Design, synthesis, inhibitory activity, and binding mode study of novel DNA methyltransferase 1 inhibitors

Article abstract of DOI:10.1016/j.bmcl.2009.12.016

To identify novel non-nucleoside DNA methyltransferase (DNMT) inhibitors, we designed and synthesized a series of maleimide derivatives. Among this series, compounds 5-8 were found to be more potent DNMT1 inhibitors than RG108, a DNMT1 inhibitor reported previously by Siedlecki et al. The binding mode analysis of compound 5 is also reported.

Full text of DOI:10.1016/j.bmcl.2009.12.016

Bioorganic & Medicinal Chemistry Letters 20 (2010) 1124–1127
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Design, synthesis, inhibitory activity, and binding mode study of novel
DNA methyltransferase 1 inhibitors
*
*
Takayoshi Suzuki , Rikako Tanaka, Shohei Hamada, Hidehiko Nakagawa, Naoki Miyata
Graduate School of Pharmaceutical Sciences, Nagoya City University, 3-1 Tanabe-dori, Mizuho-ku, Nagoya, Aichi 467-8603, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
To identify novel non-nucleoside DNA methyltransferase (DNMT) inhibitors, we designed and synthe-
sized a series of maleimide derivatives. Among this series, compounds 5–8 were found to be more potent
DNMT1 inhibitors than RG108, a DNMT1 inhibitor reported previously by Siedlecki et al. The binding
mode analysis of compound 5 is also reported.
Received 5 November 2009
Revised 30 November 2009
Accepted 3 December 2009
Available online 21 December 2009
Ó 2009 Elsevier Ltd. All rights reserved.
Keywords:
DNA methyltransferase
Inhibitor
Non-nucleoside
NH2
NH2
DNA methylation at the 5-position of cytosine in CpG dinucleo-
tides is a pivotal mechanism for the epigenetic regulation of gene
expression. Hypermethylation occurs in the CpG-rich sequence,
the so-called CpG islands, where core promoters and transcription
initiation sites are located. The hypermethylation in the CpG is-
lands leads to the silencing of genes.1–3 CpG island-specific hyper-
methylation is a characteristic common to cancer cells. This causes
the silencing of tumor suppressor genes, such as p16INK4a and hu-
man mutL homologue 1, which are involved in the tumorigenic pro-
cess, including DNA repair, cell cycle regulation, and apoptosis.4–7
Therefore, DNA methylation inhibitors are regarded as potential
anticancer agents.8
CH3
N
DNMT
SAM
N
O
N
R
O
N
R
SAH
5-methylcytosine
cytosine
Figure 1. DNA methylation at the 5-position of cytosine catalyzed by DNMT. SAH:
S-adenosyl-L-homocysteine.
COO
DNA methylation is catalyzed by DNA methyltransferases
O
(DNMTs) using S-adenosyl-
L
-methionine (SAM) as the methyl do-
H3N
O
S
N
Glu
nor (Fig. 1). At present, four mammalian DNMTs, namely, DNMT1,9
DNMT2,10 DNMT3A, and DNMT3B,11 are known. Among these,
DNMT1 is regarded as the major contributor to DNMT activity
and a maintenance methyltransferase in human cells because it
shows preference for hemi-methylated DNA substances that are
methylated in one strand and unmethylated in the other. The pri-
mary function of DNMT1 may be the copying of methylation pat-
terns from the parent DNA strand to the newly replicated
daughter strand during the DNA replication process.12
N
H3C
O
NH2
HO
O
OH
SAM
N
H
N
H
H
N
H
N
H
N
O
N
5 H
H
O
H
5
6
O
N
R
N
R
6
S
S
cytosine
Cys
The catalytic mechanism for the methylation of cytosine-5 has
been studied extensively.13–15 As depicted in Figure 2, a thiol of
the cysteine residue in the active site of DNMTs serves as a nucle-
ophile that attacks the 6-position of cytosine to generate a covalent
NH2
5 CH3
N
N
H2
CH3
5
N
N
H
:
B
6
6
S Enz
O
N
R
O
R
* Corresponding authors. Tel./fax: +81 52 836 3407 (N.M.).
E-mail addresses: suzuki@phar.nagoya-cu.ac.jp (T. Suzuki), miyata-n@phar.na-
goya-cu.ac.jp (N. Miyata).
5-methylcytosine
Figure 2. Proposed catalytic mechanism for the methylation of cytosine by DNMT.
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.12.016

T. Suzuki et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1124–1127
1125
DNA-protein intermediate that possesses nucleophilic properties
at the 5-position. This reactive intermediate accepts a methyl
group from SAM to form the 5-methyl covalent adduct and S-aden-
osyl-L-homocysteine (SAH). Following the methyl transfer, the pro-
Arg 1310
NH
Arg 1310
NH
a
b
HN
HN
HN
HN
H
H
H
H
ton at the 5-position is abstracted by a basic residue in the active
site of the enzyme that is removed from the 6-position by b-elim-
ination to generate the methylated cytosine and the free enzyme.
To date, several DNMT inhibitors have been developed.16–20
Nucleoside analogues, such as 5-azacytidine 1, 5-fluoro-20-deoxy-
cytidine 2, and zebularine 3 (Fig. 3), are DNMT inhibitors that show
antiproliferative activity against cancer cells.16,21–23 However,
many nucleoside analogues have problems, such as severe toxicity,
which are probably associated with their incorporation into DNA.24
We therefore started a search for non-nucleoside DNMT inhibitors
from the point of view of drug discovery and the discovery of a tool
for biological research. In this letter, we report the design, synthe-
sis, inhibitory activity, and binding mode study of novel non-nucle-
oside DNMT inhibitors.
In designing novel non-nucleoside DNMT inhibitors, we focused
on the structure of RG108 4 (Fig. 3), a DNMT1 inhibitor reported
previously by Siedlecki et al.25 From a computational study, it
has been predicted that the carboxylate anion and the carbonyl
of the phthalimide form hydrogen bonds with Arg 1310 and Arg
1308, respectively, and the phthalimide group lies next to Cys
1226 in the active site of DNMT1 (Fig. 4a). Based on the structure
of the RG108/DNMT1 complex, we designed maleimide derivatives
5–13 (Fig. 5). Succinimide 14 (Fig. 5) was also designed as a refer-
ence compound. These analogues were expected to undergo conju-
gate addition from a thiol of Cys 1226 in the active site of DNMT1,
which would lead to DNMT1 inhibition (Fig. 4b).
O
O
O
O
NH
NH
O
O
S
SH
Cys
Cys
1226
N
N
R1
1226
O
H
O
H
R2
NH2
NH2
Arg
1308
Arg
1308
H2N
N
H
H2N
N
H
Figure 4. Simulated binding mode of RG108 4 in the active site of DNMT1 (a), and
model for the binding of maleimide derivatives (b).
R3
O
COOH
O
NH
N
R4
R1
N
O
O
R2
5 13
-
14
5: R1 = R2 = H, R3 = COOH, R4 = 3-indolyl
6: R1 = H, R2 = CH3, R3 = COOH, R4 = 3-indolyl
COOH,
7: R1 = H, R2 = Ph, R3
=
R
4 = 3-indolyl
: R1 = R2 = CH3, R3
=
R = 3-indolyl
4
8
9
10
COOH,
: R1 = R2 = H, R3 = COOH, R4 = Ph
: R1 = R2 = H, R3
=
4 = H
COOH,
R
11: R1 = R2 = R3 = H, R4 = 3-indolyl
12: R1 = H, R2 = CH3, R3 = H, R4 = 3-indolyl
13: R1 = H, R2 = Ph, R3 = H, R4 = 3-indolyl
Figure 5. Structures of compounds 5–14.
A series of compounds modeled after RG108 were synthesized,
as shown in Schemes 1 and 2. Compounds 5–13 were synthesized
from corresponding maleic anhydrides 15a–d and amines 16a–d in
only one step via the route shown in Scheme 1. Heating maleic
anhydrides 15a–d and amines 16a–d in refluxing AcOH afforded
desired compounds 5–13. The hydrogenation of maleimide 5 using
Pd/C produced succinimide 14 (Scheme 2).
The compounds synthesized in this study were tested in an
in vitro assay using human DNMT1.26 The results are summarized
in Figure 6. In this enzyme assay, RG108 4 showed 34% inhibition
R3
O
O
R1
R2
R3
a
N
R4
O
R1
H2N
R4
O
R2
O
15a:
15b:
15c:
15d:
16a:
16b:
16c:
16d:
R
R
R
R
1 = R2 = H
R
R
R
3 = COOH, R4 = 3-indolyl
3 = COOH, R4 = Ph
5-13
1 = H, R2 = CH3
1 = H, R2 = Ph
3 = COOH, R4 = H
1 = R2
=
CH3
R
3 = H, R4 = 3-indolyl
of DNMT1 activity at the concentration of 1000 lM. Compounds 5–
8, in which the phthalimide of RG108 4 is replaced with various
maleimides, exerted more potent DNMT1 inhibitory activity than
RG108 4. Among these, compound 5 was the most potent DNMT1
Scheme 1. (a) AcOH, reflux, 5–68%.
inhibitor, displaying over 60% inhibition even at 10 lM. Com-
COOH
O
NH
pounds 9 and 10, in which the indole ring of 5 is replaced with a
phenyl ring and a hydrogen, respectively, resulted in less potent
inhibition, suggesting the importance of the indole ring. In addi-
tion, compounds 11–13 that had no carboxylic acid moiety tended
to show less potent inhibitory activity than corresponding carbox-
ylic acids 5–7. These results suggest that the interaction between
the carboxylate anion and Arg 1310 is responsible for the DNMT1
inhibitory activity. Compound 14, in which the maleimide of 5 is
replaced with a succinimide, was completely inactive. The reason
a
5
N
O
14
Scheme 2. (a) Pd/C, H2, MeOH, rt, 78%.
for the loss of activity is unclear, although it seems reasonable to
assume that it is because compound 14 does not have an ,b-
a
unsaturated keto structure and cannot undergo addition from a
thiol of Cys 1226 in the active site of DNMT1.
NH2
N
NH2
N
To study the binding mode of compound 5 in the active site, we
calculated the lowest energy conformation of 5 when it has been
docked into the model based on the crystal structure of M.Hha
I,27 a DNA methyltransferase from Haemophilus haemolyticus, using
software packages Glide 3.5 and Macromodel 8.1.28 An inspection
of the complex shows that the indole ring of compound 5 is located
in the hydrophobic region formed by the benzene ring of Tyr 254
and the methylene chains of Gly 88 and Gly 255 (Fig. 7). In addi-
tion, it is suggested that the carboxylate anion of 5 forms two
F
O
N
O
N
O
N
O
COOH
N
O
N
O
O
N
NH
OH
OH
OH
OH
O
HO
HO
HO
OH
4
2
3
1
Figure 3. Reported DNMT inhibitors.

1126
T. Suzuki et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1124–1127
160
140
120
100
80
60
40
20
0
conc.
(M)
blank
4
5
6
7
8
9
10
11
12
13
14
No inhibitor
Inhibitor
Figure 6. DNMT1 inhibitory activity of compounds 4–14.
Arg 163
NH
Gly 225
H
H2N
H
H
N
N
O
H
N
H
N
Gly 88
O
H
Arg
165
O
NH2
N
Tyr 254
O
H
NH
O
SH
Cys 81
Gln 82
Figure 7. View of the conformation of compound 5 (green) docked in the active site of M.Hha I (left) and its schematic representation (right).
5. Li, E.; Bestor, T. H.; Jaenisch, R. Cell 1992, 69, 915.
6. Okano, M.; Bell, D. W.; Haber, D. A.; Li, E. Cell 1999, 99, 247.
7. Esteller, M. Oncogene 2002, 21, 5427.
hydrogen bonds with Arg 163. It is also shown that one carbonyl of
the maleimide forms two hydrogen bonds with Arg 163 and Arg
165, and the other carbonyl forms a hydrogen bond with Gln 82.
Further, the maleimide moiety of 5 lies next to Cys 81, where the
conjugate addition of a thiol to the maleimide can occur.
In summary, we have identified novel non-nucleoside lead
structures from which more potent DNMT inhibitors can be devel-
oped. The findings of this study should pave the way for the devel-
opment of new anticancer drugs. Detailed studies of compound 5
and its analogues are under way.
8. Yu, N.; Wang, M. Curr. Med. Chem. 2008, 15, 1350.
9. Bestor, T.; Laudano, A.; Mattaliano, R.; Ingram, V. J. Mol. Biol. 1988, 203, 971.
10. Yoder, J. A.; Bestor, T. H. Hum. Mol. Genet. 1998, 7, 279.
11. Okano, M.; Xie, S.; Li, E. Nat. Genet. 1998, 19, 219.
12. Pradhan, S.; Bacolla, A.; Wells, R. D.; Roberts, R. J. J. Biol. Chem. 1999, 274,
33002.
13. Wu, J. C.; Santi, D. V. J. Biol. Chem. 1987, 262, 4778.
14. Chen, L.; MacMillan, A. M.; Chang, W.; Ezaz-Nikpay, K.; Lane, W. S.; Verdine, G.
L. Biochemistry 1991, 30, 11018.
15. O’Gara, M.; Klimasauskas, S.; Roberts, R. J.; Cheng, X. J. Mol. Biol. 1996, 261, 634.
16. Suzuki, T.; Miyata, N. Curr. Med. Chem. 2006, 13, 935.
17. Kumar, R.; Srivastava, R.; Singh, R. K.; Surolia, A.; Rao, D. N. Bioorg. Med. Chem.
2008, 16, 2276.
Acknowledgments
18. Castellano, S.; Kuck, D.; Sala, M.; Novellino, E.; Lyko, F.; Sbardella, G. J. Med.
Chem. 2008, 51, 2321.
19. Isakovic, L.; Saavedra, O. M.; Llewellyn, D. B.; Claridge, S.; Zhan, L.; Bernstein,
N.; Vaisburg, A.; Elowe, N.; Petschner, A. J.; Rahil, J.; Beaulieu, N.; Gauthier, F.;
MacLeod, A. R.; Delorme, D.; Besterman, J. M.; Wahhab, A. Bioorg. Med. Chem.
Lett. 2009, 19, 2742.
This work was supported in part by a Grant-in-Aid for Scientific
Research from the Japan Society for the Promotion of Science, and a
Grant-in Aid for Research in Nagoya City University.
20. Saavedra, O. M.; Isakovic, L.; Llewellyn, D. B.; Zhan, L.; Bernstein, N.; Claridge,
S.; Raeppel, F.; Vaisburg, A.; Elowe, N.; Petschner, A. J.; Rahil, J.; Beaulieu, N.;
MacLeod, A. R.; Delorme, D.; Besterman, J. M.; Wahhab, A. Bioorg. Med. Chem.
Lett. 2009, 19, 2747.
21. Lubbert, M. Curr. Top. Microbiol. Immunol. 2000, 249, 135.
22. Christman, J. K. Oncogene 2002, 21, 5483.
References and notes
1. Ballestar, E.; Esteller, M. Carcinogenesis 2002, 23, 1103.
2. Herman, J. G.; Baylin, S. B. N. Eng. J. Med. 2004, 350, 947.
3. Karpf, A. R.; Jones, D. A. Oncogene 2002, 21, 5496.
4. Li, E.; Beard, C.; Jaenisch, R. Nature 1993, 366, 362.
23. Egger, G.; Liang, G.; Aparicio, A.; Jones, P. A. Nature 2004, 429, 457.

T. Suzuki et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1124–1127
1127
24. Brueckner, B.; Kuck, D.; Lyko, F. Cancer J. 2007, 13, 17.
25. Siedlecki, P.; Boy, R. G.; Musch, T.; Brueckner, B.; Suhai, S.; Lyko, F.;
Zielenkiewicz, P. J. Med. Chem. 2006, 49, 678.
out at room temperature for 10 min. Thereafter, the stop solution was added to
the strip wells. The resultant mixture was transferred to a 96-well plate and
absorbance at 450 nm was measured.
26. For the in vitro methylation assay, EpiQuik™ DNA Methyltransferase Activity/
Inhibitor Assay Kit (Epigentek Group Inc., Brooklyn, NY, USA) and human DNA
(cytosine-5) methyltransferase (DNMT1) (New England Biolabs, Inc., Bevery,
MA, USA) were used. Test samples were dissolved in DMSO and the solution
was diluted by adding 10 parts of PBS. To the DHMT assay buffer in 0.5 mL
27. Zhou, L.; Cheng, X.; Connolly, B. A.; Dickman, M. J.; Hurd, P. J.; Hornby, D. P. J.
Mol. Biol. 2002, 321, 591.
28. The X-ray structure of M.Hha I (PDB code 1M0E) was used as the target
structure for docking. Protein preparation, receptor grid generation, and ligand
docking were performed using the software Glide 3.5. Compound 5 was docked
into the DNA binding site of M.Hha I. The standard precision mode of Glide was
used to determine favorable binding poses, which allowed the ligand
conformation to be flexibly explored while holding the protein as a rigid
structure during docking. Then, the predicted complex structure was fully
energy-minimized with both the protein and the ligand allowed to move using
tubes were added 1.6 mM S-adenosyl-L-methionine (AdoMet) and inhibitor
solution. Then, DNMT1 was added to the tubes. The reaction mixture was
mixed, transferred to the substrate DNA-coated strip wells, and incubated at
37 °C for 60 min. The reaction solution was discarded, methylation DNA
capture antibody was added to the strip wells, and the mixture was incubated
at room temperature for 60 min. After the capture antibody was aspirated, the
detection antibody was added to the strip wells and incubation was carried out
at room temperature for 30 min. Then, the antibody was removed, the
developing solution was added to the strip wells, and incubation was carried
Macromodel 8.1 software. The conformation of compound
5 in the DNA
binding site was minimized by MM calculation based on the OPLS-AA force
field with the following parameter set: solvent: water; method: LBFGS; max #
of iterations: 10,000; converge: gradient; convergence threshold: 0.05.