A Wrench-Shaped Synthetic Molecule that Modulates a Transcription Factor-Coactivator Interaction

Article abstract of DOI:10.1021/ja038855+

Development of synthetic molecules that provide external control over the transcription of a given gene represents a challenge in medicinal and bioorganic chemistry. Here we report design and analysis of wrenchnolol, a wrench-shaped synthetic molecule that impairs the transcription of the Her2 oncogene by disrupting association of transcription factor ESX with its coactivator Sur-2. The "jaw" part of the compound mimics the α-helical interface of the activation domain of ESX, and the "handle" region accepts chemical modifications for a range of analysis. A water-soluble handle permitted NMR study in aqueous solution; a biotinylated handle verified the selectivity of the interaction, and a fluorescent handle confirmed the cell permeability of the compound. The case study of wrenchnolol foreshadows the promise and the challenge of targeting protein-protein interactions in the nucleus and may lead to the development of unique synthetic modulators of gene transcription.

Full text of DOI:10.1021/ja038855+

Published on Web 02/24/2004
A Wrench-Shaped Synthetic Molecule that Modulates a
Transcription Factor-Coactivator Interaction
Hiroki Shimogawa,†,‡ Youngjoo Kwon,† Qian Mao,† Yoshinori Kawazoe,†
Yongmun Choi,† Shinichi Asada,† Hideo Kigoshi,*,‡ and Motonari Uesugi*,†
Contribution from The Verna and Marrs McLean Department of Biochemistry and Molecular
Biology, Baylor College of Medicine, Houston, Texas 77030, and The Department of Chemistry,
UniVersity of Tsukuba, Tsukuba, Ibaraki 305-8571, Japan
Received October 3, 2003; E-mail: muesugi@bcm.tmc.edu; kigoshi@chem.tsukuba.ac.jp
Abstract: Development of synthetic molecules that provide external control over the transcription of a given
gene represents a challenge in medicinal and bioorganic chemistry. Here we report design and analysis of
wrenchnolol, a wrench-shaped synthetic molecule that impairs the transcription of the Her2 oncogene by
disrupting association of transcription factor ESX with its coactivator Sur-2. The “jaw” part of the compound
mimics the R-helical interface of the activation domain of ESX, and the “handle” region accepts chemical
modifications for a range of analysis. A water-soluble handle permitted NMR study in aqueous solution; a
biotinylated handle verified the selectivity of the interaction, and a fluorescent handle confirmed the cell
permeability of the compound. The case study of wrenchnolol foreshadows the promise and the challenge
of targeting protein-protein interactions in the nucleus and may lead to the development of unique synthetic
modulators of gene transcription.
Introduction
scription and might serve as a potential therapeutic strategy.
Although inhibiting protein-protein interactions with small
Regulation of gene expression by transcription factors touches
many aspects of eukaryotic biology. The biological importance
of transcriptional regulation has fueled efforts aimed at discov-
ering means of controlling gene transcription by synthetic
molecules. Small-molecule ligands of nuclear receptor transcrip-
tion factors are among the most successful examples of such
molecules and have led to the development of a number of
pharmaceuticals.1,2 Inhibitors of histone deacetylases that modu-
late gene transcription have also proven to be useful in treating
certain types of human diseases.3 Furthermore, approaches that
address the interactions between DNA and transcription factors
have been tested, exemplified by the design of synthetic
polyamide molecules that display sequence-specific DNA
binding at nanomolar concentrations.4
molecules is generally difficult, direct contacts of a transcription
factor with its coactivators are often mediated by a short
R-helical segment of the activation domain,7-12 suggesting that
inhibition by small organic molecules is a reasonably tractable
problem.
A good example may be the association between the
activation domain of ESX transcription factor (ESE-1/ELF3/
ERT/Jen) and its coactivator Sur-2 (a Ras-linked subunit of the
human mediator complex).13 ESX is an epithelial-specific
transcription factor that activates Her2, an oncogene whose
overexpression occurs in ∼30% of breast cancer patients.14,15
ESX binds and strongly activates the Her2 promoter,16 and the
ESX-binding site in the Her2 promoter is required for high-
level expression of Her2 in breast cancer cells.17 The interaction
An alternative approach to controlling gene transcription is
disruption or manipulation of the interaction between a tran-
scription factor and a so-called coactivator protein. Physical
association with a coactivator enables a DNA-binding transcrip-
tion factor to stimulate transcription by recruiting the compo-
nents of transcriptional machinery, including chromatin-
remodeling proteins, to specific promoters.5,6 Pharmacological
intervention in these interactions would modulate gene tran-
(7) Uesugi, M.; Nyanguile, O.; Lu, H.; Levine, A. J.; Verdine, G. L. Science
1997, 277, 1310-1313.
(8) Uesugi, M.; Verdine, G. L. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 14801-
14806.
(9) Choi, Y.; Asada, S.; Uesugi, M. J. Biol. Chem. 2000, 275, 15912-15916.
(10) Freedman, S. J.; Sun, Z. Y.; Poy, F.; Kung, A. L.; Livingston, D. M.;
Wagner, G.; Eck, M. J. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 5367-
5372.
(11) Dames, S. A.; Martinez-Yamout, M.; De Guzman, R. N.; Dyson, H. J.;
Wright, P. E. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 5271-5276.
(12) Radhakrishnan, I.; Perez-Alvarado, G. C.; Parker, D.; Dyson, H. J.;
Montminy, M. R.; Wright, P. E. Cell 1997, 91, 741-752.
(13) Asada, S.; Choi, Y.; Yamada, M.; Wang, S. C.; Hung, M. C.; Qin, J.;
Uesugi, M. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 12747-12752.
(14) Slamon, D. J.; Clark, G. M.; Wong, S. G.; Levin, W. J.; Ullrich, A.;
McGuire, W. L. Science 1987, 235, 177-182.
(15) Slamon, D. J.; Godolphin, W.; Jones, L. A.; Holt, J. A.; Wong, S. G.;
Keith, D. E.; Levin, W. J.; Stuart, S. G.; Udove, J.; Ullrich, A.; et al. Science
1989, 244, 707-712.
(16) Chang, C. H.; Scott, G. K.; Kuo, W. L.; Xiong, X.; Suzdaltseva, Y.; Park,
J. W.; Sayre, P.; Erny, K.; Collins, C.; Gray, J. W.; Benz, C. C. Oncogene
1997, 14, 1617-1622.
† Baylor College of Medicine.
‡ University of Tsukuba.
(1) Schapira, M. Curr. Cancer Drug Targets 2002, 2, 243-256.
(2) Rosen, J.; Marschke, K.; Rungta, D. Curr. Opin. Drug DiscoV. DeV. 2003,
6, 224-230.
(3) Johnstone, R. W. Nat. ReV. Drug DiscoV. 2002, 1, 287-299.
(4) Dervan, P. B.; Burli, R. W. Curr. Opin. Chem. Biol. 1999, 3, 688-693.
(5) Ptashne, M.; Gann, A. Nature 1997, 386, 569-577.
(6) Tjian, R.; Maniatis, T. Cell 1994, 77, 5-8.
9
10.1021/ja038855+ CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 3461-3471
3461

A R T I C L E S
Shimogawa et al.
Chart 1
Structure-Activity Relationship in Cells and in Vitro. The
structure-activity relationship of adamanolol derivatives is
summarized in Table 1. In one group of derivatives (the R1
substituents in Table 1), the indole ring of adamanolol was
substituted. Replacement with a methyl group (12a) abolished
the ability of adamanolol to kill Her2-positive SK-BR3 breast
cancer cells (Table 1) and block the interaction of ESX with
Sur-2 in vitro (Figure 1), while the derivatives with indole-like
pharmacores (5a-f, 12b) retained the biological activity (Table
1). Introduction of an arylsulfonyl group, especially a tosyl group
(5a), at the N1 position of the adamanolol indole ring increased
the potency in cells and Sur-2 binding activity in vitro,
reminiscent of the improved 5-HT receptor-binding activity of
N-arylsulfonyltryptamines.19 Although full optimization of the
R1 group is still only partly achieved, the observed importance
of the indole-mimicking structures is in good agreement with
the reported contribution of a tryptophan residue to the specific
interaction of ESX with Sur-2.
In another series of derivatives (the R3 substituents in Table
1), the adamantane group of adamanolol was substituted (11a-
d). Replacement either with thienyl (11c), tolyl (11d), or methyl
(11a) groups resulted in significant loss of the biological activity
(Table 1), and the derivative with a methyl group (11a) exhibited
almost no ability to block the ESX-Sur-2 interaction in vitro
(Figure 1). The only substitutent that retained a comparable level
of activity was a biphenyl group (11b), a chemical module often
used to mimic nonpolar amino acids.20-22 The bulky, hydro-
phobic adamantane group may mimic the cluster of isoleucines
and leucines on the face of the ESX helix and perhaps
participates in binding to the hydrophobic pocket in Sur-2.
The isopropyl group extended from the urea junction of
adamanolol was also replaced by a range of bulky substituents
(the R2 substituents in Table 1). These derivatives (4a-c,e) had
biological activity comparable to that of adamanolol in cells
and inhibited the ESX-Sur-2 interaction in vitro as much as
adamanolol. By contrast, simple removal of the isopropyl group
(4d) abolished both the biological and Sur-2 binding activity.
The analogous consequences of all the bulky substituents and
the complete loss of the activity in 4d highlight the importance
of the bulkiness of the R2 substituents for activity. A bulky
substituent at the R2 position enforces the s-cis configuration
around the urea linker, which may bring the indole ring and
the adamantane group into close proximity and could form a
helix-like nonpolar surface for the interaction with Sur-2.
Design of “Wrenchnolol”. The structure-activity relation-
ship of adamanolol was translated into the design of the second-
generation compound that we named wrenchnolol (Figure 2A).
As expected, the wrench-shaped, water-soluble molecule in-
hibited the ESX-Sur-2 interaction in vitro more strongly than
adamanolol, presumably due to the presence of its N-tosyl group
(Figure 1), and was no less active than adamanolol in killing
SK-BR3 cells (IC50 ) 6.9 µM) despite its increased hydrophi-
licity (Table 1). Wrenchnolol impaired the ability of the ESX
of ESX with Sur-2 is mediated by one face of an eight-amino
acid R-helical region in the ESX activation domain (Ser-Trp-
Ile-Ile-Glu-Leu-Leu-Glu), and the tryptophan residue in the
hydrophobic face of the helix makes a unique contribution to
the specificity of the interaction.13 A small-molecule inhibitor
of the interaction has been discovered by screening a chemical
library enriched in indole-mimicking π-electron-rich pharma-
cores and named adamanolol (Chart 1).18 The drug-like com-
pound inhibits the ability of the ESX activation domain to
stimulate transcription in cells, blocks the interaction of ESX
with Sur-2 in vitro, represses the expression of Her2 gene in
cells, and impairs the viability of Her2-positive breast cancer
cell lines. Adamanolol is a unique synthetic molecule that
controls gene expression by modulating a transcription factor-
coactivator interaction.
Detailed analysis of unique inhibitors of protein-protein
interactions provides a basis for the future endeavor in the
challenging field of drug discovery. Here we describe structure-
activity relationships of adamanolol derivatives that were used
to gain insight into how it disrupts the interaction of the ESX
activation domain with Sur-2. The information gained from these
experiments was translated into the design of wrenchnolol, a
second-generation compound of adamanolol whose wrench-like
structure permitted further biochemical and biophysical analysis.
Wrenchnolol may also serve as a starting point for designing
unique synthetic modulators of gene transcription.
Results
Synthesis of Adamanolol and Its Derivatives. The synthesis
of adamanolol and its derivatives is summarized in Scheme 1.
Coupling of 4-hydroxyindole and epichlorohydrin afforded
glycidyl 4-indolyl ether (1a) quantitatively. The ether 1a was
reacted with isobutylamine, tert-butylamine, ammonia, or mono-
N-Boc-1,3-propanediamine to give the corresponding amines
2a-d. Pindolol, amines 2a-d, propranolol, N-(3-methoxy-2-
hydroxypropyl)isopropylamine, or N-[3-(1-naphthyloxy)-2-hy-
droxypropyl]isopropylamine were then coupled with amide 3
(see below) through the conversion of 3 to an imidazolide and
its N-activation by methyl iodide. The coupling resulted in the
formation of a urea linkage to afford adamanolol and its
analogues 4a-e and 12a-b. The sulfonyl analogues of ada-
manolol (5a-f) were prepared by reacting adamanolol with the
corresponding sulfonyl chlorides.
The amide 3 was obtained by coupling adamantanecarbonyl
chloride with 4,4′-bipiperidine. Reaction of acetic anhydride or
4-biphenylcarbonyl chloride instead of adamantanecarbonyl
chloride yielded amides 10a and 10b, which were converted to
11a and 11b, respectively.
(19) Russell, M. G.; Baker, R. J.; Barden, L.; Beer, M. S.; Bristow, L.;
Broughton, H. B.; Knowles, M.; McAllister, G.; Patel, S.; Castro, J. L. J.
Med. Chem. 2001, 44, 3881-3895.
(20) Qian, Y.; Vogt, A.; Sebti, S. M.; Hamilton, A. D. J. Med. Chem 1996, 39,
217-223.
(21) Vogt, A.; Qian, Y.; Blaskovich, M. A.; Fossum, R. D.; Hamilton, A. D.;
Sebti, S. M. J. Biol. Chem. 1995, 270, 660-664.
(17) Scott, G. K.; Daniel, J. C.; Xiong, X.; Maki, R. A.; Kabat, D.; Benz, C. C.
J. Biol. Chem. 1994, 269, 19848-19858.
(18) Asada, S.; Choi, Y.; Uesugi, M. J. Am. Chem. Soc. 2003, 125, 4992-
4993.
(22) Firooznia, F.; Gude, C.; Chan, K.; Tan, J.; Fink, C. A.; Savage, P.; Beil,
M. E.; Bruseo, C. W.; Trapani, A. J.; Jeng, A. Y. Bioorg. Med. Chem.
Lett. 2002, 12, 3059-3062.
9
3462 J. AM. CHEM. SOC. VOL. 126, NO. 11, 2004

Design and Analysis of Wrenchnolol
A R T I C L E S
Scheme 1 a
a Conditions: (a) Epichlorohydrin, KOH, DMSO, 45 °C. (b) NaH, TsCl, THF, rt. (c) Isobutylamine, tert-butylamine, NH3, mono-N-Boc-1,3-propanediamine,
or mono-N-Boc-1,5-pentanediamine, MeOH, 60 °C. (d) Adamantanecarbonyl chloride, acetic anhydride, or 4-biphenylcarbonyl chloride, Et3N, CHCl3, reflux.
(e) (1) CDI, THF, reflux; (2) CH3I, CH3CN, rt; (3) amine 2a-f, pindolol, N-(3-methoxy-2-hydroxypropyl)isopropylamine, or propranolol, Et3N, CH2Cl2, rt.
(f) NaH, TsCl, THF, rt. (g) TFA, CHCl3, rt. (h) Biotin-XX-NHS, DCC, DMAP, Et3N, DMSO, rt. (i) Fluorescein isothiocyanate, Et3N, THF, rt.
activation domain to stimulate the transcription of a reporter
gene in cells, whereas it had much milder effects on those of
the activation domains of VP16 and NF-κB p65, two function-
ally irrelevant activation domains structurally similar to the ESX
activation domain (Figure 2B).18 In contrast, compound 4d,
which exhibited little activity in killing SK-BR3 cells, had
almost no effects on the ESX activation domain (Figure 2B).
Western blot analysis of wrenchnolol-treated SK-BR3 cells
showed that wrenchnolol (7b) reduces the expression of the
Her2 protein in cells but not that of R-tubulin (Figure 2C).
Although expression of other genes may be influenced by
wrenchnolol, these cellular effects are consistent with its
inhibition of the ESX-Sur-2 interaction in vitro.
Synthetically, the tosyl group in wrenchnolol protected the
labile O-substituted indole from decomposition under acidic
conditions and permitted the convenient synthesis of wrench-
9
J. AM. CHEM. SOC. VOL. 126, NO. 11, 2004 3463

A R T I C L E S
Shimogawa et al.
Table 1. Structure-Activity Relationship of Adamanolol Derivatives
nolol (7b) (Scheme 1). Glycidyl 4-indolyl ether (1a) was first
tosylated to give tosylate 1b, which was then coupled with
mono-N-Boc-1,5-pentanediamine to give amine 2f. Carbonyl-
diimidazole-mediated coupling of amine 2f and amide 3 afforded
N-Boc-wrenchnolol (6b), and subsequent acid treatment of 6b
furnished wrenchnolol (7b). This stable and water-soluble
derivative of adamanolol can readily be synthesized in a large
scale and is capable of receiving further modifications through
its extended handle. The handle can be protonated when needed
or coupled with a fluorescein group or a biotinylated molecule
for mechanistic analysis (see below). Wrenchnolol is a chemi-
cally tractable bioactive compound that is suited for mechanistic
analysis and for further application.
NMR Analysis of Wrenchnolol with a Water-Soluble
Handle. When the amino group of the handle was protonated,
wrenchnolol (7b) and its aminopropyl version (7a) became
soluble up to ∼1 mM in neutral aqueous solution. The high
water solubility enabled NMR analysis of these molecules in a
9
3464 J. AM. CHEM. SOC. VOL. 126, NO. 11, 2004

Design and Analysis of Wrenchnolol
A R T I C L E S
Figure 1. Inhibition of the ESX-Sur-2 interaction by adamanolol
derivatives. The binding fractions were calculated on the basis of fluores-
cence spectra of FITC-ESX129-145 (152 nM) in the presence of GST-
Sur2352-625 (212 nM) and varied concentrations of each adamanolol
derivative.
Figure 3. NMR analysis of 7a. (A) Expanded region of a NOESY spectrum
of 7a. NOE cross-peaks among the aromatic and methyl protons of the
tosyl group and the adamantane protons were clearly observed in an aqueous
buffered solution (PBS). (B) Chemical structure of 7a with a summary of
key NOE connectivities. (C) Tertiary geometry model of 7a. A model of
the S isomer is shown.
protons and the tosyl protons, demonstrating that these two
groups are in close proximity in water (Figure 3A, B). DQF-
COSY spectra also revealed a highly constrained structure of
the arm extended from the indole ring.
Molecular modeling with the NMR constraints supported the
s-cis configuration of the urea linker and a wrench-like shape
of the molecule (Figure 3C). In the model, the nonpolar
components are clustered on one face of the molecule, and the
polar animo handle is located away from the hydrophobic face.
Although it remains unclear at this point if wrenchnolol binds
Sur-2 in the same conformation, the structural characteristics
of free wrenchnolol are analogous to those of an amphiphilic
R-helical peptide ligand. Wrenchnolol may be a nonpeptidic
organic molecule that presents chemical characteristics of an
amphiphilic R-helical peptide without mimicking the R-helical
backbone.
Figure 2. Design and activity of wrenchnolol. (A) The wrench-like structure
of wrenchnolol (7b). (B) Wrenchnolol (20 µM) selectively impairs the ability
of the ESX activation domain to activate the secreted alkaline phosphatase
reporter gene in HEK293 cells, whereas 4d (20 µM) has little effect on the
transcriptional activation. (C) Inhibition of Her2 protein expression by
wrenchnolol. SK-BR3 breast cancer cells were treated with wrenchnolol
(7 µM) or 4d (15 µM) for 24 h, and cell lysates were analyzed by western
blots.
The NMR sample of wrenchnolol (7b) was next titrated with
a GST fusion of Sur-2 protein. Selective proton signals of
wrenchnolol were lost upon the addition of GST-Sur2352-625
,
whereas titration with the corresponding amounts of GST had
little effects on those signals (Figure 4A). The signals arising
from adamantane, tosyl, and indole protons were most strongly
diminished, while the proton signals of the aminopentyl handle
and the bipiperidine linker had little or milder effects upon the
addition of GST-Sur2352-625 (Figure 4B). Signal losses are
buffered aqueous solution, and their proton signals were
completely assigned through a combination of NOESY, DQF-
COSY, and HOHAHA data sets (Table S1). NOESY spectra
of 7a showed clear long-range NOEs between the adamantane
9
J. AM. CHEM. SOC. VOL. 126, NO. 11, 2004 3465

A R T I C L E S
Shimogawa et al.
Figure 4. NMR perturbation study of wrenchnolol (7b). (A) Expanded
one-dimensional 1H NMR spectra of wrenchnolol in the presence or absence
of GST-Sur2352-625. Addition of 30 or 60 µM of GST-Sur2352-625 to the
NMR sample of wrenchnolol decreased the signal intensities of the methyl
protons of the tosyl group and those of the adamantane protons (red and
yellow arrows), whereas the signals of the pentylamine protons had little
effect (blue arrows). In contrast, addition of the corresponding amounts of
GST (top) had no significant effects on the signals. A star indicates an
internal reference signal arising from 10 µL of dimethyl sulfoxide-d6
(DMSO). (B) A summary of the differential signal intensities. The
percentage value of each signal was calculated by using the internal DMSO
peak as a reference and classified as a strong (red, < the median), medium
(yellow, > the median and < the average), and weak (blue, > the average)
effect.
Figure 5. Interaction of biotin-labeled wrenchnolol (8) with Sur-2 in cell
extracts. Conjugate 8 was incubated with cell nuclear extracts, and bound
proteins were purified by an avidin affinity column. (A) Western blot
detection of Sur-2 protein. It is evident that Sur-2 was copurified with
conjugate 8 (lane 4), whereas no Sur-2 protein was detected with the controls
of DMSO (lane 3) and compound 13 (lane 5). (B) Silver staining of the
bound proteins. The 130 KDa band indicated by an arrow matched the
western-blotted band shown in (A). Lanes 1 and 2 show protein markers
and 10% input of nuclear extracts, respectively.
observed when the exchange rate between the free and bound
states is intermediate at the NMR time scale and when the
chemical shift is perturbed due to the alteration of magnetic
environment upon binding. The Kd of wrenchnolol for Sur-2
was estimated to be in a low µM range (see wrenchnolol with
a fluorescein handle) where signal losses are often observed in
a similar type of interaction. The lost signals arise from the
protons located in the jaws of the molecule, supporting the
notion that the hydrophobic jaws of amphiphilic wrenchnolol
make direct contacts with Sur-2 protein.
Sur-2 Interaction of Wrenchnolol with a Biotinylated
Handle. To confirm the interaction of wrenchnolol with Sur-2
in a cellular context, a biotin conjugate of wrenchnolol (8) was
synthesized (Scheme 1) and incubated with cell nuclear extracts.
The bound proteins were purified through an avidin affinity
column, separated on an SDS-PAGE gel, and analyzed by
western blots with a Sur-2 antibody. As shown in Figure 5A,
Sur-2 protein was retained on the affinity column in the presence
of biotin conjugate 8, whereas no Sur-2 protein was detected
in its absence or in the presence of a control biotin molecule
(13). The retention of Sur-2 on the affinity column indicates
that wrenchnolol is capable of binding to Sur-2 in the presence
of other cellular proteins. Silver-staining detection of all the
protein bands on the gel showed a significantly reduced number
of protein bands after the affinity purification (compare lanes 2
and 4 in Figure 5B), demonstrating some degree of binding
selectivity of wrenchnolol. One of the silver-stained bands
matched the western-blotted band of Sur-2, while the identities
of the other bands remained unknown. Sur-2 (also called
DRIP130 or CRSP130) is a subunit of the human mediator
complexes:23-26 The additional bands may include those of the
other subunits copurified with Sur-2. However, all the protein
bands cannot be explained by the subunits of human mediator
complexes or Sur-2-interacting proteins, suggesting that the
selectivity of wrenchnolol remains to be further optimized.
Wrenchnolol with a Fluorescein Handle. A synthetic
compound with a molecular weight higher than 500 is generally
less likely to penetrate the cell membrane. To check the cell
permeability of wrenchnolol, whose molecular weight is 802,
its handle was coupled with a fluorescein molecule through a
(23) Naar, A. M.; Beaurang, P. A.; Zhou, S.; Abraham, S.; Solomon, W.; Tjian,
R. Nature 1999, 398, 828-832.
(24) Rachez, C.; Lemon, B. D.; Suldan, Z.; Bromleigh, V.; Gamble, M.; Naar,
A. M.; Erdjument-Bromage, H.; Tempst, P.; Freedman, L. P. Nature 1999,
398, 824-828.
(25) Ryu, S.; Zhou, S.; Ladurner, A. G.; Tjian, R. Nature 1999, 397, 446-450.
(26) Boyer, T. G.; Martin, M. E.; Lees, E.; Ricciardi, R. P.; Berk, A. J. Nature
1999, 399, 276-279.
9
3466 J. AM. CHEM. SOC. VOL. 126, NO. 11, 2004

Design and Analysis of Wrenchnolol
A R T I C L E S
The fluorescein conjugate 9 also enabled evaluation of the
Kd value of the interaction between wrenchnolol and Sur-2 in
vitro. When a GST fusion of Sur-2 protein (GST-Sur-2352-625
)
was added to a solution of conjugate 9, the fluorescence intensity
of 9 at 522 nm was decreased (Figure 6C). The clear
fluorescence perturbation by the interaction permitted an estima-
tion of the Kd value to be 1.8 µM, which is consistent with the
IC50 of wrenchnolol in cells.
Discussion
Inhibition of a Nuclear Protein-Protein Interaction by
Small Molecules. Small molecules that target a protein-protein
interaction in the nucleus need to penetrate the cell membrane,
diffuse into the nucleus, and bind selectively to a protein target
with a consequent cellular phenotype. Development of such
molecules is challenging and may require tremendous efforts
before clinical application. Nevertheless, given the importance
of the problem in enriching future opportunities, efforts in this
direction need to continue.28,29 The case study of wrenchnolol
foreshadows the promise and the challenge of targeting a
protein-protein interaction in the nucleus.
Recent evidence has suggested that protein interactions that
are mediated by short R-helical peptides are tractable to
inhibition by small synthetic molecules. For example, the
interaction of Bcl-XL with Bak, a well-documented association
that controls programmed cell death, is mediated by a short
R-helical peptide of Bak and can be inhibited by nonpeptidic
small molecules.30-33 However, inhibitors of protein-protein
interactions in general tend to be highly hydrophobic due to
the nonpolar nature of binding pockets and may lack the water
solubility required for structural or animal studies. Our results
suggest that designing a tetrasubstituted urea molecule is a
potential approach to obtaining a water-soluble inhibitor of
R-helix-mediated protein interactions. The bulky amino handle
of wrenchnolol enforced the active conformation and rendered
it water-soluble enough to permit NMR characterization in an
aqueous solution and its evaluation in mouse models of breast
cancers (currently underway). A number of bioactive peptide
ligands are amphiphilic helices with a hydrophilic face on the
opposite side of the hydrophobic binding interface. The wrench-
shape design of urea molecules could be an effective way of
mimicking an amphiphilic R-helical ligand by synthetic mol-
ecules.
Figure 6. Wrenchnolol with a fluorescein handle (9). (A) Evaluation of
cell permeability of wrenchnolol. Conjugate 9 was incubated with SK-BR3
cells for 1 h, and fluorescence images were captured by a fluorescence
microscope. (B) Costaining of the cells with nucleus-specific DAPI dye.
(C) Estimation of Kd of the wrenchnolol-Sur-2 interaction. Fluorescence
Our results suggest that the s-cis conformation of adamanolol
is an important determinant of the activity. The preorganized
hydrophobic surface of adamanolol is perhaps complementary
to the pocket in Sur-2, and its modification should influence
the activity. We synthesized two adamanolol analogues having
distinct jaw widths (Chart 2):34 in one molecule, the bipiperidine
linker was extended by inserting a benzene ring (14); in the
other, the bipiperidine linker was replaced by a shorter pipera-
spectra of conjugate 9 were monitored upon addition of GST-Sur2352-625
Shown are relative emission intensities of 9 at 522 nm (excitation: 490
nm) in the presence of varied concentrations of GST-Sur2352-625
.
.
simple thiourea formation (Scheme 1). As shown in Figure 6A,
fluorescein conjugate 9 crossed the cell membrane of breast
cancer cells and diffused both in the cytoplasm and the nucleus.
However, careful comparison of the fluorescence images of
fluorescein conjugate 9 (Figure 6A) and the nucleus-specific
DAPI dye (Figure 6B) revealed that the compound is located
more preferentially in the cytoplasm than in the nucleus, as often
observed with large synthetic organic molecules.27 Although
wrenchnolol is capable of penetrating the cell membrane,
improvement of its nuclear localization might strengthen the
biological activity and selectivity.
(28) Cochran, A. G. Curr. Opin. Chem. Biol. 2001, 5, 654-659.
(29) Toogood, P. L. J. Med. Chem. 2002, 45, 1543-1558.
(30) Kutzki, O.; Park, H. S.; Ernst, J. T.; Orner, B. P.; Yin, H.; Hamilton, A. D.
J. Am. Chem. Soc. 2002, 124, 11838-11839.
(31) Wang, J. L.; Liu, D.; Zhang, Z. J.; Shan, S.; Han, X.; Srinivasula, S. M.;
Croce, C. M.; Alnemri, E. S.; Huang, Z. Proc. Natl. Acad. Sci. U.S.A. 2000,
97, 7124-7129.
(32) Degterev, A.; Lugovskoy, A.; Cardone, M.; Mulley, B.; Wagner, G.;
Mitchison, T.; Yuan, J. Nat. Cell Biol. 2001, 3, 173-182.
(33) Tzung, S. P.; Kim, K. M.; Basanez, G.; Giedt, C. D.; Simon, J.; Zimmerberg,
J.; Zhang, K. Y.; Hockenbery, D. M. Nat. Cell Biol. 2001, 3, 183-191.
(34) Shimogawa, H.; Mao, Q.; Uesugi, M.; Kigoshi, H. Unpublished results.
(27) Belitsky, J. M.; Leslie, S. J.; Arora, P. S.; Beerman, T. A.; Dervan, P. B.
Bioorg. Med. Chem. 2002, 10, 3313-3318.
9
J. AM. CHEM. SOC. VOL. 126, NO. 11, 2004 3467

A R T I C L E S
Shimogawa et al.
Chart 2
have been made with the goal of developing a small-molecule
transcription factor. Dervan and co-workers have shown that
the DNA-binding module of a transcription factor can be
replaced by a hairpin polyamide composed of N-methylpyrrole
and N-methylimidazole amino acids that binds in the minor
groove of DNA.38-40 This cell-permeable polyamide is capable
of targeting a predetermined DNA sequence with affinities and
specificities comparable to transcription factors.44-46 Although
these studies provided a key initial step toward the goal of
engineering cell-permeable small-molecule transcription factors,
the use of peptides as an activation module prevented the
artificial transcription factors from crossing the cell membrane
and limited their application. A goal of the field is now the
replacement of the activation peptide with nonpeptidic organic
molecules in order to engineer cell-permeable compounds.
Wrenchnolol is a cell-permeable synthetic molecule that binds
to the Sur-2 subunit of the human mediator complex by
mimicking the potent activation domain of ESX. It is also
chemically tractable and capable of receiving further modifica-
tions through its handle. Wrenchnolol may serve as a nonpep-
tidic activation module and could lead to the development of
synthetic modulators of gene transcription.
dine (15). Their evaluation in SK-BR3 cells showed that either
the extension or shortening reduced the activity of adamanolol,
suggesting that the bipiperidine linker of adamanolol intrinsically
has a perfect length for binding to Sur-2. Diversification of the
width and shape in the jaws of adamanolol or wrenchnolol may
lead to the discovery of inhibitors of distinct R-helix-mediated
protein interactions.
Despite the benefits, the wrench-shape design has a draw-
back: the relatively large size of the molecules, which might
limit their cell permeability. This problem may be intrinsic to
any competitive inhibitors of protein-protein interactions,
which, unlike enzyme inhibitors, need to cover a larger binding
interface to achieve selectivity and potency. Although our
analysis showed that wrenchnolol is capable of penetrating the
cell membrane and diffusing both in the cytoplasm and in the
nucleus, its nuclear distribution is somewhat limited, as often
observed with large synthetic molecules.27 Nuclear pores are
considered to be large enough for even 10 KDa proteins to cross
passively;35 however, there may be an active transport mech-
anism to eliminate certain types of synthetic molecules from
the nucleus. Cellular distribution of small-molecule inhibitors,
especially protein-protein interaction inhibitors, is perhaps a
key determinant of their biological effectiveness, just as the
subcellular localization of signaling proteins plays an essential
role in achieving their specific interactions with the downstream
effectors. An understanding of the molecular characteristics that
govern cellular localization of large synthetic molecules would
accelerate the success of the inhibitors of protein-protein
interactions.
Experimental Section
General Procedures. SK-BR3 cells were kindly provided by Dr.
Mien-Chie Hung (M. D. Anderson Cancer Center) and grown in
DMEM/F-12 medium with 10% FBS (Invitrogen). GST-Sur2352-625
and FITC-ESX129-145 were prepared and characterized as described
before.18 Compounds 11c,d and 13 were purchased from Tripos and
Sigma, respectively. The solvents used for chemical synthesis were
dried prior to use. All other chemicals were used as received without
purification. All moisture-sensitive reactions were performed in flame-
dried and/or oven-dried glassware under a positive pressure of nitrogen
unless otherwise noted. Thin-layer chromatography was carried out with
glass TLC plates precoated with Merck silica gel 60 F254. Column
chromatography was accomplished with Fuji Silysia Chemical silica
gel BW820MH. Proton nuclear magnetic resonance spectra were
recorded in deuterated solvents at 270 or 600 MHz. High-resolution
mass spectra were obtained by Q-star spectrometer (Applied Biosys-
tems) on an ESI mode.
Transcription Modulators. The regulation of transcription
often causes drastic phenotypic changes in human bodies
through its effects on differentiation, organ development, aging,
and diseases. External manipulation of transcription by cell-
permeable synthetic molecules represents a challenge in me-
dicinal and bioorganic chemistry.4,36-43 A large number of
studies on naturally occurring transcription factors have estab-
lished the dogma that they typically have separable domains
for sequence-specific binding to DNA and for transcriptional
activation through protein-protein interactions.5,6 Efforts to
incorporate synthetic counterparts of these functional modules
Glycidyl 4-Indolyl Ether (1a). To a solution of 4-hydroxyindole
(0.50 g, 3.8 mmol) in dimethyl sulfoxide (10 mL) at 45 °C were added
potassium hydroxide (0.43 g, 7.6 mmol) and then epichlorohydrin (1.4
g, 15.2 mmol). After being stirred at 45 °C for 4 h, the reaction mixture
was diluted with a saturated solution of NH4Cl and extracted with
diethyl ether (3 × 50 mL). The combined organic layers were dried
with Na2SO4 and concentrated. The residue was purified by column
chromatography on silica gel with benzene/acetone mixtures to give
the glycidyl 4-indolyl ether (1a) (0.75 g, quant) as a colorless oil: IR
(thin film) 3402, 2923, 1497, 1362, 1244 cm-1; 1H NMR (CDCl3, 270
MHz) δH 8.17 (br. s, 1H), 7.14-7.03 (m, 3H), 6.70 (t, J ) 2.0 Hz,
1H), 6.62 (dd, J ) 1.1, 7.0 Hz, 1H), 4.35 (dd, J ) 3.5, 11.1 Hz, 1H),
4.16 (dd, J ) 5.4, 11.1 Hz, 1H), 3.46 (m, 1H), 2.94 (dd, J ) 4.1, 4.9,
1H), 2.82 (dd, J ) 2.7, 4.9, 1H); HRMS (ESI) exact mass calcd for
C11H11NO2 + H requires m/z 190.0868, found m/z 190.0842.
(35) Rout, M. P.; Aitchison, J. D. J. Biol. Chem. 2001, 276, 16593-16596.
(36) Liu, B.; Han, Y.; Corey, D. R.; Kodadek, T. J. Am. Chem. Soc. 2002, 124,
1838-1839.
(37) Nyanguile, O.; Uesugi, M.; Austin, D. J.; Verdine, G. L. Proc. Natl. Acad.
Sci. U.S.A. 1997, 94, 13402-13406.
Preparation of Glycidyl N-Tosyl-4-indolyl Ether (1b). A solution
of glycidyl 4-indolyl ether (1a) (1.0 g, 5.3 mmol) in THF (5 mL) was
cooled to 0 °C, and sodium hydride in mineral oil (60%, 0.42 g, 10.6
mmol) and tosyl chloride (1.2 g, 6.3 mmol) were added. After being
(38) Ansari, A. Z.; Mapp, A. K.; Nguyen, D. H.; Dervan, P. B.; Ptashne, M.
Chem. Biol. 2001, 8, 583-592.
(39) Arora, P. S.; Ansari, A. Z.; Best, T. P.; Ptashne, M.; Dervan, P. B. J Am
Chem Soc 2002, 124, 13067-13071.
(40) Mapp, A. K.; Ansari, A. Z.; Ptashne, M.; Dervan, P. B. Proc. Natl. Acad.
Sci. U.S.A. 2000, 97, 3930-3935.
(44) Gottesfeld, J. M.; Neely, L.; Trauger, J. W.; Baird, E. E.; Dervan, P. B.
Nature 1997, 387, 202-205.
(41) Denison, C.; Kodadek, T. Chem. Biol. 1998, 5, R129-145.
(42) Belshaw, P. J.; Ho, S. N.; Crabtree, G. R.; Schreiber, S. L. Proc. Natl.
Acad. Sci. U.S.A. 1996, 93, 4604-4607.
(45) White, S.; Szewczyk, J. W.; Turner, J. M.; Baird, E. E.; Dervan, P. B.
Nature 1998, 391, 468-471.
(43) Shi, Y.; Koh, J. T. J. Am. Chem. Soc. 2002, 124, 6921-6928.
(46) Trauger, J. W.; Baird, E. E.; Dervan, P. B. Nature 1996, 382, 559-561.
9
3468 J. AM. CHEM. SOC. VOL. 126, NO. 11, 2004

Design and Analysis of Wrenchnolol
A R T I C L E S
stirred at this temperature for 1 h, the reaction mixture was allowed to
warm to room temperature and stirred for 2 h. The reaction mixture
was diluted with water and extracted with CHCl3 (3 × 50 mL). The
combined organic layers were dried with Na2SO4 and concentrated.
The residue was purified by column chromatography on silica gel with
benzene/acetone mixtures to give the glycidyl N-tosyl-4-indolyl ether
(1b) (1.5 g, 82%) as a white solid: IR (thin film) 2930, 1362, 1189,
1137, cm-1; 1H NMR (CDCl3, 270 MHz) δH 7.75 (d, J ) 8.4 Hz, 2H),
7.61 (d, J ) 8.4 Hz, 1H), 7.46 (d, J ) 3.8 Hz, 1H), 7.23 (t, J ) 8.4
Hz, 1H), 7.21 (d, J ) 8.4 Hz, 2H), 6.80 (d, J ) 3.8 Hz, 1H), 6.63 (d,
J ) 8.4 Hz, 1H), 4.32 (dd, J ) 3.2, 11.1 Hz, 1H), 4.03 (dd, J ) 5.7,
11.1 Hz, 1H), 3.38 (m, 1H), 2.91 (t, J ) 4.9 Hz, 1H), 2.77 (dd, J )
2.7, 4.9 Hz, 1H), 2.34 (s, 3H); HRMS (ESI) exact mass calcd for C18H17-
NO4S + H requires m/z 344.0957, found m/z 344.0944.
temperature for 30 h, the mixture was concentrated and purified by
column chromatography on silica gel with CHCl3/methanol mixtures
to give adamanolol (4a) (145 mg, 67%) as a colorless oil: IR (thin
film) 3450, 2907, 1624, 1428, 1245 cm-1; 1H NMR (CDCl3, 270 MHz)
δH 8.57 (br. s, 1H), 7.11-7.04 (m, 3H), 6.59 (t, J ) 2.4 Hz, 1H), 6.52
(dd, J ) 2.7, 5.7 Hz, 1H), 5.29 (br. m, 1H), 4.53 (br. d, J ) 13.5 Hz,
2H), 4.45-4.25 (m, 2H), 4.12 (br. m, 2H), 3.44 (dd, J ) 1.4, 13.5 Hz,
1H), 3.35-3.10 (m, 2H), 2.82-2.50 (br. m, 4H), 2.03 (br. s, 3H), 1.99
(br. s, 6H), 1.72 (s, 6H), 1.80-1.44 (br. m, 4H), 1.35 (d, J ) 6.5 Hz,
3H), 1.32 (d, J ) 6.5 Hz, 3H), 1.45-0.86 (m, 6H); HRMS (ESI) exact
mass calcd for C36H52N4O4 + H requires m/z 605.4067, found m/z
605.4065.
Preparation of N-Tosyladamanolol (5a). A solution of adamanolol
(4a) (30.1 mg, 0.05 mmol) in THF (1 mL) was cooled to 0 °C, and
sodium hydride in mineral oil (60%, 12 mg, 0.50 mmol) and tosyl
chloride (9.5 mg, 0.05 mmol) were added. After being stirred at this
temperature for 0.5 h, the reaction mixture was allowed to warm to
room temperature and stirred for 2 h. The reaction mixture was diluted
with brine and extracted with CH2Cl2 (3 × 20 mL). The combined
organic layers were dried with Na2SO4 and concentrated. The residue
was purified by column chromatography on silica gel with CHCl3/
methanol mixtures to give the N-tosyladamanolol (5a) (21.5 mg, 57%)
as a colorless oil: IR (thin film) 3434, 2910, 2852, 1623, 1427, 1188
cm-1; 1H NMR (CDCl3, 270 MHz) δH 7.75 (d, J ) 8.4 Hz, 2H), 7.57
(d, J ) 8.4 Hz, 1H), 7.45 (d, J ) 3.5 Hz, 1H), 7.21 (d, J ) 8.4 Hz,
2H), 7.20 (dd, J ) 7.8, 8.4 Hz, 1H), 6.75 (d, J ) 3.5 Hz, 1H), 6.68 (d,
J ) 7.8 Hz, 1H), 5.14 (br. m, 1H), 4.55 (br. d, J ) 13.0 Hz, 2H), 4.24
(br. d, J ) 4.6 Hz, 2H), 4.32-3.94 (br. m, 2H), 2.96 (m, 2H), 2.82 (m,
1H), 2.75-2.54 (br. m, 4H), 2.34 (s, 3H), 2.03 (br. s, 3H), 1.99 (br. s,
6H), 1.72 (s, 6H), 1.80-1.44 (br. m, 4H), 1.45-0.86 (m, 6H), 1.04 (d,
J ) 6.5 Hz, 6H); HRMS (ESI) exact mass calcd for C43H58N4O6S + H
requires m/z 759.4155, found m/z 759.4138.
N-[3-(4-Indolyloxy)-2-hydroxypropyl]-isobutylamine (2a). To a
solution of glycidyl 4-indolyl ether (1a) (215 mg, 1.1 mmol) in methanol
(2 mL) was added isobutylamine (160 mg, 2.2 mmol). The solution
was heated to 100 °C and stirred for 32 h. The product was purified
by column chromatography on silica gel with CHCl3/methanol mixtures
to give the corresponding amine (2a) (120 mg, 42%) as a colorless
1
oil: IR (thin film) 3404, 2954, 1508, 1363 cm-1; H NMR (CDCl3,
270 MHz) δH 8.24 (br. s, 1H), 7.12-7.02 (m, 3H), 6.65 (t, J ) 2.4 Hz,
1H), 6.51 (dd, J ) 1.0, 7.3 Hz, 1H), 4.30 (m, 1H), 4.21-4.09 (m, 2H),
3.43 (br. s, 2H), 3.05 (dd, J ) 3.8, 12.4 Hz, 1H), 2.96 (dd, J ) 7.8,
12.4 Hz, 1H), 2.61 (dd, J ) 2.7, 11.9 Hz, 1H), 2.57 (dd, J ) 2.7, 11.9
Hz, 1H), 1.90 (m, 1H), 0.97 (d, J ) 6.8 Hz, 3H), 0.96 (d, J ) 6.8 Hz,
3H); HRMS (ESI) exact mass calcd for C15H22N2O2 + H requires m/z
263.1760, found m/z 263.1778.
N-[3-(N-Tosyl-4-indolyloxy)-2-hydroxypropyl]-N′-Boc-1,5-pen-
tanediamine (2f). The reaction of glycidyl N-tosyl-4-indolyl ether (1b)
(1.7 g, 5.0 mmol) and mono-N-Boc-1,5-pentanediamine (2.0 g, 10
mmol) as described for 2a gave the title compound (2f) (1.7 g, 63%)
as a colorless oil: IR (thin film) 3426, 2932, 1684, 1508, 1364, 1160
cm-1; 1H NMR (CDCl3, 270 MHz) δH 7.74 (d, J ) 8.4 Hz, 2H), 7.60
(d, J ) 8.4 Hz, 1H), 7.46 (d, J ) 3.8 Hz, 1H), 7.20 (d, J ) 8.4 Hz,
2H), 7.19 (t, J ) 8.4 Hz, 1H), 6.76 (d, J ) 3.8 Hz, 1H), 6.64 (d, J )
8.4 Hz, 1H), 4.54 (s, 1H), 4.07 (m, 3H), 3.10 (br. m, 2H), 2.89 (dd, J
) 3.5, 12.4 Hz, 1H), 2.78 (dd, J ) 7.0, 12.4 Hz, 1H), 2.65 (dt, J )
2.4, 4.3 Hz, 2H), 2.33 (s, 3H), 2.08 (br. s, 2H), 1.60-1.28 (m, 6H),
1.44 (s, 9H); HRMS (ESI) exact mass calcd for C28H39N3O6S + H
requires m/z 546.2638, found m/z 546.2640.
Preparation of N-Boc-Wrenchnolol with an Aminopropyl Handle
(6a). The reaction of N-[3-(N-tosyl-4-indolyloxy)-2-hydroxypropyl]-
N′-Boc-1,3-propanediamine (2e) (460 mg, 0.89 mmol) and monoamide
3 (250 mg, 0.76 mmol) as described for 4a gave the title compound
(6a) (353 mg, 53%) as a colorless oil: IR (thin film) 3441, 1633 cm-1
;
1H NMR (CDCl3, 270 MHz) δH 7.75 (d, J ) 8.4 Hz, 2H), 7.59 (d, J
) 8.4 Hz, 1H), 7.46 (d, J ) 3.8 Hz, 1H), 7.26-7.17 (m, 3H), 6.74 (d,
J ) 3.8 Hz, 1H), 6.67 (d, J ) 8.4 Hz, 1H), 5.26 (br. m, 1H), 5.10 (br.
s, 1H), 4.55 (br. d, J ) 13.0 Hz, 2H), 4.29 (br. m, 2H), 4.30-3.85 (br.
m, 2H), 3.25-3.05 (br. m, 2H), 2.95-2.78 (br. m, 2H), 2.78-2.55
(br. m, 6H), 2.34 (s, 3H), 2.03 (br. s, 3H), 1.99 (br. s, 6H), 1.90-1.55
(br. m, 6H), 1.72 (s, 6H), 1.40 (s, 9H), 1.30-0.70 (br. m, 6H); HRMS
(ESI) exact mass calcd for C48H67N5O8S + H requires m/z 874.4789,
found m/z 874.4778.
Preparation of Amide 3. To a solution of 4,4′-bipiperidine (0.84
g, 5.0 mmol) and triethylamine (2 mL) in CHCl3 (150 mL) was added
a solution of 1-adamantanecarbonyl chloride (0.50 g, 2.5 mmol) in
CHCl3 (50 mL). The resulting mixture was refluxed for 0.5 h, diluted
with water, and extracted with CHCl3 (3 × 100 mL). The combined
extracts were dried over Na2SO4 and then concentrated in a vacuum.
The residue was purified by column chromatography on silica gel with
methanol/aqueous ammonia mixtures to give amide 3 (750 mg, 90%)
Preparation of N-Boc-Wrenchnolol (6b). The reaction of N-[3-
(N-tosyl-4-indolyloxy)-2-hydroxypropyl]-N′-Boc-1,5-pentanediamine (2f)
(500 mg, 0.92 mmol) and monoamide 3 (250 mg, 0.76 mmol) as
described for 4a gave the title compound (6b) (322 mg, 47%) as a
1
as a colorless oil: IR (thin film) 3421, 2902, 1617 cm-1; H NMR
1
(CDCl3, 270 MHz) δH 4.46 (br. d, J ) 13.0 Hz, 2H), 3.02 (br. d, J )
11.6 Hz, 2H), 2.91 (br. s, 1H), 2.59 (br. t, J ) 11.9 Hz, 2H), 2.47 (br.
t, J ) 11.9 Hz, 2H), 1.95 (br. s, 3H), 1.90 (br. s, 6H), 1.63 (br. s,
10H), 1.19-1.02 (br. m, 6H); HRMS (ESI) exact mass calcd for
C21H34N2O + H requires m/z 331.2749, found m/z 331.2753.
Preparation of Adamanolol (4a). To a solution of amide 3 (120
mg, 0.36 mmol) in THF (10 mL) was added 1,1′-carbonyldiimidazole
(240 mg, 1.46 mmol). After being refluxed for 4 h, the mixture was
diluted with CHCl3 (20 mL) and washed with water (3 × 20 mL). The
organic layer was dried with Na2SO4 and concentrated to give
carbamoyl imidazole, which was used without further purification. A
solution of carbamoyl imidazole and MeI (0.5 mL) in MeCN (2 mL)
was stirred for 18 h at room temperature. Evaporation of the solvent
and excess MeI in a vacuum gave the corresponding imidazolium salt.
To a solution of imidazolium salt in CH2Cl2 (2 mL) were added pindolol
(268 mg, 1.08 mmol) and Et3N (0.5 mL). After being stirred at room
colorless oil: IR (thin film) 3445, 2933, 1684, 1617, 1428 cm-1; H
NMR (CDCl3, 270 MHz) δH 7.75 (d, J ) 8.4 Hz, 2H), 7.59 (d, J )
7.8 Hz, 1H), 7.46 (d, J ) 3.8 Hz, 1H), 7.21 (d, J ) 8.4 Hz, 2H), 7.26-
7.19 (t, J ) 7.8 Hz, 1H), 6.74 (d, J ) 3.8 Hz, 1H), 6.67 (d, J ) 7.8
Hz, 1H), 5.18 (br. m, 1H), 4.55 (br. m, 1H), 4.55 (br. d, J ) 12.7 Hz,
2H), 4.24 (br. d, J ) 4.6 Hz, 2H), 4.32-3.87 (br. m, 2H), 3.17-2.91
(m, 4H), 2.79-2.52 (m, 6H), 2.34 (s, 3H), 2.03 (br. s, 3H), 1.99 (br.
s, 6H), 1.72 (s, 6H), 1.44 (s, 9H), 1.80-0.86 (m, 16H); HRMS (ESI)
exact mass calcd for C50H71N5O8S + H requires m/z 902.5102, found
m/z 902.5100.
Wrenchnolol with an Aminopropyl Handle (7a). To a solution of
6a (350 mg, 0.40 mmol) in CHCl3 (3 mL) was added TFA. This solution
was stirred at room temperature for 15 min and then concentrated in a
vacuum, which was used without further purification to give 7a (360
mg, quant) as a colorless oil: IR (thin film) 3436, 1630 cm-1; 1H NMR
(CDCl3-CD3OD/4:1, 270 MHz) δH 7.62 (d, J ) 8.4 Hz, 2H), 7.47 (d,
9
J. AM. CHEM. SOC. VOL. 126, NO. 11, 2004 3469

A R T I C L E S
Shimogawa et al.
J ) 8.1 Hz, 1H), 7.35 (d, J ) 3.8 Hz, 1H), 7.24 (t, J ) 8.1 Hz, 1H),
7.11 (d, J ) 8.4 Hz, 2H), 6.66 (d, J ) 3.8 Hz, 1H), 6.56 (d, J ) 8.1
Hz, 1H), 5.24 (br. m, 1H), 4.40 (br. d, J ) 11.9 Hz, 2H), 4.17 (br. m,
2H), 4.05-3.72 (br. m, 2H), 3.31 (br. m, 2H), 3.04 (br. m, 2H), 2.91
(br. m, 2H), 2.78-2.42 (br. m, 4H), 2.22 (s, 3H), 2.05 (br. m, 2H),
1.91 (br. s, 3H), 1.85 (br. s, 6H), 1.70-1.30 (br. m, 4H), 1.57 (s, 6H),
1.30-0.60 (br. m, 6H); HRMS (ESI) exact mass calcd for C43H59N5O6S
+ H requires m/z 774.4264, found m/z 774.4286.
which the production of secreted alkaline phosphatase (SEAP) is under
control of an IL2 promoter carrying five GAL4 binding sites. After 10
h of incubation with wrenchnolol (20 µM), an aliquot of the culture
was assayed for SEAP activity as described.42
Effects of Wrenchnolol on the Expression Levels of Her2. SK-
BR3 cells were treated by 7 µM of wrenchnolol for 24 h, and whole
cell lysates were analyzed by western blots. The primary antibodies
used for the western analysis were a monoclonal antibody against Her2
(c-neu-Ab-3; Oncogene Science) and a monoclonal antibody against
R-tublin (Molecular Probes). The blotted membranes were then
visualized with ECL chemiluminescence detection reagents (Amersham
Bioscience).
Sur-2 Interaction of a Biotin Conjugate of Wrenchnolol (8). HeLa
cell nuclear extracts were prepared as described47 and incubated with
a biotin conjugate of wrenchnolol (8) (100 µM) at 4 °C for 24 h. The
samples were incubated with Neutravidin affinity resin (Pierce) at 4
°C for 2 h and extensively washed with PBS buffer containing 0.5%
Nonidet P-40. The bound proteins were then separated by SDS/PAGE
and analyzed by western blots with an antibody against human Sur-2
or by silver-staining.
Fluorescence Studies. To monitor the ability of compounds to inhibit
ESX-Sur-2 interaction, fluorescence emission spectra of fluorescein-
isothiocyanate-labeled ESX129-145 (FITC-ESX129-145, 152 nM, excitation
at 490 nm) were measured in the presence of GST-Sur2352-625 (212
nM) and varied concentrations of a compound. Fluorescence intensity
changes at 522 nm as a function of chemical concentrations (0 to 50
µM) were monitored. IC50 values were calculated by curve fitting of
the data. For shallow dose-effect curves, IC50 values could not be
accurately determined and were assigned with > marks. For the
evaluation of Kd of wrenchnolol, a concentrated solution of GST-
Sur2352-625 was added to a 534 nM solution of a fluorescein conjugate
of wrenchnolol (9), and the fluorescence intensity of 9 at 522 nm was
monitored. All the measurements were performed in triplicate in a PBS
solution containing 100 mM NaCl (pH 7.0) on a Perkin-Elmer LS-
50B fluorometer. The Kd of 9 was calculated by curve fitting of the
fluorescence intensity as a function of the GST-Sur2352-625 concentra-
tion, using the following equation:
Preparation of Wrenchnolol (7b). The reaction of N-Boc-wrench-
nolol (6b) (320 mg, 0.36 mmol) as described for 7a gave the title
compound (7b) (325 mg, quant) as a colorless oil: IR (thin film) 3444,
2910, 1635, 1428 cm-1; 1H NMR (CDCl3, 270 MHz) δH 7.75 (d, J )
8.1 Hz, 2H), 7.58 (d, J ) 8.1 Hz, 1H), 7.46 (d, J ) 3.8 Hz, 1H), 7.21
(d, J ) 8.1 Hz, 2H), 7.13-7.02 (dd, J ) 7.8, 8.1 Hz, 1H), 6.74 (d, J
) 3.8 Hz, 1H), 6.65 (d, J ) 7.8 Hz, 1H), 5.30 (br. m, 1H), 4.54 (br.
m, 1H), 4.54 (br. d, J ) 12.7 Hz, 2H), 4.24 (br. d, J ) 4.3 Hz, 2H),
4.32-3.81 (br. m, 2H), 3.08 (br. d, J ) 5.4 Hz, 1H), 2.85 (br. t, J )
5.4 Hz, 1H), 2.81-2.52 (br. m, 8H), 2.34 (s, 3H), 1.99 (br. s, 9H),
1.72 (br. s, 6H), 1.79-0.70 (m, 16H); HRMS (ESI) exact mass calcd
for C45H63N5O6S + H requires m/z 802.4577, found m/z 802.4550.
Biotin Conjugate Wrenchnolol (8). To a solution of wrenchnolol
(7b) (5.0 mg, 0.006 mmol) and triethylamine (0.1 mL) in DMSO (0.25
mL) were added 4-(dimethylamino)pyridine (0.1 mg), N,N′-dicyclo-
hexylcarbodiimide (2.5 mg, 0.012 mmol), and biotoin-XX-NHS (6.8
mg, 0.012 mmol). This solution was stirred at room temperature
overnight, diluted with brine, and extracted with CHCl3 (3 × 10 mL).
The combined extracts were dried over Na2SO4 and then concentrated
in a vacuum. The residue was purified by column chromatography on
silica gel with CHCl3/methanol mixtures to give the corresponding
amide (8) (7.2 mg, 95%) as a colorless oil: IR (thin film) 3284, 2933,
1
1684, 1653, 1559, 1541 cm-1; H NMR (CD3OD, 270 MHz) δH 7.79
(d, J ) 8.6 Hz, 2H), 7.60 (d, J ) 8.1 Hz, 1H), 7.56 (d, J ) 4.0 Hz,
1H), 7.31 (d, J ) 8.6 Hz, 2H), 7.21 (t, J ) 8.1 Hz, 1H), 6.79 (d, J )
4.0 Hz, 1H), 6.77 (d, J ) 8.1 Hz, 1H), 5.38 (br. m, 1H), 4.60-4.42
(br. m, 3H), 4.38-4.20 (br. m, 3H), 4.16-3.89 (br. m, 2H), 3.47 (br.
m, 2H), 3.24-3.01 (br. m, 9H), 2.91 (dd, J ) 5.0, 12.8 Hz, 1H), 2.85-
2.54 (br. m, 5H), 2.34 (s, 3H), 2.24-2.08 (br. m, 6H), 2.00 (br. s, 9H),
1.77 (br. s, 6H), 1.90-0.40 (m, 34H); HRMS (ESI) exact mass calcd
for C67H99N9O10S2 + H requires m/z 1254.7035, found m/z 1254.7010.
Fluorescein Conjugate Wrenchnolol (9). To a solution of wrench-
nolol (7b) (10 mg, 0.012 mmol) and triethylamine (0.2 mL) in THF
(1 mL) was added fluorescein isothiocyanate (5.6 mg, 0.014 mmol).
This solution was stirred at room temperature for 8 h and then
concentrated in a vacuum. The residue was purified by column
chromatography on silica gel (CHCl3/methanol) and ODS silica gel
(70%MeOH, 0.1%TFA) to give the fluorescein conjugate (9) (8.3 mg,
A ) Ao + 0.5∆ꢀ([9]o + (([9]o + [P]oKd)2 - 4[9]o[P]o)0.5
)
where A and Ao are the values of fluorescence intensity of 9 in the
presence or absence of GST-Sur2352-625, respectively, and ∆ꢀ is the
differential fluorescence intensity of 1 µM of 9 in the presence of an
infinite concentration of protein and in its absence. [9]o and [P]o are
the total concentrations of 9 and protein added, respectively.
NMR Studies. Wrenchnolol (7b) or its aminopropyl analogue (7a)
(0.5 mg) was dissolved in 0.5 mL of a phosphate-buffered saline (PBS)
solution (99.99% D2O, pD ) 6.8). Complete assignments of the proton
signals were achieved by a combination of NOESY, DQF-COSY, and
HOHAHA experiments (Table S1). A model structure of 7a was
acquired using energy minimization with CVFF force field in the Insight
2000 program (Accelrys) and adjusted on the basis of NOE connec-
tivities and J coupling constants. A concentrated PBS solution of GST-
Sur2352-625 was gradually added to an NMR sample of wrenchnolol
(7b), and a one-dimensional 1H spectrum was collected after each
addition of the protein. NMR experiments were performed on a Bruker
Avance 600 MHz spectrometer. The data were processed using FELIX
(Accelrys) and UXNMR (Bruker), and proton chemical shifts were ref-
erenced to the HOD resonance (4.70 ppm at 298 K; temperature cor-
rection factor: -0.0109 ppm/K). All one-dimensional 1H NMR spectra
were recorded with 12 ppm spectral width, 8192 data points, and 298
K temperature. Suppression of the HOD resonance was achieved by
an excitation sculpting gradient pulse. Two-dimensional experiments
of DQF-COSY, HOHAHA (88 ms mixing time), and NOESY (150
1
58%) as a yellow solid: IR (thin film) 3421, 2910, 1636 cm-1; H
NMR (CD3OD, 270 MHz) δH 8.31 (br. s, 0.6H), 8.19 (br. s, 0.4H),
7.83 (m, 1H), 7.73 (d, J ) 8.4 Hz, 2H), 7.53 (d, J ) 8.4 Hz, 1H), 7.46
(d, J ) 3.8 Hz, 1H), 7.25 (d, J ) 8.4 Hz, 2H), 7.38-6.57 (m, 10H),
5.55 (br. m, 1H), 4.45 (br. d, J ) 12.7 Hz, 2H), 4.40-4.20 (br. m,
2H), 4.20-3.91 (br. m, 4H), 3.91-3.52 (br. m, 2H), 3.40-3.20 (br.
m, 2H), 2.81-2.47 (br. m, 4H), 2.29 (s, 3H), 1.95 (br. s, 9H), 1.72 (br.
s, 6H), 1.67-0.45 (m, 16H); HRMS (ESI) exact mass calcd for
C66H74N6O11S2 + H requires m/z 1191.4935, found m/z 1191.4939.
Cell Viability Assays. SK-BR3 breast cancer cells were plated on
96-well plates at a density of 5 × 103 cells/well and maintained for 24
h. The cells were then incubated with varied concentrations of each
compound for 48 h, followed by addition of Premix WST-1 Solution
(Takara). Cell viability was estimated by measuring absorbance at 450
nm after 1 h with a SPECTRAmax microplate reader.
Transcription Reporter Gene Assay. We used the mammalian
expression vectors each encoding the activation domain of ESX, VP16,
or NF-κB p65 fused with the GAL4 DNA-binding domain (amino acids
1-94) as described. Each expression construct was cotransfected to
HEK293Tag human kidney cells together with a reporter plasmid in
(47) Dignam, J. D.; Martin, P. L.; Shastry, B. S.; Roeder, R. G. Methods Enzymol.
1983, 101, 582-598.
9
3470 J. AM. CHEM. SOC. VOL. 126, NO. 11, 2004

Design and Analysis of Wrenchnolol
A R T I C L E S
and 300 ms mixing times) were recorded with 4096 data points in the
t2 dimension and 512 increments. A 90° phase-shifted sine bell function
was used in the t2 and t1 dimensions prior to Fourier transformation of
NOESY, while a 30° phase-shifted skewed sine bell function was used
for HOHAHA and DQF-COSY. The final spectral matrices consisted
spectra were obtained in the Keck NMR facility at the University
of Houston. This work was supported by grants from the
American Cancer Society to M.U. (RGS-03-138-01-CDD), the
Welch Foundation to M.U. (Q-1490), and the Japanese Ministry
of Education, Culture, Sports, Science, and Technology to H.K.
(the 21st Century COE Program) and by fellowships from Japan
Society of the Promotion of Science to H.S., the U.S. Depart-
ment of Defense to Y.C., and the American Parkinson Disease
Association to Y.K.
1
of 1024 × 1 024 of the H frequency (F2 and F1) dimensions.
Cellular Distribution of Wrenchnolol. SK-BR3 cells grown on
cover slips were treated with 10 µM of FITC-conjugate 9 for 1 h. After
being washed with ice-cold PBS twice, the cells were fixed with 100%
methanol at -20 °C for 20 min. The cells were co-stained with DAPI.
Fluorescence images of the cells were captured by a Zeiss Axiovert
200M fluorescence microscope.
Supporting Information Available: Details of the synthetic
procedures and NMR data (Table S1 and Figures S1 and S2).
This material is available free of charge via the Internet at
http://pubs.acs.org.
Acknowledgment. We thank S. Wang and M. Hung for
supplying cell lines and assistance in fluorescence microscope
experiments, H. F. Gilbert for access to his fluorescence
spectrometer, and T. G. Wensel for editing the manuscript. NMR
JA038855+
9
J. AM. CHEM. SOC. VOL. 126, NO. 11, 2004 3471