J. Med. Chem.

2003, 46, 4259-4264 4259

Discovery, Characterization, and Structure-Activity Relationships Studies of

Proapoptotic Polyphenols Targeting B-Cell Lymphocyte/Leukemia-2 Proteins

Shinichi Kitada, Marilisa Leone, Sina Sareth, Dayong Zhai, John C. Reed,* and Maurizio Pellecchia*
The Burnham Institute, 10901 North Torrey Pines Road, La Jolla, California 92037

Received April 21, 2003

Among the most promising chemopreventive agents, certain natural polyphenols have recently
received a great deal of attention because of their demonstrated inhibitory activity against
tumorigenesis. In view of their anticancer properties, these compounds also hold great promise
as potential chemotherapeutic agents. However, to translate these chemopreventive agents
into chemotherapeutic compounds, their exact mechanisms of action must be delineated. By
using a multidisciplinary approach guided by modern nuclear magnetic resonance spectroscopy
techniques, fluorescence polarization displacement assays, and cell-based assays, we have begun
to unravel the mechanisms of actions of certain polyphenols such as Gossypol (a compound
from cotton seed extracts) and Purpurogallin (a natural compound extracted from Quercus sp.
nutgall) and their derivatives. Our findings suggest that these natural products bind and
antagonize the antiapoptotic effects of B-cell lymphocyte/leukemia-2 (Bcl-2) family proteins
such as Bcl-xL. Our in vitro and in vivo data not only open a window of opportunities for the
development of novel cancer treatments with these compounds but also provide structural
information that can be used for the design and development of novel and more effective
analogues.

Introduction ing antiapoptotic Bcl-2 family proteins as potential

therapeutic agents.6-12 However, small-molecule inhibi-
Apoptosis is essential in multicellular organisms,
tors reported thus far display only moderate in vitro
where it regulates normal development and tissue
affinity and/or poor activity in vivo and, therefore, have
homeostasis. Inhibition of apoptosis is implicated in
provided only limited data for the validation of Bcl-2
virtually every known human malignancy.1,2 This in-
and Bcl-xL antagonists as anticancer agents. The dis-
hibition provides malignant cells with a selective growth
covery of more potent and cell-permeable compounds is
advantage, allowing survival in the face of radiation or
therefore needed to validate Bcl-2-family proteins as
chemotherapy.1,2 Proteins of the Bcl-2 (B-cell lymphocyte/
anticancer targets and to provide further hope for the
leukemia-2) family are critical components of the in-
discovery of more effective anticancer therapies.
trinsic apoptotic pathway. Several homologues, as de-
fined by sequence similarity to some or all of the four Results and Discussion
Bcl-2 homology (BH) domains in Bcl-2, are found in
We screened a small library (n ) 50) of natural
humans and function as either blockers or inducers of
products by using a combination of NMR-based binding
apoptosis. Antiapopototic and proapoptotic Bcl-2 family
assays and fluorescence polarization displacement as-
proteins dimerize, negating each other’s function (re-
says (FPA). Two polyphenols were found that inhibited
viewed in ref 3). Structural studies have elucidated a
BH3 binding to Bcl-xL: Gossypol and Purpurogallin
hydrophobic crevice on the surface of antiapoptotic Bcl-
(Figure 1). In contrast, other polyphenols such as
xL proteins that binds the BH3 dimerization domain of
pelagonidine, catechin, epigallocatechin, myricetin, caf-
proapoptotic family members. Thus, molecules that
feic acid, gallocatechin, quercetin, epicatechin, mino-
mimic proapoptotic BH3 domains represent a direct
cycline, and other nonnatural polyphenols did not show
approach to overcoming the protective effects of anti-
any inhibition in our FPA with Bcl-xL.
apoptotic proteins such as Bcl-2 and Bcl-xL. Antiapo-
Gossypol is a natural product derived from cottonseed
ptotic Bcl-2 proteins are overexpressed in many human
extracts that was originally extensively investigated in
cancers. Bcl-2 is overexpressed in 80% of B-cell lym-
China as a natural male contraceptive agent.13 In
phomas, 30-60% of prostate cancers, 90% of colorectal
recent studies, Gossypol exhibited inhibitory activity
adenocarcinomas, and a wide variety of other cancers,
against a wide range of human carcinoma cell lines
and Bcl-xL is overexpressed in breast4 and lung cancers.5
derived from breast (T47D), prostate (Du-145), cervix
These observations have instigated a number of (HeLa), and pancreas (Miapaca, RwP-2), in culture and
recent studies involving a variety of approaches (includ- in tumor xenograft models.14-20 Recent studies aimed
ing computational, combinatorial, and evolutionary at the identification of the Gossypol molecular target
strategies) aimed at the discovery of compounds target- have suggested its ability to induce apoptosis even
in Bcl-2 or Bcl-xL overexpressed cells, though the
* To whom correspondence should be addressed. For J.C.R.: phone, mechanism is unknown.21 On the basis of in vitro
858-646-3140; fax, 858-646-3194; e-mail, jreed@burnham.org. For
M.P.: phone, 858-646-3159; fax, 858-646-3195; e-mail, mpellecchia@ displacement assays with a fluorescein-labeled BH3
burnham.org. peptide NLWAAQRYGRELRRMSD-K(FITC)-FVD22
10.1021/jm030190z CCC: $25.00 © 2003 American Chemical Society
Published on Web 08/22/2003
4260 Journal of Medicinal Chemistry, 2003, Vol. 46, No. 20 Kitada et al.

Figure 1. Gossypol and Purpurogallin compete for the BH3-binding pocket of Bcl-xL. Shown are the chemical structures of
Gossypol (a) and Purpurogallin (c). Results of fluorescence polarization based competitive binding assays (FPA) using a fluorescein-
labeled Bad peptide (NLWAAQRYGRELRRMSD-K(FITC)-FVD) (Synpep Corporation, Dublin, CA) are shown in part c for Gossypol
and part d for Purpurogallin. Gossypol did not show appreciable binding to BIR3 domains by NMR (see text) or by enzymatic
assays.
(Synpep Corporation, Dublin, CA), we found that Gos- the chemical-shift differences upon binding into the
sypol directly interacts with Bcl-xL and is able to three-dimensional structure of Bcl-xL.22 The binding of
displace BH3 peptides with IC50 ) 0.5 µM (Figure 1). Gossypol and Purpurogallin to 15N-labeled Bcl-xL in-
In contrast, Gossypol did not inhibit protein interactions duces similar large variations in the [15N,1H] correlation
or enzyme activity in other assays tested (including spectrum of the protein. At a 1:1 molar ratio of Gossypol/
NMR-based binding to 15N-labeled proteins such as Bcl-xL, many resonances are broadened beyond detec-
BIR3 domain and BID, as well as FPA with BIR3 tion, characteristic of binding with intermediate ex-
domain). As controls for our fluorescence polarization change rates on the NMR time scale, which translates
assays, we have tested several polyphenolic compounds to a dissociation constant in the low micromolar to
as negative controls and we have tested a series of submicromolar range (Figure 2a). A higher concentra-
benzylidene-R-isopropyl-4-oxo-2-thioxo-3-thiozolidineace- tion of Purpurogallin is required to obtain a similar
tic acid compounds (Chembridge Corp.) that were effect (3:1 ratio; not shown), indicating that at NMR
recently reported to bind Bcl-xL in the micromolar concentrations used (millimolar) the binding of Pur-
range7 as positive controls (not shown). Here, our data purogallin to Bcl-xL appears somehow weaker than
are in good agreement with the data reported in the Gossypol, in agreement with our fluorescence polariza-
literature.7 As an additional control, we also tested a tion displacement assays.
polyphenylurea (compound 1396-34), a specific small- Mapping of the observed changes in the NMR spectra
molecule inhibitor of XIAP23 developed in our labora- upon titration with Gossypol and Purpurogallin into the
tories, within its active dose range. This compound did three-dimensional structure of Bcl-xL reveals that the
not bind to Bcl-xL in our FPA assays. Finally, we tested binding of the polyphenols mostly affects residues in the
p-hydroxybenzaldehyde (PHBA) to verify that the activ- BH3-binding pocket, although the changes are rather
ity shown is not merely due to Gossypol aldehyde widespread throughout a larger region surrounding this
groups. When tested within the same dose range used pocket (Figure 2b). As controls, Gossypol did not bind
for Gossypol, PHBA did not show significant competition to the XIAP BIR3 domain23 in a similar NMR binding
for BH3 peptides binding to Bcl-xL in our FPA assays assay with 15N-labeled protein and to the proapoptotic
(not shown). Bcl-2 member BID (not shown). Protein-based experi-
A second compound that inhibited Bcl-xL is Purpuro- ments require generally high micromolar to millimolar
gallin, an antioxidant compound used in edible oils.24 concentrations of protein and ligand, which could result
Purpurogallin has also been shown to inhibit tyrosine- in nonspecific binding interactions. For this reason, we
specific protein kinase25 and DNA synthesis in U-87 MG have also measured ligand-based NMR binding assays
glioblastoma cells in vitro.26 In our studies, Purpuro- by observing changes in ligand nuclear spin relaxations
gallin is a moderately potent inhibitor of Bcl-xL, displac- properties in the presence of a substoichiometric amount
ing the FITC-BH3 peptide with an IC50 of 2.2 µM, of Bcl-xL.27 We measured the T1F28 and saturation
strongly suggesting that its mechanism of antitumor transfer effects on ligands using WaterLOGSY experi-
action may be linked at least in part to its ability to ments.29 T1F experiments with 10 µM Bcl-xL and 100
inactivate Bcl-2 proteins. µM Gossypol showed a complete loss in signals (Figure
To provide further insights into the mechanism of 2c), which is characteristic of binding in the low micro-
action of these polyphenols at the structural level, we molar to submicromolar range. The broadening effects
studied the binding of Gossypol and Purpurogallin to on Purpurogallin are less dramatic, but binding is
Bcl-xL by heteronuclear NMR spectroscopy techniques.27 clearly appreciable in T1F experiments with a nearly
Use of recombinant 15N-labeled Bcl-xL allowed us to map complete loss of signal intensity at 300 ms (Figure 2d).
Proapoptotic Polyphenols Journal of Medicinal Chemistry, 2003, Vol. 46, No. 20 4261

Figure 2. NMR binding studies. (a) 2D [15N,1H] transverse

relaxation-optimized spectra for Bcl-xL in the apo (left) and
Gossypol-bound (right) forms. (b) Chemical-shift mapping of
Gossypol into the three-dimensional structure of Bcl-xL in
complex with Bak peptide (PDB code 1BXL). The peptide is
displayed in yellow. Regions affected by the binding of Gos-
sypol are in red. (c, d) T1F experiments (300 ms relaxation time)
with Gossypol (c) and Purpurogallin (d): blue, without protein;
red, with 10 µM Bcl-xL. Peaks shown in part c represent the
isopropyl and the methyl groups in Gossypol. In part d, the
peak marked with an asterisk represents residual imidazole Figure 3. Molecular modeling studies. (a) Surface represen-
present in the protein preparation. tation of Bcl-xL with the docked structure of Gossypol obtained
by FlexX. The surface is depicted according to cavity depth
Similar conclusions were reached from WaterLOGSY (blue, surface exposed; yellow, buried) representation. (b)
experiments (not shown). Detailed interactions between (-)-Gossypol (white and red
Docking studies with FlexX software30 implemented sticks) and Bcl-xL. Only selected amino acids are shown and
in Sybyl (Tripos) using the Bcl-xL conformation found labeled for clarity. Hydrogen bonds are shown as dashed lines,
and van der Waals interactions involving the isopropyl groups
in the complex with Bak peptide showed an optimal are highlighted. (c) Superposition of compound 5D1 (green) and
location for Gossypol in the deep hydrophobic cleft (-)-Gossypol (white with red for oxygen atoms).
normally occupied by the Bak helical BH3 peptide in
the complex (Figure 3a). We docked both the (+)- and chains, respectively, as highlighted in Figure 3b. We are
the (-)-stereoisomers of Gossypol because these exhib- currently testing the validity of this model by synthesiz-
ited different activity in previous cell-based assays that ing some analogues in which the aldehyde functional
showed that (-)-Gossypol is 10 times more effective groups, which do not seem to be deeply involved in the
than (+)-Gossypol as a cytotoxic agent.31 The goodness interaction with the protein, are either replaced or
of the fit as measured by a scoring function,32 and the derivatized. We also plan to produce single point
intermolecular energy after minimization with the mutants of Bcl-xL that, on the basis of the model, are
DOCK routine of Sybyl, was considerably better for (-)- predicted to abolish binding to Gossypol.
Gossypol (-32.7 kcal/mol) versus (+)-Gossypol (-25 Docking studies with Purpurogallin produced am-
kcal/mol), in agreement with these observations. The biguous results, given the large surface area of the BH3-
structure of (-)-Gossypol is shown (Figure 3), but the binding pocket of Bcl-xL, compared with the small size
overall positioning of both stereoisomers of Gossypol is of Purpurogallin, thus leading to several possible solu-
very similar. In the proposed model, the naphthalene tions. However, placement of Purpurogallin into the
moieties work as scaffolds to position the two isopropyl binding pocket of Bcl-xL could be attempted by using
groups and some of the hydroxyl groups to make the (-)-Gossypol structure as a template structure. To
favorable contacts with hydrophobic and polar side this end, we also further studied the inhibitory proper-
4262 Journal of Medicinal Chemistry, 2003, Vol. 46, No. 20 Kitada et al.

Table 1. Structure-Activity Relationships of Purpurogallin Derivatives

IC50 (µM)
compd R1 R2 R3 R4 R5 Bcl-xL a HeLa
Purpurogallin -OH -OH -OH -OH -H 2.2 6.5
5D1 -H -OH -OH -OH -COOC2H5 73 51.5
1163 -H -OH -OH -OH -COOCH3 2.6 ∼30
1142 -H -OH -OH -OH -COOH 7.4 22.9
6A1 -OCH3 -OCH3 -OCH3 -OCH3 -H >100 >100
6A7 -OCH3 -OCH3 -OH -OCH3 -H >100 >100
a IC
50 values against Bcl-xL were derived from dose-response curves in a fluorescence polarization assay with a FITC-labeled BH3
peptide.

Figure 4. Inhibitory effect of compounds on cancer cell survival. The effects of Gossypol on viability of tumor cells in culture
were monitored by using XTT assays with (a) MCF7 and (c) ZR75-1 cell lines (black circles). As a negative control, a generic
polyphenolic compound was also tested (open circles). Low-passage HeLa cells (between passage number 10 and 20) were transfected
with pcDNA3-Bcl-xL (black circles) or control pcDNA3 plasmids (open circles). (b) Immunoblot analysis confirmed overexpression
of Bcl-xL in the cells transfected with pcDNA3-Bcl-xL compound to pcDNA3-control transfectants. Cell lysates were normalized
for total protein content, 25 µg per lane. (d) HeLa transfectants were treated with various doses of Gossypol (0, 1, 3, 10, and 100
µM). Data shown represent the mean ( standard deviation (n ) 4).

ties of several Purpurogallin analogues (Table 1) in an activities using XTT dye reduction assays using two
attempt to identify the essential pharmacophoric sub- breast cancer cell lines: MCF7 (high expressor of Bcl-
structures common to Gossypol and Purpurogallin. 2/Bcl-xL) and ZR75-1 (low expressor of Bcl-2/Bcl-xL).
From these studies (Table 1), we concluded that (a) only Gossypol is a cytotoxic agent for MCF7 and ZR75-1 cells
two hydroxyl phenols groups are essential for binding, (Figure 4a,c), reducing cell viability in a dose-dependent
(b) substitutions in position R5 in Purpurogallin are manner, with IC50 values of 13.2 and 8.4 µM, respec-
tolerated only by small groups, given the steric hin- tively. Purpurogallin, however, did not show appreciable
drance from the methyl of the methyl esters, and (c) the activity in these assays, potentially because of its
hydrogen bond donor properties of the phenols are hydrophilic character (ClogP ≈ 0.7). Consistent with this
important. On the basis of these data, we propose a observation, a Purpurogallin derivative 5D1 that is
model of the superposition of Purpurogallin and its predicted to have better cell-membrane permeability
derivatives to Gossypol that reflects the observed SAR properties (based on its ClogP of ∼2.5) reduced cell
studies (Figure 3c). For example, in this model, Pur- viability in a dose-dependent manner, with an IC50
purogallin fits rather well into the binding pocket of Bcl- value of ∼50 µM in the ZR75-1 cell line (not shown).
xL with a docking energy of -21 kcal/mol, whereas For these reasons, we further evaluated the cellular
docking of compound 5D1 required its ethyl group to activity of our compounds in HeLa cells (Table 1), which
be folded into a gauche conformation to be accom- are known to be less selective for compound uptake. The
modated into the Bcl-xL binding pocket, which may inhibition data obtained with HeLa cell viability assays
reflect its weaker affinity (Table 1). parallel the in vitro binding data with Bcl-xL (Table 1),
To evaluate the cytotoxic activity of our compounds with a correlation coefficient of r ) 0.9 (p ) 0.001).
on human tumors cells, we tested their biological Finally, to further explore whether Bcl-xL is the
Proapoptotic Polyphenols Journal of Medicinal Chemistry, 2003, Vol. 46, No. 20 4263

primary target of Gossypol, we used HeLa transfectants saturation was performed with a train of selective IBURP2
overexpressing Bcl-xL. Low-passage HeLa cells were pulses of 7 ms durations spaced by a 10 ms delay. Total
transfected with pcDNA3-Bcl-xL or control pcDNA3 saturation time used was 2.5 s. The T1F series were measured
with a spin-lock pulse of variable length. Measurements were
plasmids (Figure 4b,d) and then treated with various then performed with 1, 10, 50, 150, 200, 250, and 300 ms spin-
doses of compounds (0, 1, 3, 10, and 100 µM). The dose- lock time with 100 µM compounds in the absence and presence
response curve of the Bcl-xL transfectants was signifi- of 10 µM protein. In all experiments, dephasing of residual
cantly shifted to the right (p < 0.01, two-way ANOVA), water signals was obtained with a WATERGATE sequence.
compared with the neocontrol transfectants, indicating Molecular Modeling. Molecular modeling studies were
that a higher amount of Gossypol is required to induce conducted on several R12000 SGI Octane workstations with
apoptosis of Bcl-xL overexpressing cells, corroborating the software package Sybyl, version 6.9 (Tripos). The docked
our findings that Gossypol targets Bcl-xL. Also consis- structure of Gossypol was initially obtained by FlexX30 as
implemented in Sybyl. Two calculations were performed. In
tent with these observations are the measured IC50 the first, all binding-site torsion angles were kept fixed, while
values of Gossypol for MCF-7 and ZR75-1 cells (Figure in the second side chain torsion angles were free to change.
4), with MCF-7 expressing higher levels of endogenous The average scoring function for the 30 best solutions was only
Bcl-xL than ZR75-1 and consequently tolerating a higher slightly lower when the side chains were free to rotate. The
amount of Gossypol. position of the side chains in the model did not change
substantially from the initial values. The scoring function for
Conclusions (+)-Gossypol was inferior to that of (-)-Gossypol, but the
overall positioning of both steroisomers was very similar. The
Bcl-2 proteins are known to induce chemoresistance resulting best scoring structures were subsequently energy-
in cancer patients, and Bcl-2/Bcl-xL-mediated chemore- minimized by using the routine DOCK of Sybyl, keeping the
sistance has been a subject of intense investigation. Our site rigid. The energy of the ligands after the DOCK minimiza-
finding that certain naturally occurring polyphenols are tion was within 5 kcal/mol from their global minimum of
potent Bcl-xL inhibitors provides for the first time a energy. Superposition of compounds was obtained by the
routine MULTIFIT of Sybyl. Color figures showing three-
plausible molecular explanation for the anticancer
dimensional structures were prepared with the programs
activity of these natural products. Our observations also Sybyl and MOLMOL.33
constitute a first example of a small-molecule compound Inhibitory Effect of Compounds on Cancer Cell Sur-
(Gossypol) targeting a Bcl-2 family protein that has been vival. The effects of the compounds studied in this paper on
tested in human clinical trials of cancer patients.19,20 viability of tumor cells in culture were monitored by using
The pharmacophore model developed here, based on XTT34 assays with MCF7 and ZR75-1 cell lines. MCF7 cells
experimental NMR binding data and structure-activity were grown in DMEM containing 10% fetal bovine serum,
relationships, provides an important framework on penicillin/streptomycin, supplemented with 10-10 M insulin,
1 mM sodium pyruvate, and glutamine. ZR75-1 cells were
which to advance the design of even more potent and
grown in RPMI containing 10% fetal bovine serum, penicillin/
selective anticancer drugs targeting Bc1-2 family pro- streptomycin, supplemented with HEPES buffer, 1 mM sodium
teins. pyruvate, and glutamine. Cells were regularly tested for
mycoplasma contamination. Cells were seeded in triplicate at
Experimental Section an initial cell density of 1000 cells per well. Blank wells
Fluorescence Polarization Assays (FPA). FPA assays received no cells. Gossypol, Purpurogallin, and compound 5D1
were conducted with a fluorescein-labeled Bad peptide (NL- were added at final concentrations of 0, 1, 10, and 100 µM and
WAAQRYGRELRRMSD-K(FITC)-FVD) (Synpep Corporation, incubated for 3 days. Relative numbers of viable cells were
Dublin, CA) using a LJL Analyst HT (Molecular Devices Co., determined by XTT assay. Briefly, in a 96-well plate, we added
Sunnyvale, CA). Dilution buffer for all stocks and samples was 50 µL of a mixture of 1 mg/mL XTT34 (2,3-bis(2-methoxy-4-
50 mM Tris-Bis, pH 7.4, 0.01% bovine γ-globulin. A series of nitro-5-sulfophenyl)-5-[(phenylamino)carbonyl]-2H-tetrazoli-
2-fold dilutions of Gossypol were prepared, i.e., 100, 50, and um hydroxide) (Polysciences, Washington, PA) containing
0.1 µM in dilution buffer. To each tube was added a solution 0.025 mM PMS (phenazine methosulfate) to each well. The
containing 30 nM Bcl-xL and 4 nM fluoresceinated peptide. 96-well plates were reincubated for an additional 4 h to allow
The tubes were incubated for 5 min at room temperature, and for XTT formazan production. Then the contents of each plate
20 µL of each of reaction mixture was transferred to a 96-well were mixed and optical densities were determined at a
black PS, HE microplate (LJL Biosystems Co.). All assays were wavelength of 450 nm (OD450). Net OD450 was determined after
performed in quadruplicate, with blank wells receiving no subtracting the OD450 of blank wells. Low-passage HeLa cells
Gossypol. Then the plate was read for total intensity and (between passage number 10 and 20) were transfected with
polarization (in mP units) was measured. Controls included pcDNA3-Bcl-xL or control pcDNA3 plasmids using Lipo-
dose-response measurements in the absence of the proteins fectamine Plus reagent (Invitrogen) and selected in medium
to assess any interactions between the compounds and the containing 800 µg/mL G418. Immunoblot analysis of Bcl-xL
FITC-BH3 peptide. Eventual effects were taken into account was accomplished as described by Krajanski in Cancer Re-
by subtraction. search, 1996. HeLa transfectants were treated with various
NMR Spectroscopy. 2D [15N,1H] transverse relaxation- doses of Gossypol, Purpurogallin, and their derivatives (0, 1,
optimized spectra27,32 for Bcl-xL were obtained with 0.5 mM 3, 10, and 100 µM).
samples of 15N-labeled Bcl-xL. 15N-labeled and unlabeled Bcl- Chemicals. Pure polyphenols were obtained from Sigma
xL samples were prepared and purified as described in ref 22. (Gossypol and Purpurogallin) and/or from Microsource Dis-
For chemical-shift mapping and docking studies, we used the covery Systems (Purpurogallin derivatives). Reference com-
three-dimensional structure of Bcl-xL in complex with Bak pounds were obtained from Chembridge Corp. (San Diego, CA).
peptide (PDB code 1BXL). In addition to chemical-shift map- Gossypol was tested as a racemic mixture of (+)- and (-)-
ping with labeled proteins, T1F measurements28 and saturation isomers. Compounds were dissolved in DMSO at 100 mM and
transfer experiments such as WaterLOGSY experiments29 stored at -20 °C. NMR analysis was periodically performed
were also performed to further validate the binding of the on the compounds as a quality control prior to further dilution
studied compounds to Bcl-xL. All experiments were performed for binding and displacement assays. Reactivity of Gossypol
with a 500 MHz Varian Unity+ spectrometer or a 600 MHz was tested with a 15N-labeled test protein (BIR3 domain of
Bruker Avance600 spectrometer, both equipped with four rf XIAP). A solution containing 1 mM Gossypol and 200 µM 15N-
channels and z-axis pulse-field gradients. Selective water labeled BIR3 was incubated for 2 h, and the [15N,1H] correla-
4264 Journal of Medicinal Chemistry, 2003, Vol. 46, No. 20 Kitada et al.

tion spectrum was recorded and compared with the spectrum (17) Wang, X. Cytotoxic effect of gossypol in the treatment of metastic
of the apo-Bir3. No appreciable differences in the spectra were adrenal cancer. J. Clin. Endocrinol. Metab. 1993, 76, 1019-1024.
observed. (18) Leblanc, M. L.; Russo, J.; Kudelka, A. P.; Smith, J. A. An in
Vitro Study of Inhibitory Activity of Gossypol, a Cottonseed
Acknowledgment. We thank Dr. Gerhard Wagner Extract, in Human Carcinoma Cell Lines. Pharmacol. Res. 2002,
46, 551-555.
for the backbone resonance assignments of Bcl-xL, and (19) Stein, R. C.; Joseph, A. E.; Matlin, S. A.; Cunningham, D. C.;
we thank the NIH for support (Grant CA78040-05). M.L. Ford, H. T.; Coombes, R. C. A preliminary clinical study of
is supported by a fellowship from the University of gossypol in advanced human cancer. Cancer Chemother. Phar-
macol. 1992, 30, 480-482.
Naples “Federico II”. (20) Van Poznak, C.; Seidman, A. D.; Reidenberg, M. M.; Moasser,
M. M.; Sklarin, N.; et al. Oral gossypol in the treatment of
References patients with refractory metastatic breast cancer: a phase I/II
(1) Reed, J. C. Regulation of apoptosis by bcl-2 family proteins and clinical trial. Breast Cancer Res. Treat. 2001, 66, 239-248.
its role in cancer and chemoresistance. Curr. Opin. Oncol 1995, (21) Wang, X.; Wang, J.; Wong, S. C.; Chow, L. S.; Nicholls, J. M.; et
7, 541-546. al. Cytotoxic effect of gossypol on colon carcinoma cells. Life Sci.
(2) Kelekar, A.; Thompson, C. B. Bcl-2-family proteins: the role of 2000, 67, 2663-2671.
the BH3 domain in apoptosis. Trends Cell Biol. 1998, 8, 324- (22) Sattler, M.; Liang, H.; Nettesheim, D.; Meadows, R. P.; Harlan,
330. J. E.; et al. Structure of Bcl-xL-Bak peptide complex: recognition
(3) Johnstone, R. W.; Ruefli, A. A.; Lowe, S. W. Apoptosis: a link between regulators of apoptosis. Science 1997, 275, 983-986.
between cancer genetics and chemotherapy. Cell 2002, 108, 153- (23) Deveraux, Q. L.; Reed, J. C. IAP family proteinsssuppressors
164. of apoptosis. Genes Dev. 1999, 13, 239-252.
(4) Olopade, O. I.; Adeyanju, M. O.; Safa, A. R.; Hagos, F.; Mick, (24) Anderson, W. A. Hydroxylic aromatic compounds as additives
R.; et al. Overexpression of BCL-x protein in primary breast for rubber based, composite solid propellant. U.S. Patent 4,181,-
cancer is associated with high tumor grade and nodal me-
545, 1980 (United Technologies Corporation).
tastases. Cancer J. Sci. Am. 1997, 3, 230-237.
(5) Reeve, J. G.; Xiong, J.; Morgan, J.; Bleehen, N. M. Expression (25) Abou-Karam, M.; Shier, W. T. Inhibition of oncogene product
of apoptosis-regulatory genes in lung tumour cell lines: relation- enzyme activity as an approach to cancer chemoprevention.
ship to p53 expression and relevance to acquired drug resistance. Tyrosine-specific protein kinase inhibition by purpurogallin from
Br. J. Cancer 1996, 73, 1193-1200. Quercus sp. nutgall. Phytother. Res. 1999, 13, 337-340.
(6) Wang, J. L.; Liu, D.; Zhang, Z. J.; Shan, S.; Han, X.; et al. (26) Fung, K. P.; Wu, T. W.; Lui, C. P. Purpurogallin inhibits DNA
Structure-based discovery of an organic compound that binds synthesis of murine fibrosarcoma L-929 and human U-87 MG
Bcl-2 protein and induces apoptosis of tumor cells. Proc. Natl. glioblastoma cells in vitro. Chemotherapy 1996, 42, 199-205.
Acad. Sci. U.S.A. 2000, 97, 7124-7129. (27) Pellecchia, M.; Sem, D. S.; Wüthrich, K. NMR in drug discovery.
(7) Degterev, A.; Lugovskoy, A.; Cardone, M.; Mulley, B.; Wagner, Nat. Rev. Drug Discovery 2002, 1, 211-219.
G.; et al. Identification of small-molecule inhibitors of interaction (28) Hajduk, P. J.; Olejniczak, E. T.; Fesik, S. W. One-dimensional
between the BH3 domain and Bcl-xL. Nat. Cell Biol. 2001, 3, relaxation- and diffusion-edited NMR methods for screening
173-182. compounds that bind to macromolecules. J. Am. Chem. Soc.
(8) Tzung, S. P.; Kim, K. M.; Basanez, G.; Giedt, C. D.; Simon, J.; 1997, 119, 12257-12261.
et al. Antimycin A mimics a cell-death-inducing Bcl-2 homology
(29) Dalvit, C.; Pevarello, P.; Tato, M.; Veronesi, M.; Vulpetti, A.; et
domain 3. Nat. Cell Biol. 2001, 3, 183-191.
(9) Enyedy, I. J.; Ling, Y.; Nacro, K.; Tomita, Y.; Wu, X.; et al. al. Identification of compounds with binding affinity to proteins
Discovery of small-molecule inhibitors of Bcl-2 through structure- via magnetization transfer from bulk water. J. Biomol. NMR
based computer screening. J. Med. Chem. 2001, 44, 4313-4324. 2000, 18, 65-68.
(10) Kaneko, M.; Nakashima, T.; Uosaki, Y.; Hara, M.; Ikeda, S.; et (30) Kramer, B.; Rarey, M.; Lengauer, T. Evaluation of the FLEXX
al. Synthesis of tetrocarcin derivatives with specific inhibitory incremental construction algorithm for protein-ligand docking.
activity towards Bcl-2 functions. Bioorg. Med. Chem. Lett. 2001, Proteins 1999, 37, 228-241.
11, 887-890. (31) Qiu, J.; Levin, L. R.; Buck, J.; Reidenberg, M. M. Different
(11) Chin, J. W.; Schepartz, A. Design and Evolution of a Miniature pathways of cell killing by gossypol enantiomers. Exp Biol Med
Bcl-2 Binding Protein. Angew. Chem., Int. Ed. 2001, 40, 3806- (Maywood) 2002, 227, 398-401.
3809. (32) Pervushin, K.; Riek, R.; Wider, G.; Wüthrich, K. Attenuated T2
(12) Kutzki, O.; Park, H. S.; Ernst, J. T.; Orner, B. P.; Yin, H.; et al. relaxation by mutual cancellation of dipole-dipole coupling and
Development of a potent Bcl-x(L) antagonist based on R-helix chemical shift anisotropy indicates an avenue to NMR structures
mimicry. J. Am. Chem. Soc. 2002, 124, 11838-11839. of very large biological macromolecules in solution. Proc. Natl.
(13) Wu, D. An overview of the clinical pharmacology and therapeutic
Acad. Sci. U.S.A. 1997, 94, 12366-12371.
potential of gossypol as a male contraceptive agent and in
gynaecological disease. Drugs 1989, 38, 333-341. (33) Koradi, R.; Billeter, M.; Wüthrich, K. MOLMOL: a program for
(14) Gilbert, N. E.; O’Reilly, J. E.; Chang, C. J.; Lin, Y. C.; Bruegge- display and analysis of macromolecular structures. J. Mol.
meier, R. W. Antiproliferative activity of gossypol and gossy- Graphics 1996, 14, 51-55.
polone on human breast cancer cells. Life Sci. 1995, 57, 61-67. (34) Weislow, O. S.; Kiser, R.; Fine, D. L.; Bader, J.; Shoemaker, R.
(15) Wu, Y. W.; Chik, C. L.; Knazek, R. A. An in vitro and in vivo H.; et al. New soluble-formazan assay for HIV-1 cytopathic
study of antitumor effects of gossypol on human SW-13 adreno- effects: application to high-flux screening of synthetic and
cortical carcinoma. Cancer Res. 1989, 49, 3754-3758. natural products for AIDS-antiviral activity. J. Natl. Cancer Inst.
(16) Thomas, M.; von Hagen, V.; Moustafa, Y.; Montmasson, M. P.; 1989, 81, 577-586.
Monet, J. D. Effects of gossypol on the cell cycle phases in T-47D
human breast cancer cells. Anticancer Res. 1991, 11, 1469-1475. JM030190Z