Molecular insights into azumamide E histone deacetylases inhibitory activity

Article abstract of DOI:10.1021/ja0686256

Azumamide E, a cyclotetrapeptide isolated from the sponge Mycale izuensis, is the most powerful carboxylic acid containing natural histone deacetylase (HDAC) inhibitor known to date. In this paper, we describe design and synthesis of two stereochemical va

Full text of DOI:10.1021/ja0686256

Published on Web 02/21/2007
Molecular Insights into Azumamide E Histone Deacetylases
Inhibitory Activity
Nakia Maulucci,† Maria Giovanna Chini,‡ Simone Di Micco,‡ Irene Izzo,*,†
Emiddio Cafaro,† Adele Russo,† Paola Gallinari,§ Chantal Paolini,§
Maria Chiara Nardi,§ Agostino Casapullo,‡ Raffaele Riccio,‡ Giuseppe Bifulco,*,‡
and Francesco De Riccardis*,†
Contribution from the Department of Chemistry, UniVersity of Salerno, Via Ponte Don Melillo,
84084 Fisciano, Salerno, Italy, Department of Pharmaceutical Sciences, UniVersity of Salerno,
Via Ponte Don Melillo, 84084 Fisciano, Salerno, Italy, and Department of Biochemistry,
Istituto di Ricerche di Biologia Molecolare “P. Angeletti”, Via Pontina, Km 30,600,
00040 Pomezia, Rome, Italy
Received December 7, 2006; E-mail: dericca@unisa.it; bifulco@unisa.it; iizzo@unisa.it
Abstract: Azumamide E, a cyclotetrapeptide isolated from the sponge Mycale izuensis, is the most powerful
carboxylic acid containing natural histone deacetylase (HDAC) inhibitor known to date. In this paper, we
describe design and synthesis of two stereochemical variants of the natural product. These compounds
have allowed us to clarify the influence of side chain topology on the HDAC-inhibitory activity. The present
contribution also reveals the identity of the recognition pattern between azumamides and the histone
deacetylase-like protein (HDLP) model receptor and reports the azumamide E unprecedented isoform
selectivity on histone deacetylases class subtypes. From the present studies, a plausible model for the
interaction of azumamides with the receptor binding pocket is derived, providing a framework for the rational
design of new cyclotetrapeptide-based HDAC inhibitors as antitumor agents.
Introduction
HDAC inhibitors are divided in six distinct groups of zinc
chelators.8 Among them, azumamide A-E (1-5), recently
isolated by Nakao et al. from the marine sponge Mycale izuensis,
collocate in the small family of hydrophobic cyclotetrapeptides
(Figure 1).9
Histone deacetylases (HDACs) are a family of evolutionarily
conserved metalloenzymes that, interfering with posttranslational
chromatin modification,1 modulate cell differentiation,2 apop-
tosis,3 cell-cycle progression,4 and angiogenesis.5
Structurally, azumamides include three D-R-amino acids (D-
Phe, D-Tyr, D-Ala, D-Val) and a unique â-amino acid assigned
as (Z,2S,3R)-3-amino-2-methyl-5-nonenedioic acid 9-amide
(amnaa), in azumamides A (1), B (2), and D (4), and (Z,2S,3R)-
3-amino-2-methyl-5-nonenedioic acid (amnda), in azumamides
C (3) and E (5). When compared with the other natural
cyclopeptide HDAC inhibitors (e.g., trapoxin A (6),10 apicidin
D1,11 FR-235222,12 FR-225497,13 chlamydocin,14 HC toxin,15
WF-3161,16 Cyl-217), azumamides show an intriguing retro-
enantio arrangement,18 displaying an inverse direction of the
For their epigenetic effects on gene expression,6 HDAC
inhibitors have been considered promising anticancer agents and
many natural and synthetic inhibitors have entered in clinical
trials as possible antitumor agents.7
† Department of Chemistry, University of Salerno.
‡ Department of Pharmaceutical Sciences, University of Salerno.
§ Istituto di Ricerche di Biologia Molecolare “P. Angeletti”.
(1) Bayle, J. H.; Crabtree, G. R. Chem. Biol. 1997, 4, 885-888.
(2) Munster, P. N.; Troso-Sandoval, T.; Rosen, N.; Rifkind, R.; Marks, P. A.;
Richon, V. M. Cancer Res. 2001, 61, 8492-8497.
(3) Guo, F.; Sigua, C.; Tao, J.; Bali, P.; Gorge, P.; Li, Y.; Wittmann, S.;
Moscinski, L.; Atadja, P.; Bhalla, K. Cancer Res. 2004, 64, 2580-2589.
(4) (a) Sambucetti, L.; Fischer, D. D.; Zabludoff, S.; Kwon, P. O.; Chamberlin,
H.; Trogani, N.; Xu, H.; Cohen, D. J. Biol. Chem. 1999, 274, 34940-
34947. (b) Sato, N.; Ohta, T.; Kitagawa, H.; Kayahara, M.; Ninomiya, I.;
Fushida, S.; Fujimura, T.; Nishimura, G.; Shimizu, K.; Miwa, K. Int. J.
Oncol. 2004, 24, 679-685.
(10) (a) Itazaki, H.; Nagashima, K.; Sugita, K.; Yoshida, H.; Kawamura, Y.;
Yasuda, Y.; Matsumoto, K.; Ishii, K.; Uotani, N.; Nakai, H.; Terui, A.;
Yoshimatsu, S.; Ikenishi, Y.; Nakagawa, Y. J. Antibiot. 1990, 43, 1524-
1532. (b) Kijima, M.; Yoshida, M.; Sugita, K.; Horinouchi, S.; Beppu, T.
J. Biol. Chem. 1993, 268, 22429-22435. (c) Taunton, J.; Hassig, C. A.;
Schreiber, S. L. Science 1996, 272, 408-411.
(11) Singh, S. B.; Zink, D. L.; Polishook, J. D.; Dombrowski, A. W.; Darkin-
Rattray, S. J.; Schmatz, D. M.; Goetz, M. A. Tetrahedron Lett. 1996, 37,
8077-8080.
(12) (a) Mori, H.; Urano, Y.; Abe, F.; Furukawa, S.; Tsurumi, Y.; Sakamoto,
K.; Hashimoto, M.; Takase, S.; Hino, M.; Fujii, T. J. Antibiot. 2003, 56,
72-79. (b) Mori, H.; Urano, Y.; Abe, F.; Furukawa, S.; Furukawa, S.;
Sakai, F.; Hino, M.; Fujii, T. J. Antibiot. 2003, 56, 80-86.
(13) Mori, H.; Fumie, A.; Seiji, Y.; Shigehiro, T.; Motohiro, H. PCT Application
WO 00/08048, 2000.
(5) Deroanne, C. F.; Bonjean, K.; Servotte, S.; Devy, L.; Colige, A.; Clausse,
N.; Blacher, S.; Verdin, E.; Foidart, J.-M.; Nusgens, B. V.; Castronovo, V.
Oncogene 2002, 21, 427-436.
(6) Biel, M.; Wascholowski, V.; Giannis, A. Angew. Chem. 2005, 117, 3248-
3280; Angew. Chem., Int. Ed. 2005, 44, 3186-3216.
(7) Minucci, S.; Pelicci, P. G. Nat. ReV. Cancer 2006, 6, 38-51.
(8) Miller, T. A.; Witter, D. J.; Belvedere, S. J. Med. Chem. 2003, 46, 5097-
5116.
(9) (a) Isolation: Nakao, Y.; Yoshida, S.; Matsunaga, S.; Shindoh, N.; Terada,
Y.; Nagai, K.; Yamashita, J. K.; Ganesan, A.; van Soest, R. W. M.; Fusetani,
N. Angew. Chem., Int. Ed. 2006, 45, 7553-7557. (b) Synthesis: Izzo, I.;
Maulucci, N.; Bifulco, G.; De Riccardis, F. Angew. Chem., Int. Ed. 2006,
45, 7557-7560.
(14) Closse, A.; Huguenin, R. HelV. Chim. Acta 1974, 57, 533-545.
(15) Pringle, R. B. Plant Physiol. 1970, 46, 45-49.
9
10.1021/ja0686256 CCC: $37.00 © 2007 American Chemical Society
J. AM. CHEM. SOC. 2007, 129, 3007-3012
3007

A R T I C L E S
Maulucci et al.
Figure 1. Structures of azumamides A-E (1-5).
Figure 3. Formulas and schematic representation of the two target (+)-
azumamide E (5) stereoisomers: (2R,3S)-azumamide E (7) and (-)-
azumamide E (8).
Figure 2. Schematic representation of (+)-azumamide E (5) and trapoxin
A (6).
between azumamide E (5)22 and the HDLP binding pocket.20a,23
Subsequently, with the aim to probe the effects of the side
chains’ relative orientation and the role of the peptide bond
direction on the binding efficiency, we synthesized two carefully
chosen stereochemical variants of natural (+)-azumamide E
(5): (2R,3S)-azumamide E (7) and (-)-azumamide E (8, Figure
3), carried out the docking study, and verified their HDAC
inhibitory activities.
The first selected target, compound 7, presents a configura-
tionally inverted amnda residue24 so that the C3-linked esenoic
side chain results topologically off-register. The second target,
8, being the enantiomer of azumamide E, displays complete
amide bonds and contemporary opposite absolute configuration
at the side chain-bearing stereogenic centers (Figure 2).
Though a zinc-chelating residue is necessary for the HDAC
inhibitory activity, there is strong evidence that the tetrapeptide
three-dimensional appendage orientation significantly influences
the ligand-protein recognition pattern.19 However, our under-
standing of the HDAC-ligand interaction is limited to small
molecule inhibitors,20 being published only one contribution
analyzing the binding contacts between the histone deacetylase-
like protein (HDLP) and the tetracyclopeptide framework.21
The scarce information on the role played by the cyclic
tetrapeptide cap group in the HDAC recognition process
prompted us to explore the molecular details of the interaction
(22) The reported biological activities were performed on synthetic azumamide
E. The analytical data ([R]D, electrospray mass spectrometry, 1H and 13C
NMR spectrometry) measured on the synthetic sample were virtually
indistinguishable from those reported for the natural product.
(23) Since there is no X-ray structure of human HDAC1 available now, we
have used the monomer form of the HDLP X-ray structure as receptor
model. A sequence alignment shows a 35.2% sequence identity of HDLP
and human HDAC1. The residues around the HDLP zinc-binding site are
completely conserved in human HDAC1 and all of the hydrophobic residues
that make up the 11 Å channel in HDLP are identical in HDAC1. Most of
the residues making up the 14 Å internal cavity are either identical or
conservatively substituted in HDAC1 and, as can be expected from the
high degree of sequence similarity, HDAC has the same structural features
of HDLP described above.
(24) Both the â-amino acid stereogenic centers were inverted to retain the same
relative configuration of the 3-amino-2-methyl-5-nonendioic acid.
(25) (a) Gru¨enewald, J. Marahiel, M. A. Microbiol. Mol. Biol. ReV. 2006, 70,
121-146. (b) Hamada, Y.; Shioiri, T. Chem. ReV. 2005, 105, 4441-4482.
(26) In a recently reported SAR study on cyclic hydroxamic acid containing
cyclopeptides (designed as hybrid of trichostatin and trapoxin, see:
Komatsu, Y.; Tomizaki, K.-Y.; Tsukamoto, M.; Kato, T.; Nishino, N.; Sato,
S.; Yamori, T.; Tsuruo, T.; Furumai, R.; Yoshida, M.; Horinouchi, S.;
Hayashi, H. Cancer Res. 2001, 61, 4459-4466), the HDAC inhibitory
potency of different amino acid D/L combinations was directly related to
the hydrophobicity of the molecules (higher hydrophobicity meant higher
membrane permeability). This assumption contrasted with the weak activity
(2 orders of magnitude lower) observed for one of the synthesized
derivatives (showing HPLC retention time comparable with those of the
cognate compounds). We believe that biological activity should be, more
appropriately, attributable to the tetrapeptide-receptor interaction modes.
Our studies clarify some aspects of the binding at molecular level.
(16) Umehara, K.; Nakahara, K.; Kiyoto, S.; Iwami, M.; Okamoto, M.; Tanaka,
H.; Kohsaka, M.; Aoki, H.; Imanaka, H. J. Antibiot. 1983, 36, 478-483.
(17) Hirota, A.; Suzuki, A.; Suzuki, H.; Tamura, S. Agric. Biol. Chem. 1973,
37, 643-647.
(18) (a) Shemyakin, M. M.; Ovchinnikov, Yu, A.; Ivanov, V. T. Angew. Chem.,
Int. Ed. 1969, 8, 492-499. (b) Prelog, V.; Gerlach, H. HelV. Chim. Acta
1964, 47, 2288-2294. (c) Goodman, M.; Chorev, M. Acc. Chem. Res. 1979,
12, 1-7.
(19) (a) Colletti, S. L.; Myers, R. W.; Darkin-Rattray, S. J.; Gurnett, A. M.;
Dulski, P. M.; Galuska, S.; Allocco, J. J.; Ayer, M. B.; Li, C.; Lim, J.;
Crumley, T. M.; Cannova, C.; Schmatz, D. M.; Wyvratt, M. J.; Fisher,
M. H.; Meinke, P. T. Bioorg. Med. Chem. Lett. 2001, 11, 113-117. (b)
Furumai, R.; Komatsu, Y.; Nishino, N.; Khochbin, S.; Yoshida, M.;
Horinouchi, S. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 87-92. (c) Kim,
S.-D. J. Biochem. Mol. Biol. 1995, 28, 227-231. (d) Colletti, S. L.; Myers,
R. W.; Darkin-Rattray, S. J.; Gurnett, A. M.; Dulski, P. M.; Galuska, S.;
Allocco, J. J.; Ayer, M. B.; Li, C.; Lim, J.; Crumley, T. M.; Cannova, C.;
Schmatz, D. M.; Wyvratt, M. J.; Fisher, M. H.; Meinke, P. T. Bioorg. Med.
Chem. Lett. 2001, 11, 107-111.
(20) (a) Finnin, M. S.; Donigian, J. R.; Cohen, A.; Richon, V. M.; Rifkind,
R. A.; Marks, P. A.; Breslow, R.; Pavletich, N. P. Nature 1999, 401, 188-
193. (b) Vannini, A.; Volpari, C.; Filocamo, G.; Casavola, E. C.; Brunetti,
M.; Renzoni, D.; Chakravarty, P.; Paolini, C.; De Francesco, R.; Gallinari,
P.; Steinku¨hler, C.; Di Marco, S. Proc. Natl. Acad. Sci. U.S.A. 2004, 101,
15064-15069.
(21) Rodriquez, M.; Terracciano, S.; Cini, E.; Settembrini, G.; Bruno, I.; Bifulco,
G.; Taddei, M.; Gomez-Paloma, L. Angew. Chem., Int. Ed. 2006, 45, 423-
427.
9
3008 J. AM. CHEM. SOC. VOL. 129, NO. 10, 2007

Azumamide E Histone Deacetylases Inhibitory Activity
A R T I C L E S
Scheme 1. Synthesis of Compound 7a
Scheme 2. Synthesis of Compound 8a
a Reagents and conditions: (a) EDC, HOBt, DIPEA, CH2Cl2, 72%; (b)
TFA/CH2Cl2; 1:1, quant.; (c) FDPP, DIPEA, DMF, 3d, 54%; (d) LiOH,
H2O/THF, 0 °C, 75%.
by azumamide E (5) against the HeLa cell nuclear extract and
against class I (HDACs 1-3 and 8) and class II (HDACs 4-7
and 9) histone deacetylase subtypes28 and those of 7 and 8
against the HeLa cell nuclear extract.
Results and Discussion
Synthesis of (2R,3S)-Azumamide E (7) and (-)-Azuma-
mide E (8). The elaboration of (2R,3S)-azumamide E (7,
Scheme 1) was accomplished by a route previously reported
for the successful (+)-azumamide E (5) synthesis.9b The
“unnatural” (Z,2R,3S)-3-[(tert-butyloxycarbonyl)amino]-2-meth-
yl-5-nonenedioic acid, 9-ethyl ester (16, Scheme 1) residue,
whose synthesis started from commercially available 3-benzyl-
oxy propanol and relied on the Brown’s crotylboration reac-
tion,29 was joined to the linear all-D tripeptide 19, synthesized
following the classic solution-phase protocol.9b
a Reagents and conditions: (a) i. oxalyl chloride, DMSO, Et3N, CH2Cl2,
-78 °C; ii. (-)-Ipc2BOMe, (E)-2-butene, t-BuOK, n-BuLi, BF3‚Et2O, THF,
-78 °C, then aq NaOH, H2O2; iii. Ac2O, Py, CH2Cl2, 68%, over three
steps; (b) i. O3, CH2Cl2, then PPh3; ii. NaBH4, EtOH; iii. K2CO3, MeOH,
64% over three steps; (c) TPSCl, DMAP, Py, CH2Cl2, 93%; (d) i. MsCl,
Et3N, THF; ii. NaN3, DMF, 60 °C, iii. H2, Pt2O, AcOEt; iiii. Boc2O, Et3N,
CH2Cl2, 71% over four steps; (e) i. H2, Pd/C, EtOH; ii. oxalyl chloride,
DMSO,Et3N,CH2Cl2,-78°C,80%overtwosteps.(f)i.BrPh3PCH2CH2CH2CO2Et,
KHMDS, THF, -78 °C to r.t., 94%; (g) i. HF/Py, Py; ii. TEMPO, phospate
buffer, NaClO2, NaClO, CH3CN, 59%, over two steps; (h) EDC, HOBt,
DIPEA, CH2Cl2, 87%; (i) TFA/CH2Cl2 1:1, quant.; (j) FDPP, DIPEA, DMF,
3d, 29%; (k) LiOH, H2O/THF, 0 °C, 79%.
Contemporary terminal carboxy and amino group deprotec-
tion, followed by an FDPP-mediated macroclactamization
reaction,30 furnished the (Z,2R,3S)-azumamide E ethyl ester (20),
which liberated target compound 7 upon hydrolysis in the
presence of LiOH (2.7% overall yield, starting from 9).
The synthesis of 8 proceeded uneventfully and followed the
same strategy reported for its enantiomer 5 (Scheme 2).9b This
time, the bis-protected â-amino acid 16 was linked to the linear
side chain reversal. However, for the intrinsic stereochemical
arrangement of the azumamides, if compound 8 is turned in
the plane 180°, it shows a side chain topology comparable to
that of natural azumamide E (Val and Phe result in positionally
exchanged and configurationally inverted) and a reversed peptide
bond orientation (similar to that of trapoxin A (6) and the related
natural cyclotetrapeptides, see Figure 2).
Cyclopeptides isolated from natural sources represent the
latest products of a constant phylogenetic molecular morphology
optimization.25 This means that the effects of stereochemical/
framework alterations could, in principle, conspicuously affect
biological activity. With the synthesis and docking studies on
stereoisomeric azumamide derivatives 7 and 8, we probe the
morphological role of the residue positioning and the impact
of the amide bond peptide core orientation on biological
activity.26
As a consequence of this study, we evidence a rare case in
which the enantiomer of a putative peptide ligand, though acting
on a chiroselective receptor, retains biological activity.27
This contribution reports on the synthesis of (2R,3S)-
azumamide E (7) and (-)-azumamide E (8). It presents a 3D
model for the interaction of 5, 7, and 8, with the HDLP active
site, and fully discloses the HDAC inhibitory activities exerted
(27) The term “chiroselective” is used in opposition to “nonspecific”. Typical
nonspecific biological actions are exerted on membranes. For the action
of enantiomers of bioactive compounds on nonspecific target, see, for
example: Juvvadi, P.; Vunnam, S.; Merrifield, R. B. J. Am. Chem. Soc.
1996, 118, 8989-8997. Vunnam, S.; Juvvadi, P.; Rotondi, K. S.; Merrifield,
R. B. J. Pept. Res. 1998, 51, 38-44. Shemyakin, M. M.; Ovchinnikov,
Y. A.; Ivanov, V. T.; Evstratov, A. V. Nature 1967, 213, 412-413.
Shemyakin, M. M.; Ovchinnikov, Y. A.; Ivanov, V. T. Angew. Chem., Int.
Ed. 1969, 8, 492-499. For the action of enantiomers of bioactive
compounds on chiroselective target, see, for example: Bednarek, M. A.;
Silva, M. V.; Arison, B.; MacNeil, T.; Kalyani, R. N.; Huang, R.-R. C.;
Weinberg, D. H. Peptides 1999, 20, 401-409 (in which is described a
significant drop of potency of the enantio-analog of the cyclic peptide MT-
II). Guichard, G.; Benkirane, N.; Zeder-Lutz, G.; Van Regenmortel,
M. H. V.; Briand, J.-P.; Muller, S. Proc. Natl. Acad. Sci. U.S.A. 1994, 91,
9765-9769 (in which is described the immunoglobulin IgG3 response on
an enantiomeric form of an immunogenic peptide. Interestingly, no cross-
response was observed for IgG1, IgG2a, and IgG2b antibodies).
(28) Johnstone, R. W. Nat. ReV. Drug DiscoVery 2002, 1, 287-299.
(29) Brown, H. C.; Bhat, K. S. J. Am. Chem. Soc. 1986, 108, 293-294.
(30) (a) Chen, S.; Xu, J. Tetrahedron Lett. 1991, 32, 6711-6714. (b) Dudash,
J., Jr.; Jiang, J.; Mayer, S. C.; Joullie, M. M. Synth. Commun. 1993, 23,
349-356.
9
J. AM. CHEM. SOC. VOL. 129, NO. 10, 2007 3009

A R T I C L E S
Maulucci et al.
Figure 4. NMR solution-state conformations of 5 and 7 obtained by
restrained MD calculations, using ROESY (tmix)100 ms, 300 K) data
collected at 600 MHz in [D6]DMSO (see Supporting Information).
Figure 5. 3D model of the interaction between azumamide E (5) and the
HDLP binding site. The protein is represented by molecular surface and
sticks and balls. 5 is depicted by sticks (yellow) and balls (by atom type:
C, gray; polar H, sky blue; N, dark blue; O, red). The figure highlights
essential interactions: the phenyl group is located in a hydrophobic pocket
and the tetrapeptide core interacts with a shallow cavity on the receptor
surface. The amnda side chain establishes interactions with the 11 Å
hydrophobic channel and with the zinc ion contained at the bottom.
all-L tripeptide 21. After that, (-)-azumamide E (8) was formed
in two steps and 4.0% overall yield (starting from 9).
NMR and Docking Studies. Docking studies were performed
on (+)-azumamide E (5), (2R,3S)-azumamide E (7), and (-)-
azumamide E (8), with the HDLP20a,31 binding pocket, using
AutoDock 3.0.5 software,32 which has been successfully used
in the interpretation of the inhibitory activity of several HDAC
ligands.33 The DMSO solution-state structures were obtained
by 2D-NMR spectroscopy (Figure 4 a, b) and were used as the
initial conformations in the docking study. The NMR structure
backbone analysis of 7 (Figure 4b), in analogy with that reported
for 5 (Figure 4a),9b did not reveal the presence of any identifiable
turns or defined secondary structure. On the other side, it showed
a marked difference in the peptide bond spatial arrangement
and an unexpected hydrogen bond between the C9 carboxy
group and the NH of the D-Ala residue.
The partial charges of the three ligands (5, 7, and 8), of the
zinc ion, and of the amino acids involved in the catalytic center
(A169, H170, D168, D258) have been calculated at DFT B3LYP
level and 6-31G(d) basis set using the ChelpG34 method for
population analysis and have been used in the subsequent
docking calculations.
In the theoretical studies, we have considered the amnda
residue carboxy group functionality as dissociated at physi-
ological pH and, consequently, deprotonated in the calculations.
The carboxylate moiety is of primary importance for its
interactions network: it coordinates the zinc ion and establishes
hydrogen bonds with Hꢀ2 of H131 and H132, and OH of Y297
(not shown).35 Moreover, the C3 linked esenoic chain makes
stabilizing hydrophobic contacts with the zinc-containing tubular
pocket. The macrolactam portion is accommodated in a shallow
groove, establishing van der Waals interactions and hydrogen
bonds with the receptor counterpart, formed by H170, Y196,
A197, F198, and F200 residues (Figure 5).
delimited by the Y264, L265, S266, and K267 residues,
enveloping the phenyl ring. The D-Val side chain points outside
the receptor.
The (2R,3S)-azumamide E (7) is characterized by a configu-
rationally inverted amnda residue and topologically repositioned
D-Phe and D-Val side chains. In contrast with a single and
defined family of conformations representing the efficient
binding mode for natural azumamide E (5), this time we
obtained three different conformation families, accounting for
three independent, low affinity, binding modes (Figure 6 a-c).
In the first conformation (Figure 6a), which has the best
binding energy among the three, the macrolactam ring interacts
with a shallow pocket delimited by E92, H170, F141, F198,
G140, L265, and H132.35 The D-Val side chain enjoys van der
Waals interactions with H170, F198, and L265. The D-Phe side
chain establishes stabilizing hydrophobic interactions with F198
and F200 residues. Although the above-reported hydrophobic
interactions are maintained, the spatial positioning of the amnda
chain prevents, in this arrangement, an efficient coordination
of the carboxylate group to the zinc ion, as witnessed by the
distance of about 7 Å between the two groups, suggesting lack
in the binding activity.
In the second conformation, 7 coordinates the zinc ion, but
the interactions expected for the phenyl and the isopropyl groups
are lost (Figure 6b). In the third model (Figure 6c), the phenyl
group is placed in the same hydrophobic cavity found for 5,
but a minor number of interactions are observed; moreover, only
one of the two oxygens of the carboxylate function coordinates
the zinc ion.
The tetrapeptide core extends its hydrophobic contacts thanks
to the D-Phe side chain, which is accommodated in a deep pocket
(31) Wang, D-F.; Wiest, O.; Helquist, P.; Lan-Hargest, H-Y.; Wiech, N. L.
J. Med. Chem. 2004, 47, 3409-3417.
(32) Morris, G. M.; Goodsell, D. S.; Halliday, R. S.; Huey, R.; Hart, W. E.;
Belew, R. K.; Olson, A. J. J. Comp. Chem. 1998, 19, 1639-1662.
(33) (a) Wang, D-F.; Helquist, P.; Wiech, N. L.; Wiest, O. J. Med. Chem. 2005,
48, 6936-6947. (b) Park, H.; Lee, S. J. Comput.-Aided Mol. Des. 2004,
18, 375-388.
(34) Breneman, C. M.; Wiberg, K. B. J. Comp. Chem. 1990, 11, 361-373.
(35) H131, H132, and Y297 are not shown in Figures 5-7 because they are
too deep in the channel. They are reported (in the 3D models of the
interaction between 5 and 8 with the HDLP binding site) in Figure S3 and
S4 of the Supporting Information.
In (-)-azumamide E (8), the complete configuration inversion
maintains the three-dimensional side chains’ relative positioning.
Our docking studies indicate that the amnda chain and the
tetrapeptide core of the two docked enantiomers (8 and 5) fill
equivalent spaces, and the valine and phenylalanine residues
inverted positions give minor perturbations (Figure 7). More-
over, the amnda side chain exerts the same set of interactions
observed in 5, with the tubular hydrophobic pocket and the zinc-
9
3010 J. AM. CHEM. SOC. VOL. 129, NO. 10, 2007

Azumamide E Histone Deacetylases Inhibitory Activity
A R T I C L E S
Figure 6. 3D model of the interaction between 7 and the HDLP binding site. The protein is represented by molecular surface and sticks and balls. 7 is
shown in three different docking poses (by sticks and balls): (a) in red is the best binding energy conformation but it does not coordinate the Zn; (b) in
orange is the best binding energy conformation that coordinates zinc ion; (c) in yellow is the best docking energy conformation.
Figure 8. 5 and 8 superimposition in the zinc-binding site. The green mesh
represents the hydrophobic pocket of the protein. The Zn2+ is represented
by a CPK sphere in dark green. 5 is depicted by sticks (yellow) and balls
(by atom type: C, gray; polar H, sky blue; N, blue; O, red); 8 is shown by
sticks (dark blue) and balls (by atom type: C, gray; polar H, sky blue; N,
Figure 7. 3D model of the interaction between 8 and the HDLP binding
site. The protein is represented by molecular surface and sticks and balls.
8 is depicted by sticks (blue) and balls (by atom type: C, gray; polar H,
blue; O, red). The figure highlights the similar interactions with the 11 Å
sky blue; N, dark blue; O, red). The figure highlights essential interac-
deep hydrophobic channel of the amnda side chains, whereas the phenyl
tions: the phenyl establishes hydrophobic interactions and the tetrapeptide
groups of 5 and 8 are located on opposite sides of the receptor surface
core interacts with a shallow cavity on the receptor surface The amnda
side chain establishes interactions with the 11 Å hydrophobic channel and
hydrophobic pocket.
with the zinc ion contained at the bottom.
Table 1. Biological Activities of 5, 7, and 8 on HeLa NE
compound
5
7
8
IC50 (µM)a
0.134
((0.024)
N.A. at
50 µM
26.0
((4.4)
coordinating carboxylate group, forming hydrogen bonds with
Hꢀ2 of H131 and H132 and OH of Y297.35
The macrolactam accommodates in the shallow pocket
delimited by H170, Y196, A197, L265, F198, and F200 amino
acids. In this case, the L-Val isopropyl group is placed in a
similar position with respect to the L-Phe side chain observed
in natural 5 (see Figure 5). Even though such a group does not
perfectly interact with the receptor, as observed for the D-Phe
side chain of the natural enantiomer, it establishes hydrophobic
contacts with Y196, L265, S266, and K267. Further hydrophobic
interactions are formed by L-Phe side chain that finds a
counterpart delimited by residues P22, Y91, E92, and F141.
a Values are means of at least two independent experiments.
to the results of biological essays (IC50 values of 0.134 µM and
26 µM, respectively; see Table 1).
Biological Activities. (+)-Azumamide E (5), (2R,3S)-azuma-
mide E (7), and (-)-azumamide E (8) showed distinct effects
on the HDAC enzymatic activity associated with nuclear extracts
(NE) prepared from HeLa cells (Table 1).
As predicted by the docking studies, the (2R,3S)-azumamide
E (7) showed no measurable inhibitory activity in this assay up
to a concentration of 50 µM. In contrast, both (+)-azumamide
E (5) and (-)-azumamide E (8) displayed inhibitory activities,
albeit the enantiomer showed a significant loss in potency, thus
confirming that the side chain topology is an important
determinant for the biological activity of these cyclotetrapep-
tides. Extensive profiling of (+)-azumamide E (5) revealed that
this compound is an isoform-selective HDAC inhibitor, display-
ing potent inhibition of HDAC 1-3, weaker activity on HDAC8,
and more severe decrease in potency on HDAC 4-7, 9
(Table 2).
The different binding modes of 8 and 5, with respect to the
HDLP active site, are highlighted in Figure 8. From this
superimposition, it can be well appreciated the inverse arrange-
ment of the peptide bonds and the resemblant positioning of
the enantiomers’ cap groups.
The different arrangement of the (-)-azumamide E (8) and
the suboptimal hydrophobic interactions are responsible for a
predicted decrease in the binding affinity to the receptor of about
40-fold (Ki of azumamide E 3.5 × 10-9 vs Ki of the enantiomer
1.4 × 10-8). Such a decrease is in a good qualitative accordance
9
J. AM. CHEM. SOC. VOL. 129, NO. 10, 2007 3011

A R T I C L E S
Maulucci et al.
9
Table 2. Biological Activities of 5 on HDAC Isoforms
HDAC subtype
1
2
3
4
5
6
7
8
IC50
0.050
((0.001)
0.100
((0.033)
0.080
((0.007)
30% at
50 µM
10.0
((0.7)
12.0
((2.7)
9.7
((1.2)
3.7
((0.7)
28.0
((4.9)
(µM)a
a Values are means of at least two independent experiments.
In summary, 5 is a potent, selective inhibitor of class I
HDACs 1-3. The inhibitory activity on these recombinant
enzymes is similar to that measured on HeLa NE. This is
consistent with the observation that endogenous HDAC 1, 2,
and 3 account for most of the enzymatic activity associated with
NE (unpublished results). While it is documented that HDAC
6 (class IIb) is relatively resistant to the inhibition by cyclo-
peptide compounds,19b the significant loss of potency shown
by class IIa HDACs (4, 5, 7, and 9) is, to our knowledge,
unprecedented.36
tetrapeptides scaffold, with their ample set of surface contacts,
can, in principle, target enzyme specificity through proper
modulation of the amino acid configurational and structural
assortment. Our integrated studies demonstrate the relevance
of the side chain topology and the limited impact of the peptide
bond orientation in terms of binding affinity and HDAC
inhibitory activity. In particular, in a rational design of new
peptidomimetic HDAC inhibitors, novel targets should be
selected among those presenting proper appendages morphology.
In conclusion, we have designed and stereoselectively
synthesized appropriate topochemical variants of natural (+)-
azumamide E (5). Accurate docking studies and concurrent
biological testing allowed us to derive a plausible model for
the interaction of azumamides with the HDLP binding pocket
and to highlight the critical features necessary for the optimal
contact modes.
Conclusion
Cyclic peptides constitute the most structurally complex and
diverse class of HDAC inhibitors. Our combined efforts clarify
the action of the most powerful natural carboxylic acid contain-
ing HDAC-inhibitor known to date37 and unravel, at molecular
level, the high HDLP binding affinity conferred by the unique
azumamides cyclotetrapeptide scaffold cap group.
Acknowledgment. Financial support from the MIUR (“Sintesi
di Sostanze Naturali e di loro analoghi sintetici con attivita`
antitumorale”) and the Universita` degli Studi di Salerno. Both
the institutions are gratefully acknowledged. We also thank Dr.
Patrizia Iannece for mass spectra and Miss Chiara De Cola for
experimental work.
Accumulating data suggests that each subtype of the HDAC
family plays a different role in the gene expression,38 and cyclic
(36) For non-cyclopeptide compounds showing selective class I HDAC inhibitory
activity, see: Jones, P.; Altamura, S.; Chakravarty, P. K.; Cecchetti, O.;
De Francesco, R.; Gallinari, P.; Ingenito, R.; Meinke, P. T.; Petrocchi, A.;
Rowley, M.; Scarpelli, R.; Serafini, S.; Steinku¨hler, C. Bioorg. Med. Chem.
Lett. 2006, 16, 5948-5952.
(37) The most powerful, carboxylic acid containing inhibitor known is semi-
synthetic and is reported in ref 19d (IC50: 15 nM, HDAC derived from
partially purified extracts of human HeLa cells). Usually, synthetic
carboxylic acid containing inhibitors show activity in µM range. The only
exception (IC50: 200 ( 100 nM) can be found in Woo, S. H.; Frechette,
S.; Khalil, E. A.; Bouchain, G.; Vaisburg, A.; Bernstein, N.; Moradei, O.;
Leit, S.; Allan, M.; Fournel, M.; Trachy-Bourget, M.-C.; Li, Z.; Besterman,
J. M.; Delorme, D. J. Med. Chem. 2002, 45, 2877-2885.
(38) (a) Brehm, A.; Miska, E. A.; McCance, D. J.; Reid, J. L.; Bannister, A. J.;
Kouzarides, T. Nature 1998, 391, 597-601. (b) Magnaghi, J. L.; Groisman,
R.; Naguibneva, I.; Robin, P.; Lorain, S.; Le, Villain. J. P.; Troalen, F.;
Trouche, D.; Harel-Bellan. A. Nature 1998, 391, 601-605. (c) Lemercier,
C.; Verdel, A.; Galloo, B.; Curtet, S.; Brocard, M. P.; Khochbin, S. J. Biol.
Chem. 2000, 275, 15594-15599.
Supporting Information Available: Experimental procedures
for all new compounds (7, 8, 10-24); 1H and 13C NMR, HSQC
for 7 and 8; 2D-ROESY for 5 and 7; table of the ROESY cross-
peak volumes; list of significant distances in the 3D docking
model between 5, 7, and 8 and the HDLP binding site; values
of inhibition constants and energies associated with the calcu-
lated complex; ChelpG calculated charges for 5, 7, and 8. This
material is available free of charge via the Internet at
http://pubs.acs.org.
JA0686256
9
3012 J. AM. CHEM. SOC. VOL. 129, NO. 10, 2007