Identification of a potent non-hydroxamate histone deacetylase inhibitor by mechanism-based drug design

Article abstract of DOI:10.1016/j.bmcl.2004.10.074

In order to find novel non-hydroxamate histone deacetylase (HDAC) inhibitors, we synthesized several suberoylanilide hydroxamic acid (SAHA)-based compounds designed on the basis of the catalytic mechanism of HDACs. Among these compounds, mercaptoacetamide 5b was found to be as potent as SAHA. Kinetic enzyme assays and molecular modeling are also reported. In order to find novel non-hydroxamate histone deacetylase (HDAC) inhibitors, we synthesized several suberoylanilide hydroxamic acid (SAHA)-based compounds designed on the basis of the catalytic mechanism of HDACs. Among these compounds, 5b was found to be as potent as SAHA. Kinetic enzyme assays and molecular modeling suggested that the mercaptoacetamide moiety of 5b interacts with the zinc in the active site of HDACs and removes a water molecule from the reactive site of the deacetylation.

Full text of DOI:10.1016/j.bmcl.2004.10.074

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Bioorganic & Medicinal Chemistry Letters 15 (2005) 331–335
Identification of a potent non-hydroxamate histone
deacetylase inhibitor by mechanism-based drug design
Takayoshi Suzuki,* Azusa Matsuura, Akiyasu Kouketsu,
Hidehiko Nakagawa and Naoki Miyata*
Graduate School of Pharmaceutical Sciences, Nagoya City University, 3-1 Tanabe-dori, Mizuho-ku, Nagoya, Aichi 467-8603, Japan
Received 10 September 2004; revised 25 October 2004; accepted 25 October 2004
Available online 13 November 2004
Abstract—In order to find novel non-hydroxamate histone deacetylase (HDAC) inhibitors, we synthesized several suberoylanilide
hydroxamic acid (SAHA)-based compounds designed on the basis of the catalytic mechanism of HDACs. Among these compounds,
5b was found to be as potent as SAHA. Kinetic enzyme assays and molecular modeling suggested that the mercaptoacetamide moi-
ety of 5b interacts with the zinc in the active site of HDACs and removes a water molecule from the reactive site of the deacetylation.
Ó 2004 Elsevier Ltd. All rights reserved.
Histone deacetylases (HDACs) catalyze the deacetyl-
ation of the acetylated e-amino groups of specific
histone lysine residues,1,2 and are involved in the expres-
sion of a number of genes.3 In addition, HDACs have
also been implicated in certain disease states such as
cancer.4–7 For this reason, there is a growing interest
in the generation of potent small-molecule inhibitors
of HDACs. Thus far, several classes of small-molecule
HDAC inhibitors have been recognized.8 Most of these
are hydroxamic acid derivatives, typified by suberoylani-
lide hydroxamic acid (SAHA) (Fig. 3), and are thought
to chelate the zinc ion in the active site.9,10 Although
hydroxamic acids are responsible for various potent
inhibitors, they generally have many problems associ-
ated with their use such as low oral availability, poor
in vivo stability, and undesirable side effects.11,12 Thus,
it has become increasingly desirable to find replacement
groups that possess strong inhibitory action against
HDACs. We and other groups have searched for a suit-
able hydroxamic acid replacement for HDAC inhibitors
by structure-based drug design (SBDD)13–15 ever since
the crystal structure of an archaebacterial HDAC
homologue (HDAC-like protein, HDLP)/SAHA com-
plex was first reported.9 However, SBDD has not yet
led to the discovery of potent non-hydroxamate HDAC
inhibitors, and the non-hydroxamates found with
SBDD are approximately 10–1000-fold less potent than
their corresponding hydroxamates. We therefore
decided to search for hydroxamic acid replacements by
an alternative approach, namely, mechanism-based drug
design. In this paper, we report the mechanism-based
design, synthesis, enzyme inhibition, and binding mode
of non-hydroxamate HDAC inhibitors.
The crystal structures of the HDLP/hydroxamates and
HDAC8/hydroxamates complexes have led to a solid
understanding of not only the three dimensional struc-
ture of the active site of HDACs but also the catalytic
mechanism for the deacetylation of acetylated lysine
substrate.9,10 It is proposed that the carbonyl oxygen
of this substrate could bind the zinc, and the carbonyl
could be attacked by a zinc-chelating water molecule
(Fig. 2a), which would result in the production of
deacetylated lysine via a tetrahedral carbon-containing
transition state (Fig. 1a). On the basis of the proposed
catalytic mechanism, we attempted to design non-hydr-
oxamate HDAC inhibitors. First, we designed transi-
tion-state (TS) analogues. The TS of HDAC
deacetylation was estimated to include a tetrahedral car-
bon (Fig. 1a) as with other zinc proteases.16 To date,
there has been only one report on TS analogue inhibi-
tors of HDACs, namely, phosphorus-based SAHA ana-
logues.17 However, these analogues have a potency
about 1000-fold less than that of SAHA. We focused
attention on sulfone derivative TS analogues because it
has been suggested that the sulfonamide moiety has
strong similarity with the TS of amide bond hydrolysis,
Keywords: Histone deacetylase inhibitor; Non-hydroxamate.
*
Corresponding authors. Tel./fax: +81 52 836 3407; e-mail addresses:
suzuki@phar.nagoya-cu.ac.jp; miyata-n@phar.nagoya-cu.ac.jp
0960-894X/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.10.074

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
332
T. Suzuki et al. / Bioorg. Med. Chem. Lett. 15 (2005) 331–335
ero atomto the zinc ion, and might behave as HDAC
inhibitors (Fig. 2b).
Ph
O
O
HN
H
N
Tyr
Tyr
The compounds prepared for this study are shown in
Table 1. The routes used for synthesis of the compounds
are indicated in Schemes 1–3. Scheme 1 shows the prep-
aration of sulfonamide 1, a TS analogue. The condensa-
tion of dicarboxylic acids 8a–c with an equivalent
amount of aniline gave mono-anilides 9a–c. Carboxylic
O
H
O
O
H
O
NH
Me
X
S
Zn
Zn
O
Me
O
(a)
(b)
Figure 1. The transition state proposed for HDACs (a), and models
for the binding of sulfone derivatives (b).
Table 1. HDAC inhibition data for SAHA, SAHA-based transition
state analogues, and substrate analoguesa
H
N
18
R
n
both froma steric and an electronic point of view.
O
Compounds 1 and 2, in which a hydroxamic acid of
SAHA was replaced by a sulfonamide and a sulfone,
respectively, were designed as TS analogues (Figs. 1b,
and 3). Our second approach was based on the proposed
Compd
SAHAb
R
n
% Inhbtn
at 100lM
IC50 (lM)
–CONHOH
6
100
0.28
deacetylation mechanism whereby
a zinc-chelating
water molecule activated by His142 and His143
(HDAC8 numbering) makes a nucleophilic attack on
the carbonyl carbon of acetylated lysine substrate (Fig.
2a). With this mechanism, if the water molecule is forci-
bly removed from the zinc ion, the HDACs would sup-
posedly be inhibited. We then designed hetero atom
containing substrate analogues 3–5 (Fig. 3). These ana-
logues would be recognized as substrates by HDACs
and would be easily taken into the active site where they
could force the water molecule off the zinc ion and the
reactive site of the deacetylation by chelation of the het-
1
2
–NHSO2Me
–SO2Me
5
6
10
33
7500
230
3
–NHCOCH2NH2
–NHCOCH2OH
–NHCOCH2SH
–NHCOCH2SH
–NHCOCH2SH
–NHCOCH2SAc
–NHCOCH2CH2SH
5
5
6
5
4
5
5
6
0
>100
>100
3.0
4
5a
5b
5c
6
96
99
88
72
78
0.39
11
22
7
24
a Values are means of at least three experiments.
b Prepared as described in Ref. 25.
O
H
Ph
O
Ph
O
HN
N
HN
O
NH
CH3
O
NH
O
X
NH
H
Zn
:
Zn
:
Zn
O
H
X
O
:N
H
H
H
O
H
N
H
His
(a)
(b)
Figure 2. The mechanism proposed for the deacetylation of acetylated lysine substrate (a), and a model for the binding of hetero atom containing
substrate analogues to zinc ion (b).
O
H
N
H
N
OH
R
N
H
O
O
3: R = -NHCOCH2NH2
2: R = -CH2SO2Me 4: R = -NHCOCH2OH
5: R = -NHCOCH2SH
1: R = -NHSO2Me
SAHA
Figure 3. Structures of SAHA, SAHA-based transition state analogues 1 and 2, and hetero atomcontaining substrate analogues 3–5 designed on the
basis of the deacetylation mechanism.

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
T. Suzuki et al. / Bioorg. Med. Chem. Lett. 15 (2005) 331–335
333
O
O
a
b - d
HOOC
COOH
n
Ph
NH2
n
Ph
COOH
n
N
H
N
H
8a: n = 6
8b: n = 5
8c: n = 4
9a: n = 6
9b: n = 5
9c: n = 4
10a: n = 6
10b: n = 5
10c: n = 4
O
H
e
Ph
N
Me
10b
N
H
S
5
O
O
1
Scheme 1. Reagents and conditions: (a) aniline, 180°C, 46–54%; (b) diphenylphosphoryl azide, Et3N, benzene, reflux; (c) BnOH, reflux, 63–94% (two
steps); (d) H2, 5% Pd–C, MeOH, rt, 72–96%; (e) MsCl, pyridine, rt, 71%.
equivalents of m-chloroperoxybenzoic acid gave sulfone
2. Hetero atomcontaining substrate analogues 3–7 were
prepared fromamines 10 obtained above by the proce-
dure outlined in Scheme 3. The amine 10b was reacted
with an appropriate carboxylic acid in the presence of
EDCI and HOBt in DMF to give compounds 4, 7 and
14. The N-Boc group of compound 14 was removed
by treating with trifluoroacetic acid to give aminoacet-
amide 3. Coupling between amines 10a–c and bromoace-
tyl chloride and subsequent treatment with potassium
thioacetate afforded compounds 6, 15, and 16 and the
deacetylation of these compounds in the presence of
K2CO3 in MeOH gave mercaptoacetamides 5a–c.
O
O
d
a - c
O
EtOOC
Br
6
Ph
SMe
6
Ph
Br
6
N
H
N
H
11
12
13
O
O
e
Ph
S
N
H
Me
6
2
Scheme 2. Reagents and conditions: (a) LiOH Æ H2O, THF, EtOH,
H2O, rt, 99%; (b) (COCl)2, DMF, CH2Cl2, rt; (c) aniline, Et3N,
CH2Cl2, rt, 87%; (d) 15% aq NaSMe, EtOH, rt, 99%; (e) m-
chloroperoxybenzoic acid, CH2Cl2, rt, 70%.
The compounds prepared for this study were evaluated
using an HDAC enzyme inhibition assay19 (Table 1).
In the case of TS analogues, sulfone 2 showed anti-
HDAC activity and the IC50 value was 230lM, which
was greater than those of phosphorus-based SAHA ana-
logues.17 However, sulfone 2 was approximately 820-
fold less effective than SAHA. Next, we examined hetero
atomcontaining substrate analogues. While 3 and 4 did
not possess HDAC inhibitory activities,20 potent inhibi-
tion was observed with mercaptoacetamide 5b. Com-
pound 5b exhibited an IC50 of 0.39lM, and its activity
largely surpassed those of phosphorus compounds17
and was comparable to those of SAHA and previously
reported non-hydroxamates.21,22 The potency of merca-
ptoacetamide 5a–c was directly related to chain length,
and the most potent compound was 5b, where n = 5.
As expected, thiol transformation into thioacetate (6)
led to a 55-fold less potent inhibitor. This result suggests
that thiolate anion generated under physiological condi-
tions has an intimate involvement in the interaction with
the zinc ion in the active site. The conversion of merca-
ptoacetamide to mercaptopropionamide (7) reduced po-
tency as compared to compound 5b.
O
H
a
Ph
N
10b
N
H
R
5
O
14: R = -NHBoc
3 : R = -NH2
4 : R = -OH
b
7 : R = -CH2SH
O
H
O
H
N
n
c, d
e
Ph
N
N
Ph
10a-c
SAc
N
H
SH
n
H
O
O
15: n = 6
6 : n = 5
16: n = 4
5a: n = 6
5b: n = 5
5c: n = 4
Scheme 3. Reagents and conditions: (a) RCH2COOH, EDCI, HOBt,
DMF, rt, 35–99%; (b) trifluoroacetic acid, CH2Cl2, rt, 84%; (c)
bromoacetyl chloride, Et3N, CH2Cl2, rt, 23–56%; (d) AcSK, EtOH, rt,
50–89%; (e) K2CO3, MeOH, rt, 28–75%.
acids 9a–c were converted to amines 10a–c in three
steps: Curtius rearrangement of carboxylic acids 9a–c,
treatment of the resulting isocyanates with benzyl alco-
hol, and cleavage of the Z group by hydrogenolysis.
Coupling between amine 10b and methanesulfonyl chlo-
ride afforded sulfonamide 1. Preparation of 2, the other
TS analogue, is shown in Scheme 2. 7-Bromoheptanoic
acid ethyl ester 11 was converted to 12 in three steps
by hydrolysis of the ester of 11, acid chloride formation
by oxalyl chloride, and condensation with aniline. Bro-
mide 12 was allowed to react with sodiummethanethio-
late to give sulfide 13, after which treatment with two
Next, we studied the inhibition mechanism of mercapto-
acetamide 5b. Although the mercaptoacetamide group
of 5b was designed to make use of its chelation of the
zinc ion in the active site, there is a possibility that
mercaptoacetamide 5b inhibits HDACs by forming a
covalent disulfide bond with cysteine residues of these
enzymes. We examined this possibility using a Line-
weaver–Burk plot (a double reciprocal plot of 1/V versus
1/[substrate] at varying concentrations of inhibitor 5b)

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
334
T. Suzuki et al. / Bioorg. Med. Chem. Lett. 15 (2005) 331–335
0.04
0.03
0.02
0.01
0
In summary, in order to find novel non-hydroxamate
HDAC inhibitors, we prepared several SAHA-based
compounds whose designs were based on the proposed
HDAC catalytic mechanism. Although transition state
analogues were weakly active against HDACs, merca-
ptoacetamide 5b, one of the hetero atomcontaining sub-
strate analogues, was found to be as potent as SAHA.
Mercaptoacetamide 5b exhibits strong competitive inhi-
bition versus acetylated lysine substrate. As far as we
could determine, this is the first report of HDAC inhib-
itors with mercaptoacetamide. Since mercaptoacet-
amides are reported as potent, long-lived, and low-
toxic matrix metalloproteinase inhibitors,23,24 we believe
that our findings in this study will provide the basis for
the development of ideal HDAC inhibitors free of the
problems associated with hydroxamates. Further
detailed structure–activity relationship studies are cur-
rently under way and the next stage of evaluations per-
taining to mercaptoacetamides 5 has begun.
-0.01
0
0.01
0.02
1/[substrate] (µM-1)
Figure 4. Reciprocal rate versus reciprocal acetyl lysine substrate
concentration in the presence of 1 (d), 0.3 (m), 0.1 (j), and 0 (s) lM
of 5b.
(Fig. 4), and the data fromthis study established that
mercaptoacetamide 5b engages in competitive inhibition
versus acetylated lysine substrate, with an inhibition
constant (Ki) of 0.78lM. Since cysteine is not a compo-
nent in the construction of the active site of HDACs, the
mercaptoacetamide group of 5b likely interacts with the
zinc in the active site.
Acknowledgements
This work was supported in part by grants fromthe
Health Sciences Foundation of the Ministry of Health,
Labor and Welfare of Japan, and the Mochida Memo-
rial Foundation for Medical and Pharmaceutical
Research.
Since mercaptoacetamide 5b was proven to act in the
HDAC active center, we studied its binding mode in this
site. The low energy conformation of 5b was calculated
when docked in the model based on the crystal structure
of HDAC8 (PDB code 1T64, 1T67, 1T69, and 1VKG)
using Macromodel 8.1 software. An inspection of the
HDAC8/5b complex showed that the sulfur atom and
˚
˚
References and notes
oxygen atomof 5b were located 2.44A and 2.04A from
the zinc ion, respectively, and that a water molecule,
which is required for the deacetylation of acetylated ly-
1. Grozinger, C. M.; Schreiber, S. L. Chem. Biol. 2002, 9,
3.
2. Hassig, C. A.; Schreiber, S. L. Curr. Opin. Chem. Biol.
1997, 1, 300.
3. Taunton, J.; Hassig, C. A.; Schreiber, S. L. Science 1996,
272, 408.
4. Yoshida, M.; Horinouchi, S.; Beppu, T. BioEssays 1995,
17, 423.
˚
sine substrate, was positioned 4.95A apart fromthe zinc
ion (Fig. 5). This calculation suggests that 5b inhibits
HDACs by chelating the zinc ion in a bidentate fashion
through its sulfur and oxygen atoms, and by removing a
water molecule from the zinc and the reactive site of the
deacetylation, without being hydrolyzed by HDACs.
Figure 5. View of the conformation of 5b (ball and stick) docked in the HDAC8 catalytic core. Residues around the zinc ion and a water molecule are
displayed as wires and tubes, respectively.