Terephthalamide derivatives as mimetics of helical peptides: Disruption of the Bcl-xL/Bak interaction

Article abstract of DOI:10.1021/ja0446404

A series of Bcl-x_L/Bak antagonists, based on a terephthalamide scaffold, was designed to mimic the α-helical region of the Bak peptide. These molecules showed favorable in vitro activities in disrupting the Bcl-x_L/Bak BH3 domain comp

Full text of DOI:10.1021/ja0446404

Published on Web 03/25/2005
Terephthalamide Derivatives as Mimetics of Helical Peptides:
Disruption of the Bcl-xL/Bak Interaction
Hang Yin,† Gui-in Lee,† Kristine A. Sedey,‡ Johanna M. Rodriguez,†
Hong-Gang Wang,‡ Said M. Sebti,‡ and Andrew D. Hamilton*,†
Contribution from the Department of Chemistry, Yale UniVersity, P.O. Box 208107,
New HaVen, Connecticut 06520-8107, and Drug DiscoVery Program,
H. Lee Moffitt Cancer Center and Research Institute, Departments of Oncology and
Biochemistry and Molecular Biology, UniVersity of South Florida, Tampa, Florida 33612
Received September 3, 2004; E-mail: andrew.hamilton@yale.edu
Abstract: A series of Bcl-xL/Bak antagonists, based on a terephthalamide scaffold, was designed to mimic
the R-helical region of the Bak peptide. These molecules showed favorable in vitro activities in disrupting
the Bcl-xL/Bak BH3 domain complex (terephthalamides 9 and 26, Ki ) 0.78 ( 0.07 and 1.85 ( 0.32 µM,
respectively). Extensive structure-affinity studies demonstrated a correlation between the ability of
terephthalamide derivatives to disrupt Bcl-xL/Bak complex formation and the size of variable side chains
on these molecules. Treatment of human HEK293 cells with the terephthalamide derivative 26 resulted in
disruption of the Bcl-xL/Bax interaction in whole cells with an IC50 of 35.0 µM. Computational docking
simulations and NMR experiments suggested that the binding cleft for the BH3 domain of the Bak peptide
on the surface of Bcl-xL is the target area for these synthetic inhibitors.
Introduction
The development of small-molecule modulators of protein-
protein interactions is regarded as a challenging goal since the
large interfaces involved, typically around 1600 Å2 of buried
area (around 170 atoms), pose a serious hurdle for any small
molecule to be competitive.9 The binding regions of protein
partners are often discontiguous and thus cannot be mimicked
by simple synthetic peptides with linear or extended conforma-
tions. Conventional methods for identifying inhibitors of
protein-protein interactions require much input in the prepara-
tion and screening of a chemical library in order to discover
lead compounds. An alternative approach is to design synthetic
recognition scaffolds that reproduce features of the protein
secondary structure at the interface. We have previously reported
functionalized terphenyls as mimetics of R-helices.10,11 However,
the challenging syntheses and physical properties of terphenyls
prompted us to search for simpler scaffolds that could similarly
mimic the side chain presentation on an R-helix.12 We have
recently reported a group of Bcl-xL inhibitors based on a
terephthalamide scaffold, designed to mimic the R-helical region
Proteins in the B-cell lymphoma-2 (Bcl-2) family play a
critical role in determining whether a cell survives or dies
through a programmed cell death known as apoptosis.1 The
Bcl-2 protein family, comprised of both pro-apoptotic and anti-
apoptotic members, acts as a checkpoint downstream of the
tumor suppressor protein p53,2 and upstream of mitochondrial
membrane rupture and caspase cysteine proteases, which
transduce the apoptotic signal.1 Previous studies showed that
oncogenic mutations induced apoptosis defects through a Bcl-
2-dependent pathway.3 Overexpression of the anti-apoptotic
proteins, such as Bcl-2 and Bcl-xL, can inhibit the potency of
many currently available anticancer drugs by blocking the
apoptotic pathway.4 All of the pro-apoptotic subfamily proteins
possess the minimal death domain BH3. These molecules (Bak,
Bax, Bad, Bid) are able to induce apoptosis through het-
erodimerization with the anti-apoptotic Bcl-2 family members.5
Several low-molecular-weight inhibitors of Bcl-2 (Bcl-xL) have
been identified by screening diverse chemical libraries.6,7 The
rational design of agents that directly mimic the death-promoting
region, the BH3 domain of the pro-apoptotic subfamily of Bcl-2
proteins, is an important alternative to screening as it allows
structure-based optimization of initial hits.8
(6) Tzung, S. P.; Kim, K. M.; Basanez, G.; Giedt, C. D.; Simon, J.; Zimmerberg,
J.; Zhang, K. Y. J.; Hockenbery, D. M. Nat. Cell Biol. 2001, 3, 183-191.
Enyedy, I. J.; Ling, Y.; Nacro, K.; Tomita, Y.; Wu, X. H.; Cao, Y. Y.;
Guo, R. B.; Li, B. H.; Zhu, X. F.; Huang, Y.; Long, Y. Q.; Roller, P. P.;
Yang, D. J.; Wang, S. M. J. Med. Chem. 2001, 44, 4313-4324. Wang, J.
L.; Liu, D. X.; Zhang, Z. J.; Shan, S. M.; Han, X. B.; Srinivasula, S. M.;
Croce, C. M.; Alnemri, E. S.; Huang, Z. W. Proc. Natl. Acad. Sci. U.S.A.
2000, 97, 7124-7129. Lugovskoy, A. A.; Degterev, A. I.; Fahmy, A. F.;
Zhou, P.; Gross, J. D.; Yuan, J. Y.; Wagner, G. J. Am. Chem. Soc. 2002,
124, 1234-1240.
† Yale University.
‡ University of South Florida.
(1) Adams, J. M.; Cory, S. Science 1998, 281, 1322-1326. Reed, J. C.; Nature
1997, 387, 773-776.
(7) Degterev, A.; Lugovskoy, A.; Cardone, M.; Mulley, B.; Wagner, G.;
Mitchison, T.; Yuan, J. Y. Nat. Cell Biol. 2001, 3, 173-182.
(8) Adams, J. M.; Cory, S. Trends Biochem. Sci. 2001, 26, 61-66.
(9) Stites, W. E. Chem. ReV. 1997, 97, 1233-1250.
(2) Lane, D. P. Nature 1992, 358, 15-16.
(3) Graeber, T. G.; Osmanian, C.; Jacks, T.; Housman, D. E.; Koch, C. J.;
Lowe, S. W.; Giaccia, A. J. Nature 1996, 379, 88-91. Fearon, E. R.;
Vogelstein, B. Cell 1990, 61, 759-767.
(10) Orner, B. P.; Ernst, J. T.; Hamilton, A. D. J. Am. Chem. Soc. 2001, 123,
5382-5383.
(4) Strasser, A.; Huang, D. C. S.; Vaux, D. L. Biochim. Biophys. Acta-ReV.
Cancer 1997, 1333, F151-F178.
(11) Kutzki, O.; Park, H. S.; Ernst, J. T.; Orner, B. P.; Yin, H.; Hamilton, A. D.
J. Am. Chem. Soc. 2002, 124, 11838-11839.
(5) Chao, D. T.; Korsmeyer, S. J. Annu. ReV. Immunol. 1998, 16, 395-419.
9
10.1021/ja0446404 CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 5463-5468
5463

A R T I C L E S
Yin et al.
Figure 2. (A) Generic structure of terephthalamide helical mimetics. (B)
The superimposition of 2 on the i, i+4, i+7 positions of an R-helix (RMSD
) 1.03 Å). (C) The superimposition of terphenyl 1 and terephthalamide 2
(RMSD ) 0.34 Å). The functionalized side chains are highlighted.
Figure 1. (A) NMR structure of the Bak BH3 domain complex.14 The
side chains of Val74, Leu78, and Ile81 residues (shown in yellow) play
critical roles in the Bcl-xL/Bak complex formation. (B) Generic structure
of terphenyl R-helical mimetics. An energy-minimized structure shows that
the spatial arrangement of 3,2′,2′′-side chains on the terphenyl scaffold has
a good agreement with those at the i, i+4, and i+7 positions on an R-helix.15
of the phenyl rings, due to the restricted rotation of the amide
bonds. Another conformational constraint in the molecule was
imposed by an intramolecular hydrogen bond between the amide
-NH and the alkoxy oxygen atom, to influence the position of
the amino acid side chain R4 (Figure 2A). The modularity of
the inhibitor synthesis is strengthened since terphenyl 1 requires
difficult C-C bond formations to attach the side chains to the
scaffold and to assemble functionalized phenyl building blocks.10,11
Terephthalamide derivatives, on the other hand, were synthe-
sized using O-alkylation and amide formations, accelerating the
process of preparing these compounds.
The newly introduced carboxamide groups also increase the
polarity of the inhibitors. One limitation of the terphenyl
derivatives is their high hydrophobicity. Calculations using the
QikProp program predicted that terephthalamide 26 has a logP
(partition coefficient for n-octanol/water) of 4.42, compared to
of the Bak peptide.13 Using a fluorescence polarization assay,
we have observed high in vitro inhibition potencies in disrupting
the Bcl-xL/Bak BH3 domain complex and a significant im-
provement in water solubility relative to the terphenyl deriva-
tives. In this report, we present an expanded structure-affinity
study that demonstrates a correlation between the potency of
the Bcl-xL/Bak disruption and the size of side chains on these
molecules. Computational docking simulations and NMR
experiments are used to confirm the binding mode of the
terephthalamide inhibitors and suggest that the binding cleft for
the Bak BH3 domain on the surface of Bcl-xL is the target area.
Treatment of human HEK293 cells with one of the terephthala-
mide derivatives resulted in inhibition of the association of Bcl-
xL with Bax in whole cells.
9.25 for terphenyl 1 (R1
) R3 ) iBu, R2 ) 1-naphthylmeth-
ylene), suggesting that terephthalamides should have more
favorable properties within an acceptable log P range (-2 to
6).16,17 Structure-based solubility prediction showed that 26 has
a log S (aqueous solubility) of -5.35, compared to -10.76 for
1, suggesting that terephthalamide derivatives should have
improved solubility in aqueous solution.18
The preferred conformation of terephthalamide derivatives
in solution was investigated using energy minimization with
the MM2 force field (Macromodel 7.0). Parts B and C of Figure
2 show, respectively, superimpositions of the energy-minimized
conformation of 2 on the i, i+4, i+7 side chains of an R-helix
(with a root-mean-square deviation (RMSD) value of 1.03 Å)
and on a terphenyl scaffold (RMSD ) 0.34 Å), suggesting good
stereochemical similarity between the two pairs.19
Results and Discussion
Inhibitor Design. The NMR-derived structure of the Bcl-
xL/Bak BH3 domain complex indicates that the Bak BH3
peptide is an amphipathic R-helix that interacts with Bcl-xL by
projecting hydrophobic side chains (Val74, Leu78, Ile81, and
Ile85) on one face of the R-helix into a hydrophobic cleft of
Bcl-xL (Figure 1A).14 We have reported terphenyl derivatives
(1) as structural mimetics of the Bak BH3 helical peptide as
they adopt a staggered conformation and project appropriately
positioned substituents in a manner similar to an R-helix (Figure
1B).11
The design of a next generation of helical mimetics should
maintain the similarity between the arrangement of the i, i+4,
i+7 side chains of an R-helix and the substituents on 3,2′,2′′-
positions on terphenyl 1, while minimizing the structural
complexity and increasing the rigidity and solubility of the
inhibitors. This strategy of simplifying a proven proteomimetic
was accomplished by using terephthalamide 2 as the scaffold.
The flanking phenyl rings in 1 were replaced by two function-
alized carboxamide groups, which retain the planar geometry
Two strategies were used to orient the interacting N-alkyl
group R2 of the lower tertiary amide into the desired Z-
conformation. The first was to avoid the problem by using
identical substituents (R1 ) R2) on the tertiary amide nitrogen.
(15) Energy minimization was carried out using MM2 force field within
Macromodel V7.0.
(16) Jorgensen, W. L.; Duffy, E. M. AdV. Drug DeliVer. ReV. 2002, 54, 355-
366; Jorgensen, W. L.; Duffy, E. M. Bioorg. Med. Chem. Lett. 2000, 10,
1155-1158.
(17) Jorgensen, W. L. Science 2004, 303, 1813-1818.
(18) The log P and log S calculations were carried out using QikProp V2.1
(Shrodinger Inc., New York, 2003). Huuskonen et al. have shown that 85%
of drugs have log S values between -1 and -5, and virtually none has a
value below -6 (J. Chem. Inf. Comput. Sci. 1998, 38, 450-456).
(19) Ernst, J. T.; Becerril, J.; Park, H. S.; Yin, H.; Hamilton, A. D. Angew.
Chem.-Int. Edit. 2003, 42, 535-550.
(12) Lipinski, C. A.; Lombardo, F.; Dominy, B. W.; Feeney, P. J. AdV. Drug
DeliVer. ReV. 1997, 23, 3-25.
(13) Yin, H.; Hamilton, A. D. Bioorg. Med. Chem. Lett. 2004, 14, 1375-1379.
(14) Sattler, M.; Liang, H.; Nettesheim, D.; Meadows, R. P.; Harlan, J. E.;
Eberstadt, M.; Yoon, H. S.; Shuker, S. B.; Chang, B. S.; Minn, A. J.;
Thompson, C. B.; Fesik, S. W. Science 1997, 275, 983-986.
9
5464 J. AM. CHEM. SOC. VOL. 127, NO. 15, 2005

Terephthalamides as Mimetics of Helical Peptides
A R T I C L E S
In a second series of derivatives, such as 22, steric differentiation
of the substituents favored the placement of the less bulky group
R3 in the Z-position. The consequence of these constraints is
that in low-energy, accessible conformations of 22, the three
substituents project from one face of the terephthalamide
scaffold in a manner analogous to the terphenyl helix mimetics
(e.g. Figure 2A).
Conformational Studies. The intramolecular hydrogen bond
between the amide -NH and the alkoxy oxygen atom of 9
ensures that the 2-isopropoxy group and the upper isobutyl side
chain are positioned on the same side of the terephthalamide.
This hydrogen bond was confirmed by proton NMR experi-
ments, which showed very little change in the amide -NH
resonance (δ ) 8.54 ppm) on heating (∆δ ) 1.54 ppb/K) or
changing concentration. As a comparison, 2-isopropylamino-
terephthalamide 17 showed both concentration (7.36 ppm, 0.5
M in CDCl3; 6.58 ppm, 0.05 M in CDCl3; 6.46 ppm, 0.005 M
in CDCl3, 298 K) and temperature (∆δ ) 5.5 ppb/K) depen-
dence of the aniline proton, suggesting inter- rather than
intramolecular hydrogen bonding.
The conformation of the lower tertiary amide in 22 in solution
1
was probed by computational simulations and H NMR spec-
troscopy. MM2 energy minimization using Macromodel sug-
gested that the Z-conformation is favored by 8.01 kJ/mol in
water solution (Figure 3A), and indeed NMR integration
indicated that 72% of 22 adopted the Z-conformation in
chloroform solution. Computer simulation showed that Hb and
Hc have similar distances to the ortho-proton Ha in the Z- and
E-conformations (2.52 and 2.62 Å, respectively). However, only
the nuclear Overhauser effect (NOE) cross-peaks between Hb
and the ortho-aryl proton Ha were detected, while no significant
NOE effect could be seen between Hc and Ha, suggesting the
Z- is the major conformation in solution. Rotational Overhauser
effect spectroscopy (ROESY) confirmed the presence of both
Z- and E-conformations. Correlations corresponding to the
chemical exchange of Hb and Hc were observed in the ROESY
experiment (Figure 3B), which indicated that both conformations
of 22 exist in DMSO solution at 298 K. Furthermore, the signals
of both Ha and Hb, which are split at room temperature,
coalesced at 353 K. These combined experimental results
suggest that both Z- and E-amide conformations are present with
the Z-conformation being favored.
In Vitro Fluorescence Polarization Assay. The binding af-
finity of the terephthalamide molecules for Bcl-xL was assessed
by a previously reported fluorescence polarization assay using
a fluorescently labeled 16-mer Bak-peptide (Fl-GQVGRQLAI-
IGDDINR-CONH2).14 Displacement of this probe through
competitive binding of the terephthalamide into the hydrophobic
cleft of Bcl-xL leads to a decrease in its fluorescence polariza-
tion. Computational regression analysis was conducted to
determine the IC50 values, which in turn can be related to the
known affinity of the 16-mer Bak peptide (Kd ) 120 nM) to
acquire the inhibitory constant (Ki) of the inhibitors.11,20 To test
the validity of this assay, we used a nonlabeled Bak-peptide as
the competitive inhibitor to bind Bcl-xL, giving a Ki of 122 nM,
which closely corresponds to the Kd value obtained from the
saturation titration experiment.
Figure 3. (A) Energy minimized Z- and E-isomers of 22. (B) ROESY
1H-1H NMR experiments showed cross-peaks corresponding to the
chemical exchange of Hc.
terephthalamide solution in 10 mM PBS buffer (pH ) 7.4, 298
K) with less than 0.1% DMSO, a testament to the good solubility
of terephthalamides in water. Table 1 shows that terephthalamide
9 has good affinity for Bcl-xL with a Ki value of 0.78 ( 0.07
µM. By screening compounds with a range of side chains on
the upper carboxamide, we found that the isobutyl group as the
upper substituent provided the best inhibition results (9, 10, 29,
34). The newly introduced stereogenic center in the tereph-
thalamide did not affect the affinity, as seen by comparing 26
and 27. The optimal alkoxy group in the 2-position of tereph-
thalamide was found to be isopropoxy (9, 10, 34, 37), which
closely mimics the size of Leu78 of the Bak peptide; both larger
(29, 31) and smaller (26, 27) substituents gave decreased
affinities. The N,N-alkyl positions on the lower carboxamide
were shown to favor the medium to small substituents since
N,N-dimethyl (34), -diethyl (37), and -diisopropyl (9, 10)
terephthalamide analogues have low micromolar Ki values while
most of the affinity was lost when the alkyl substituents were
replaced by phenyl groups (42-45). Comparison of the tereph-
thalamide derivatives with the free amino acid in the upper
carboxamide moiety and their methyl esters suggested that the
ester group did not significantly affect the binding (9, 10, and
26-28). However, for less hydrophilic compounds, such as 31
and 32, only the free acid derivatives are soluble in aqueous
solution. The 2-isopropylamino terephthalamide 17 showed
A series of terephthalamides with varied side chains were
prepared. All the assays were carried out with 0.01-1000 µM
(20) Cheng, Y.; Prusoff, W. H. Biochem. Pharmacol. 1973, 22, 3099-3108.
9
J. AM. CHEM. SOC. VOL. 127, NO. 15, 2005 5465

A R T I C L E S
Yin et al.
Table 1. Results of the Fluorescence Polarization Competition
Assay of Terephthalamide Derivatives as the Antagonists of
Bcl-xL/Bak Complex
R1
X
R2
R3
entry Ki ± SD (µ
M)a
-iPr -OiPr
-iBu
-iBu
-Me
-iPr
-Bn
-Me
-H
9
0.78 ( 0.07
10 0.84 ( 0.17
-Me 23 11.0 ( 1.9
-Me 24 5.85 ( 2.07
-Me 25 10.9 ( 5.17
-OMe
-OPh
(S)-iBu -Me 26 1.85 ( 0.32
(R)-iBu -Me 27 1.79 ( 0.39
(S)-iBu -H
-iBu
-iBu
-iBu
28 2.31 ( 0.33
29 1.01 ( 0.05
-H
-Me 30 not soluble
-H
-Me 32 13.2 ( 8.8
-H
-O(2-naphthyl)
-O(p-nitrophenyl) -iBu
-NH(iPr)
-Me -OiPr
31 7.63 ( 1.02
-iBu
-Me
-iBu
-Bn
-H
-Me
-iBu
-Me
-iBu
-iPr
-Me
-iBu
-iBu
-Bn
17 3.31 ( 0.35
-Me 33 8.34 ( 1.61
-Me 34 3.14 ( 0.31
-H
35 8.92 ( 3.59
H
-Me 36 no affinity
-Me 37 2.44 ( 0.95
-Me 38 11.4 ( 5.1
-Me 39 8.76 ( 2.78
-Me 40 8.36 ( 3.80
-Me 41 138 ( 50
-Et -OiPr
-iBu -OiPr
-Ph -OiPr
-H
-H
42 260 ( 58
43 780 ( 174
-Me 44 no affinity
-H
45 no affinity
R1
R2
R3
R4
Entry
Ki ± SD (µM)
-Me -iBu -iPr
-H -H -H
-iPr -iBu -iPr
-H
-H
-H
-H
22 no affinity
46 no affinity
47 450 ( 231
48 no affinity
-iBu -H
-iBu -H
-iBu -H
-iBu -H
-Ph
-Ph
-Me 49 not soluble
50 82.1 ( 11.0
-(2-naphthyl) -Me 51 not soluble
-(2-naphthyl) -H
Figure 4. (A) 15N HSQC spectrum recorded on 15N-labeled Bcl-xL protein
(2 × 10-3 M solution in PBS) in the absence (blue) and in the presence of
0.5 equiv (cyan) and 1.0 equiv (red) of terephthalamide 26 added. Residues
A89, G94, Y101, R102, L108, L112, A119, R139, and A142 show
significant chemical shift change upon the addition of 26. (B) Superimposi-
tion of top ranked result of the molecular-docking studies of 26 and Bak
peptide. The residues that showed significant and moderate chemical shift
change upon the addition of 26 are shown in magenta and yellow,
respectively. The critical binding residues V74, L78, and I81 of the Bak
peptide are shown in stick representation.
affinity 4-fold less than its 2-isopropoxy analogue 10, suggesting
the intramolecular hydrogen bond in 10 helps to orient the side
chain and in turn to enhance binding. The importance of
hydrophobic side chains was further confirmed by the weak
binding of 36, which lacks the key substituents. The series of
terephthalamide derivatives with asymmetric substituents on the
lower carboxamide group did not show activity in disrupting
the Bcl-xL/Bak interaction (22, 46-49, 51) with an exception
of 50, possibly due to a different binding mode. These assay
results confirmed that the terephthalamide derivatives retained
the high in vitro affinity of the original terphenyl scaffold while
reducing its complexity.
(15N,1H)-2D-Amide Chemical Shift Perturbation Experi-
ments. The displacement of Bak peptide by terephthalamide in
the fluorescence polarization suggests that the inhibitor and the
peptide bind to the same surface area of Bcl-xL protein. To probe
the binding mode of the terephthalamide, (15N,1H)-2D-amide
chemical shift perturbation mapping with 15N-labeled Bcl-xL
was performed. Addition of 26 led to shifts in a number of
residues on the surface of Bcl-xL. The residues of A89, G94,
Y101, R102, L108, L112, A119, R139, and A142 showed
significant chemical shift changes upon the addition of the
synthetic inhibitor 26 (shown in magenta in Figure 4B). Some
other residues, including E96, H113, I114, V127, I140, S145,
and V161 (shown in yellow in Figure 4B), showed moderate
chemical shift change under the same conditions. These affected
residues all lie in the shallow cleft on the protein into which
the Bak helix binds. Overlay of 26 and the Bak peptide within
the binding pocket suggests that the terephthalamide is indeed
mimicking the cylindrical shape of the helix with the substituents
making a series of hydrophobic contacts with the protein (Figure
4B). The residues V74, L78, and I81 of Bak BH3, which the
terephthalamide has been designed to mimic, are within 4 Å of
residues F97, R102, L108, I140, and A142 of Bcl-xL, all of
which showed significant chemical shift changes upon addition
of helical mimetic 26. The effect on F97 was unclear due to
overlap with N5 although it seems to shift significantly. These
results confirmed that terephthalamide 26 targets the same cleft
which the Bak BH3 peptide recognizes.
1
Comparison of these results with 15N, H-HSQC mapping
of the previously reported terphenyl 1 (R1 ) R3 ) iBu, R2 )
1-naphthylmethylene) suggested a similar binding mode to Bcl-
xL for both proteomimetics.11 The residues most affected by
both of the inhibitors are G94 in BH3, R102 and 108 in the
C-terminal region of BH2, and R139, I140, and A142 in BH1
9
5466 J. AM. CHEM. SOC. VOL. 127, NO. 15, 2005

Terephthalamides as Mimetics of Helical Peptides
A R T I C L E S
Scheme 1. Modular Synthesis of Terephthalamide Derivatives 9
and 10a
Figure 5. Results of the double transfection cellular assay (each percentage
value is an average of three independent experiments).
domain of Bcl-xL. These residues are all found in the conserved
hydrophobic pocket on Bcl-xL where the Bak BH3 helix binds.14
The chemical shift perturbation of these or their neighboring
residues listed above indicated that both inhibitors and the Bak
BH3 peptide target the same area on the Bcl-xL exterior surface,
suggesting that terephthalamide is a successful alternative
scaffold to terphenyl as a mimetic of R-helices.
a Reagents and conditions: (a) 2-iodopropane, K2CO3, acetone, reflux;
(b) NaNO2, H2SO4, MeOH, H2O, 0 °C; (c) KI, Cu (bronze), reflux; (d)
tributyl(vinyl)tin, Pd(PPh3)4, toluene, reflux; (e) NaOH(aq), MeOH; (f)
(COCl)2, DMF, CH2Cl2; (g) (iPr)2NH, CH2Cl2; (h) OsO4, NaIO4, tBuOH:
CCl4:H2O (2:1:1); (i) pyridinium dichromate, DMF; (j) L-leucine methyl
ester hydrochloride, 1-hydroxybenzotriazole hydrate, 1-[3-(dimethylami-
no)propyl]-3-ethylcarbodiimide hydrochloride; (k) KOH, MeOH.
Computational Docking Studies. A computational docking
study using the Autodock 3.0 program lent further support to
the proposal that the binding cleft for the BH3 domain of the
Bak peptide on the surface of Bcl-xL is the target area for the
synthetic inhibitors. Over 90% of the conformational search
results showed the terephthalamide docked to this region. Figure
4B shows the overlay of the top-ranked docking result, which
was predicted to have a favored -10.28 kcal/mol mean docked
energy, with the BH3 domain of the Bak peptide in the Bcl-
xL/Bak complex, suggesting that the side chains of the tereph-
thalamide scaffold have an analogous spatial arrangement to
the three key alkyl side chains of the Bak BH3 peptide.
Inhibition of Bcl-xL/Bax Association in Intact Human
Cells. Human HEK293 cells were used to determine if tereph-
thalamide derivatives inhibit the Bcl-xL/BH3 domain interaction
in a cellular environment. HEK293 cells transfected with both
HA-Bcl-xL and flag-Bax, an analogue to Bak, were treated with
terephthalamide derivatives. After 24 h incubation, the cells were
harvested and lysed. HA-tagged Bcl-xL protein was collected
via immunoprecipitation with HA antibody. The resulting
immunoprecipitates were loaded on to a 12.5% SDS-PAGE
gel, and proteins transferred to nitrocellulose for western blots
analysis. The presence of Bax protein was probed with anti-
flag antibody. The inhibitory potencies of the terephthalamide
compounds were determined by measuring the relative intensity
of Bax protein bound to Bcl-xL. Treatment of HEK293 cells
with terephthalamide 26 (100 µM) resulted in inhibition of the
association of Bcl-xL with Bax with an average value of 81.1
( 13.5% (Figure 5). Treatment of HEK293 cells with various
concentrations of 26 resulted in an IC50 of 35 µM. Terphenyl 1
(R1 ) R3 ) iBu, R2 ) 1-naphthylmethylene) inhibited the
Bcl-xL/Bax association by only 50.9 ( 34.5% at a concentration
of 100 µM. This result suggests that terephthalamide 26 is able
to perturb Bcl-xL/Bax interaction in a cellular environment and
that it has good membrane uptake properties.
confirmed by computational simulations and NMR spectroscopy.
Experiments using intact human cells showed that the tereph-
thalamides are able to disrupt the Bcl-xL/Bax BH3 domain
interaction in cell-based assays.
Experimental Section
Inhibitor Preparation. A modular synthesis of terephthalamide
derivative 10 is shown in Scheme 1. The 2-isopropoxy group was
introduced by O-alkylation. Sandmeyer reaction was used to introduce
the iodo-substituent in 5. A vinyl group was installed through Stille
coupling, followed by hydrolysis of the methyl ester to generate a
carboxylic acid at the 4-position. The lower amide bond formation was
accomplished using diisopropylamine to attack the corresponding acid
chloride intermediate to afford 7. The 1-vinyl group was converted to
a carboxylic acid by Lemieux-Johnson and Corey-Schmidt oxidation.
Terephthalamide 9 was obtained by using standard peptide coupling
of 8 and L-leucine.
Fluorescence Polarization Assays. Fluorescence polarization ex-
periments were conducted on a Photon Technology International
instrument using a 0.3 cm path length cuvette. Spectra were measured
at 25 °C using 10.0 nm slit widths. Fluoroscein-labled Bak peptide
(Fl-GQVGRQLAIIGDDINR-CONH2) was purchased from the HHMI
Biopolymer/Keck Foundation Biotechnology Resource Center at the
Yale University School of Medicine (New Haven, CT). The N-terminus
of the peptide was capped with the fluorophore and the C-terminus
was amidated. Bcl-xL was expressed and purified as previously
described.7 Excitation at 495 nm was used for the fluorescein-containing
peptide and the excimer emission maximum at 535 nm was monitored.
Polarization measurements were recorded upon titration of inhibitors
(ca. 10 mM stock solutions in DMSO) at varying concentrations into
a solution of 15 nM Fl-Bak and 184 nM Bcl-xL (25 °C, 10.0 mM PBS,
pH 7.4). Regression analysis was carried out using SigmaPlot 2001
(Systat Co.) ligand binding macro module. Experimental data were fitted
into eq 1 to determine the IC50 values, which in turn can be related to
the known affinity of the 16-mer Bak peptide (Kd ) 120 nM) to acquire
the inhibitory constant Ki.11,20
In conclusion, a novel family of Bcl-xL antagonists, based
on a terephthalamide mimic of R-helical presentation of side
chains, has been developed. A library of terephthalamide
derivatives was prepared in a modular fashion and good in vitro
affinities were observed using fluorescence polarization binding
assays. The binding mode of a potent terephthalamide was
g IC50
y ) min + (max - min)/(1 + 10x-l
°
)
(1)
(y ) total binding, x ) log concentration of ligand, min ) nonspecific
binding, max ) maximum binding in absence of ligand)
9
J. AM. CHEM. SOC. VOL. 127, NO. 15, 2005 5467

A R T I C L E S
Yin et al.
(15N,1H)-Amide Chemical Shift Pertubation Experiments. (15N,1H)-
2D-HSQC spectra were recorded on a Varian DPX-600 spectrometer.
The concentration of Bcl-xL was 2.0 mM, with 0, 1.0 mM, and 2.0
mM 26, respectively, in 10% DMSO/D2O (25 °C, 10.0 mM PBS, pH
7.0).
(6 µL) was added dropwise into 100 µL of serum free RPMI medium
and incubated at room temperature for 20 min. A total of 2 µg of
pcDNA3 HA-Bcl-xL and 2 µg of pCMV2 Flag-Bax were added to the
diluted transfection reagent and mixed by gentle pipetting. The
transfection reagent/DNA complex was added dropwise to the cells
and the cells were gently rocked. The cells were then incubated for 24
h with media solution containing various concentrations of tereph-
thalamide compounds. The cells were scraped in PBS and lysed in
NP-40 lysis buffer. HA-tagged Bcl-xL protein was collected via
immunoprecipitation with HA antibody, washed and resuspended in
Laemmli buffer (2×). The resulting mixture was loaded on to a 12.5%
SDS-PAGE gel for protein separation, then transferred to nitrocellulose
for western blots analysis. The presence of Bax was probed with anti-
flag antibody. The inhibitory potency of the terephthalamide compounds
was determined by measuring the relative intensity of Bax protein bound
to Bcl-xL.
Docking Studies. The docking studies were performed using
AutoDock 3.0.21 A genetic algorithm (Lamarckian genetic algorithm,
or LGA for short) was used and the torsion angles of the ligand were
varied using AUTOTORS. All other procedures for the docking
experiment were followed as described in the users manual for the
AUTODOCK program. Docked conformations were ranked automati-
cally by the AUTODOCK program using a force field scoring function.
A total of 100 distinct conformational clusters were found out of 100
runs using an rmsd tolerance of 1.0 Å. Among those, one of the highest
ranked docked structures was used for molecular visualization.
Disruption of Bcl-xL/Bax Association in Whole Cells. The anti-
flag antibody (M2 mouse monoclonal) was purchased from Sigma-
Aldrich. HEK293 cells were plated at an appropriate density 24 h prior
to transfection and incubated overnight. Mirus LT-1 transfection reagent
Acknowledgment. We thank the National Institutes of Health
for financial support of this work.
Supporting Information Available: Scheme S1-S3, general
synthesis procedures, compound characterization. This material
is available free of charge via the Internet at http://pubs.acs.org.
(21) Morris, G. M.; Goodsell, D. S.; Huey, R.; Olson, A. J. J. Comput.-Aided
Mol. Des. 1996, 10, 293-304. Morris, G. M.; Goodsell, D. S.; Halliday,
R. S.; Huey, R.; Hart, W. E.; Belew, R. K.; Olson, A. J. J. Comput. Chem.
1998, 19, 1639-1662. Goodsell, D. S.; Olson, A. J. Proteins: Struct., Funct.
Genet. 1990, 8, 195-202.
JA0446404
9
5468 J. AM. CHEM. SOC. VOL. 127, NO. 15, 2005