Synthesis, evaluation and molecular modeling of cyclic tetrapeptide histone deacetylase inhibitors as anticancer agents

Article abstract of DOI:10.1002/psc.2392

Histone deacetylase inhibitors (HDACIs) are a promising class of anticancer agents. To examine whether a slight change in the recognition domain could alter their inhibitory activity, we synthesized a series of cyclo(-l-Am7(S2Py)-Aib-l-Phe(n-Me)-d-Pro)derivatives and evaluated their HDAC inhibitory and anticancer activities. The peptides exhibited potent HDAC inhibitory activity and inhibited three human cancer cell lines with IC₅₀ in the micromolar range. Docking and molecular dynamics simulation were conducted to explore the interaction mechanisms of class I and II HDACs with these inhibitors. It revealed that the zinc ion in the active site coordinated five atoms of HDACs and the sulfur atom of the inhibitor. The metal binding domains of these compounds interacted with HDAC2, and the surface recognition domains of these compounds interacted with HDAC4 through hydrogen bonding. The hydrophobic interactions also provided favorable contributions to stabilize the complexes. The results obtained from this study would be helpful for us to design some novel cyclic tetrapeptides that may act as potent HDACIs.

Full text of DOI:10.1002/psc.2392

Research Article
Received: 14 July 2011
Revised: 24 October 2011
Accepted: 1 December 2011
Published online in Wiley Online Library: 17 January 2012
(wileyonlinelibrary.com) DOI 10.1002/psc.2392
Synthesis, evaluation and molecular modeling
of cyclic tetrapeptide histone deacetylase
inhibitors as anticancer agents
Dawei Huang,a Xiaohui Li,a* Lei Sun,a Zhilong Xiua and Norikazu Nishinob
Histone deacetylase inhibitors (HDACIs) are a promising class of anticancer agents. To examine whether a slight change in the
recognition domain could alter their inhibitory activity, we synthesized a series of cyclo(ÀL-Am7(S2Py)-Aib-L-Phe(n-Me)-D-Pro)
derivatives and evaluated their HDAC inhibitory and anticancer activities. The peptides exhibited potent HDAC inhibitory activity
and inhibited three human cancer cell lines with IC50 in the micromolar range. Docking and molecular dynamics simulation were
conducted to explore the interaction mechanisms of class I and II HDACs with these inhibitors. It revealed that the zinc ion in the
active site coordinated five atoms of HDACs and the sulfur atom of the inhibitor. The metal binding domains of these compounds
interacted with HDAC2, and the surface recognition domains of these compounds interacted with HDAC4 through hydrogen
bonding. The hydrophobic interactions also provided favorable contributions to stabilize the complexes. The results obtained
from this study would be helpful for us to design some novel cyclic tetrapeptides that may act as potent HDACIs. Copyright ©
2012 European Peptide Society and John Wiley & Sons, Ltd.
Keywords: histone deacetylase; histone deacetylase inhibitor; anticancer agent; cyclic tetrapeptide; docking; molecular dynamics
such as apicidin [13], FK228 [14–16] and chlamydocin [17]
(Figure 1). The zinc-binding site of FK228 is the sulfhydryl group,
Introduction
Acetylation and deacetylation of e-amino groups of lysine
residues in the N-terminal domain of histone catalyzed by
histone acetyltransferase and histone deacetylase (HDAC) play a
fundamental role in gene expression because acetylation of
histone tails cirelates with an open chromatin configuration and
transcriptional activation, and deacetylation induces transcrip-
tional repression through chromatin condensation [1]. In general,
HDAC activity correlates to transcriptional repression, and
abnormal increase in HDAC activity has been associated with the
development of some human cancers [2]. HDACs are classified into
four different phylogenetic classes according to their cellular locali-
zations, structural and functional differences: class I (HDAC1–3 and
HDAC8) is closely related to yeast Rpd3; class II (HDAC4–7, HDAC9
and HDAC10) have domains that are similar to yeast Hda1; class
IV (HDAC11) displays properties of both class I and class II HDACs
[3,4]. All the aforementioned HDACs are zinc-dependent proteases
and are referred to as ‘classical’ HDACs. There are seven members of
class III HDACs in mammals, and these have been identified on the
basis of sequence homology with Sir2, a yeast transcription repres-
sor, and require NAD+ for their deacetylase activities [5]. The crystal
structures of human HDACs (HDAC2 [6], HDAC4 [7] and HDAC8 [8])
show that all have a deep narrow pocket, and a zinc ion is
positioned near the bottom of the pocket.
and FK228 is converted to its active form (RedFK228) by cellular
reducing activity. FK228 (Romidepsin international nonpropri-
etary name, trade name Istodax (Gloucester Pharmaceuticals,
Cambridge, Mass., USA)) was approved by the U.S. Food and Drug
Administration for the treatment of cutaneous T-cell lymphoma,
after 5 years in the agency’s fast track development program.
Chlamydocin was originally isolated from the fungus Diheterospora
chlamydosporia. It contains Aib (a-aminoisobutyric acid), L-Phe, D-Pro
and L-Aoe (L-amino-8-oxo-9,10-epoxydecanoic acid), which react and
inhibit HDACs [18]. Hydrophobicity at the surface recognition domain
of HDACIs is crucial for their activities, and the cyclic tetrapeptide
framework has a significant structural role in the specific hydro-
phobic interaction with the surface of HDACs. The role and the
appropriate position of the aromatic ring of L-Phe in the chlamy-
docin macrocycle have been investigated, and the data indicated
that L-Phe was important for interacting with the surface binding
region of HDACs [19].
In order to find some specific inhibitors, we designed and
synthesized several novel cyclic tetrapeptides with methyl L-Phe
in their macrocyclic frameworks. The importance of L-Phe has
stimulated our interest to study whether slight change to the
Some studies have shown that inhibition of HDACs produces
anticancer effects in several tumor cell lines. Normal cells are
more resistant to the cell-death-inducing effect of HDAC inhibi-
tors (HDACIs) than cancer cells [9,10]. HDACIs have been
developed for anticancer chemotherapy in the recent years.
Some natural and synthetic compounds have been reported to
act as HDACIs, such as trichostatin A (TSA) [11], vorinostat (suber-
oylamilide hydroxamic acid) [12] and some cyclic tetrapeptides,
*
Correspondence to: Xiaohui Li, School of Life Science and Biotechnology,
Dalian University of Technology, Dalian 116024, China. E-mail: lxhxh@dlut.
edu.cn
a School of Life Science and Biotechnology, Dalian University of Technology,
Dalian 116024, China
b Graduate School of Life Science and Systems Engineering, Kyushu Institute of
Technology, Kitakyushu 808-0196, Japan
J. Pept. Sci. 2012; 18: 242–251
Copyright © 2012 European Peptide Society and John Wiley & Sons, Ltd.

CYCLIC TETRAPEPTIDES AS ANTICANCER AGENTS
Experiments and Computational
Methodology
Synthesis
Synthesis of Boc-L-a-amino-7-bromoalkanoic (Boc-L-Ab7-OH) and
Boc-L-methylphenylalanines (Boc-L-Phe(n-Me)-OH)
Boc-L-Ab7-OH and Boc-L-Phe(n-Me)-OH were synthesized according
to the procedure reported in [22], and the overall yield was
35%–40%.
Synthesis of cyclo(ÀL-Am7(S2Py)-Aib-L-Phe-D-Pro)
This compound was synthesized according to conventional solu-
tion phase method. H-D-Pro-OBzl (2 mmol), Boc-L-Phe-OH
(2.2 mmol), HOBtÁH2O (2 mmol) and DCC (2 mmol) were added
in DMF (4 ml), and the solution was stirred overnight. After evapo-
ration of DMF, the residue was extracted and purified using silica
gel chromatography (CHCl3 : MeOH = 99 : 1) to yield Boc-L-Phe-D-
Pro-OBzl (1.91 mmol). Boc group was then removed using 4 N
HCl/dioxane (6 ml) and then the N-terminus free dipeptide
(1.91 mmol), Boc-Aib-OH (2.2 mmol), HBTU (2.2 mmol) and
HOBtÁH2O (2 mmol) were added in DMF (4ml) and stirred overnight.
Boc-Aib-L-Phe-D-Pro-OBzl (1.18 mmol) was obtained in the same
manner as described earlier. The N-terminus of the tripeptide was
deprotected using 4 N HCl/dioxane (3.6 ml), and the N-terminus free
tripeptide (1.18 mmol) was coupled with Boc-L-Ab7-OH (1.3 mmol)
using HOBtÁH2O (1.3 mmol) and DCC (1.3 mmol) to yield tetrapep-
tide. Boc-L-Ab7-Aib-L-Phe-D-Pro-OBzl (0.84 mmol) was obtained after
purification using silica gel chromatography (CHCl3 :MeOH= 99 : 1).
After the C-terminal benzyl protection was removed using catalytic
hydrogenation, the N-terminal Boc group was removed using treat-
ment with TFA (2.5 ml). Cyclization reaction was carried out in DMF
(840 ml), the linear tetrapeptide (0.84 mmol), HATU (1.3 mmol) and
DIEA (2.1 mmol) was added, and the solution was stirred for 3 h.
The yield of cyclic tetrapeptide was 61% (0.51 mmol) after purifica-
tion using silica gel chromatography (CHCl3 : MeOH = 99 : 1). The
cyclic tetrapeptide (0.51 mmol) containing L-Ab7 was reacted with
potassium thioacetate (0.75 mmol) for 5 h to convert the bromide
to thioacetate ester, cyclo(ÀL-Am7(SAc)-Aib-L-Phe-D-Pro) (0.47 mmol).
To a solution of cyclo(ÀL-Am7(SAc)-Aib-L-Phe-D-Pro) in DMF (2 ml)
under argon atmosphere, the 2,2′-dipyridyl disulfide (0.9mmol)
and 40% NH2CH3/MeOH (2.5 mmol) were added and stirred for
5 h. The target compound, cyclo(ÀL-Am7(S2Py)-Aib-L-Phe-D-Pro)
was obtained after purification using silica gel chromatography
(CHCl3 : MeOH = 99 : 1). Its overall yield was 7.4%. (High resolution)
HR-FAB-MS [M+ H]+ 598.2478 for C30H39O4N5S2 (calculated
598.2522); 1H NMR (CDCl3, 400MHz, d, ppm) 1.27 (2H, m), 1.34
(3H, s), 1.42 (2H, m), 1.61 (2H, m), 1.71 (2H, m), 1.77 (3H, s), 1.88
(2H, m), 2.17 (1H, m), 2.25 (1H, m), 2.78 (1H, m), 2.96 (1H, m), 3.23
(2H, m), 4.19 (2H, m), 4.66 (2H, d, J = 6.2 Hz), 5.17 (1H, m), 6.06
(1H, s), 7.11 (1H, m), 7.21 (1H, d, J= 4.1 Hz), 7.25 (5H, m), 7.51 (1H, d,
J= 9.9 Hz), 7.66 (1H, m), 7.73 (1H, d, J= 8.0 Hz), 8.47(1H, d,
J=5.6Hz); 13C NMR (CDCl3, 400 MHz, d, ppm) 23.57, 24.75, 25.02,
25.09, 26.47, 28.06, 28.61, 28.83, 35.82, 38.66, 46.99, 53.43, 54.30,
57.78, 58.81, 119.62, 120.55, 126.72, 128.62, 128.62, 129.04, 129.04,
137.01, 137.04, 149.60, 160.54, 171.85, 172.84, 174.33, 175.63.
Figure 1. Structures of some HDACIs.
aromatic ring of L-Phe could alter the inhibitory activity of
inhibitor against HDAC. The metal binding domain of these com-
pounds was the sulfhydryl group that was protected as disulfide
hybrid (Figure 2), but it could be reduced to sulfhydryl group by
intracellular L-glutathione.
Cyclic tetrapeptides are the most structurally complex HDACIs,
but the interaction between these inhibitors and HDACs is only
partly known. Studies analyzing the binding between HDAC-like
protein (HDLP), which shares 35% sequence identity with human
HDAC1, and the cyclic tetrapeptides, FR235222 and azumamide
E, have been published [20,21], but the detail of interaction
between human HDACs and cyclic tetrapeptide type HDACIs is
still not available.
Conformational analysis of HDACs with their inhibitors is
very helpful for elucidating the structure–activity relationship
behavior between enzyme and inhibitor, and for designing new
drug candidates. Molecular dynamics (MD) simulation is a very
useful technique for exploring the structure and the dynamic
behavior of molecules of biochemical interest. To explore the
mechanism by which HDACs interact with these sulfur-containing
cyclic tetrapeptide inhibitors and to compare the different
modes by which these inhibitors bind to class I and II HDACs,
docking and MD simulations were conducted for HDAC2, which
shares 82% sequence identity with HDAC1, and HDAC4 with
these inhibitors.
O2
O3
N
HN
HN
S
R4
R2
S
N
NH
O1
R3
O
R1
1. R1=H, R2=H, R3=H, R4=H cyclo(-L-Am7(S2Py)-Aib-L-Phe-D-Pro-)
Synthesis of cyclo(ÀL-Am7(S2Py)-Aib-L-Phe(2-Me)-D-Pro)
This compound was synthesized according to cyclo(ÀL-Am7(S2Py)-
Aib-L-Phe-D-Pro) using Boc-L-Phe(2-Me)-OH instead of Boc-L-Phe-
OH. Its overall yield was 8.2%. HR-FAB-MS [M + H]+ 612.2722 for
2. R1=CH3, R2=H, R3=H, R4=H cyclo(-L-Am7(S2Py)-Aib-L-Phe(2-Me)-D-Pro-)
3. R1=H, R2=H, R3=CH3, R4=H cyclo(-L-Am7(S2Py)-Aib-L-Phe(4-Me)-D-Pro-)
4. R1=H, R2=CH3, R3=H, R4=CH3 cyclo(-L-Am7(S2Py)-Aib-L-Phe(3,5-2Me)-D-Pro-)
Figure 2. Structures of synthesized cyclic tetrapeptides.
J. Pept. Sci. 2012; 18: 242–251 Copyright © 2012 European Peptide Society and John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/jpepsci

HUANG ET AL.
C31H41O4N5S2 (calculated 612.2679); 1H NMR (CDCl3, 400 MHz,
d, ppm) 1.08 (1H, m), 1.32 (1H, m), 1.34 (3H, s), 1.37 (3H, m), 1.60
(1H, m), 1.66 (3H, s), 1.70 (2H, m), 1.77 (2H, s), 1.93 (2H, m), 2.31
(1H, m), 2.39 (3H, s), 2.78 (1H, t), 2.96 (1H, m), 3.23 (1H, m), 3.44
(1H, m), 3.84 (1H, m), 4.18 (1H, m), 4.68 (1H, m), 5.19 (1H, m), 6.01
(1H, s), 7.11 (1H, m), 7.13 (1H, m), 7.26 (4H, s), 7.52 (1H, d,
J = 10.1Hz), 7.65 (1H, m), 7.71 (1H, m), 8.46 (1H, d, J = 4.3Hz); 13C
NMR (CDCl3, 400 MHz, d, ppm) 19.72, 23.70, 24.86, 25.08, 26.46,
28.17, 28.72, 29.35, 33.02, 34.08, 38.72, 46.97, 52.35, 54.44, 57.87,
58.87, 119.69, 120.20, 126.20, 126.91, 129.43, 130.55, 135.35,
136.60, 137.15, 149.71, 160.62, 171.97, 173.03, 174.53, 175.80.
was then purified using preparative HPLC, and the yield of
reduced compound 1 was 25.7%. Reduced compounds 2, 3 and
4 were obtained using the same method, and their yields were
27.8%, 26.1% and 25.2%, respectively.
The HDAC inhibitory activities of all compounds were assayed
with the HDAC Fluorometric/Drug Discovery Assay Kits (AK511
and AK500; BioMol (Farmingdale, NY 11735, USA)) according to
the manufacturer’s instructions. Each compound was assayed in
triplicate, and the assay was repeated three times. The IC50 values
were calculated using SPSS 17.0 (SPSS Inc., Chicago, IL, USA).
Inhibition Against Cancer Cell Growth
Synthesis of cyclo(ÀL-Am7(S2Py)-Aib-L-Phe(4-Me)-D-Pro)
Growth inhibition was determined using the MTT 3-(4,5)-
dimethylthiahiazo (-z-y1)-3,5-di-phenytetrazoliumromide assay.
Human breast cancer cells MCF-7 (human breast cancer), HeLa
(human cervix cancer) and 7721 (human liver cancer) cells were
all cultured in RPMI 1640 medium supplemented with 10%
fetal bovine serum. Cell cultures were incubated in a humidified
atmosphere of 5% CO2 in air. Cells were diluted to 5–9 Â 104
cells/ml with the corresponding medium and were plated in 96-well
microplate. A serial dilution of the compound to be evaluated was
added, and the plate was incubated for 48 h. After incubation,
200 ml MTT (0.5 mg/ml) reagent diluted in serum-free medium
(0.5 mg/ml) was added to each well, and the cells were incubated
for an additional 4 h followed by addition of 200ml DMSO to dis-
solve the dark blue crystal (formazan). The optical density of the
plate was measured at 570 nm with a microplate spectrophotome-
ter. All experiments were performed in triplicate and repeated three
times. The IC50 values were calculated using SPSS 17.0 (SPSS Inc.).
This compound was synthesized according to cyclo(ÀL-Am7
(S2Py)-Aib-L-Phe-D-Pro) using Boc-L-Phe(4-Me)-OH instead of
Boc-L-Phe-OH. Its overall yield was 9.5%. HR-FAB-MS [M + H]+
612.2680 for C31H41O4N5S2 (calculated 612.2679); 1H NMR (CDCl3,
400 MHz, d, ppm) 0.86 (1H, m), 1.28 (3H, m),1.34 (3H, s), 1.42 (2H, m),
1.70 (3H, m), 1.78 (3H, s), 1.80 (2H, m), 2.17 (1H, m), 2.26 (3H, s), 2.33
(1H, m), 2.78 (1H, t), 2.88 (1H, m), 3.17 (1H, m), 3.27 (2H, m), 4.17
(1H, m), 4.66 (1H, d, J = 6.1Hz), 5.15 (1H, m), 5.96 (1H, s), 6.84
(3H, s), 7.09 (2H, m), 7.27 (1H, s), 7.49 (1H, d, J = 10.2Hz), 7.66
(1H, t), 7.72 (1H, d, J = 8.2 Hz), 8.46 (1H, d, J = 4.1Hz); 13C NMR (CDCl3,
400 MHz, d, ppm) 21.05, 23.57, 24.76, 25.04, 25.1, 26.53, 28.08, 28.62,
28.79, 35.37, 38.67, 47.00, 53.49, 54.26, 57.78, 58.84, 119.61, 120.54,
128.88, 128.88, 129.30, 129.30, 133.89, 136.21, 136.99, 149.62,
160.56, 171.85,172.90, 174.29, 175.63.
Synthesis of cyclo(ÀL-Am7(S2Py)-Aib-L-Phe(3,5-2Me)-D-Pro)
This compound was synthesized according to cyclo(ÀL-Am7
(S2Py)-Aib-L-Phe-D-Pro) using Boc-L-Phe(3,5-Me)-OH instead of
Boc-L-Phe-OH. Its overall yield was 8.7%. HR-FAB-MS: [M + H]+
626.2814 for C32H43O4N5S2 (calculated 626.2835); 1H NMR (CDCl3,
400 MHz, d, ppm) 0.80 (1H, m), 1.28 (2H, m), 1.34 (3H, s), 1.42 (2H,
m), 1.60 (1H, m), 1.69 (3H, m), 1.77 (3H, s), 2.17 (1H, m), 2.30 (6H, s),
2.33 (1H, m), 2.78 (2H, t), 2.90 (1H, m), 3.22 (2H, m), 3.87 (1H, m),
4.17 (1H, m), 4.65 (1H, s), 5.14 (1H, m), 5.92 (1H, s), 6.84 (3H, s),
7.09 (1H, m), 7.27 (2H, s), 7.49 (1H, d, J = 10.4 Hz), 7.66 (1H, t),
7.72 (1H, d, J = 8.2 Hz), 8.46 (1H, d, J = 4.4 Hz); 13C NMR (CDCl3,
400 MHz, d, ppm) 21.26, 21.26, 23.57, 24.76, 25.05, 25.09, 26.53,
28.07, 28.62, 28.79, 35.64, 38.68, 47.01, 53.40, 54.25, 57.79, 58.84,
119.62,120.53, 126.82, 128.82, 128.33, 136.9, 136.97, 137.99,
137.99, 149.61, 160.57, 171.84, 172.91, 174.29, 175.57.
Docking Studies
Docking studies were conducted with AutoDock4.0 (The Scripps
Research Institute (TSRI), La Jolla, California, USA) program [23,24]
using a Lamarckian genetic algorithm. AutoDock docking proto-
col and scoring function have been successfully applied in the in-
terpretation of the inhibitory activity of several HDACIs [21,25].
Initial structure of HDAC2 and HDAC4 were modeled from the
atom coordinates of the X-ray crystal structure (protein data bank
(PDB) code: 3MAX and 2VQJ, respectively). The active site of
HDACs was covered using a grid box size of 70 Â 70 Â 70 points
with a spacing of 0.375Å between the grid points. For the cyclic
tetrapeptides, all single bonds except the amide bonds and cyclic
bonds were treated as active torsional bonds. For each inhibitor,
200 independent dockings, i.e. 200 runs, were performed using
genetic algorithm searches. A maximum number of 250 000000
energy evaluations and a maximum number of 10000 generations
were implemented during each genetic algorithm run. The default
non-bonded zinc parameters in AutoDock4.0 [23] were employed.
The LigPlot program (Wallace AC, Laskowski RA and Thornton JM,
London, UK) [26] was also employed to analyze the docking results
focusing on hydrogen bonds and hydrophobic contacts.
Conformation Studies using CD and NMR
The CD spectra of all compounds were carried out in methanol
with peptide concentration of 0.1 mM. The solution conformation
1
of compound 2 was studied using H NMR in CDCl3. Complete
assignments were made using COSY and NOESY spectra. The
JNH-HCa values were obtained from NMR charts. The structure
of compound 2 with minimum energy was generated using
molecular operating environment (MOE) program.
HDAC Inhibition Assay
MD Simulation
The disulfide bond of each compound was reduced using dithio-
threitol (DTT) to give a sulfhydryl group that acts as metal binding
domain before the assay. Compound 1 (598 mg) and DTT
(385 mg) were added to a mixture of water (10 ml) and acetoni-
trile (3 ml), and the mixture was left at room temperature for
24 h. The acetonitrile was removed via evaporation. The mixture
All the molecular mechanics and dynamics calculations were car-
ried out with the Amber9 [27] package. The standard amber ff99
force field [28] was used as the parameters for the protein and
water atoms, and the general amber force field [29] and austin
model 1-bond charge correction (AM1-BCC) charges [30] were used
for the ligands. Zinc was modeled using the Stote non-bonded
wileyonlinelibrary.com/journal/jpepsci Copyright © 2012 European Peptide Society and John Wiley & Sons, Ltd. J. Pept. Sci. 2012; 18: 242–251

CYCLIC TETRAPEPTIDES AS ANTICANCER AGENTS
model (q = +2eÀ, r = 1.7Å, e = 0.67kcal/mol) [31]. Each initial struc-
ture for the simulation was prepared from the docked conforma-
tions of HDAC2-compound 3 and HDAC4-compound 3 complexes.
The local hydrogen bonding network around the histidine residues
was checked. His183 of HDAC2 and His198 of HDAC4 were
assigned as HIE (histidine with hydrogen on its epsilon nitrogen),
and other histidine residues as HID (histidine with hydrogen on its
delta nitrogen). The force field parameters of compound 3 were
prepared using Antechamber module [32] of Amber9 package.
Hydrogen atoms were added to the crystallographic protein with
the Amber9 Leap module. Sodium counterions were added to
neutralize the system. The system was then solvated with an
octahedral box of TIP3P water molecules [33]. The minimum
distance from the surface of the protein to the faces of the box
was 10 Å. The particle mesh Ewald method [34] was used to treat
long-range electrostatic interactions. The cutoff distance for the
long-range electrostatic and the van der Waals energy terms were
set at 12.0Å. All covalent bonds to hydrogen atoms were
constrained using the SHAKE algorithm [35].
Figure 3. CD spectra of compounds 1–4.
Energy minimization was achieved in three steps. In the first
step, movement was allowed only for the water molecules and
ions. In the second step, the inhibitor and the enzyme residues
were all allowed to move and the water molecules, together with
ions, were constrained. In the final step, all atoms were permitted
to move freely. In each step, energy minimization was executed
by the steepest descent method for the first 5000 steps and the
conjugated gradient method for the subsequent 2500 steps.
Periodic boundary conditions were used. The time steps were
2 fs during the production dynamics. The temperature was
maintained by rescaling the velocities using the Berendsen
weak-coupling algorithm [36], with a time constant of 2 ps for
the heat bath. After the system was heated to 300 K from the
initial temperature of 0 K using the constant number, volume
and temperature (NVT) ensemble in 120 ps, molecular dynamics
were performed at a constant temperature of 300 K. After a
50 ps position-restrained dynamics, each simulation proceeded
for 2 ns under periodic boundary conditions with constant number,
pressure and temperature (NPT) ensemble at 101325 pascal and at
300 K. The convergence of energies, temperatures, pressures of the
systems and the atomic root mean square deviations (RMSD) of the
enzyme and the inhibitor, were used to verify the stability of the
systems. Trajectories were analyzed using the Ptraj modules [37].
In the present MD simulations, the overall structure of all complexes
appeared to be equilibrated after 1.5 ns. Hence, atom coordinates
were collected at an interval of 5 ps for the last 500 ps to analyze
the structure in detail. The series of snapshots between 1.5 and
2 ns of the equilibrium phase was used for further analysis.
ellipticities at 210 and 245 nm, and a weak positive ellipticity at
230 nm. Compound 4 had a sharp ellipticity because of containing
two methyl groups (L-Phe(3,5-Me)). These similarities on CD spectra
suggested that the peptide backbone of these compounds had
similar structures. This explained the fact that the change of methyl
positions in L-Phe(n-Me) could not affect their conformations. The
energy-minimized structure of compound 2 was shown in Figure 4.
HDAC Inhibition Assay
All compounds were tested for their HDAC inhibitory activity
in vitro, and TSA was used as positive control. The dose–response
curves of all synthesized compounds are shown in Figure 5, and
the IC50 values of all compounds against HDAC1 and HDACs are
presented in Table 1. All compounds showed potent HDAC inhib-
itory activities that were comparable with the activity of TSA. The
similar HDAC inhibition abilities of all compounds revealed that
introduction of methyl groups on the aromatic ring of L-Phe in
this series of HDACIs did not interfere with the contacts between
inhibitors and enzymes.
Inhibition Against Cancer Cell Growth
Compounds 1–4 were tested for in vitro antiproliferative activity
against MCF-7, Hela and 7721 cell lines using MTT assays. As the
dose–response curves of all compounds (Figure 6) and the
Results and Discussion
Chemistry
In order to synthesize potent HDACIs that had the similar cyclic
framework to chlamydocin, we prepared the non-natural amino
acids Boc-L-Ab7-OH and Boc-L-Phe(n-Me)-OH. Target cyclic tetra-
peptides were synthesized according to conventional solution
phase method. All the compounds shown were confirmed using
HR-FAB-MS, 1H NMR and 13C NMR.
Conformation Studies
The CD spectrum of each synthesized compound was similar at
Figure 4. The energy-minimized structure of compound 2 using NMR
190–260 nm (Figure 3). These compounds had two negative
calculation.
J. Pept. Sci. 2012; 18: 242–251 Copyright © 2012 European Peptide Society and John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/jpepsci

HUANG ET AL.
Figure 5. Dose–response curves of all synthesized compounds for inhib-
itory activity against HDAC1 and HDACs.
IC50 values (Table 1) showed, all compounds inhibited the prolif-
eration of these cancer cells at low concentration, thus exhibiting
good anticancer activity. There were no obvious differences in
antiproliferative activity among the different compounds, but all
seemed to exhibit about 20-fold more selectivity than 7721 in the
inhibition of Hela cells.
Docking Studies
Figure 6. Dose–response curves of all synthesized compounds for anti-
proliferative activity against cancer cell lines.
The results of docking for compounds 1–4 into HDAC2 and
HDAC4 are shown in Figure 7. All compounds bound to the active
sites of HDACs with a similar pattern. The aliphatic chain occu-
pied the long and narrow channel, the sulfhydryl group was at
the bottom of this channel, and the large cap domain interacted
with the external surface of the enzyme. In HDAC2-inhibitor com-
plexes, the aromatic rings of L-Phe were all at the groove formed
by Tyr209 and Phe210, and in HDAC4-inhibitor complexes, the
aromatic rings interacted with Phe227. The adjacent methylene
group caused the aromatic ring to move to the appropriate posi-
tion, so no difference in binding to HDACs was detected among
the compounds, though the introduced methyl groups could
enhance the hydrophobicity and steric hindrance of L-Phe.
The distances between sulfur atoms of all compounds and zinc
ion in the active site of HDAC2 were 2.5, 2.2, 2.4 and 2.3 Å,
respectively, and between sulfur atoms of all compounds and
zinc ion in the active site of HDAC4 were 2.7, 2.3, 2.2 and 2.3 Å,
respectively.
Binding energy study showed that no real difference in binding
energy occurred between the complexes formed by inhibitor and
HDAC2 and those formed by inhibitor and HDAC4 (Table 2), and this
was in accordance with the fact that these compounds had similar
affinities for enzymes. It could be concluded that the introduction of
methyl groups on the aromatic ring of L-Phe did not affect the
ability of these compounds to bind to HDACs, so the slight changes
in hydrophobicity and steric hindrance on the large surface recogni-
tion domain of cyclic tetrapeptide HDACIs had little effect on their
interactions with the surface binding region of HDACs.
The interactions between all compounds and HDACs from
docking results were almost the same, so only one of the com-
pounds, compound 3, was chosen for further analysis. Hydrogen
bonds and hydrophobic contacts between HDACs and com-
pound 3 are shown in Figure 8. In HDAC2–compound 3 complex,
the sulfur atom of compound 3 established hydrogen bonds with
the OH atom of Tyr308 and the ND1 atom of His183, and in
HDAC4–compound 3 complex, the sulfur atom of compound 3
formed hydrogen bond with the NE2 atoms of both His158 and
His159. In docked HDLP–cyclic tetrapeptide inhibitor complexes,
the hydrogen bonds were formed between the carboxyl group
Table 1. The IC50 values of all synthesized compounds against
enzymes and cancer cell lines
Compound
Enzymes (nM)
HDAC1 HDACs
Cell lines (mM)
MCF-7
Hela
7721
TSA
1
8.19
4.30
4.90
4.57
4.08
1.77
2.85
3.52
2.56
2.81
—
—
—
3.60
3.86
4.12
3.30
0.54
0.51
0.46
0.48
11.51
11.17
12.08
11.76
2
3
4
wileyonlinelibrary.com/journal/jpepsci Copyright © 2012 European Peptide Society and John Wiley & Sons, Ltd. J. Pept. Sci. 2012; 18: 242–251

CYCLIC TETRAPEPTIDES AS ANTICANCER AGENTS
Figure 7. Views of the active site regions of (a) HDAC2–Compound 3; (b) HDAC2–Compound 1–4; (c) HDAC4–Compound 3; (d) HDAC4–Compound
1–4 from docking results.
and hydroxyl group in the metal binding domain of the inhibitors
and the corresponding His131, His132, Asp168 and Tyr297 of
HDLP [20,21]. It suggested that the hydrogen bonds formed by
the metal binding domain were important for the cyclic tetrapep-
tide HDACIs with powerful inhibitory activity.
Compound 3 also made contact with some other amino acid
residues of HDACs, such as Phe155, Glu208, Tyr209, Phe210 and
Leu276 of HDAC2, His198, Phe226, Phe227, Pro298 and Leu299 of
HDAC4 via hydrophobic interaction. Hydrophobic interaction also
occurred between cyclic tetrapeptide HDACIs and HDLP [20,21],
and this interaction was also crucial for stabilizing the complexes.
converged and were stable as indicated by their energetic and
structural properties. The potential energies and RMSD remained
stable during the last 500 ps of the simulations for the two
complexes. The RMSD between Ca, C and N atoms of the
structures obtained during the trajectories and initial structures
are shown in Figure 9.
The root mean squared fluctuation (RMSF) of Ca, C and N atoms
for these complexes showed that the structures of the two com-
plexes shared different RMSF distribution (Figure 10), but the con-
formational changes of the amino acid residues in the active site
were all very small, especially for those residues at the bottom of
the channel (His145, His146, Asp181, His183 and Asp269 of HDAC2,
His158, His159, Asp196, His198 and Asp290 of HDAC4) (Figure 11).
This indicated that binding modes of compound 3 in the active site
of HDAC2 and HDAC4 were similar, and the binding between
HDACs and the inhibitor directly caused the rigidity of the channel.
MD Simulations
The MD simulations were performed for the HDAC2–compound
3 and HDAC4–compound 3 complexes. Both of the complexes
Table 2. Binding energies of all compounds to HDACs from docking
Energy (kcal/mol)
HDAC2
HDAC4
1
2
3
4
1
2
3
4
Ebinding
À5.66
À7.67
À7.26
À0.41
À0.89
2.20
À5.86
À7.99
À7.44
À0.55
À0.62
2.20
À5.74
À7.79
À7.29
À0.50
À0.75
2.20
À5.76
À7.87
À7.35
À0.52
À0.72
2.20
À5.75
À7.69
À7.33
À0.36
À0.82
2.20
À5.88
À7.65
À7.27
À0.38
À0.79
2.20
À5.94
À7.86
À7.51
À0.35
À0.88
2.20
À5.85
À7.71
À7.39
À0.32
À0.84
2.20
Eintermolecular
EvdW + Hbond + desolv
Eelectrostatic
Etotal internal
Etorsional
Eunbound
À0.61
À0.65
À0.61
À0.63
À0.62
À0.66
À0.60
À0.63
J. Pept. Sci. 2012; 18: 242–251 Copyright © 2012 European Peptide Society and John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/jpepsci

HUANG ET AL.
Figure 8. Hydrogen bonds and hydrophobic contacts between (a) HDAC2 and (b) HDAC4 and compound 3. Blue, amino acids that create hydrogen
bonds; yellow, amino acids that form hydrophobic interaction.
In the crystal structure of HDAC–inhibitor complex, the zinc ion
that lies at the end of the hydrophobic channel is bound to the
carboxylate oxygens of Asp181, Asp269 and the nitrogen atom
of His183 of HDAC2 and corresponding Asp169, Asp290, and
His198 of HDAC4. The other two coordination sites are occupied
by two atoms of the inhibitors [6,7]. The average distances be-
tween the zinc ion and the nearby atoms during the last 500 ps
in MD simulations are shown in Figure 12, where it can be seen
that the zinc atom coordinated with six atoms. In HDAC2–
compound 3 complex, the atoms included all carboxylate
oxygens of Asp181 and Asp269, a nitrogen atom of His183 and
a sulfur atom of the inhibitor; in HDAC4–compound 3 complex,
there were two carboxylate oxygens of Asp196, a carboxylate
oxygen of Asp290, nitrogen atoms of His158 and His198 and a
sulfur atom of the inhibitor. This showed that all the aspartic acid
and histidine residues in the bottom of active site of HDACs could
coordinate zinc ion in the MD simulation. The coordination states
of zinc ion in both HDAC2–compound 3 and HDAC4–compound
wileyonlinelibrary.com/journal/jpepsci Copyright © 2012 European Peptide Society and John Wiley & Sons, Ltd. J. Pept. Sci. 2012; 18: 242–251

CYCLIC TETRAPEPTIDES AS ANTICANCER AGENTS
Figure 9. RMSD of Ca, C and N atoms of the systems in MD simulations with respect to the starting structure.
Figure 10. RMSF for all residues during the last 500 ps in MD simulations.
Figure 11. RMSF for residues in the active site during the last 500 ps in MD simulations.
Figure 12. Average zinc ion coordination distances for (a) HDAC2–compound 3 and (b) HDAC4–compound 3 complexes.
J. Pept. Sci. 2012; 18: 242–251 Copyright © 2012 European Peptide Society and John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/jpepsci

HUANG ET AL.
Table 3. Hydrogen bonds between compound 3 and HDACs in MD simulations
Enzyme
Donor
AcceptorH
Acceptor
Occupied (%)
Distance (Å)
HDAC2
HDAC4
Compound 3:S
Compound 3:O2
Compound 3:O3
Tyr308:HH
His198:HE2
Phe227:H
Tyr308:OH
His198:NE2
Phe227:N
100.00
44.57
42.86
2.89
3.08
3.22
3 complexes were the same, the sulfur atom of compound 3
occupied one coordination site, and this may be the precondition
for the inhibitory activities of these sulfur-containing cyclic
tetrapeptide HDACIs.
References
1 Verdin E, Histone Deacetylases-Transcriptional Regulation and other
Cellular Functions. Humana Press Inc.: Totowa, NJ, 2006.
2 Drummond DC, Noble GK, Kirpotin DB, Guo Z, Scott GK, Benz CC. Clin-
ical development of histone deacetylase inhibitors as anticancer
agents. Annu. Rev. Pharmacol. Toxicol. 2005; 45: 495–528.
3 De Ruijter AJ, van Gennip AH, Caron HN, Kemp S, van Kuilenburg AB.
Histone deacetylases (HDACs): characterization of the classical HDAC
family. Biochem. J. 2003; 370: 737–749.
4 Voelter-Mahlknecht S, Ho AD, Mahlknecht U. Chromosomal organiza-
tion and localization of the novel class IV human histone deacetylase
11 gene. Int. J. Mol. Med. 2005; 16: 589–598.
5 Michan S, Sinclair D. Sirtuins in mammals: insights into their biological
function. Biochem. J. 2007; 404: 1–13.
6 Bressi JC, Jennings AJ, Skene R, Wu YQ, Melkus R, Jong RD, Shawn O,
Grimshaw CE, Navre M, Gangloff AR. Exploration of the HDAC2 foot
pocket: synthesis and SAR of substituted N-(2-aminophenyl)benza-
mides. Bioorg. Med. Chem. Lett. 2010; 20: 3142–3245.
Hydrogen bonding between ligand and protein play an essen-
tial role in stabilizing the complex. The hydrogen bonds formed
between compound 3 and HDACs in the active pocket were in-
vestigated. As shown in Table 3, the sulfur atom of compound
3 was strongly hydrogen bonded to Tyr308 of HDAC2. The
corresponding amino acid residue of Tyr308 is a proline in
HDAC4, and there was no hydrogen bond established between
the sulfur atom of compound 3 and HDAC4, but the O2 and O3
atoms in surface recognition domain of compound 3 formed
two weaker hydrogen bond to NE2 atom of His198 and N atom
of Phe227, respectively. The hydrogen bonds may be crucial for
stabilizing the complexes, and this may be the reason for the
similar inhibitory activities to class I and II HDACs of these cyclic
tetrapeptides. (The IC50 values of compound 1 were 3.9 nM
against HDAC1 and 1.8 nM against HDAC4, respectively [38].)
7 Bottomley M, Lo SP, Di Giovine P, Cirillo A, Scarpelli R, Ferrigno F,
Jones P, Neddermann P, De Francesco R, Steinkuhler C, Gallinari P,
Carfi A. Structural and functional analysis of the human HDAC4
catalytic domain reveals a regulatory zinc-binding domain. J. Biol.
Chem. 2008; 283: 26694–26704.
8 Somoza JR, Skene RJ, Katz BA, Mol C, Ho JD, Jennings AJ, Luong C,
Arvai A, Buggy JJ, Chi E, Tang J, Sang BC, Verner E, Wynands R, Leahy
EM, Dougan DR, Snell G, Navre M, Knuth MW, Swanson RV, McRee DE,
Tari LW. Structural snapshots of human HDAC8 provide insights into
the class I histone deacetylases. Structure 2004; 12: 1325–1334.
9 Johnstone RW. Histone-deacetylase inhibitors: novel drugs for the
treatment of cancer. Nat. Rev. Drug Discov. 2002; 1: 287–299.
10 Marks PA, Dokmanovic M. Histone deacetylase inhibitors: discovery
and development as anticancer agents. Expert Opin. Investig. Drugs
2005; 14: 1497–1511.
11 Yoshida M, Kijima M, Akita M, Beppu T. Potent and specific inhibition
of mammalian histone deacetylase both in vivo and in vitro by tri-
chostatin A. J. Biol. Chem. 1990; 265: 17174–17179.
12 Richon VM, Emiliani S, Verdin E, Webb Y, Breslow R, Rifkind RA, Marks
PA. A class of hybrid polar inducers of transformed cell differentiation
inhibits histone deacetylases. Proc. Natl. Acad. Sci. U.S.A. 1998; 95:
3003–3007.
13 Singh SB, Zink DL, Polishook JD, Dombrowski AW, Darkin-Rattray SJ,
Schmatz DM, Goetz MA. Apicidins: novel cyclic tetrapeptides as
coccidiostats and antimalarial agents from Fusarium pallidoroseum.
Tetrahedron Lett. 1996; 37: 8077–8080.
14 Ueda H, Nakajima H, Hori Y, Fujita T, Nishimura M, Goto T, Okuhara M.
Action of FR901228, a novel antitumor bicyclic depsipeptide pro-
duced by Chromobacterium violaceum no. 968, on Ha-ras transformed
NIH3T3 cells. J. Antibiot. 1994; 47: 301–310.
15 Ueda H, Manda T, Matsumoto S, Mukumoto S, Nishigaki F, Kawamura I,
Shimomura K. FR901228, a novel antitumor bicyclic depsipeptide
produced by Chromobacterium violaceum no. 968. III. Antitumor activities
on experimental tumors in mice. J. Antibiot. 1994; 47: 315–323.
16 Furumai R, Matsuyama A, Kobashi N, Lee KH, Nishiyama M, Nakajima
H, Tanaka A, Komatsu Y, Nishino N, Yoshida M, Horinouchi S. FK228
(depsipeptide) as a natural prodrug that inhibits class I histone deace-
tylases. Cancer Res. 2002; 62: 4916–4921.
Conclusions
We have successfully synthesized several sulfur-containing cyclic
tetrapeptides that can act as HDACIs, and all of them showed
potent HDAC inhibitory and anticancer activities. The slight
change of the surface recognition domain of these inhibitors by
introduction of methyl groups on the aromatic ring of the L-Phe
residue within the peptide appeared to have no obvious change
in their HDAC inhibitory and anticancer activities. The molecular
modeling studies gave us the information about how this series
of HDACIs interacted with class I and II HDACs, and it showed that
all compounds bind to HDACs were almost in the same mode.
The interaction mechanisms of HDAC2 and HDAC4 with these
inhibitors were similar but had some differences. Coordinating
the zinc ion in the active site of HDACs by the sulfur atom of
these inhibitors was the precondition for their inhibitory activ-
ities. Both the surface recognition and metal binding domains
of these inhibitors could form hydrogen bond with HDACs, so
exhibiting similar affinity for class I and II HDACs. The hydro-
phobic interactions between HDACs and these cyclic tetrapep-
tides were also very important for stabilizing the enzyme–inhibitor
complexes. The data generated from this study would allow us to
design some novel cyclic tetrapeptides with different surface
recognition and metal binding domains that may act as potent
HDACIs.
Acknowledgements
17 Closse A, Hugenin R. Isolation and structural clarification of chlamydo-
cin. Helv. Chim. Acta 1974; 57: 533–545.
This work was supported by The National High Technology
Research and Development Program of China (863 program),
2007AA021604. We thank Alan K Chang for his help in the revi-
sion of the manuscript.
18 De Shepper S, Bruwiere H, Steller U, Andries L, Wouters W, Janicot M, Arts
J, Van Heusden J. Inhibition of histone deacetylases by chlamydocin
induces apoptosis and proteasome-mediated degradation of survivin.
J. Pharmacol. Exp. Ther. 2003; 304: 881–888.
wileyonlinelibrary.com/journal/jpepsci Copyright © 2012 European Peptide Society and John Wiley & Sons, Ltd. J. Pept. Sci. 2012; 18: 242–251

CYCLIC TETRAPEPTIDES AS ANTICANCER AGENTS
19 Gururaj MS, Satoshi A, Tamaki K, Norikazu N, Satoko M, Tomonori GN
and Minoru Y. Molecular design of histone deacetylase inhibitors by
aromatic ring shifting in chlamydocin framework. Bioorgan. Med.
Chem. 2007; 15: 7830–7839.
28 Wang JM, Cieplak P, Kollman PA. How well does a restrained electro-
static potential (RESP) model perform in calculating conformational
energies of organic and biological molecules? J. Comput. Chem.
2000; 21: 1049–1074.
20 Rodriquez M, Terracciano S, Cini E, Settembrini G, Bruno I, Bifulco G,
Taddei M, Gomez-Paloma L. Total synthesis, NMR solution structure,
and binding model of the potent histone deacetylase inhibitor
FR235222. Angew. Chem. Int. Ed. 2006; 118: 437–441.
21 Maulucci N, Chini MG, Micco SD, Izzo I, Cafaro E, Russo A, Gallinari P,
Paolini C, Nardi MC, Casapullo A, Riccio R, Bifulco G, Riccardis FD. Mo-
lecular insights into azumamide E histone deacetylases inhibitory ac-
tivity. J. Am. Chem. Soc. 2007; 129: 3007–3012.
22 Li XH, Wang G, Li JX, Xiu ZL, Nishino N. Synthesis and resolution of
Boc-L-methylphenylalanines. Chem. J. Chinese U. 2008; 29: 1363–1366.
23 Huey R, Morris GM, Olson AJ, Goodsell DS. A semiempirical free en-
ergy force field with charge-based desolvation. J. Comput. Chem.
2007; 28: 1145–1152.
24 Morris GM, Goodsell DS, Halliday RS, Huey R, Hart WE, Belew RK, Olson AJ.
Automated docking using a Lamarckian genetic algorithm and an empir-
ical binding free energy function. J. Comput. Chem. 1998; 19: 1639–1662.
25 Wang DF, Helquist P, Wiech NL, Wiest O. Toward selective histone
deacetylase inhibitor design: homology modeling, docking studies,
and molecular dynamics simulations of human class I histone deace-
tylases. J. Med. Chem. 2005; 48: 6936–6947.
29 Wang J, Wolf RM, Caldwell JW, Kollamn PA, Case DA. Development
and testing of a general amber force field. J. Comput. Chem. 2004;
25: 1157–1174.
30 Jakalian A, Bush BL, Jack DB, Bayly CI. Fast, efficient generation of
high-quality atomic charges. AM1-BCC model: I. Method. J. Comput.
Chem. 2000; 21: 132–146.
31 Stote RH, Karplus M. Zinc binding in proteins and solution: a simple
but accurate nonbonded representation. Proteins 1995; 23: 12–31.
32 Wang JM, Wang W, Kollman PA. ANTECHAMBER: an accessory soft-
ware package for molecular mechanical calculations. Abstr. Pap. Am.
Chem. Soc. 2001; 222: 403.
33 Jorgensen WL, Chandrasekhar J, Madura JD, Impey RW, Klein ML.
Comparison of simple potential functions for simulating liquid water.
J. Chem. Phys. 1983; 79: 926–935.
34 Darden T, York D, Pedersen L. Particle mesh Ewald-an N.Log(N)
method for Ewald sums in large systems. J. Chem. Phys. 1993; 98:
10089–10092.
35 Ryckaert JP, Ciccotti G, Berendsen HJC. Numerical integration of the
Cartesian equations of motion of a system with constraints: molecular
dynamics of n-alkanes. J. Comput. Phys. 1977; 23: 327–341.
36 Berendsen HJC, Postma JPM, Vangunsteren WF, Dinola A, Haak JR.
Molecular dynamics with coupling to an external bath. J.Chem. Phys.
1984; 81: 3684–3690.
26 Wallace AC, Laskowski RA, Thornton JM. LIGPLOT: a program to gen-
erate schematic diagrams of protein-ligand interactions. Protein Eng.
1995; 8: 127–134.
27 Case DA, Darden TA, Cheatham TE, Simmerling CL, Wang J, Duke RE,
Luo R, Merz KM, Pearlman DA, Crowley M, Walker RC, Zhang W, Wang
B, Hayik S, Roitberg A, Seabra G, Wong KF, Paesani F, Wu X, Brozell S,
Tsui V, Gohlke H, Yang L, Tan C, Mongan J, Hornak V, Cui G, Beroza P,
Mathews DH, Schafmeister C, Ross WS, Kollman PA. AMBER version 9,
University of California: San Francisco, CA, 2006.
37 Honig B, Nicholls A. Classical electrostatics in biology and chemistry.
Science 1995; 268: 1144–1149.
38 Hirashima Y, Kato T, Nishino N, Maeda S, Nishino TG, Yoshida M.
Design and synthesis of histone deacetylase inhibitors by aromatic
ring shifting in chlamydocin framework. In Proceedings of 29th
European Peptide Symposium (Peptides), 2006; 758–759.
J. Pept. Sci. 2012; 18: 242–251 Copyright © 2012 European Peptide Society and John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/jpepsci