A credit-card library approach for disrupting protein-protein interactions

Article abstract of DOI:10.1016/j.bmc.2005.11.052

Protein-protein interfaces are prominent in many therapeutically important targets. Using small organic molecules to disrupt protein-protein interactions is a current challenge in chemical biology. An important example of protein-protein interactions is provided by the Myc protein, which is frequently deregulated in human cancers. Myc belongs to the family of basic helix-loop-helix leucine zipper (bHLH-ZIP) transcription factors. It is biologically active only as heterodimer with the bHLH-ZIP protein Max. Herein, we report a new strategy for the disruption of protein-protein interactions that has been corroborated through the design and synthesis of a small parallel library composed of 'credit-card' compounds. These compounds are derived from a planar, aromatic scaffold and functionalized with four points of diversity. From a 285 membered library, several hits were obtained that disrupted the c-Myc-Max interaction and cellular functions of c-Myc. The IC₅₀ values determined for this small focused library for the disruption of Myc-Max dimerization are quite potent, especially since small molecule antagonists of protein-protein interactions are notoriously difficult to find. Furthermore, several of the compounds were active at the cellular level as shown by their biological effects on Myc action in chicken embryo fibroblast assays. In light of our findings, this approach is considered a valuable addition to the armamentarium of new molecules being developed to interact with protein-protein interfaces. Finally, this strategy for disrupting protein-protein interactions should prove applicable to other families of proteins.

Full text of DOI:10.1016/j.bmc.2005.11.052

Bioorganic & Medicinal Chemistry 14 (2006) 2660–2673
A credit-card library approach for disrupting
protein–protein interactions
Yang Xu,a Jin Shi,b Noboru Yamamoto,a Jason A. Moss,a
Peter K. Vogtb,* and Kim D. Jandaa,*
aDepartment of Chemistry and Immunology, The Skaggs Institute for Chemical Biology and Worm Institute for Research
and Medicine (WIRM). The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla CA 92037, USA
bDepartment of Molecular and Experimental Medicine, The Scripps Research Institute,
10550 North Torrey Pines Road, La Jolla, CA 92037, USA
Received 23 November 2005; accepted 23 November 2005
Available online 27 December 2005
Abstract—Protein–protein interfaces are prominent in many therapeutically important targets. Using small organic molecules to
disrupt protein–protein interactions is a current challenge in chemical biology. An important example of protein–protein interac-
tions is provided by the Myc protein, which is frequently deregulated in human cancers. Myc belongs to the family of basic
helix–loop–helix leucine zipper (bHLH-ZIP) transcription factors. It is biologically active only as heterodimer with the bHLH-
ZIP protein Max. Herein, we report a new strategy for the disruption of protein–protein interactions that has been corroborated
through the design and synthesis of a small parallel library composed of Ôcredit-cardÕ compounds. These compounds are derived
from a planar, aromatic scaffold and functionalized with four points of diversity. From a 285 membered library, several hits were
obtained that disrupted the c-Myc–Max interaction and cellular functions of c-Myc. The IC50 values determined for this small
focused library for the disruption of Myc–Max dimerization are quite potent, especially since small molecule antagonists of pro-
tein–protein interactions are notoriously difficult to find. Furthermore, several of the compounds were active at the cellular level
as shown by their biological effects on Myc action in chicken embryo fibroblast assays. In light of our findings, this approach is
considered a valuable addition to the armamentarium of new molecules being developed to interact with protein–protein interfaces.
Finally, this strategy for disrupting protein–protein interactions should prove applicable to other families of proteins.
ꢀ 2005 Elsevier Ltd. All rights reserved.
1. Introduction
do not target a specific protein–protein interaction,
probably because of the complexity of the recognition
mechanism.8,9 To address such issues, the use of
combinatorial chemical libraries containing small organ-
ic molecules holds promise as a powerful approach in
the identification of novel lead compounds.7,10–15
The biological function of a protein is often determined
by its interactions with other protein molecules.1–4
Because of the critical role of such protein–protein inter-
actions in many cellular processes, the development of
modulators able to interfere with specific interactions
has emerged as an important goal. Ultimately, specific
control of protein functions may offer therapeutic
benefits. Cell-permeable small organic molecules are of
particular interest as modulators of protein–protein
interactions, but the discovery of effective compounds
is challenging.5–7 Chemical entities that have been
designed to interact with specific binding motifs often
Myc and its partner protein Max belong to the basic
helix–loop–helix leucine zipper (bHLH-LZ) protein
family.16–23 Myc is a transcription factor with oncogenic
and apoptotic potential.24–26 The role of Myc in tumor-
igenesis has been linked to its transcriptional activities,
promoting cell growth and repressing differentiation.
Myc shows gain of function in numerous and diverse
human cancers including BurkittÕs lymphoma, neuro-
blastoma, lung, breast, and colon carcinomas. The
oncogenic activity of Myc depends on dimerization with
Max via the helix–loop helix and leucine zipper
domains.27–30 Myc cannot form homodimers, and in
order to bind to its DNA target sequence, the E box
Keywords: Credit-card library; Myc–Max; Protein–protein interac-
tions; Oncogenic transformation; Transcriptional regulation.
*
Corresponding authors. Tel.: +1 858 784 2516; fax: +1 858 784
2595; e-mail: kdjanda@scripps.edu
0968-0896/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2005.11.052

Y. Xu et al. / Bioorg. Med. Chem. 14 (2006) 2660–2673
2661
element (CACGTG), Myc must associate with Max.
The inhibition of Myc–Max dimerization is an enticing
drug target, since inhibitors may prevent tumorigenesis
and, more generally, could serve as a proof-of-principle
for the design of small molecule inhibitors of protein–
protein interactions.
communication at the interface. Herein, we report the
design and parallel synthesis of a credit-card library.
Furthermore, we show the usefulness of the library for
the discovery of small molecule inhibitors of Myc–
Max dimerization and, importantly, the biological effect
of these inhibitors on Myc activity in chicken embryo
fibroblast assays.
The structural stability of protein–protein interactions
˚
2
derives from large (typically 1000–3000 A ), but relative-
ly shallow, interfaces.31–34 The difficulty of disrupting
such expansive interactions with small molecules has
been linked to the area of the more buried surfaces.
Breakthroughs in breaching such interactions have oc-
curred with the identification of Ôhot spotsÕ.35,36 These
domains have been characterized as shallow loci of
2. Results
2.1. Library design
To design and generate the library, we chose a multiple-
component reaction, namely the Ugi reaction.39–42 Mul-
tiple-component reactions use three or more reactants in
a one-pot procedure to form a product that contains
structural aspects of each of the components.43 The
reactions are versatile and can be carried out in solution
or on a solid support in parallel fashion. For the Ugi
reaction, we employed a naphthalene-based template44
as a rigid, planar, aromatic core (Scheme 1) and a set
of components that would introduce additional diversity
(Fig. 1).
2
˚
about 600 A found at, or near, the geometric center
of the protein–protein interface, and certain amino acids
in hot spot regions contribute significantly to the stabil-
ity of the protein–protein complexes. These amino acids
are usually aromatic, such as tryptophan, tyrosine, and
histidine, along with other hydrophobic residues.37,38
In considering structural information, modeling analy-
sis, and the recent identification of peptidomimetic and
small molecule inhibitors of protein–protein interac-
tions, we designed a new scaffold aimed at hot spot spec-
ificity and, hence, the disruption of protein–protein
interactions. We viewed hot spots at interfaces between
proteins as aromatic, slot-like regions or Ôcard readersÕ
characteristically different from the deeper binding clefts
often found within proteins. Logically, insertion of a
Ôcredit-cardÕ into the locale of the reading interface
should trigger an event at the interface. This conceptual
approach led to what we term Ôcredit-cardÕ compounds
upon which to base a small molecule library.
Hence, appended a-acylamino amide diversity would be
attached to the naphthyl scaffold 2 to generate general
structure 5 (Scheme 1). Parallel solution-phase method-
ology was employed to generate the library because of
its medium size, non-limiting scale, and ease of resynthe-
sis of hits in a short period of time.7,11,45 The general
building blocks for this library in total would allow an
extended structure into both two- and three-dimensional
space. To initially explore structure–activity relation-
ships surrounding the planar template, we functional-
ized the naphthalene core with motifs that spanned a
wide range of size, polarity, aromaticity, and hydro-
gen-bonding capability. Our hope was that a high degree
of diversity in scaffold functionalization would allow a
thorough investigation of the chemical basis for small
molecule inhibition of protein–protein interactions.
As the name implies, the credit-card library members are
simply planar, aromatic core structures elaborated with
chemical diversity. However, while this notion is rather
simplistic, we note that our overall structural design of
this library should provide favorable enthalpic contribu-
tions from van der Waals interactions, p-stacking,
possible desolvation, and favorable entropy gains
from hydrophobic effects. In total, the compounds are
intended to function as inhibitors of protein–protein
interactions or otherwise alter/disrupt the necessary
2.2. Library preparation
A two- or three-step synthesis was used as our general
procedure (Scheme 1). The key component to our credit-
Scheme 1. General synthetic design of the Ôcredit-cardÕ library. Reagents and conditions: (a) n-BuLi (2 equiv), THF, ꢀ78 ꢁC, then DMF; (b) R1Br,
K2CO3, DMF, 60 ꢁC; (c) R2NC, R3NH2, R4CO2H, MeOH (and CHCl3), reflux; (d) TFA, CH2Cl2, 0 ꢁC.

2662
Y. Xu et al. / Bioorg. Med. Chem. 14 (2006) 2660–2673
Figure 1. Components utilized in the library preparation.
card design was 6-hydroxy-naphthalene-2-carbaldehyde
2 that was obtained from commercially available
6-bromo-2-naphthol 1, using n-BuLi and DMF as an
electrophilic trapping agent (Scheme 1). Treatment of
2 with a variety of halides (R1X) provided 3 in good
to moderate yields under basic conditions. The Ugi reac-
tion and our library synthesis were conducted using a
combination of isocyanides (R2NC), amines (R3NH2),
and carboxylic acids (R4CO2H) with naphthol deriva-
tives 2 or 3. As anticipated, the Ugi reactions proceeded
well, and in virtually all cases the desired multi-compo-
nent products eluted between by-products and starting
materials on silica gel. Hence, products were easily puri-
fied by short column chromatography using 2 mL of sil-
ica gel in a 5 mL polypropylene fritted syringe, upon
eluting with 10–50% of EtOAc/hexane. When R1 was
the acetic acid tert-butyl ester, the product 4 was treated
with TFA to provide acid derivatives 5. Finally, all
compounds, 40 were selected. These 40 compounds were
studied in greater detail by electrophoretic mobility shift
assay (EMSA). Four representative examples are shown
in Figure 2.
DNA-binding by Myc requires dimerization with Max,
and inhibitors of dimerization are ipso facto inhibitors
of DNA-binding. EMSA was performed following
published techniques.11 The target oligonucleotide with
the consensus binding site of Myc–Max dimers was
50AGTTGACCACGTGGTCTGGG30. It was labeled
with c-32P-ATP by T4 polynucleotide kinase. The basic
helix–loop–helix leucine zipper domains of Myc (50 nM)
and of Max (27.5 nM), purified from Ni columns, were
first allowed to dimerize in the presence or absence of
various amounts of the compound for 1 h at room tem-
perature and were then incubated with the target oligo-
nucleotide at a concentration of 60 pg/lL (50,000 cpm)
for 30 min. Protein–DNA complexes were resolved on
4% acrylamide gels and detected by autoradiography.
From the four representative compounds shown in
Figure 3, dose–response curves using EMSA were deter-
mined and yielded IC50 values from 17 to 36 lM (Fig. 4
and Table 1).
1
library members were characterized by H NMR and
produced in high purities (greater than 95%) with
yields ranging from 30% to 99% on a multi-milligram
scale (typically 20–100 mg).
2.3. Screening for inhibitors of Myc–Max dimerization
Initial screening for compounds that interfere with the
Myc–Max interaction was carried out by fluorescence
resonance energy transfer (FRET) using the 96-well
microplate platform and following previously described
methods.11 Briefly, the basic helix–loop–helix leucine
zipper domains of Myc and of Max, linked to a cyan
fluorescent protein and yellow fluorescent protein,
respectively, were produced in bacteria and purified on
Ni columns using a His-Tag motif incorporated in the
constructs. Interaction of the purified proteins produced
a strong FRET signal characterized by an emission peak
at 525 nm. All compounds from the library were exam-
ined at 7.6 lM. A reduction in fluorescence at 525 nm by
25% was considered a positive hit and was selected for
further study. From the library of 285 credit card-like
The four selected compounds were reexamined in single-
cuvette FRET assays to confirm the EMSA results and
at 7.6 lM inhibited dimerization to extents of between
66% (NY2276) and 36% (NY2280) (data not shown).
In contrast, three control compounds representing
the core scaffolds 6–8 (Fig. 5) failed to inhibit the
Myc–Max FRET signal.
2.4. Biophysical analysis for putative denaturation of
Myc–Max complex by selected library members
To confirm that selected compounds disrupted the
Myc–Max complexation by a mechanism other than
non-specific denaturation, a synthetic Myc–Max dimer
was prepared by chemical synthesis; its helicity was then

Y. Xu et al. / Bioorg. Med. Chem. 14 (2006) 2660–2673
2663
Figure 2. Inhibition of Myc–Max DNA-binding as determined by EMSA. Left panel: control reactions showing the mobility shift induced by Myc–
Max, inhibition of the shift by a 100-fold excess of unlabeled probe, and failure of Myc–Max to bind to a mutated probe (50AGTTGACTAC
GTAGTCTGGG30). The four panels to the right depict EMSA inhibition achieved at various concentrations of four selected compounds (see Fig. 3)
from the 285-membered library.
Figure 3. Selected compounds from EMSA.
studied by far-UV circular dichroism in the presence and
absence of the selected compounds. For this purpose, a
covalent, ÔtetheredÕ Myc–Max dimer was prepared using
thioether ligation chemistry from separately synthesized
Myc and Max peptides bearing mutually reactive func-
tionalities. This chemistry was chosen because it can
be performed chemoselectively in the context of all other
functionalities in the Myc and Max peptides in aqueous
solution allowing for high solubility and facile monitor-
ing by reversed-phase (RP) HPLC. Hodges and

2664
Y. Xu et al. / Bioorg. Med. Chem. 14 (2006) 2660–2673
120
100
80
60
40
20
0
100
IC50 = 37 μM
IC50 = 17 μM
80
60
40
20
0
0
10
20
30
40
0
10
20
30
40
50
60
70
Concentration of NY2267 (μM)
Concentration of NY2276 (μM)
100
80
60
40
20
0
100
80
60
40
20
0
IC50 = 27 μM
IC50 = 28 μM
0
10
20
30
40
50
0
10
20
30
40
50
60
Concentration of NY2279 (μM)
Concentration of NY2280 (μM)
Figure 4. Dose–response determination by EMSA. EMSA was performed with serial dilutions of each compound and the degree of inhibition was
measured by densitometry of the shifted band. Densitometry was carried out by using the NIH Image 1.63 software program. IC50 values were
determined arithmetically from the slopes of the dose–response curves.
Table 1. IC50 values determined for the four compounds selected by
EMSA
Mpa-Max were prepared by automated, stepwise solid-
phase peptide synthesis (SPPS) using custom-modified
IC50 (lM)a
in situ neutralization protocols for Boc chemistry. Both
peptides were easily solubilized to 10 mg/mL in N2-
purged 0.5 M triethanolamine, pH 8.0, containing 6 M
guanidine hydrochloride. After 16 h reaction at ambient
temperature, RP-HPLC/MS analysis indicated that this
reaction was complete, and the product was purified by
preparative RP-HPLC (Fig. 7). It should be noted that
the product elutes only 0.2 min later than the unreacted
Mpa-Max peptide, which is commensurate with previ-
ous work46,47 MS analysis was critical in confirming
the absence of any contaminating unreacted material
in the full-length product.
Compound
NY2267
NY2276
NY2279
NY2280
36.5
17.3
26.9
28.1
a DNA-binding EMSA was performed with serial dilutions of each
compound and the degree of inhibition determined by densitometry
of the shifted band. Densitometry was performed using NIH Image
1.63 software program.
co-workers46,47 have prepared a covalent, tethered Myc–
Max dimer using a disulfide linkage and verified the
resulting conjugateÕs helicity. However, whereas the
formation of mixed homo- and heterodimers is often
problematic in the case of disulfide linkage chemistry,
thioether ligation is performed under reducing
conditions, thereby affording the desired compound in
near-quantitative yield.
The purified, lyophilized Myc–S–Max construct was
then reconstituted in phosphate-buffered saline (PBS),
pH 7.4, at 100 lM. This solution was then diluted
10-fold into separately prepared 25 lM solutions of
NY2267, NY2276, NY2279, and NY2280 in PBS
containing 5 vol % DMSO. After 14 h equilibration,
these four solutions were subjected to far-UV circular
dichroism analysis and compared with a solution of
The final step in the synthesis of the covalent Myc–S–
Max conjugate is shown in Figure 6. BrAc-Myc and
Figure 5. Control compounds, 6–8, used in the FRET, EMSA, and transcriptional regulation assays.

Y. Xu et al. / Bioorg. Med. Chem. 14 (2006) 2660–2673
2665
Figure 6. Synthesis of the Myc–S–Max construct by thioether ligation.
Figure 7. (A) Analytical RP-HPLC and (B) raw ESI-MS data for the purified Myc–S–Max construct. M/Z ion charge states are accurate to 0.05%
of theory (see Section 5).
Myc–S–Max only. After overnight incubation, no gross
change in the helicity of the Myc–S–Max construct was
observed (Fig. 8), both with regard to intensity and the
wavelength of global and local minima (208 and 221 nm,
respectively). This implies that at their IC50 concentra-
tion, these compounds do not exert their effects on the
Myc–Max dimer by a denaturation-type mechanism.
However, this CD experiment can not clarify if in the
presence of the four selected compounds, at their IC50
concentrations, the Myc–S–Max construct still exists
as a ÔtetheredÕ dimer. In total, these compounds may dis-
rupt the inter-helix contacts between the Myc and Max
sequences, thereby yielding a similar per-residue normal-
ized helicity signal while being spatially separated from
one another.
2.5. Effect of NY2267, NY2276, NY2279, and NY2280 on
oncogenic transformation
The four selected compounds inhibiting the Myc–Max
interactions in EMSA and FRET assays were investigat-
ed for their effects on oncogenic transformation induced
by retroviruses in chicken embryo fibroblasts (CEF).
Figure 8. Circular dichroism study of the synthetic Myc–S–Max
construct, alone and in the presence of the four NY compounds.

2666
Y. Xu et al. / Bioorg. Med. Chem. 14 (2006) 2660–2673
CEF were infected with viruses expressing the Myc, Jun,
P3K, and Src oncoproteins, and were then treated for
the duration of the experiment with 20 lM of the NY
and control compounds. Transformed cell foci were
counted after three weeks, and the efficiencies of trans-
formation were determined for the oncoprotein–com-
pound combinations (Table 2). In viewing Table 2, the
following conclusions can be made: (1) NY2267 and
NY2280 (see Fig. 9) strongly inhibited oncogenic trans-
formation induced by Myc; (2) NY2276 and NY2279
were not effective in transformation assays, although
NY2276 caused some cytotoxicity that varied in extent
with different avian embryos and appeared to affect pref-
erentially Myc-transformed cells; (3) NY2280 had a
modest inhibitory effect on focus formation by Jun
(Fig. 9); (4) three of the four compounds reduced focus
counts in cultures infected with P3K-expressing virus,
but similar reductions were seen with a control com-
pound, and we view this mild effect as non-specific.
2.6. Effect of NY2267, NY2276, NY2279, NY2280 and
simple core scaffolds 6–8 on transcriptional regulation
We tested NY2267, NY2276, NY2279, NY2280 and
three negative controls, 6–8, for their ability to affect
rates of transcription in reporter assays. Cells were
transfected with an expression construct of a transcrip-
tional activator (Myc, Jun, or NFkB) and with a lucifer-
ase reporter construct whose expression is activated by
the transfected activator. The luciferase values obtained
for an activator–reporter combination in the absence of
an experimental compound were set as 1.00. Luciferase
values obtained in the presence of test compounds were
expressed as fractions of these activator–reporter con-
trols (Table 3). Myc- and Jun-induced transcriptional
activation was inhibited by the four NY compounds.
There was no significant effect on transactivation by
NFkB with any of the compounds. Furthermore, the
negative control compounds did not interfere with the
transcriptional activation by Myc, Jun or NFkB.
Table 2. Efficiency of oncogenic transformation induced by various
oncoproteins in the presence of selected library compounds and
controls
3. Discussion
Compound
Oncoprotein
Jun
The conceptual and practical approaches toward organ-
ic reactions and drug discovery are in the midst of a rev-
olution. The discrete synthesis and biological screening
of individual compounds that often takes years to find
and optimize leads are being supplanted by emerging
technologies. On one hand, structure-based design has
entered a new era driven by advances in X-ray instru-
mentation, automation, and computation with a focus
on proteins derived from the genome project. On the
other hand, combinatorial chemistry has provided ac-
cess to a variety of compound libraries in a short period
of time, and the technologies are generally more readily
available.48–52 In conjunction with the latter approach,
Myc
Src
P3K
NY2267
NY2279
NY2276
NY2280
7
0.05 0.01 0.85 0.03 1.02 0.28 0.38 0.05
0.41 0.01 0.92 0.01 1.05 0.30 0.31 0.05
1.07 0.01 ND 0.42 0.01 0.95 0.20
0.06 0.01 0.80 0.03 0.28 0.05 0.32 0.03
1.06 0.02 1.24 0.32 ND
0.62 0.02 1.27 0.35 ND
1.37 0.63
0.40 0.25
8
Efficiency of transformation is defined as the focus forming titer of the
oncoprotein expression vector in the presence of compound over the
titer in the absence of compound. Strong inhibition of Myc-induced
oncogenicity is indicated by bold numbers.
ND = not determined.
Figure 9. Formation of transformed cell foci in monolayer cultures of CEF in the absence or presence of 20 lM NY2280.

Y. Xu et al. / Bioorg. Med. Chem. 14 (2006) 2660–2673
2667
Table 3. Effect of NY and control compounds on transcriptional
regulation
tional groups can efficiently interact with the protein
and account for large binding energies.
Compound
Transcriptional activation
In spite of the perceived difficulties associated with bind-
ing a vast surface area at a protein–protein interface, the
binding energy of the interactions within this space is
not evenly dispersed. The identification of protein hot
spots has provided new insight into the interface do-
main. As demonstrated by Wells and co-workers, one
can generate a comparative map of mutant proteins
where mutations that contribute most to the loss of
affinity are near each other on the protein surface and
that the residues are not necessarily contiguous in the
primary sequence of the protein.62 Surrounding the
hydrophobic hot spot are regions of residues that con-
tribute less to the stabilization of the complex. Thus,
Myca
Junb
NFkBc
None (vector control) 1.00
1.00
1.00
NY2267
NY2276
0.38 0.05 0.26 0.01 0.93 0.03
0.27 0.03 0.35 0.02 1.02 0.09
0.40 0.05 0.35 0.03 1.05 0.09
0.51 0.04 0.35 0.04 0.86 0.12
1.11 0.10 1.17 0.23 1.51 0.29
NY2279
NY2280
6
7
8
ND 1.26 0.27 0.78 0.17
0.88 0.11 1.21 0.15 1.12 0.15
a HEK293 cells were transfected with the pCVMV3HuMyc construct
which expresses human Myc from the CMV promoter in a pcDNA3
vector backbone. The reporter construct was pGL2M4, which con-
tains the luciferase gene controlled by four Myc binding sites in the
SV40 promoter of the pGL2 vector. Vector control represents
transfection of the Myc and reporter vectors without experimental
compound; it was arbitrarily set as 1. Standardization was carried
out with the Renilla vector pRLCM.
b JEG-3 cells were transfected with pHygEF2-c-Jun in which expres-
sion of full length c-Jun is controlled by the promoter of elongation
factor 1a. The reporter vector used was 3 · 11#/pRNGP-luciferase.
Vector control represents transfection of the Jun and reporter con-
structs without experimental compound. Standardization was carried
out with the pRLCMV Renilla expression vector.
c HEK293 cells were transfected with the CMX-p50 and CMX-p65
constructs expressing NFkB p50 and p65, respectively. The reporter
construct was NFkB-Luc containing five NFkB binding sites in the
promoter controlling the firefly luciferase gene. Vector control rep-
resents transfection of the NFkB and reporter vectors without
experimental compound; it was arbitrarily set as 1. Standardization
was carried out with Renilla vector pRLCMV.
2
˚
the ꢁ600 A area of the hot spot may be the size critical
to make a water-excluded seal and promote energetically
favorable contributions of hydrogen-bonding and elec-
trostatic interactions. Furthermore, just as solvation of
hydrophobic residues can be important in the unfolding
of proteins, solvation at the interface could disrupt pro-
tein–protein interactions. Hence, rather than dealing
with the formidable task of specifically binding the en-
tire protein–protein interface, we chose only to concep-
tualize the fundamental characteristics of the hot spot as
a relatively planar, hydrophobic region surrounded by
elements of protein specificity.
From active compounds discovered to disrupt protein–
protein interactions, several conclusions can be drawn.
With regard to peptides, virtually all structures possess
tyrosine, tryptophan, and/or arginine.32,33 For non-pep-
tide molecules, most of them contain multiple aromatic
rings connected by relatively rigid linkers that might be
expected to prevent intramolecular hydrophobic col-
lapse.33 Based on these observations, we chose a planar,
aromatic scaffold as an appropriate framework for our
credit-card design. Using this scaffold, we envisioned
that molecular diversity would be introduced in the form
of pendant polar/hydrophobic functionalities. There-
fore, the entire credit-card structure should provide the
necessary combination of planarity, p-stacking interac-
tions, hydrophilicity, and favorable entropy gain from
hydrophobic effects to cover the necessary surface area
and thus binding interactions with varying degrees of
specificity. To satisfy these conditions, we considered
diversity-oriented synthesis (DOS)48,50,63–65 and the sim-
ple appending of building blocks to a common molecu-
lar scaffold.66,67 We chose the latter, since it can be
highly efficient and amenable to a variety of chemistries.
While it may be argued that this approach provides lim-
ited access to three-dimensional space, our intent was to
create a small library (less than 300) to examine quickly
our proposed concept. Using multi-component coupling
reactions like the Ugi reaction provides ease of access to
inputs that allow the credit-card design with structural
variability in the final library.
high-throughput screening methods can rapidly identify
lead compounds and yield information for subsequent
optimization.53–56 The development and application of
combinatorial chemical libraries derived from split-mix
and/or parallel synthesis are now widely utilized in the
search for new drugs.
To exploit fully the interplay between proteins discov-
ered from the genome and combinatorial technology,
drug design concepts must be evolved by examining
alternative ways of thinking. A case in point can be
found in the disruption of protein–protein interactions
using small synthetic molecules that is currently consid-
ered a challenging target for drug discovery.57–61 One of
the most cited difficulties for a synthetic molecule in dis-
rupting protein–protein interactions is that relatively
large protein surface areas—on the order of 1000–
2
˚
3000 A in size and stabilized by multiple interac-
tions—must be Ôout-competedÕ by the target molecule.
The nature of the binding surface between two proteins
may be a more practical concern. These surfaces lack
distinct pockets and are largely devoid of discrete bind-
ing sites that could accommodate small molecules. The
situation is quite distinct from that of mimicking small
molecule ligands for enzymes or protein receptors in
which the combining sites are most often found at the
hydrophobic interior of proteins with little or no aque-
ous solvation. Within the confines of a highly organized
pocket evolved to bind a small substrate, ligand func-
The importance of the transcription factor Myc in onco-
genic transformation cannot be overstated. This protein
contributes directly to a number of human cancers, and
since its activity is dependent on binding to the

2668
Y. Xu et al. / Bioorg. Med. Chem. 14 (2006) 2660–2673
activation partner Max, inhibitors of the protein–protein
interaction between Myc and Max are of therapeutic
interest. The research groups of Boger and Prochownik
have studied small molecule inhibitors of Myc–Max inter-
actions.7,59 Boger and co-workers have used an elegant
solution-phase library approach to create and screen a li-
brary of 7000 members, while Prochownik screened a
10,000 member commercial library from Chembridge
using a yeast two-hybrid assay. Both groups have found
molecules that block Myc–Max interaction at the micro-
molar level. However, a direct comparison between the
two approaches cannot be made as the assay systems
are different. We note, however, that both groups found
molecules that we define as having a credit-card structure,
built on such frameworks as an isoindoline, 2-phenyl-
chroman, and a carbazol.
tive compounds NY2276 and NY2279 failed to affect
Myc focus formation. As judged by their ability to inter-
fere with transcriptional activation, these compounds
are capable of entering cells. They may have off-target
effects that cancel inhibition of transformation. The
resistance of Src-induced transformation to the Myc
inhibitors is unexpected, as Src transformation has been
reported to depend on Myc.68,69 However, the inhibitors
cause only a partial reduction in Myc-dependent tran-
scriptional regulation; the residual activity may be suffi-
cient to support Src-induced transformation. Previously
described inhibitors of Myc–Max dimerization also
failed to affect Src. The activity of the Myc inhibitors
in reporter assays measuring transcriptional activation
is in concert with the results in EMSA and FRET.
The negative control compounds, 6–8, had no effect on
the transcriptional activities, which demonstrated that
while the naphthyl-scaffold is a necessary component,
it also requires structural points of diversity to be effec-
tive. The inhibition of Jun-dependent transcriptional
activation requires further study, notably a determina-
tion of target molecules and binding sites. It is conceiv-
able that the leucine zipper domains of Myc–Max and of
Jun–Fos, Jun–Fra or Jun–ATF share some structural
features that would make them common targets of the
selected compounds. A possible interplay between Jun
and Myc also remains to be investigated. More informa-
tion on transcriptional effects of the compounds can also
be obtained by measuring expression levels of known
specific Myc, Jun or NFkB targets. Such studies are in
progress.
Validation of our approach has been achieved by the use
of credit-card compounds based on a naphthyl-scaffold.
However, a myriad of other scaffolds fit within our con-
ceptual framework. We prepared a small 285-membered
library using the Ugi reaction with four points of chem-
ical diversity. The components were readily obtained
from commercial sources, and because the reactions
were essentially one pot, work-up, isolation, and
resynthesis of ÔhitÕ compounds made the entire process
efficient. Based on a FRET assay between cyan fluores-
cent protein-tagged Myc and yellow fluorescent protein-
tagged Max, 40 compounds were selected from the
library that reduced dimer association and produced a
negative effect on the FRET signal at a concentration
of 7.6 lM. Four compounds advanced through the
screening process when a more stringent and time-con-
suming EMSA was used to detect disruption of Myc–
Max dimers. Using a synthetic Myc–Max dimer, the
spectra of a denaturation mechanism for disruption of
dimerization has been largely ruled out. Furthermore,
a potential explanation for these compoundsÕ mode of
action was proposed, namely, that they bind at the inter-
face between the neighboring Myc and Max sequences,
thus impeding their co-association without detriment
to helicity.
Taken in total these observations suggest that the
FRET, EMSA, and transcriptional reporter assays used
in concert can identify small-molecule inhibitory com-
pounds for disrupting protein–protein interactions
which we have termed Ôcredit-cardsÕ, and several of these
molecules form a subset of anti-oncogenic compounds.
Future investigations will look at the generality of such
an approach to other protein–protein interactions as
well as refinement of the specificity of these interactions.
The four compounds identified by EMSA (Fig. 2) are
composed of aromatic and other hydrophobic motifs.
Importantly, these compounds were potent enough to
allow for IC50 values to be obtained from the EMSA
(Fig. 3). Notably, NY2276 is one of the most potent
small molecule inhibitors discovered to date for block-
ing Myc–Max dimerization in vitro.7,59 Moreover, the
compound was identified from a library more than 10-
fold smaller than previously screened libraries. While
conclusions about structure–activity are premature, we
can make the following comments: (1) The tert-butyl es-
ter functionality at the R1 position is found as a com-
mon motif shared by all four compounds; (2) The R2
and R3 positions utilized hydrophobic cyclohexyl, ben-
zyl, p-methoxy-benzyl, and butyl functionalities; and
(3) The R4 position invoked an aromatic functionality.
4. Conclusion
The development of small molecules that can recognize
protein surfaces and disrupt protein–protein interac-
tions is of great interest, and such compounds have been
difficult to identify. The most challenging aspect of this
problem is the need to recognize relatively large and flat
surface areas as opposed to well-organized clefts, as
found, for example, in enzyme active sites. We have
shown how a small well-designed library based on what
we term Ôcredit-cardÕ structures can effectively inhibit
protein–protein interactions. The success of our ap-
proach is grounded upon a planar, aromatic scaffold
that allowed the introduction of molecular diversity to
achieve coverage and specificity at the interface domain.
Validation of the design was accomplished by disrupting
the Myc–Max interaction. Myc is a transcriptional reg-
ulator known to induce lymphoid tumors in animals and
its deregulation is associated with numerous types of
human cancers. The importance of this work is that it
Two of the four selected compounds, NY2267 and
NY2280, inhibited Myc-induced oncogenic transforma-
tion in cell culture. It is not clear why the EMSA-posi-

Y. Xu et al. / Bioorg. Med. Chem. 14 (2006) 2660–2673
2669
identifies a small molecule strategy that can disrupt pro-
tein–protein interactions and provides a thought pro-
voking hypothesis for previous/future reports of small
molecule activity at protein–protein interfaces.
K2CO3 (6.9 g, 50 mmol) and tert-butyl bromoacetate
(2.34 g, 12 mmol). After stirring for 1 h at 60 ꢁC, the
mixture was cooled to room temperature and diluted
with water. The product was extracted with diethyl
ether twice. The combined organic layers were washed
with water and then with brine. The organic layer
was dried over MgSO4 and concentrated in vacuo,
the residue purified by silica gel column chromatogra-
phy (20% EtOAc/hexane) to produce (6-formyl-naph-
thalen-2-yloxy)-acetic acid tert-butyl ester (2.63 g, 92%
yield) as white crystals. This material was utilized as
compound 3 for the preparation of NY2267, NY2276,
NY2279, and NY2280.
5. Experimental procedures
5.1. General procedure
The 1H NMR and 13C NMR spectra were recorded on a
Bruker AMX-400 or Varian Inova-400 instrument. The
following abbreviations were used to explain the multi-
plicities: s = singlet, d = doublet, t = triplet, q = quartet,
m = multiplet, and br = broad. High-resolution mass
spectra (HRMS) were recorded at The Scripps Research
Institute on a VG ZAB-ZSE mass spectrometer using
MALDI. All reactions were monitored by thin-layer
chromatography (TLC) carried out on 0.25 mm E.
Merck silica plates (60F-254), with fractions being visu-
alized by UV light. Column chromatography was
carried out with Mallinckrodt SilicAR 60 silica gel
(40–60 lM). Reagent grade solvents for chromatogra-
phy were obtained from Fisher Scientific. Reagents were
purchased at the highest commercial quality and used
without further purification. All reactions were carried
out under an argon atmosphere, unless otherwise noted.
Reported yields were determined after purification for a
homogeneous material. Compounds 7 and 8 were pur-
chased from Aldrich chemical company.
1H NMR (CDCl3, 400 MHz): d 1.51 (s, 9H), 4.67 (s,
2H), 7.10 (d, J = 2.4 Hz, 1H), 7.32 (dd, J = 2.4 Hz,
11.2 Hz, 1H), 7.78 (d, J = 8.4 Hz, 1H), 7.93 (d,
J = 8.4 Hz, 1H), 7.93 (d, J = 11.2 Hz, 1H) 8.21 (s, 1H),
10.1 (d, J = 0.4 Hz, 1H).
13C NMR (CDCl3, 100 MHz): d 28.0, 65.6, 82.7, 107.2,
119.8, 123.6, 127.8, 131.3, 137.9, 158.4, 167.4, 192.0.
MALDI-FTMS Calcd for C17H19O4 (M++H): 287.1278.
Found: 287.1283.
5.4. General procedure for the preparation of
Ugi 4-component reaction library
To a solution of the naphthal derivative 3 (0.2 mmol,
1.0 equiv) in MeOH were added acid (0.4 mmol,
2.0 equiv), amine 0.4 mmol, (2.0 equiv), and isocyanide
(0.4 mmol, 2.0 equiv). After stirring for 24 h at reflux,
the mixture was cooled to room temperature and con-
centrated in vacuo. The residue was purified by a
short silica gel column packed in a 5 mL Teflon syr-
inge with 10–50% EtOAc/hexane gradient to afford
the product. The products were analyzed by 1H
NMR and HRMS.
5.2. tert-Butyl 2-(naphthalene-2-yloxy)acetate (6)
To a solution of 2-naphthol (100 mg, 0.69 mmol) in
DMF (5 mL) were added K2CO3 (500 mg, 3.5 mmol)
and tert-butyl bromoacetate (135 mg, 0.69 mmol). After
stirring for 1 h at 60 ꢁC, the mixture was cooled to room
temperature and diluted with water. The product was
extracted with diethyl ether twice. The combined organ-
ic layers were washed with water and then with brine.
The organic layer was dried over MgSO4 and concen-
trated in vacuo. The residue was purified by silica gel
column chromatography (20% EtOAc/hexane) to pro-
duce compound 6 (130 mg, 55% yield) as a light yellow
oil.
5.5. (6-{Cyclohexylcarbamoyl-[(4-methoxy-benzyl)-(pyri-
dine-2-carbonyl)-amino]-methyl}-naphthalen-2- yloxy)-
acetic acid tert-butyl ester (NY2267)
NY2267 was produced as a colorless oil (59.5 mg, 47%
yield).
1H NMR (CDCl3, 400 MHz): d 1.51 (s, 9H), 4.64 (s,
2H), 7.01 (s, 1H), 7.25 (t, J = 8.84 Hz, 1H), 7.36 (t,
J = 7.45 Hz, 1H), 7.45 (t, J = 7.24 Hz, 1H), 7.71 (d,
J = 8.16, 1H), 7.77 (m, 2H).
1H NMR (DMSO-d, 400 MHz) d 0.85–1.25 (m, 5H),
1.43 (s, 9H), 1.48–1.75 (m, 5H), 3.54 (s, 2H), 3.58 (s,
3H), 4.30–4.52 (m, 1H), 4.76 (s, 2H), 5.83 (s, 1H), 6.32
(s, 1H), 6.43 (d, J = 7.77 Hz, 1H), 6.51 (d, J = 8.09 Hz,
1H), 6.62 (d, J = 7.75 Hz, 1H), 6.808 (d, J = 7.9 Hz,
1H), 7.16–8.67 (m, 10H).
13C NMR (CDCl3, 100 MHz): d 28.03, 65.75, 82.40,
106.99, 118.63, 123.93, 126.41, 126.79, 127.63, 129.29,
129.61, 134.24, 155.80, 167.91.
13C NMR (DMSO-d, 100 MHz) d 24.36, 25.08, 27.62,
32.02, 54.77, 55.03, 60.85, 64.15, 64.99, 81.37, 106.86,
112.70, 112.80, 118.52, 122.84, 124.57, 126.60, 127.38,
127.77, 128.12, 129.52, 130.45, 130.57, 131.26, 131.57,
133.36, 136.79, 137.41, 154.28, 155.79, 157.37, 167.60,
167.96, 169.50.
MALDI-FTMS Calcd for C16H18O3Na (M++Na):
281.1148. Found: 281.1146.
5.3. (6-Formyl-naphthalen-2-yloxy)-acetic acid tert-butyl
ester
To a solution of 6-hydroxy-naphthalene-2-carbalde-
hyde (1.72 g, 10 mmol) in DMF (30 mL) were added
MALDI-FTMS Calcd for C38H44N3O6 (M++H):
638.3224. Found: 638.3234.

2670
Y. Xu et al. / Bioorg. Med. Chem. 14 (2006) 2660–2673
5.6. [6-(Cyclohexylcarbamoyl-{(4-methoxy-benzyl)-[2-(4-
methoxy-phenyl)-acetyl]-amino}-methyl)-naphthalen-2-
yloxy]-acetic acid tert-butyl ester (NY2276)
J = 2.0 Hz, 8.8 Hz, 1H), 7.66 (d, J = 8.8 Hz, 1H), 7.67
(d, J = 8.8 Hz, 1H).
13C NMR (CDCl3, 100 MHz): d 13.4, 19.9, 28.0, 31.89,
40.0, 43.6, 47.2, 55.2, 62.9, 65.7, 82.5, 106.8, 114.0,
119.1, 126.9, 127.3, 127.5, 127.6, 128.3, 128.4, 128.5,
129.0, 129.8, 129.9, 130.7, 133.9, 138.0, 156.4, 158.4,
167.8, 169.9, 172.4.
NY2276 was produced as a colorless oil (132 mg, 99%).
1H NMR (CDCl3, 400 MHz): d 0.94–1.40 (m, 5H), 1.49
(s, 9H), 1.50–1.96 (m, 5H), 3.54–3.67 (m, 2H), 3.70 (s,
3H), 3.74–3.82 (m, 1H), 3.76 (s, 3H), 4.47 (d,
J = 17.2 Hz, 1H), 4.605 (s, 2H), 4.68 (d, J = 17.2 Hz,
1H), 5.82–5.90 (m, 1H), 5.97 (s, 1H), 6.67 (d,
J = 8.4 Hz, 2H,), 6.81 (d, J = 8.0 Hz, 2H), 6.93 (d,
J = 8.4 Hz, 2H), 6.99 (br s, 1H), 7.09 (d, J = 8.0 Hz,
2H), 7.20 (dd, J = 1.6 Hz, 9.2 Hz, 1H), d 37 (d,
J = 8.8 Hz, 1H), 7.57 (d, J = 8.8 Hz, 1H), 7.63 (d,
J = 9.2 Hz, 1H), 7.70 (s, 1H).
MALDI-FTMS Calcd for C42H44N27Na (M++Na):
711.3041. Found: 711.3016.
5.9. Synthetic Myc–S–Max dimer
N-terminal-bromoacetyl-Myc (BrAc-Myc) and N-ter-
minal-3-mercaptopropionyl-Max (Mpa-Max) peptides
were prepared by stepwise SPPS on a 1.0 mmol scale
using in situ neutralization protocols for Boc chemistry
custom-written for a CSBio 136 automated peptide
synthesizer (available by request from author). All
Boc amino acids and HBTU were obtained from Senn
Chemicals (Dielsdorf, Switzerland). p-Methylbenzhyd-
rylamine (MBHA) resin was prepared by Advanced
Chemtech as a custom synthesis at a loading of
0.64 mmol NH2/g (100–200 mesh). Side chain protec-
tions were as follows: Ser/Thr (Bzl), Asp/Glu (OcHex),
His (Bom and Dnp, vide infra), Tyr (2-BrZ), Lys (2-
ClZ), Arg (Tos), and Asn (Xan); all other amino acids
were incorporated without side chain protection. Tri-
fluoroacetic acid (Biograde) was from Halocarbon
(River Edge, NJ), N,N-dimethylformamide (BioAna-
lyzed) was from J. T. Baker (St. Louis, MO) and
EMD (Gibbstown, NJ), N,N-diisopropylethylamine
(Atofina EDIPA) was from Aldrich (Milwaukee,
WI), and anhydrous hydrogen fluoride (CP grade)
was from Matheson Gas (Cucamonga, CA). HF cleav-
age was performed in a Type II vacuum-driven HF
apparatus from Peptide Institute (Minoh, Osaka, Ja-
pan). All other reagents, solvents, and chemicals were
of the highest purity commercially available and used
as received. RP-HPLC was performed employing bina-
ry gradients of solvents A and B, where A is 0.1%
TFA in water and B is 0.09% TFA in acetonitrile.
Analytical RP-HPLC was performed using a Vydac
218TP5415 column at a flow rate of 1 mL/min, with
detection at 214 nm during a linear gradient of 10–
50%B over 30 min. Preparative RP-HPLC was per-
formed using a Vydac 218TP101522 column at a flow
rate of 10 mL/min, with detection at 230 nm during a
linear gradient of 25–45% B over 30 min. In all cases,
fractions were analyzed off-line using an ABI/Sciex
150EX single quadrupole mass spectrometer and
judged for purity after a consistent summing of 50
scans in multi-channel analysis (MCA) mode. For pre-
parative purification purposes, fractions that contained
no consistent charged species which accounted for
more than 10% of the total ion intensity were desig-
nated ÔpureÕ and pooled; the homogeneity of this pool
was verified by analytical RP-HPLC and was >95%.
All purification and manipulation steps were per-
formed at ambient temperature unless otherwise
indicated.
13C NMR (CDCl3, 100 MHz) d 24.6, 24.7, 25.4, 27.9,
32.6, 40.4, 48.4, 49.7, 55.1, 63.1, 65.6, 77.2, 82, 106.7,
113.7, 113.9, 126.7, 127.2, 127.4, 127.5, 128.78, 128.82,
129.3, 129.6, 129.7, 129.9, 130.7, 131.0, 133.8, 156.3,
158.4, 158.5, 167.7, 168.5, 173.1.
MALDI-FTMS Calcd for C41H48N2O7Na (M++Na):
703.3354. Found: 703.3380.
5.7. [6-(Benzylcarbamoyl-{(4-methoxy-benzyl)-[2-(4-
methoxy-phenyl)-acetyl]-amino}-methyl)-naphthalen-2-
yloxy]-acetic acid tert-butyl ester (NY2279)
NY2279 was produced as a colorless oil (133 mg, 99%).
1H NMR (CDCl3, 400 MHz): d 1.50 (s, 9H), 3.59–3.65
(m, 1H), 3.70 (s, 3H), 3.75 (s, 3H), 4.34–4.46 (m, 2H),
4.51 (d, J = 17.2 Hz, 1H), 4.70 (d, J = 17.2 Hz, 1H),
6.07 (s, 1H), 6.55 (m, 1H), 6.66 (d, J = 8.4 Hz, 2H),
6.79 (d, J = 8.4 Hz, 2H), 6.91 (d, J = 8.4 Hz, 2H), 6.99
(br s, 1H), 7.07 (d, J = 8.4 Hz, 2H), 7.15–7.32 (m, 6H),
7.37 (d, J = 9.2 Hz), 7.56 (d, J = 8.8 Hz, 1H), 7.59 (d,
J = 9.2 Hz, 1H), 7.69 (s, 1H).
13C NMR (CDCl3, 100 MHz): d 28.3, 40.6, 43.8, 50.1,
55.4, 63.5, 66.0, 82.7, 107.0, 114.1, 114.2, 114.3, 126.9,
127.48, 127.55, 127.7, 127.9, 128.8, 129.1, 130.0, 130.2,
130.6, 130.7, 134.2, 138.3, 156.7, 158.8, 158.9, 168.0,
170.0, 173.6.
MALDI-FTMS Calcd for C42H44N2O7Na (M++Na):
711.3041. Found: 711.3016.
5.8. [6-(Benzylcarbamoyl-{butyl-[2-(4-methoxy-phenyl)-
acetyl]-amino}-methyl)-naphthalen-2-yloxy]-acetic acid
tert-butyl ester (NY2280)
NY2280 was produced as a colorless oil (109 mg, 87%).
1H NMR (CDCl3, 400 MHz): d 0.66 (t, J = 7.2 Hz, 2H),
0.90–1.10 (m, 2H), 1.34–1.56 (m, 3H), 1.50 (s, 3H), 3.30–
3.37 (m, 2H), 3.50–3.80 (m, 2H), 3.78 (s, 3H), 4.42 (dd,
J = 5.6 Hz, 15.2 Hz, 1H), 4.48 (dd, J = 5.6 Hz, 15.2 Hz,
1H), 4.63 (s, 2H), 6.00 (s, 1H), 6.55 (t, J = 5.6 Hz,
1H), 6.76–6.86 (m, 2H), 6.94–7.34 (m, 10H), 7.40 (dd,

Y. Xu et al. / Bioorg. Med. Chem. 14 (2006) 2660–2673
2671
For BrAc-Myc, the N-terminus was bromoacetylated
using 5 equiv bromoacetic anhydride. This reagent was
prepared immediately prior to use by reacting 10 mmol
bromoacetic acid and 5 mmol N,N0diisopropylcarbodii-
mide in 10 mL CH2Cl2; after 10 min pre-activation, this
gelatinous mixture was diluted with one volume of
DMF and added to the pre-neutralized peptide-resin.
Ninhydrin analysis indicated that this coupling was
complete in 20 min, after which the peptide-resin was
dried in vacuo overnight and then treated with HF
(15 mL/g peptide-resin) containing 2.5 wt % each of p-
cresol and resorcinol for 1 h at 0 ꢁC. It should be noted
that His(Bom) was used for this synthesis because on-
resin thiolytic deprotection of His(Dnp) would have de-
stroyed the bromoacetyl moiety. After evaporation of
HF, the peptide was extracted from the resin with
TFA; this filtrate was then concentrated by centrifugal
evaporation and the peptide recovered and washed
using three successive triturations with 10 volumes of
diethyl ether (chilled to ꢀ20 ꢁC). The crude peptide
was then solubilized in aqueous acetonitrile (containing
0.1% TFA) and lyophilized for purification. For Mpa-
Max, S-trityl-mercaptopropionic acid (3 equiv) was cou-
pled to the N-terminus using BOP (4 equiv) with excess
DIEA in DMF. Ninhydrin analysis indicated that this
coupling was complete in 30 min, after which the
His(Dnp) protection was removed by three successive
Max protein to cyan fluorescent protein (CFP) or yellow
fluorescent protein (YFP) were described previously
(Berg, 2003). They were expressed by the histidine-tag
vector pET28a (Novagen) to produce MycCFP and
MaxYFP. A fusion construct of the Myc bHLHZip do-
main to green fluorescent protein (GFP) was produced
following the same protocols and was used in the elec-
trophoretic mobility shift assays. The bHLHZip domain
(amino acids 13–93) of rat Max with C-terminal AU1
tag was also cloned into pET28a. Proteins were ex-
pressed in Escherichia coli strain BL21(DE3), purified
by affinity chromatography on nickel columns, and dia-
lyzed against buffer containing 200 mM Hepes (pH 7.0),
500 mM KCl, 30 mM MgCl2, 2 mM DTT, and 10 mM
EDTA (referred to as 1· Max buffer).
5.12. Fluorescence resonance energy transfer
The bHLHZip domains of Myc and of Max were fused
to the N termini of CFP (MycCFP) and YFP (Max-
YFP), respectively. The fusions were expressed in E.
coli, purified, and allowed to dimerize, followed by exci-
tation of CFP at wavelength 433 nm. Dimerization gen-
erated a FRET spectrum characterized by a strong
emission signal of YFP at 525 nm. In control experi-
ments, the ratios of fluorescence intensities at 525 nm
over 475 nm were 1.7 at complete dimerization of
MycCFP with MaxYFP and 0.4 for the monomeric
state of Myc CFP. One hundred percent inhibition of
MycCFP/MaxYFP dimerization was achieved by addi-
tion of 100X molar excess of the bHLHZip domain of
Max, functioning as a competitive inhibitor of Myc–
Max dimerization, resulting in a 525/475 nm ratio of
slightly more than 0.4. To screen for potential inhibitors
of dimerization, compounds were added to a final con-
centration of 7.6 lM and 7.6% DMSO, and the mixture
of each compound with MycCFP and MaxYFP (at
80 nM monomer concentration) was incubated in
1· Max buffer for 1 h at room temperature. After excita-
tion of CFP at 433 nm, both the CFP fluorescence at
475 nm and the YFP fluorescence at 525 nm were mea-
sured in a 96-well fluorescence plate reader (Molecular
Devices). Dimerization generates a FRET signal that
causes the emission of CFP at 475 nm to decrease, while
enhancing the emission of YFP at 525 nm. Compounds
that dissociate MycCFP/MaxYFP dimers increase the
emission of CFP and decrease the emission of YFP. Pre-
sumptive inhibitors were retested in single-cuvettes on a
Perkin-Elmer LS 50B to confirm the fluorescence data.
10 min
batchwise
treatments
with
b-mercap-
toethanol:DIEA:DMF (2:1:7 by volume). The N-termi-
nal trityl thioether was then deprotected by three
successive 2 min batchwise treatments with TFA con-
taining 2 vol % each of TIS and 3,6-dioxa-1,8-octane-
dithiol. The peptide-resin was then dried in vacuo
overnight; HF cleavage and isolation of the peptide were
performed as in the case of BrAc-Myc, with the excep-
tion that 5 wt % p-cresol was used as the sole scavenger.
The Myc–S–Max dimer was prepared by a chemoselec-
tive thioether ligation strategy. Forty milligrams of each
peptide was dissolved in 4 mL N2-purged 0.5 M trietha-
nolamine containing 6 M GuHCl, buffered at pH 8.0.
After 16 h, RP-HPLC indicated that the reaction was
complete; this material was then purified by preparative
RP-HPLC and used in CD experiments. MW 7477.5,
ESI-MS M/Z ions (theory): M4+ 1870.4, M5+ 1496.5,
M6+ 1247.3, M7+ 1069.2, M8+ 935.7, M9+ 831.8; (ob-
served): M4+ 1870.6, M5+ 1496.7, M6+ 1247.1, M7+
1069.6, M8+ 936.0, M9+ 832.2.
5.10. CD experiments
Far-UV circular dichroism spectra were obtained on an
Aviv Model 202 instrument with a 1 cm path length
fused silica cell at 25 ꢁC. Purified Myc–S–Max was
reconstituted at 100 lM and diluted 10-fold into four
separate solutions of selected library members at
25 lM in phosphate-buffered saline, pH 7.4, containing
5 vol % DMSO. Data are corrected for blank absorbance
of each compound and are the average of three scans.
5.13. Electrophoretic mobility shift assay
MycGFP and Max were mixed and incubated with
screening compounds for 1 h at room temperature. The
candidate compounds were tested at the concentrations
indicated in the figure legends. All test mixtures contained
10% DMSO. A double-stranded DNA oligonucleotide
with the consensus binding site of Myc–Max dimers
(50AGTTGACCACGTGGTCTGGG 30) was then add-
ed. The DNA–protein interaction was allowed to proceed
for 30 min. As control, a mutated probe was used
(50AGTTGACTACGTAGTCTGGG30). Final concen-
trations were: 50 nM MycGFP, 27.5 nM Max, 200 mM
5.11. Recombinant proteins
The fusion constructs of the basic helix–loop–helix leu-
cine zipper (bHLHZip) domains of the Myc or the

2672
Y. Xu et al. / Bioorg. Med. Chem. 14 (2006) 2660–2673
Schandorff, S.; Shewnarane, J.; Vo, M.; Taggart, J.;
Goudreault, M.; Muskat, B.; Alfarano, C.; Dewar, D.;
Lin, Z.; Michalickova, K.; Willems, A. R.; Sassi, H.;
Nielsen, P. A.; Rasmussen, K. J.; Andersen, J. R.;
Johansen, L. E.; Hansen, L. H.; Jespersen, H.; Podtelejni-
kov, A.; Nielsen, E.; Crawford, J.; Poulsen, V.; Sorensen,
B. D.; Matthiesen, J.; Hendrickson, R. C.; Gleeson, F.;
Pawson, T.; Moran, M. F.; Durocher, D.; Mann, M.;
Hogue, C. W. V.; Figeys, D.; Tyers, M. Nature 2002, 415,
180–183.
Hepes (pH 7.0), 500 mM KCl, 30 mM MgCl2, 2 mM
DTT, 10 mM EDTA, 5% glycerol, 40 ng/lL salmon testis
DNA (Sigma), and 60 pg/lL 32P-labeled oligonucleotide
probe. Protein–DNA complexes were resolved on 4%
acrylamide gels and detected by autoradiography. The
concentration of compounds causing 50% inhibition of
Myc–Max dimerization was determined by densitometry
of the shifted band. Densitometry was performed by using
the NIH Image 1.63 software program.
2. Gavin, A. C.; Bosche, M.; Krause, R.; Grandi, P.;
Marzioch, M.; Bauer, A.; Schultz, J.; Rick, J. M.; Michon,
A. M.; Cruciat, C. M.; Remor, M.; Hofert, C.; Schelder,
M.; Brajenovic, M.; Ruffner, H.; Merino, A.; Klein, K.;
Hudak, M.; Dickson, D.; Rudi, T.; Gnau, V.; Bauch, A.;
Bastuck, S.; Huhse, B.; Leutwein, C.; Heurtier, M. A.;
Copley, R. R.; Edelmann, A.; Querfurth, E.; Rybin, V.;
Drewes, G.; Raida, M.; Bouwmeester, T.; Bork, P.;
Seraphin, B.; Kuster, B.; Neubauer, G.; Superti-Furga,
G. Nature 2002, 415, 141–147.
3. Berg, T. Angew. Chem., Int. Ed. 2003, 42, 2462–2481.
4. Veselovsky, A. V.; Ivanov, Y. D.; Ivanov, A. S.; Archa-
kov, A. I.; Lewi, P.; Janssen, P. J. Mol. Recognit. 2002, 15,
405–422.
5. Arkin, M. R.; Wells, J. A. Nat. Rev. Drug Disc. 2004, 3,
301–317.
6. Berg, T. Chembiochem 2004, 5, 1051–1053.
7. Berg, T.; Cohen, S. B.; Desharnais, J.; Sonderegger, C.;
Maslyar, D. J.; Goldberg, J.; Boger, D. L.; Vogt, P. K.
Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 3830–3835.
8. Arkin, M. R.; Randal, M.; DeLano, W. L.; Hyde, J.;
Luong, T. N.; Oslob, J. D.; Raphael, D. R.; Taylor,
L.; Wang, J.; McDowell, R. S.; Wells, J. A.; Braisted,
A. C. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 1603–
1608.
5.14. Luciferase assays
HEK293 cells were seeded at 5 · 104 in 24-well plates and
transfected on the next day with PolyFect (Qiagen) and
appropriate combinations of reporter and expression
vectors in the following amounts: 100 ng Myc-response
luciferase reporter vector pGLM4, 2 ng Renilla control
vector pRLCMV, and either 300 ng pcDNA3 or 300 ng
pCMV3HuMyc expressing the human Myc protein
JEG-3 cells were transfected with PolyFect and the fol-
lowing vectors: 100 ng Jun-dependent luciferase reporter
vector pRBGB + c-Jun-site, 2 ng Renilla control vector
pRLCMV, and either 300 ng c-Jun expression vector
pEF2hyg-cJun or control vector pEF2hyg. Luciferase
activity was assayed 48 h after transfection using the
dual-luciferase reporter assay system (Promega). Where
noted, compound was added to cells 24 h after transfec-
tion. All assays were normalized for transfection efficien-
cy by using a co-transfected-Renilla expression vector
pRLCMV vector. Assays were performed in triplicate.
5.15. Focus assays
9. Carlson, H. A. Curr. Opin. Chem. Biol. 2002, 6, 447–452.
10. Greenbaum, D. C.; Arnold, W. D.; Lu, F.; Hayrapetian,
L.; Baruch, A.; Krumrine, J.; Toba, S.; Chehade, K.;
Bromme, D.; Kuntz, I. D.; Bogyo, M. Chem. Biol. 2002, 9,
1085–1094.
11. Cheng, S.; Tarby, C. M.; Comer, D. D.; Williams, J. P.;
Caporale, L. H.; Myers, P. L.; Boger, D. L. Bioorg. Med.
Chem. 1996, 4, 727–737.
12. Terrett, N. K.; Gardner, M.; Gordon, D. W.; Kobylecki,
R. J.; Steele, J. Tetrahedron 1995, 51, 8135–8173.
13. Janda, K. D. Proc. Natl. Acad. Sci. U.S.A. 1994, 91,
10779–10785.
14. Gallop, M. A.; Barrett, R. W.; Dower, W. J.; Fodor, S. P.
A.; Gordon, E. M. J. Med. Chem. 1994, 37, 1233–1251.
15. Pavia, M. R.; Sawyer, T. K.; Moos, W. H. Bioorg. Med.
Chem. Lett. 1993, 3, 387–396.
CEF were seeded at 1.5 · 105 cells per 12-well tissue cul-
ture plate in HAMÕs F10 containing 10% FBS. One day
after seeding, the cells were infected with 10-fold serial
dilutions of oncogenic retroviruses. These viruses were:
(i) the RCAS viral vector expressing chicken cellular
Myc, (ii) the Prague strain of Rous sarcoma virus, coding
for the Src oncoprotein, and (iii) the avian sarcoma virus
17, expressing the oncoprotein v-Jun. The cultures were
then overlaid with nutrient agarose consisting of 57.5%
(v/v) of media (75% F10 2x, 5% FBS, 2% chicken serum,
15% tryptose–phosphate broth, 1.5% of L-glutamine/
penicillin/streptomycin solution, and screening com-
pounds in DMSO, final concentration of DMSO:
0.3%) and 42.5% (v/v) of 1.5% Sea Plaque Agarose. After
three weeks, cultures were stained with 2% crystal violet.
16. Duesberg, P. H.; Vogt, P. K. Proc. Natl. Acad. Sci. U.S.A.
1979, 76, 1633–1637.
17. Duesberg, P. H.; Bister, K.; Vogt, P. K. Proc. Natl. Acad.
Sci. U.S.A. 1977, 74, 4320–4324.
18. Amati, B.; Littlewood, T. D.; Evan, G. I.; Land, H.
EMBO J. 1993, 12, 5083–5087.
Acknowledgments
19. Atchley, W. R.; Fitch, W. M. Proc. Natl. Acad. Sci.
U.S.A. 1997, 94, 5172–5176.
20. Dang, C. V.; Lewis, B. C. J. Biomed. Sci. 1997, 4, 269–278.
21. Dang, C. V.; McGuire, M.; Buckmire, M.; Lee, W. M. F.
Nature 1989, 337, 664–666.
This work was supported by the Skaggs Institute for
Chemical Biology and by grants from the National Can-
cer Institute and the California Breast Cancer Research
Program.
22. Davis, L. J.; Halazonetis, T. D. Oncogene 1993, 8, 125–
132.
References and notes
23. Fisher, D. E.; Parent, L. A.; Sharp, P. A. Proc. Natl. Acad.
Sci. U.S.A. 1992, 89, 11779–11783.
1. Ho, Y.; Gruhler, A.; Heilbut, A.; Bader, G. D.; Moore, L.;
Adams, S. L.; Millar, A.; Taylor, P.; Bennett, K.;
Boutilier, K.; Yang, L. Y.; Wolting, C.; Donaldson, I.;
24. Dang, C. V.; Resar, L. M. S.; Emison, E.; Kim, S.; Li, Q.;
Prescott, J. E.; Wonsey, D.; Zeller, K. Exp. Cell Res. 1999,
253, 63–77.

Y. Xu et al. / Bioorg. Med. Chem. 14 (2006) 2660–2673
2673
25. Morrish, F.; Giedt, C.; Hockenbery, D. Genes Dev. 2003,
17, 240–255.
49. Edwards, P. J. Comb. Chem. High Throughput Screening
2003, 6, 11–27.
26. Nesbit, C. E.; Tersak, J. M.; Grove, L. E.; Drzal, A.;
Choi, H.; Prochownik, E. V. Oncogene 2000, 19, 3200–
3212.
27. Amati, B.; Brooks, M. W.; Levy, N.; Littlewood, T. D.;
Evan, G. I.; Land, H. Cell 1993, 72, 233–245.
28. Blackwood, E. M.; Eisenman, R. N. Science 1991, 251,
1211–1217.
29. Blackwood, E. M.; Luscher, B.; Eisenman, R. N. Genes
Dev. 1992, 6, 71–80.
30. Blackwood, E. M.; Luscher, B.; Kretzner, L.; Eisenman,
R. N. Cold Spring Harb. Symp. Quant. Biol. 1991, 56, 109–
117.
50. Spring, D. R.; Krishnan, S.; Blackwell, H. E.; Schreiber, S.
L. J. Am. Chem. Soc. 2002, 124, 1354–1363.
51. Bhattacharyya, S. Curr. Med. Chem. 2001, 8, 1383–1404.
52. Sternson, S. M.; Wong, J. C.; Grozinger, C. M.; Schreiber,
S. L. Org. Lett. 2001, 3, 4239–4242.
53. Ozbal, C. C.; LaMarr, W. A.; Linton, J. R.; Green, D. F.;
Katz, A.; Morrison, T. B.; Brenan, C. J. H. Assay Drug
Dev. Technol. 2004, 2, 373–381.
54. Cheon, H. G.; Lee, C. M.; Kim, B. T.; Hwang, K. J.
Bioorg. Med. Chem. Lett. 2004, 14, 2661–2664.
55. Parniak, M. A.; Min, K. L.; Budihas, S. R.; Le Grice, S. F.
J.; Beutler, J. A. Anal. Biochem. 2003, 322, 33–39.
56. Koehler, R. T.; Villar, H. O. J. Comput. Chem. 2000, 21,
1145–1152.
31. Lo Conte, L.; Chothia, C.; Janin, J. J. Mol. Biol. 1999,
285, 2177–2198.
32. Stites, W. E. Chem. Rev. 1997, 97, 1233–1250.
33. Jones, S.; Thornton, J. M. Proc. Natl. Acad. Sci. U.S.A.
1996, 93, 13–20.
34. Lawrence, M. C.; Colman, P. M. J. Mol. Biol. 1993, 234,
946–950.
35. Hu, Z. J.; Ma, B. Y.; Wolfson, H.; Nussinov, R.
Proteins—Struct. Funct. Genet. 2000, 39, 331–342.
36. Bogan, A. A.; Thorn, K. S. J. Mol. Biol. 1998, 280, 1–9.
37. DeLano, W. L. Curr. Opin. Struct. Biol. 2002, 12, 14–20.
38. Ma, B. Y.; Wolfson, H. J.; Nussinov, R. Curr. Opin.
Struct. Biol. 2001, 11, 364–369.
39. Domling, A.; Ugi, I. Angew. Chem., Int. Ed. 2000, 39,
3169–3210.
40. Ugi, I. Angew. Chem., Int. Ed. Engl. 1962, 74, 9.
41. Ugi, I. Angew. Chem., Int. Ed. Engl. 1959, 71, 386.
42. Ugi, I.; Meyr, R. Angew. Chem., Int. Ed. Engl. 1958, 70,
702–703.
43. Posner, G. H. Chem. Rev. 1986, 86, 831–844.
44. Kolasa, T.; Gunn, D. E.; Bhatia, P.; Basha, A.; Craig, R.
A.; Stewart, A. O.; Bouska, J. B.; Harris, R. R.;
Hulkower, K. I.; Malo, P. E.; Bell, R. L.; Carter, G. W.;
Brooks, C. D. W. J. Med. Chem. 2000, 43, 3322–3334.
45. Boger, D. L.; Desharnais, J.; Capps, K. Angew. Chem.,
Int. Ed. 2003, 42, 4138–4176.
46. Lavigne, P.; Kondejewski, L. H.; Houston, M. E.;
Sonnichsen, F. D.; Lix, B.; Sykes, B. D.; Hodges, R. S.;
Kay, C. M. J. Mol. Biol. 1995, 254, 505–520.
47. Lavigne, P.; Crump, M. P.; Gagne, S. M.; Hodges, R. S.;
Kay, C. M.; Sykes, B. D. J. Mol. Biol. 1998, 281, 165–181.
48. Burke, M. D.; Schreiber, S. L. Angew. Chem., Int. Ed.
2004, 43, 46–58.
57. Li, L.; Thomas, R. M.; Suzuki, H.; De Brabander, J. K.;
Wang, X. D.; Harran, P. G. Science 2004, 305, 1471–1474.
58. Muller, O.; Gourzoulidou, E.; Carpintero, M.; Karaguni,
I. M.; Langerak, A.; Herrmann, C.; Moroy, T.; Klein-
Hitpass, L.; Waldmann, H. Angew. Chem., Int. Ed. 2004,
43, 450–454.
59. Yin, X. Y.; Giap, C.; Lazo, J. S.; Prochownik, E. V.
Oncogene 2003, 22, 6151–6159.
60. Cochran, A. G. Chem. Biol. 2000, 7, R85–R94.
61. Ferrer, M.; Kapoor, T. M.; Strassmaier, T.; Weissenhorn,
W.; Skehel, J. J.; Oprian, D.; Schreiber, S. L.; Wiley, D.
C.; Harrison, S. C. Nat. Struct. Biol. 1999, 6, 953–960.
62. Pearce, K. H.; Ultsch, M. H.; Kelley, R. F.; deVos, A. M.;
Wells, J. A. Biochemistry 1996, 35, 10300–10307.
63. Taylor, S. J.; Taylor, A. M.; Schreiber, S. L. Angew.
Chem., Int. Ed. 2004, 43, 1681–1685.
64. Schreiber, S. L. Abstr. Papers Am. Chem. Soc. 2003, 226,
U238.
65. Schreiber, S. L. Abstr. Papers Am. Chem. Soc. 2002, 224,
U177.
66. Dewitt, S. H.; Kiely, J. S.; Stankovic, C. J.; Schroeder, M.
C.; Cody, D. M. R.; Pavia, M. R. Proc. Natl. Acad. Sci.
U.S.A. 1993, 90, 6909–6913.
67. Bunin, B. A.; Ellman, J. A. J. Am. Chem. Soc. 1992, 114,
10997–10998.
68. Bowman, T.; Broome, M. A.; Sinibaldi, D.; Wharton, W.;
Pledger, W. J.; Sedivy, J. M.; Irby, R.; Yeatman, T.;
Courtneidge, S. A.; Jove, R. Proc. Natl. Acad. Sci. U.S.A.
2001, 98, 7319–7324.
69. Johnson, D.; Agochiya, M.; Samejima, K.; Earnshaw, W.;
Frame, M.; Wyke, J. Cell Death Differ. 2000, 7, 685–696.