Novel histone deacetylase inhibitors: Design, synthesis, enzyme inhibition, and binding mode study of SAHA-Based non-hydroxamates

Article abstract of DOI:10.1016/j.bmcl.2003.09.048

In order to find novel non-hydroxamate histone deacetylase (HDAC) inhibitors, a series of compounds modeled after suberoylanilide hydroxamic acid (SAHA) were designed and synthesized as (i) substrate (acetyl lysine) analogues (compounds 3-7), (ii) analogu

Full text of DOI:10.1016/j.bmcl.2003.09.048

Bioorganic & Medicinal Chemistry Letters 13 (2003) 4321–4326
Novel Histone Deacetylase Inhibitors: Design, Synthesis, Enzyme
Inhibition, and Binding Mode Study of SAHA-Based
Non-hydroxamates
Takayoshi Suzuki,a Yuki Nagano,a Azusa Matsuura,a Arihiro Kohara,b
Shin-ichi Ninomiya,b Kohfuku Kohdaa and Naoki Miyataa,*
aGraduate School of Pharmaceutical Sciences, Nagoya City University, 3-1 Tanabe-dori, Mizuho-ku, Nagoya,
Aichi 467-8603, Japan
bDaiichi Pure Chemicals Co., Ltd., 2117 Muramatsu, Tokai, Ibaraki 319-1182, Japan
Received 26 August 2003; revised 19 September 2003; accepted 22 September 2003
Abstract—In order to find novel non-hydroxamate histone deacetylase (HDAC) inhibitors, a series of compounds modeled after
suberoylanilide hydroxamic acid (SAHA) were designed and synthesized as (i) substrate (acetyl lysine) analogues (compounds 3–7),
(ii) analogues bearing various functional groups expected to chelate zinc ion (compounds 8–15), and (iii) analogues bearing nucleo-
philic functional groups which could bind covalently to HDACs (compounds 16–18). In this series, semicarbazide 8b and bromo-
acetamides 18b,c were found to be potent HDAC inhibitors for non-hydroxamates.
# 2003 Elsevier Ltd. All rights reserved.
Histone deacetylases (HDACs) are known to play an
important role in the regulation of gene expression1 by
catalyzing the deacetylation of the acetylated e-amino
groups of specific histone lysine residues.2ꢀ4 In addition,
HDAC inhibitors such as trichostatin A (TSA), sub-
eroylanilide hydroxamic acid (SAHA), and Trapoxin B
(TPX B) (Fig. 1) have been reported to inhibit cell
growth,5ꢀ8 induce terminal differentiation in tumor
cells,5,6 and prevent the formation of tumors in mice.9
Therefore, HDACs have emerged as an attractive target
for the development of new anticancer drugs. A number
of structurally diverse HDAC inhibitors have been
reported10 and most of them belong to hydroxamic acid
derivatives, typified by TSA and SAHA, which can
interact with zinc in the active site.11 Although hydro-
xamic acids are frequently employed as zinc-binding
groups, they often present metabolic and pharmaco-
kinetic problems such as glucoronidation and sulfation
that result in a short in vivo half-life.12,13 Many hydrox-
amates are unstable in vivo, and are prone to hydrolysis
giving hydroxylamine which has potential mutagenic
properties.14 Because of such concerns with the
metabolic stability and the toxicity associated with
hydoxamic acids, it has become increasingly desirable to
find replacement groups that possess strong inhibitory
action against HDACs. In addition, in terms of biolo-
gical research, the discovery of specific irreversible
HDAC inhibitors by the replacement of a hydroxamic
acid with reactive functional groups will lead to the
identification of a new type of biological probe for the
isolation and cloning of an HDAC.15 To date, with the
exception of natural macrocyclic products which are not
suitable for oral administration, o-aminoanilides16ꢀ18
*Corresponding author. Tel./fax: +81-52-836-3407; e-mail: miyata-n
@phar.nagoya-cu.ac.jp
Figure 1. HDAC inhibitors.
0960-894X/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2003.09.048

4322
T. Suzuki et al. / Bioorg. Med. Chem. Lett. 13 (2003) 4321–4326
Figure 2. Synthetic non-hydroxamate HDAC inhibitors.
and trifluoromethylketones19 (Fig. 2) have been repor-
ted as non-hydroxamate HDAC inhibitors, and MS-27-
27516 bearing o-aminoanilide is currently undergoing
evaluation in clinical trials. However, o-aminoanilides
and trifluoromethylketones have reduced potency as
compared to hydroxamate inhibitors, and what is
worse, trifluoromethylketones have a metabolic prob-
lem in that they are readily reduced to inactive alcohols
in vivo, even within cells.19 We therefore initiated a
search for replacement groups of hydroxamic acid,
other than o-aminoanilide and trifluoromethylketone,
with the goal of drug discovery as well as finding new
tools for biological research. In this letter, we report the
design, synthesis, HDAC inhibition, and binding mode
analysis of non-hydroxamates based on the structure of
SAHA.
equivalent amount of aniline gave mono-anilides 20a–d.
Curtius rearrangement of carboxylic acids 20b–d using
diphenylphosphoryl azide20 provided the isocanates,
which on treatment with an appropriate alcohol or
amine gave carbamates 3, 4, urea 7, and semicabazides
8a–c.
Compounds 5, 6, 16, 17, and 18a–d were prepared from
carboxylic acids 20a–d obtained above by the procedure
outlined in Scheme 2. Carboxylic acids 20a–d were con-
verted to amines 21a–d with a three step sequence:
Curtius rearrangement of carboxylic acids 20a–d, treat-
ment of the resulting isocyanates with benzyl alcohol,
and cleavage of the Z group by hydrogenation. Coupl-
ing between amines 21a–d and an appropriate acid
anhydride or bromoacetyl bromide afforded compounds
5, 6, and 18a–d. The amine 21b was reacted with pro-
pargyl bromide in the presence of K2CO3 to give mono-
and di-alkylated compounds 16 and 17.
A series of compounds modeled after SAHA were syn-
thesized as schematized in Schemes 1–4. Compounds 3,
4, 7, and 8a–c were synthesized from the corresponding
dicarboxylic acids 19b–d by the route shown in Scheme 1.
The condensation of dicarboxylic acids 19a–d with an
Compounds 9–12, and 13 were obtained using standard
methodology (Scheme 3). Carboxylic acid 20b was treated
Scheme 2. (a) Diphenylphosphoryl azide, Et3N, benzene, reflux; (b)
BnOH, reflux, 53–94% (two steps); (c) H2, 5%Pd/C, MeOH, 72–99%;
(d) TFAA or (CCl3CO)2O, Et3N, CHCl3, 80% for 5, 38% for 6; (e)
bromoacetyl bromide, Et3N, THF, 22–41%; (f) propargyl bromide,
K2CO3, MeOH, 58% (16), 17% (17).
Scheme 1. (a) Aniline, 180 ꢁC, 46–54%; (b) diphenylphosphoryl azide,
Et3N, benzene, reflux; (c) R0OH or R0NH2, reflux, 24–97% (two
steps).

T. Suzuki et al. / Bioorg. Med. Chem. Lett. 13 (2003) 4321–4326
4323
Table 1. HDAC inhibition data for SAHA and SAHA-based non-
hydroxamatesa
Entry Comp
R
n
% Inhib at
100 mM
IC50
(mM)
1
2
SAHAb
–CONHOH
6
6
100
48
0.28
12
1
0
3
4
5
6
7
8
9
10
11
12
13
3
4
5
6
7
8a
8b
8c
9
10
11
–NHCOOMe
–NHCOOEt
–NHCOCF3
–NHCOCCl3
–NHCONH2
6
6
6
6
6
6
5
4
6
6
6
12
19
3
4
7
18
35
12
5
4
3
>100
>100
>100
>100
>100
650
–NHCONHNH2
–NHCONHNH2
–NHCONHNH2
–COOCH2CH2OH
–CONHCH2CH2OH
–CONHCH2CH2NH2
150
Scheme 3. (a) ClCOOMe, Et3N, THF; (b) ethyleneglycol or amino-
ethanol, CH2SO4, 84% for 9, 48% for 10 (two steps); (c) ethylene-
diamine, EDCI, NHS, CHCl3, 49%; (d) 2-aminopyridine, 1H-
benzotriazol-1-yloxytripyrrolidinophosphonium hexafluorophosphate
(PyBOP1), DMF, 85%; (e) 2-aminothiazole, EDCI, HOBt, DMF,
95%.
>1000
>100
>100
>100
14
15
16
17
18
19
12
13
14
15
16
17
6
6
6
6
6
6
8
9
>100
>100
>100
>100
>100
>100
15
0
Scheme 4. (a) N,O-Dimethylhydroxylamine hydrochloride, Et3N,
EDCI, HOBt, DMF, 85%; (b) 2-bromopyridine, n-BuLi, THF,
ꢀ78 ꢁC, 88%; (c) hydroxylamine hydrochloride, pyridine, EtOH, 90%.
16
8
with methyl chloroformate followed by ethyleneglycol
or aminoethanol to yield ester 9 and amide 10. Amides
11, 12, and 13 were obtained by the condensation of
carboxylic acid 20b with an appropriate amine in the
presence of condensing reagents such as EDCI and
HOBt.
20
21
22
23
18a
18b
18c
18d
–NHCOCH2Br
–NHCOCH2Br
–NHCOCH2Br
–NHCOCH2Br
7
6
5
4
40
79
78
47
300
14
17
121
aValues are means of at least three experiments.
bPrepared as described in ref 31.
Synthesis of ketone 14 was accomplished via Weinreb
amide21 22 as shown in Scheme 4. Carboxylic acid 20b
was converted to Weinreb amide 22 using the procedure
employed in the preparation of 13. Compound 22 was
allowed to react with 2-pyridinyllithium at ꢀ78 ꢁC to
give ketone 14 and subsequent oxime formation affor-
ded compound 15. The stereochemistry of oxime 15 was
determined to be E by following the Beckmann rear-
rangement method.22,23 More specifically, treatment of
15 with PCl5 in dichloromethane at room temperature
yielded amide 12.
synthesized acetyl lysine-like derivatives 3–7 expected to
inhibit HDACs competitively were found to be totally
inactive or only weakly potent against HDACs (entries
3–7).
The crystal structure of an archaebacterial HDAC
homologue (HDAC-like protein, HDLP)/SAHA com-
plex made it clear that the hydroxamic acid group
coordinates the zinc ion through its CO and OH groups
and also forms three hydrogen bonds between its CO,
NH, OH groups and Tyr 297, His 132, His 131, respec-
tively.11 From these data, semicarbazide 8b was synthe-
sized and tested as an HDAC inhibitor because it is
possible for semicarbazide to chelate metal ions and
form hydrogen bonds with Tyr and His. As was expec-
ted, 8b showed anti-HDAC activity and the IC50 value
was comparable to that of o-aminoanilide 1 (entry 9).
The compounds synthesized in this study were tested
with an in vitro assay using a HeLa nuclear extract, rich
in HDAC activity.24 The results are summarized in
Table 1.
The IC50 values of SAHA and o-aminoanilide 1 were
0.28 and 120 mM, respectively (entries 1 and 2). Newly

4324
T. Suzuki et al. / Bioorg. Med. Chem. Lett. 13 (2003) 4321–4326
Figure 3. Superposition of the low energy conformations of 8b (tube) and SAHA (wire) (left). The HDLP pocket is not shown for the sake of clarity.
View of the conformation of 8b (tube) docked into the HDLP catalytic core (right). Residues 5 A from the ligand are displayed in the wire graphic.
Figure 4. Superposition of the low energy conformations of 18c (tube) and SAHA (wire) (left). The HDLP pocket is not shown for the sake of
clarity. View of the conformation of 18c (ball and stick) docked into the HDLP catalytic core (right). Residues within 5 A of the zinc ion are dis-
played in the tube graphic.
The potency of semicarbazides 8a–c was directly related
to chain length, and the most potent compound was 8b,
where n=5 (entries 8–10). To find more potent zinc-
binding groups, we examined the simplified analogues
(9–11) of o-aminoanilide 1 in which the benzene ring is
replaced by an ethylene chain, but these compounds had
lost HDAC inhibitory activities (entries 11–13). These
results suggest that the benzene ring of 1 is required to
fix the position of its amino group or to interact with
amino acid residues of the enzyme. Since the effect of
the benzene ring of 1 proved to be influential, we intro-
duced heteroaryl rings as zinc-binding groups (12–15).
Although arylamide, the 2-acylated heteroaryl ring, and
the oxime have been reported to coordinate zinc ion in a
bidentate fashion through the hetero atom of the het-
eroaryl ring and the carbonyl/oxime group,14,25,26 all
heteroaryl compounds proved to be less effective than o-
aminoanilide 1 in inhibiting the enzyme at 100 mM
(entries 14–17).
We turned our attention to irreversible HDAC inhibi-
tors. TPX B is an irreversible HDAC inhibitor,8 but
finding more specific and simpler irreversible HDAC
inhibitors is useful for the isolation and cloning of an
HDAC.15 As described above, the crystal structure of
HDLP/SAHA complex revealed that the hydroxamic
acid group forms three hydrogen bonds with Tyr 297,
His 132, and His 131, and furthermore, zinc ion is

T. Suzuki et al. / Bioorg. Med. Chem. Lett. 13 (2003) 4321–4326
4325
coordinated by His 170, Asp 168, and Asp 258. Since
the phenol group of Tyr, the imidazole group of His,
and the carboxyl group of Asp are able to react with
electrophiles, we prepared analogues bearing propargyl
amino (16, 17) and bromoacetamide (18b) which could
bind covalently to Tyr, His, and Asp of the enzyme, and
evaluated their anti-HDAC activities. While propargyl
amino compounds 16 and 17 did not possess HDAC
inhibitory activities, potent inhibition was observed
with bromoacetamide 18b (entries 18, 19, and 21). Bro-
moacetamide 18b exhibited an IC50 of 14 mM and its
activity largely surpassed that of o-aminoanilide 1.
HDAC inhibition was distinctly dependent on chain
length within the bromoacetamide series, with n=7
(18a) and n=4 (18d) resulting in less potent inhibitors.
However, 18c, in which n=5, proved to be at least
equally effective to 18b, in which n=6 (entries 20–23).
amides 18b and 18c proved to be the strongest HDAC
inhibitors of all non-hydroxamate inhibitors prepared
for this study. HDAC inhibition was also distinctly
dependent on chain length within the bromoacetamide
series, with n=5 and 6 optimal. The binding mode
analysis of 18c suggests that 18c may inhibit HDACs by
forming a covalent bond with His 132. More specific
irreversible HDAC inhibitors can be used for the isola-
tion and cloning of an HDAC.15 Further investigation
and a detailed inhibitory mechanism analysis pertaining
to 8b and 18b,c are under way.
Acknowledgements
This work was supported in part by a grant from the
Health Sciences Foundation from the Ministry of
Health, Labor and Welfare of Japan.
To study the binding mode of compounds 8b and 18c in
the active site, we calculated the low energy conforma-
tions of 8b and 18c when they are docked into the model
based on the crystal structure of HDLP (PDB code
1C3R) using the software packages Macromodel 8.0,27
and compared the results with those obtained with
SAHA. First, the binding mode of 8b was investigated.
The anilide group and alkyl chain of 8b and SAHA are
essentially superimposed in the binding pocket (Fig. 3,
left), and the binding mode of 8b is similar to that of
SAHA (Fig. 3, right). An inspection of the HDLP/8b
complex shows that the semicarbazide moiety of 8b was
slightly removed from the zinc ion (CO-Zn, 3.1 A;
NH2–Zn, 2.1 A), whereas the hydroxamic acid group of
SAHA was positioned at the optimal bond distance
(CO–Zn, 2.7 A; OH–Zn, 1.8 A). Next, we investigated
the binding mode of 18c. As is the case with 8b, the
anilide group and alkyl chain of 18c and SAHA are
superimposed in the binding pocket (Fig. 4, left). In the
case of 18c, only the carbonyl group of bromoacetamide
chelates zinc ion as shown in Figure 4, right. Because 18c
is thought to be an irreversible inhibitor, 18c may inhibit
HDACs by forming a covalent bond with His 132lying
next to the reactive site of the bromoacetamide.
References and Notes
1. Davie, J. R.; Chadee, D. N. J. Cell. Biochem. 1998, 30–
31 (Suppl), 203.
2. Grunstein, M. Nature 1997, 389, 349.
3. Ng, H. H.; Bird, A. Trends Biochem. Sci. 2000, 25, 121.
4. Cheung, W. L.; Briggs, S. D.; Allis, C. D. Curr. Opin. Cell.
Biol. 2000, 12, 326.
5. Yoshida, M.; Horinouchi, S.; Beppu, T. BioEssays 1995,
17, 423.
6. Richon, V. M.; Webb, Y.; Merger, R.; Sheppard, T.; Jursic,
B.; Ngo, L.; Civoli, F.; Breslow, R.; Rifkind, R. A.; Marks,
P. A. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 5705.
7. Richon, V. M.; Emiliani, S.; Verdin, E.; Webb, Y.; Breslow,
R.; Rifkind, R. A.; Marks, P. A. Proc. Natl. Acad. Sci. U.S.A.
1998, 95, 3003.
8. Kijima, M.; Yoshida, M.; Sugita, K.; Horinouchi, S.;
Beppu, T. J. Biol. Chem. 1993, 268, 22429.
9. Cohen, L. A.; Amin, S.; Marks, P. A.; Rifkind, R. A.;
Desai, D.; Richon, V. M. Anticancer Res. 1999, 19, 4999.
10. Jung, M. Curr. Med. Chem. 2001, 8, 1505.
11. 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.
In summary, in order to find novel non-hydroxamate
HDAC inhibitors, we designed and prepared SAHA-
based compounds as (i) substrate (acetyl lysine) analo-
gues (compounds 3–7), (ii) analogues bearing various
functional groups expected to chelate zinc ion (com-
pounds 8–15), and (iii) analogues bearing nucleophilic
functional groups which could bind covalently to
HDACs (compounds 16–18). In zinc coordination-
oriented series, semicarbazide 8b was found to be as
potent as o-aminoanilide 1, and we have shown that the
potency is related to chain length, with n=5 optimal.
Molecular modeling suggests that the binding mode of
8b is similar to that of SAHA. As far as we could
determine, this is the first report of semicarbazide as a
zinc-dependent enzyme inhibitor. The semicarbazide
group may be useful as a source of other zinc-dependent
enzyme inhibitors such as matrix metalloproteinase
inhibitors,14,28 tumor necrosis factor-a converting
enzyme inhibitors,29 and aggrecanase inhibitors.30 In
the irreversible inhibition-oriented series, bromoacet-
12. Mulder, G. J.; Meerman, J. H. Environ. Health Perspect.
1983, 49, 2 7.
13. Vassiliou, S.; Mucha, A.; Cuniasse, P.; Georgiadis, D.;
Lucet-Levannier, K.; Beau, F.; Kannan, R.; Murphy, G.;
Knaeuper, V.; Rio, M. C.; Basset, P.; Yiotakis, A.; Dive, V. J.
Med. Chem. 1999, 42, 2610.
14. Whittaker, M.; Floyd, C. D.; Brown, P.; Gearing, A. J.
Chem. Rev. 1999, 99, 2735.
15. Taunton, J.; Hassig, C. A.; Schreiber, S. L. Science 1996,
272, 408.
16. Saito, A.; Yamashita, T.; Mariko, Y.; Nosaka, Y.; Tsu-
chiya, K.; Ando, T.; Suzuki, T.; Tsuruno, T.; Nakanishi, O.
Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 4592.
17. Wong, J. C.; Hong, R.; Schreiber, S. L. J. Am. Chem. Soc.
2003, 125, 5586.
18. Haggarty, S. J.; Koeller, K. M.; Wong, J. C.; Butcher,
R. A.; Schreiber, S. L. Chem. Biol. 2003, 10, 383.
19. Frey, R. R.; Wada, C. K.; Garland, R. B.; Curtin, M. L.;
Michaelides, M. R.; Li, J.; Pease, L. J.; Glaser, K. B.; Mar-
cotte, P. A.; Bouska, J. J.; Murphy, S. S.; Davidsen, S. K.
Bioorg. Med. Chem. Lett. 2002, 12, 3443.

4326
T. Suzuki et al. / Bioorg. Med. Chem. Lett. 13 (2003) 4321–4326
20. Shioiri, T.; Ninomiya, K.; Yamada, S. J. Am. Chem. Soc.
1972, 94, 6203.
25. Auge, F.; Hornebeck, W.; Decarme, M.; Laronze, J.-Y.
Bioorg. Med. Chem. Lett. 2003, 13, 1783.
21. Nahm, S.; Weinreb, S. M. Tetrahedron Lett. 1981, 22,
3815.
26. Cooperman, B. S.; Lloyd, G. J. J. Am. Chem. Soc. 1971,
93, 4883.
22. Huntress, E. H.; Walter, H. C. J. Am. Chem. Soc. 1948,
70, 3702.
23. Negi, S.; Matsukura, M.; Mizuno, M.; Miyake, K.; Min-
ami, N. Synthesis 1996, 991.
24. HDAC activity assay was performed using the HDAC
fluorescent activity assay/drug discovery kit (AK-500, BIO-
MOL Research Laboratories): HeLa Nuclear Extract (0.5 mL/
well) were incubated (37 ꢁC) with 25 mM of Fluor de LysTM
substrate and various concentrations of samples. Reactions
were stopped after 30 min. with Fluor de LysTM Developer
and fluorescence was measured on a fluorometric reader with
excitation set at 360 nm and emission detection set at 460 nm.
27. Mohamadi, F.; Richards, N. G. J.; Guida, W. C.; Lis-
kamp, R.; Lipton, M.; Caufield, C.; Chang, G.; Hendrickson,
T.; Still, W. C. J. Comput. Chem. 1990, 11, 440.
28. Skiles, J. W.; Gonnella, N. C.; Jeng, A. Y. Curr. Med.
Chem. 2001, 8, 425.
29. Chen, J. M.; Jin, G.; Sung, A.; Levin, J. I. Bioorg. Med.
Chem. Lett. 2002, 12, 1195.
30. Yao, W.; Chao, M.; Wasserman, Z. R.; Liu, R.-Q.; Cov-
ington, M. B.; Newton, R.; Christ, D.; Wexler, R. R.; Decicco,
C. P. Bioorg. Med. Chem. Lett. 2002, 12, 101.
31. Mai, A.; Esposito, M.; Sbardella, G.; Massa, S. Org. Prep.
Proced. Int. 2001, 33, 391.