Two-face, two-turn α-helix mimetics based on a cross-acridine scaffold: Analogues of the bim BH3 domain

Article abstract of DOI:10.1002/cbic.201402040

The design of a cross-acridine scaffold mimicking the i, i+3, i+5, and i+7 residues distributed over a two-face, two-turn α-helix is described. Docking studies and 2D ¹H,¹⁵N HSQC NMR spectroscopy provide compelling evidence that compound 3-d accurately reproduces the arrangement of four hotspots in the Bim BH3 peptide to permit binding to the Mcl-1 and Bcl-2 proteins (K_i 0.079 and 0.056 μM, respectively). Furthermore, the hotspot mutation could also be mimicked by individual or multiple deletions of side chains on the scaffold. Building site: A cross-acridine scaffold was designed to project functional groups with spatial and angular geometries that accurately mimic the i, i+3, i+5, and i+7 side chains on a two-turn, two-face section of an α-helix. The binding mode of the most potent compound, 3-d, to Mcl-1 was confirmed by ¹H, ¹⁵N HSQC NMR.

Full text of DOI:10.1002/cbic.201402040

CHEMBIOCHEM
COMMUNICATIONS
DOI: 10.1002/cbic.201402040
Two-Face, Two-Turn a-Helix Mimetics Based on a Cross-
Acridine Scaffold: Analogues of the Bim BH3 Domain
Xiangqian Li,[a] Ziqian Wang,[a] Yingang Feng,[b] Ting Song,[a] Pengchen Su,[a]
Chengbin Chen,[c] Gaobo Chai,[c] Ying Yang,[d] and Zhichao Zhang*[a]
The design of a cross-acridine scaffold mimicking the i, i+3,
i+5, and i+7 residues distributed over a two-face, two-turn a-
1
helix is described. Docking studies and 2D H,15N HSQC NMR
spectroscopy provide compelling evidence that compound 3d
accurately reproduces the arrangement of four hotspots in the
Bim BH3 peptide to permit binding to the Mcl-1 and Bcl-2 pro-
teins (Ki 0.079 and 0.056 mm, respectively). Furthermore, the
hotspot mutation could also be mimicked by individual or
multiple deletions of side chains on the scaffold.
Protein–protein interactions (PPIs) control essential cellular pro-
cesses, and misregulation of these interactions is often impli-
cated in disease states.[1] Artificial mimetics of a-helices offer
a basis for the development of PPI inhibitors because most
multiprotein complexes feature helices at the interface; these
interfaces are considered chemically tractable.[2] Proteomimet-
ics reproduce the arrangement of key side chains (“hotspots”)
in an a-helix and match their topography rather than replicat-
Scheme 1. Structure of terphenyl derivatives and series 1–4.
ing the helical conformation observed in helix stabilizers and
foldamers.[2b] Such proteomimetics might be more “lead-like”
and could overcome the inherent limitations of functional
small-molecule mimetics developed through modular ap-
proaches, which are not transferable to different PPIs.[3,4]
side chains of an a-helix.[5] Based on the terphenyl scaffold,
a number of a-helix-like scaffolds, such as diphenylindane,[6]
terephthalamide,[3] and terpyridine,[7] have been designed and
evaluated. However, none of these yielded a clinical drug.
Notably, proteomimetics typically impart functionality from
one face of the a-helix. Approximately 40% of helical interfa-
ces feature hot-spot residues on two or three faces.[1b] These
helices feature a hydrophilic surface on the opposite side of
a hydrophobic surface, and a-helical segments in peptide hor-
mones rely strongly on this amphipathicity for maximal inter-
action with target partner proteins.[8] Hence, mimics lacking
a hydrophilic region can result in not only suboptimal affinity
to target proteins but also lower selectivity, which significantly
limits the utility of these mimetics (potential promiscuity).
Previously, we reported the two-faced Bim mimetic scaf-
fold 1 (Scheme 1).[9] Based on 1, we designed a scaffold with
high structural rigidity; it can easily install up to four functional
groups to mimic the i, i+3, i+5, and i+7 residues located on
two faces over two turns of an a-helix. In addition, hotspot
mutations can be mimicked by deleting individual or multiple
functional groups from this scaffold. Our design concept will
help circumvent the significant chasm between elegantly de-
signed helix mimetics and the biological applications of these
mimetics.
The most attractive proteomimetic template, based on a ter-
phenyl scaffold, was first proposed by Hamilton in 2001
(Scheme 1).[4] A library of a-helix mimetics was subsequently
assembled (400 mixtures of 20 compounds, 8000 compounds)
in which the added subunits were designed to mimic all possi-
ble permutations of the naturally occurring i, i+4 (i+3), i+7
[a] X. Li, Z. Wang, T. Song, P. Su, Prof. Z. Zhang
State Key Laboratory of Fine Chemicals, School of Chemistry
Dalian University of Technology
2 Linggong Road, Dalian 116012 (P.R. China)
E-mail: zczhang@dlut.edu.cn
[b] Prof. Y. Feng
Shandong Provincial Key Laboratory of Energy Genetics
Qingdao Institute of Bioenergy and Bioprocess Technology
Chinese Academy of Sciences
189 Songling Road, Qingdao, Shandong 266101 (P.R. China)
[c] C. Chen, G. Chai
School of Life Science and Technology, Dalian University of Technology
2 Linggong Road, Dalian 116024 (P.R. China)
[d] Y. Yang
Public Laboratory of Bioenergy and Biofuels, Qingdao Institute of Bioenergy
and Bioprocess Technology
189 Songling Road, Qingdao, Shandong 266101 (P.R. China)
In complexes between BH3 a-helix and Bcl-2-like proteins,
critical interactions are found on two faces of the helix, at side
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cbic.201402040.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2014, 15, 1280 – 1285 1280

CHEMBIOCHEM
COMMUNICATIONS
www.chembiochem.org
scaffold, it is difficult for a flexible small-molecule scaffold to
be suspended with an ideal conformation in aqueous solution.
In other words, it is a challenge for small molecules to span
the cavity between D67 and I65. A group that is rigidly con-
nected to the skeleton might best mimic D67 and support the
skeleton. Thus, our efforts started with two series of com-
pounds (3 and 4 (control)) in which the rigid pyrogallol and
the relatively flexible formic acid were connected to 2, to
mimic D67 (i+5) or as a control, respectively (Scheme 1).
Briefly, bromination[12] of commercially available 3,4,5-tri-
methoxybenzoic acid gave the bromo-derivative 5, then 6 was
obtained by a substitution reaction,[13] and 7 was obtained by
subsequent chlorination.[14] These chloro derivatives were cou-
pled with the appropriate aniline to give 8, and then with ben-
zoyl chloride to give series 9;[15,16] the latter compounds were
demethylated with excess BBr3 in CH2Cl2 to afford the target
compounds 3 (Scheme 2).[17] The synthesis of 4 began with
Table 1. Structure and binding affinities of a-helix mimetics with Mcl-
l and Bcl-2 determined by FPA.
Compounds
Ki ÆSD (Mcl-1) [mm]
Ki ÆSD (Bcl-2) [mm]
3a
3b
3c
3d
11 a
11 b
12a
12b
12c
0.417Æ0.135
0.313Æ0.084
0.247Æ0.049
0.079Æ0.022
0.432Æ0.144
0.151Æ0.033
0.469Æ0.168
0.360Æ0.124
0.210Æ0.079
0.454Æ0.056
0.287Æ0.034
0.260Æ0.032
0.266Æ0.089
0.224Æ0.053
0.207Æ0.066
0.056Æ0.018
0.668Æ0.086
0.188Æ0.028
0.534Æ0.067
0.446Æ0.078
0.274Æ0.045
1.350Æ0.253
0.760Æ0.178
0.431Æ0.058
Figure 1. A) Representation of an a-helix in Bim showing L62 (i), I65 (i+3),
D67 (i+5) and F69 (i+7) residues. B) Lowest-energy conformation of a-helix
mimetic 2. C) Docking of Mcl-1 to 3a. D) Overlay of 3a and a-helix in Bim.
E) Docking of Mcl-1 to 3d. F) Overlay of 3d and a-helix in Bim.
13a
13b
(À)-gossypol
chains of i, i+3, i+5, and i+7 residues distributed over two
turns.[1b,10] The relative positions of these groups in Bim, L62 (i),
I65 (i+3), D67 (i+5), and F69 (i+7), are shown in Figure 1A,
and the proteomimetic compound 2 was designed based on
this structure (Scheme 1, Figure 1B). In the lowest-energy con-
formation of 2, four substituents project from the scaffold with
angular relationships and distances similar to those of the i,
i+3, i+5, and i+7 a-carbons in the a-helix.
treatment of the appropriate aniline with diethyl ethoxymeth-
ylenemalonate (EMME) to obtain the enamine derivatives, fol-
lowed by thermal cyclization in diphenyl ether (Scheme 3).
Refluxing these quinolone derivatives in phosphoryl chloride
afforded the chloro derivatives of quinolone,[18] which were
coupled with the appropriate aniline and benzyl bromide, fol-
lowed by hydrolysis to give the desired compound 4a.[19]
Fluorescence polarization assays (FPAs) revealed that 3a
binds to Mcl-1 and Bcl-2 with Ki values of 0.417 and 0.266 mm,
respectively (Table 1, Figure S1A and B in the Supporting Infor-
mation). In contrast, 4a at up to 100 mm exhibited no apprecia-
ble binding to the targets. Consistent with our experimental
results, the predicted binding model (Figure 1C and D) shows
that the pyrogallol group in 3a can mimic D67 and form a hy-
drogen bond with R263 in Mcl-1, whereas the very flexible
D67-mimicking formic acid in 4a cannot, thus confirming the
requirement for a rigid structure to span the two sides of the
a-helix. Strong support was also provided by molecular model-
ing studies; these revealed that the three substituted positions
(R1, R2, and R3) of 3a pointed to the p2, p3, and p4 pockets of
Mcl-1, respectively, and to residue H224 (Figure 1C).
As the i, i+3, and i+7 residues in an a-helix are not equi-
distant, the a-helix mimetic should have a corresponding
design.[11] The scaffold studied in this work has unequal distri-
butions of R1, R2, and R3 (the distance between R1 and R2 is
shorter than that between R2 and R3), thus it more accurately
reproduces the i, i+3, i+7 arrangement. Additionally, this
design avoids steric clashes of H224 between the p3 and p4
pockets in Mcl-1 and imparts structure features to the scaffold
to maintain binding in the groove of Mcl-1. Thus, this design
circumvents the limitations of terphenyl-like scaffolds, which
cannot recognize i+3 and i+4 because of head and tail flips.
Moreover, when terphenyl-like scaffolds reverse, i and i+7 also
switch.
In the Bim a-helix, the hydrophilic residue D67 is on the op-
posite side of the helix to the three hydrophobic residues. A
rigid skeleton is necessary to bridge the aqueous gap between
D67 and I65. Unlike for a large, amphipathic, and rigid helix
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2014, 15, 1280 – 1285 1281

CHEMBIOCHEM
COMMUNICATIONS
www.chembiochem.org
Scheme 2. Synthesis of a-helix mimetics 3a–3d. a) Br2, H2O, CHCl3; b) CuI, K2CO3, N2, EtOH; c) POCl3; d) HCl, EtOH; e) NaH, DMF; f) BBr3, CH2Cl2.
We performed heteronuclear
single quantum coherence
(HSQC) NMR spectroscopy by
using 15N-labeled Mcl-1 to identi-
fy the binding site of 3d. 2D
1H,15N HSQC NMR spectra of Mcl-
1 were recorded in the absence
(black) and presence (red) of 3d
(Figure S2). Figure 2A shows
a plot of chemical shift perturba-
tion against Mcl-1 residue. When
3d was added, approximately
80% of the residues that
showed chemical shift changes
of more than 0.05 ppm were lo-
cated in the BH3 binding
domain of Mcl-1 (Figure 2B).
Scheme 3. Synthesis of a-helix mimetic 4a. a) EMME; b) Ph2O, reflux; c) POCl3, reflux; d) C6H5NH2, HCl;
e) C6H5CH2Br, K2CO3,DMF; f) NaOH, C2H5OH; g) HCl, H2O.
To more closely reproduce the arrangement of the i, i+3,
and i+7 residues in Bim, compounds 3b–d were designed,
synthesized, and evaluated by FPAs (Table 1, Figure S1A and
B). Increasing steric bulk of the substituents (3b<3c<3d) re-
sulted in a corresponding increase in affinity (0.313, 0.247, and
0.079 mm for Mcl-1; 0.224, 0.207, and 0.056 mm for Bcl-2). The
most potent compound, 3d, exhibited an approximately five-
fold enhancement in Ki toward Mcl-1 and Bcl-2 over 3a. Dock-
ing studies showed that 3d maintained the mimicry of D67
and achieved functional mimicry of three additional residues,
L62, I65, and F69, which occupy the p2, p3, and p4 pockets in
Mcl-1 (Figure 1E). As shown in the overlay of 3d on the Bim a-
helix (Figure 1 f), the compound exactly reproduced the spatial
arrangement of L62, I65, D67, and F69, which are distributed
on two faces and two turns of the Bim a-helix.
Residue R263 and neighboring residues V253, D256, G257, and
N260 showed significant changes in chemical shift upon the
addition of 3d. H224, F228, L235, V249, V253, T266, and F270
in the p2 and p3 pockets and V216, G217, G219, V220 and
N223 in the p4 pocket were also significantly affected. Notably,
D218, Q221, R222, D236, and K244, which do not contact 3d,
showed significant chemical shift perturbations, probably as
a result of changes in the chemical environment as the groove
opened to accommodate 3d. The NMR results demonstrate
that 3d accurately mimics the L62, I65, F69, and D67 residues
of Bim.
BH3-only proteins (which include Bim, Bad, Bmf, Bid, Bik,
Puma, Noxa, and Hrk) have a striking ability to interact with
the binding grooves of multiple prosurvival proteins.[20] The
ability of BH3-only proteins to engage multiple binding part-
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2014, 15, 1280 – 1285 1282

CHEMBIOCHEM
COMMUNICATIONS
www.chembiochem.org
Figure 2. A) Chemical shift perturbations of Mcl-1 residues when bound to 3d. B) NMR-derived structure of 3d bound to Mcl-1. Mcl-1 residues with chemical
shift changes >0.05 ppm are shown.
ners and the local conformational adaptations that enable
binding are partially attributable to differences in crucial resi-
dues.[21] For example, no significant effects of the point muta-
tions L62A and F69A (or the L62A/F69A combination) on the
binding of Bim to Mcl-1 were observed, but these mutations
drastically affected binding of Bim to Bcl-xL, Bcl-w, and Bcl-2.[10]
In addition, the sequence or geometry of the area near the p3
and p4 pockets in Mcl-1 determines its specificity for Noxa and
Bim.[22] Mutation of F69 (p4 pocket) showed that this residue
confers Bim BH3 peptide-binding specificity for Mcl-1 but not
for Bcl-xL.[23]
Thus, we attempted to delete some functional groups of R1,
R2, and R3, to mimic a residue deletion mutant. Small mole-
cules that mimic hotspot mutations could probe the detailed
local structural environment that fine-tunes binding specificity
to multiple targets of prosurvival proteins. We also attempted
a proof-of-concept demonstration that residue mimics could
be easily installed and combined on our scaffold. Compounds
11–13 (Scheme 4) were designed with different combination
of functional groups. Series 11 mimics L62, I65, and D67,
whereas series 12 mimics I65, D67, and F69. Series 13 only
mimics I65 and D67.
These compounds were synthesized and evaluated by FPA
(Scheme 4, Table 1). Briefly, gallic acid derivatives 6a and 6c
were cyclized in polyphosphoric acid to give acridine deriva-
tives 10a–b,[13] which were converted to 11 a–c by acylation
and demethylation. Compounds 12a–c and 13a–b were pre-
pared from 8a–c and 10a–b by demethylation in HBr solu-
tion.[24]
Although the affinities of these smaller molecules toward
targets were weaker than for 3d, they retained sub-micromolar
affinities for Bcl-2 and Mcl-1 (Table 1, Figure S1A and B). Com-
pounds 11 b and 12c behaved similarly: two to five times less
potent than 3d. Compound 13b bound Mcl-1 and Bcl-2 with
Ki values of 0.287 and 0.760 mm, respectively (3 and 14 times
less potent than 3d, respectively). These results indicate that
removal of L62 or F69 from Bim mimetics has similar effects
on binding to Mcl-1 and Bcl-2. However, removal of both L62
and F69 in the mimetics had a more significant impact on
Scheme 4. Synthesis of a-helix mimetics 11 a–11 b, 12a–12c, and 13a–13b.
a) polyphosphoric acid; b) NaH, DMF; c) BBr3, CH2Cl2; d) 48% HBr.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2014, 15, 1280 – 1285 1283

CHEMBIOCHEM
COMMUNICATIONS
www.chembiochem.org
binding to Bcl-2 than on binding to Mcl-1. These variations are
consistent with experimental observations of the binding of
the Bim L62A/F69A double mutant to Mcl-1 and Bcl-2.[23]
The minimal scaffold 13a (mimicking only D67 in Bim)
bound to Mcl-1 and Bcl-2 with Ki values of 0.454 and 1.350 mm,
respectively (6 and 24 times less potent than 3d, respectively).
This indicates that D67 mimicry plays a more important role in
binding to Mcl-1 than in binding to Bcl-2; affinity for Bcl-2 is
more dependent on the mimicry of hydrophobic groups. This
is also consistent with the results of alanine scanning experi-
ments: only the D67A mutant had a significant effect on Mcl-
1 binding, whereas ABT-737 (mimicking L62 and F69 but not
D67) did not bind Mcl-1.[23,25] The more minimal scaffolds syn-
thesized in this study (groups deleted individually or in combi-
nation) could mimic hotspot mutations. The data also confirm
the one-to-one correspondence between the R1, R2, and R3
groups on our scaffold and the i, i+3, and i+7 residues on
Bim.
Acknowledgements
We thank Prof. Paramjit Arora (New York University) for insightful
discussion. The work was supported by the National Natural Sci-
ence Foundation of China (21372036, 81272876, and 81273436).
Keywords: acridines
·
helical structures
·
Mcl-1
·
peptidomimetics · two-face
[1] a) D. P. Ryan, J. M. Matthews, Curr. Opin. Struct. Biol. 2005, 15, 441–446;
b) B. N. Bullock, A. L. Jochim, P. S. Arora, J. Am. Chem. Soc. 2011, 133,
14220–14223.
[2] a) A. L. Jochim, P. S. Arora, ACS Chem. Biol. 2010, 5, 919–923; b) V. Az-
zarito, K. Long, N. S. Murphy, A. J. Wilson, Nat. Chem. 2013, 5, 161–173.
[3] H. Yin, G.-i. Lee, K. A. Sedey, J. M. Rodriguez, H.-G. Wang, S. M. Sebti,
A. D. Hamilton, J. Am. Chem. Soc. 2005, 127, 5463–5468.
[4] B. P. Orner, J. T. Ernst, A. D. Hamilton, J. Am. Chem. Soc. 2001, 123,
5382–5383.
[5] A. Shaginian, L. R. Whitby, S. Hong, I. Hwang, B. Farooqi, M. Searcey, J.
Chen, P. K. Vogt, D. L. Boger, J. Am. Chem. Soc. 2009, 131, 5564–5572.
[6] I. C. Kim, A. D. Hamilton, Org. Lett. 2006, 8, 1751–1754.
[7] J. M. Davis, A. Truong, A. D. Hamilton, Org. Lett. 2005, 7, 5405–5408.
[8] S. Marimganti, M. N. Cheemala, J.-M. Ahn, Org. Lett. 2009, 11, 4418–
4421.
To further identify the direct binding ability of these a-helix
mimetics to Mcl-1, isothermal titration calorimetry (ITC) was
performed (Figure S3A–D). The most potent compound, 3d,
exhibited a Kd value of 0.107 mm with Mcl-1. The corresponding
Kd values for 11 b, 12c, and 13b were 0.342, 0.507, and
0.781 mm, respectively. Consistent with this, the Ki values
(Table 1) show that the affinity of 3d was three to seven times
improved over that of 11 b, 12c, and 13b.
[9] Z. Zhang, X. Li, T. Song, Y. Zhao, Y. Feng, J. Med. Chem. 2012, 55,
10735–10741.
[10] E. F. Lee, P. E. Czabotar, M. F. van Delft, E. M. Michalak, M. J. Boyle, S. N.
Willis, H. Puthalakath, P. Bouillet, P. M. Colman, D. C. S. Huang, W. D. Fair-
lie, J. Cell Biol. 2008, 180, 341–355.
Subsequently, we tested the affinities of 3d (maximal) and
13a (minimal) with an R263A mutant of Mcl-1. ITC showed
both compounds lost most of the affinity: 3.06 mm for 3d
(nearly 30 times weaker than with wild-type Mcl-1, Figure S4),
and no appreciable binding for 13a at up to 100 mm. It further
confirmed that mimicking D67 to form a hydrogen bond with
R263 plays the most important role in binding to Mcl-1. This
explains why the naked scaffold 13a exhibits relatively good
affinity (sub-micromolar), because of capture of this very “hot”
residue. Also, the significant decrease in affinity with the one-
residue mutation rules out the possibility that the molecules
developed in this study are promiscuous binders that might
bind to many other proteins.
[11] a) M. Peters, M. Trobe, H. Tan, R. Kleineweischede, R. Breinbauer, Chem.
Eur. J. 2013, 19, 2442–2449; b) M. Peters, M. Trobe, R. Breinbauer, Chem.
Eur. J. 2013, 19, 2450–2456.
[12] A. I. Meyers, J. R. Flisak, R. A. Aitken, J. Am. Chem. Soc. 1987, 109, 5446–
5452.
[13] T. L. Su, B. Kohler, T. C. Chou, M. W. Chun, K. A. Watanabe, J. Med. Chem.
1992, 35, 2703–2710.
[14] C. Wang, S. Rakshit, F. Glorius, J. Am. Chem. Soc. 2010, 132, 14006–
14008.
[15] H. T. Nguyen Thi, C.-Y. Lee, K. Teruya, W.-Y. Ong, K. Doh-ura, M.-L. Go,
Bioorg. Med. Chem. 2008, 16, 6737–6746.
[16] C. Gao, Y. Jiang, C. Tan, X. Zu, H. Liu, D. Cao, Bioorg. Med. Chem. 2008,
16, 8670–8675.
[17] H. Gao, J. Kawabata, Bioorg. Med. Chem. 2005, 13, 1661–1671.
[18] a) R. d. R. Reis, E. C. Azevedo, M. C. B. V. de Souza, V. F. Ferreira, R. C.
Montenegro, A. J. Araffljo, C. Pessoa, L. V. Costa-Lotufo, M. O. de Moraes,
J. D. .B. M. Filho, A. M. T. de Souza, N. C. de Carvalho, H. C. Castro, C. R.
Rodrigues, T. R. A. Vasconcelos, Eur. J. Med. Chem. 2011, 46, 1448–1452;
b) Y. F. Suen, L. Robins, B. Yang, A. Verkman, M. H. Nantz, M. J. Kurth,
Bioorg. Med. Chem. Lett. 2006, 16, 537–540.
[19] a) I. P. Hçglund, S. Silver, M. T. Engstrçm, H. Salo, A. Tauber, H.-K. Kyyrç-
nen, P. Saarenketo, A.-M. HoffrØn, K. Kokko, K. Pohjanoksa, J. Sallinen, J.-
H. Savola, S. Wurster, O. A. Kallasta, J. Med. Chem. 2006, 49, 6351–6363;
b) G. Hiltensperger, N. G. Jones, S. Niedermeier, A. Stich, M. Kaiser, J.
Jung, S. Puhl, A. Damme, H. Braunschweig, L. Meinel, M. Engstler, U.
Holzgrabe, J. Med. Chem. 2012, 55, 2538–2548; c) D. Kaminsky, R. I.
Meltzer, J. Med. Chem. 1968, 11, 160–163.
The effect of 3d on plasmid DNA pBR322 was measured, be-
cause some acridines are well known DNA intercalators (Fig-
ure S5). No significant differences in superhelical content or
band shape were found (agarose gel electrophoresis) in the
presence or absence of 3d. This implies that 3d does not in-
teract with DNA in vitro. Therefore, it seemed very likely these
a-helix mimetics are selective Bcl-2/Mcl-1 protein inhibitors.
In conclusion, we designed a cross-acridine scaffold that
projects functional groups with spatial and angular geometries
that accurately mimic the i, i+3, i+5, and i+7 side chains on
a two-turn, two-face section of an a-helix. In addition, hotspot
mutations were mimicked by different parts of the scaffold
and combinations of side chains. The affinities of these com-
pounds toward Bcl-2 and Mcl-1 were determined by FPA, and
the binding mode of the most potent compound, 3d, to Mcl-
[20] G. J. P. Rautureau, C. L. Day, M. G. Hinds, Int. J. Mol. Sci. 2010, 11, 1808–
1824.
[21] a) C. L. Day, L. Chen, S. J. Richardson, P. J. Harrison, D. C. S. Huang, M. G.
Hinds, J. Biol. Chem. 2005, 280, 4738–4744; b) Q. Liu, T. Moldoveanu, T.
Sprules, E. Matta-Camacho, N. Mansur-Azzam, K. Gehring, J. Biol. Chem.
2010, 285, 19615–19624; c) L. Chen, S. N. Willis, A. Wei, B. J. Smith, J. I.
Fletcher, M. G. Hinds, P. M. Colman, C. L. Day, J. M. Adams, D. C. S.
Huang, Mol. Cell 2005, 17, 393–403.
[22] P. E. Czabotar, E. F. Lee, M. F. van Delft, C. L. Day, B. J. Smith, D. C. S.
Huang, W. D. Fairlie, M. G. Hinds, P. M. Colman, Proc. Natl. Acad. Sci. USA
2007, 104, 6217–6222.
1
1 was confirmed by H,15N HSQC NMR.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2014, 15, 1280 – 1285 1284

CHEMBIOCHEM
COMMUNICATIONS
www.chembiochem.org
[23] E. F. Lee, P. E. Czabotar, B. J. Smith, K. Deshayes, K. Zobel, P. M. Colman,
W. D. Fairlie, Cell Death Differ. 2007, 14, 1711–1713.
M. K. Joseph, S. Kitada, S. J. Korsmeyer, A. R. Kunzer, A. Letai, C. Li, M. J.
Mitten, D. G. Nettesheim, S. Ng, et al., Nature 2005, 435, 677–681.
[24] I. Kawasaki, K. Matsuda, T. Kaneko, Bull. Chem. Soc. Jpn. 1971, 44, 1986–
1987.
Received: February 17, 2014
[25] T. Oltersdorf, S. W. Elmore, A. R. Shoemaker, R. C. Armstrong, D. J.
Augeri, B. A. Belli, M. Bruncko, T. L. Deckwerth, J. Dinges, P. J. Hajduk,
Published online on May 18, 2014
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2014, 15, 1280 – 1285 1285