. Rev. 1993, 93, 1281.