Design, synthesis and biological evaluation of 4-Amino-N-(4-aminophenyl) benzamide analogues of quinoline-based SGI-1027 as inhibitors of DNA methylation

Article abstract of DOI:10.1002/cmdc.201300420

Quinoline derivative SGI-1027 (N-(4-(2-amino-6-methylpyrimidin-4-ylamino) phenyl)-4-(quinolin-4-ylamino)benzamide) was first described in 2009 as a potent inhibitor of DNA methyltransferase (DNMT) 1, 3A and 3B. Based on molecular modeling studies, performed using the crystal structure of Haemophilus haemolyticus cytosine-5 DNA methyltransferase (MHhaI C5 DNMT), which suggested that the quinoline and the aminopyridimine moieties of SGI-1027 are important for interaction with the substrates and protein, we designed and synthesized 25 derivatives. Among them, four compounds - namely the derivatives 12, 16, 31 and 32 - exhibited activities comparable to that of the parent compound. Further evaluation revealed that these compounds were more potent against human DNMT3A than against human DNMT1 and induced the re-expression of a reporter gene, controlled by a methylated cytomegalovirus (CMV) promoter, in leukemia KG-1 cells. These compounds possessed cytotoxicity against leukemia KG-1 cells in the micromolar range, comparable with the cytotoxicity of the reference compound, SGI-1027. Structure-activity relationships were elucidated from the results. First, the presence of a methylene or carbonyl group to conjugate the quinoline moiety decreased the activity. Second, the size and nature of the aromatic or heterocycle subsitutents effects inhibition activity: tricyclic moieties, such as acridine, were found to decrease activity, while bicyclic substituents, such as quinoline, were well tolerated. The best combination was found to be a bicyclic substituent on one side of the compound, and a one-ring moiety on the other side. Finally, the orientation of the central amide bond was found to have little effect on the biological activity. This study provides new insights in to the structure-activity relationships of SGI-1027 and its derivative. Ep-(igenet)-ic! Guided by modeling studies, derivatives of the known DNA methyltransferase (DNMT) inhibitor SGI-1027 were designed, synthesized and evaluated. Structure-activity relationships were derived from the results, leading to the identification of derivatives with improved potency and potential for further development.

Full text of DOI:10.1002/cmdc.201300420

CHEMMEDCHEM
FULL PAPERS
DOI: 10.1002/cmdc.201300420
Design, Synthesis and Biological Evaluation of 4-Amino-N-
(4-aminophenyl)benzamide Analogues of Quinoline-Based
SGI-1027 as Inhibitors of DNA Methylation
Elodie Rilova,[a] Alexandre Erdmann,[a] Christina Gros,[a] Vꢀronique Masson,[a]
Yannick Aussagues,[a] Valꢀrie Poughon-Cassabois,[a] Arumugam Rajavelu,[b] Albert Jeltsch,[b]
Yoann Menon,[a] Natacha Novosad,[a] Jean-Marc Gregoire,[a] Stꢀphane Vispꢀ,[a]
Philippe Schambel,[c] Frꢀderic Ausseil,[a] FranÅois Sautel,[a] Paola B. Arimondo,*[a] and
Frꢀdꢀric Cantagrel*[a]
Dedicated to the memory of M. Pierre Fabre (1926–2013), founder of the Pierre Fabre Laboratories and a visionary pharmacist
Quinoline derivative SGI-1027 (N-(4-(2-amino-6-methylpyrimi-
din-4-ylamino)phenyl)-4-(quinolin-4-ylamino)benzamide) was
first described in 2009 as a potent inhibitor of DNA methyl-
transferase (DNMT) 1, 3A and 3B. Based on molecular modeling
studies, performed using the crystal structure of Haemophilus
haemolyticus cytosine-5 DNA methyltransferase (MHhaI C5
DNMT), which suggested that the quinoline and the aminopyr-
idimine moieties of SGI-1027 are important for interaction with
the substrates and protein, we designed and synthesized 25
derivatives. Among them, four compounds—namely the deriv-
atives 12, 16, 31 and 32—exhibited activities comparable to
that of the parent compound. Further evaluation revealed that
these compounds were more potent against human DNMT3A
than against human DNMT1 and induced the re-expression of
a reporter gene, controlled by a methylated cytomegalovirus
(CMV) promoter, in leukemia KG-1 cells. These compounds pos-
sessed cytotoxicity against leukemia KG-1 cells in the micro-
molar range, comparable with the cytotoxicity of the reference
compound, SGI-1027. Structure–activity relationships were elu-
cidated from the results. First, the presence of a methylene or
carbonyl group to conjugate the quinoline moiety decreased
the activity. Second, the size and nature of the aromatic or het-
erocycle subsitutents effects inhibition activity: tricyclic moiet-
ies, such as acridine, were found to decrease activity, while bi-
cyclic substituents, such as quinoline, were well tolerated. The
best combination was found to be a bicyclic substituent on
one side of the compound, and a one-ring moiety on the
other side. Finally, the orientation of the central amide bond
was found to have little effect on the biological activity. This
study provides new insights in to the structure–activity
relationships of SGI-1027 and its derivative.
Introduction
DNA methylation is an epigenetic modification that is involved
in the control of gene expression.[1] It plays an important role
in tumorigenesis and tumor maintenance.[2] Concomitantly to
a global hypomethylation contributing to genetic instability,
promoter hypermethylation of tumor suppressor genes is com-
monly observed in cancers, such that it can be used as a bio-
marker.[3] As with the other epigenetic marks, DNA methylation
is reversible and constitutes a promising target in cancer treat-
ment. DNA methylation is catalyzed by DNA methyltransferas-
es (DNMT), which transfer the methyl group from the cofactor,
S-adenosyl-l-methionine (AdoMet), to the position 5 of the de-
oxycytidine in a CpG context.[4] Two nucleosides analogues of
cytidine, 5-azacytidine 2 (5aza) and 5-azadeoxycytidine 3
(5azadC), have been approved by the US Food and Drug Ad-
ministration (FDA) for the treatment of myelodysplastic syn-
dromes (MDS) and acute myeloide leukemia (AML).[5]
[a] E. Rilova, A. Erdmann, C. Gros, V. Masson, Y. Aussagues,
V. Poughon-Cassabois, Y. Menon, N. Novosad, Dr. J.-M. Gregoire,
Dr. S. Vispꢀ, Dr. F. Ausseil, F. Sautel, Dr. P. B. Arimondo, Dr. F. Cantagrel
USR CNRS-Pierre Fabre No. 3388 ETaC
Centre de Recherche et de Dꢀveloppement Pierre Fabre (CRDPF)
3 Ave Hubert Curien, 31035 Toulouse Cedex 01 (France)
E-mail: paola.arimondo@etac.cnrs.fr
frederic.cantagrel@etac.cnrs.fr
[b] A. Rajavelu, Prof. A. Jeltsch
Institute of Biochemistry, Faculty of Chemistry, University Stuttgart
Pfaffenwaldring 55, 70569 Stuttgart (Germany)
[c] P. Schambel
Institut de Recherches Pierre Fabre
17 Rue Jean Moulin, 81106 Castres Cedex (France)
Beside nucleoside inhibitors, many efforts are made to de-
velop new non-nucleoside inhibitors to overcome the limits of
these azanucleosides, such as chemical instability and incorpo-
ration into DNA for activity.[6] In 2009, Datta et al. published
a quinolone derivative, SGI-1027 (1), that depletes DNMT1.[7] In-
terestingly, its bisquaternary salt was first described for its anti-
leukemic activity.[8] Inhibitory activity of SGI-1027 was de-
scribed against DNMT1, 3A and 3B[7]. The compound was
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201300420.
ꢁ 2014 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA.
This is an open access article under the terms of the Creative Commons
Attribution Non-Commercial NoDerivs License, which permits use and
distribution in any medium, provided the original work is properly cited,
the use is non-commercial and no modifications or adaptations are
made.
ꢁ 2014 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 590 – 601 590

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
shown to degrade the enzymes, inhibit DNA methylation in
colon cancer cell lines, and reactivate P16, MLH1 and TIMP3
genes.
accomodates the amino acid of AdoMet, and the other accom-
odates the cytidine of the DNA that is flipped out of the DNA
double helix into the catalytic pocket of the enzyme. Our dock-
ing studies of SGI-1027 (1) in MHhaI C5 DNMT (Figure 1)
showed that the compound fits within the adenine binding
pocket of the cofactor through the aminopyrimidine group
(part C of SGI-1027) and within the cytidine binding pocket
through the quinoline moiety (part A of SGI-1027). In our
model, the orientation of the molecule seems to forbid any in-
teraction with the amino acid binding pocket.
Here, we describe the conception of new derivatives of SGI-
1027 guided by a molecular modeling study. A total of 25 de-
rivatives were synthetized and screened for their ability to in-
hibit the catalytic domain of human DNMT3A. Selectivity
against several methyltransferases was assessed for the most
potent inhibitors, as was their ability to reactivate gene expres-
sion in an epigenetic reporter system in a leukemia cell line.
Results and Discussion
Docking
To design new analogues of SGI-
1027 and test their activity
against the catalytic domain of
hDNMT3A, we started by carry-
ing out a docking study of SGI-
1027 in the catalytic pocket of
DNMTs. Recently, the crystal
structure of the murine catalytic
complex
Dnmt3A/3L
(PDB:
2QRV[9]) and several crystal struc-
tures of DNMT1 have been pub-
lished (PDB: 3PTA,[10] 3OS5,
4DA4[11] and 3PT6[10]), together
with molecular docking and
pharmacophore modelling stud-
ies based on these struc-
tures.[12–14]
Concerning
the
DNMT1 structures, we chose not
to use them since the N-terminal
domain of the C5 DNA methyl-
transferases is not well con-
served and, in particular, DNMT1
Figure 1. SGI-1027 molecular docking in Haemophilus haemolyticus cytosine-5 DNA methyltransferase (MHhaI C5
DNMT; PDB: 2HR1[16]). The co-crystalized S-adenosyl-l-homocysteine (AdoHcy) product of the methylation reaction
is depicted in yellow and the cytidine substrate in green. The docked SGI-1027 is in cyan. The three parts of the
molecule (A, B and C) explored here are encircled in colors in the zoom of SGI-1027.
contains an autoinhibition linker that is lacking in the DNMT3
isoforms[10,11, 15] confering very specific properties to the inter-
action with the substrates and affecting inhibition, as observed
for SGI-1027.[13,14] Concerning the murine Dnmt3A catalytic
domain co-crystallized with C-terminal Dnmt3L (PDB: 2QRV[9]),
the substrate cytosine is not resolved in the crystal structure,
only the cofactor (here the product S-adenosyl-l-homocys-
teine, AdoHcy). Interestingly, the DNMT3A catalytic pocket is
well superimposable with that of the crystal structure of the
bacterial Haemophilus haemolyticus cytosine-5 DNA methyl-
transferase (MHhaI C5 DNMT; PDB: 2HR1),[16] in particular for
the AdoHcy molecule (shown in Figure S1 in the Supporting
Information). We chose to conduct our docking studies on bac-
terial MHhaI C5 DNMT, since the catalytic pocket is well con-
served among the C5 DNMTs and in the crystal structure of
MHhaI C5 DNMT, both the co-factor (here the product AdoHcy)
and the DNA substrate (deoxycytidine) are well resolved. Sche-
matically, the catalytic pocket of the C5 DNA methyltransferas-
es can be considered formed of three binding pockets (Figure
1): one pocket accomodates the adenine of AdoMet, another
Noteworthy, the finding that part C fits into the adenine
pocket of the cofactor is in agreement with the results ob-
tained by induced-fit docking of SGI-1027 in the MTase domain
of mDnmt3A,[14] even though part A is differently positioned
compared with our docking result. In particular, the pyrimidine
moiety (C) of SGI-1027 was found well superimposable to the
six-membered ring of the adenine of AdoHcy (Figure 1, right-
hand zoom). In addition, the amino group on the C4 of the ad-
enine of AdoHcy formed hydrogen bonds with the carboxylic
residue of Asp60 (corresponding to Asp682 in DNMT3A), and
in parallel, the amino group of the pyrimidyl moiety of the in-
hibitor interacted with the amide residue of Asn39 (corre-
sponding to Ser659 in DNMT3A). On the contrary, little interac-
tions with the quinoline moiety A were observed in the cyto-
sine pocket, except for some superimposition with the ribose
of the cytidine. This can explain the differences observed for
part A between our model and the induced-fit docking model
of SGI-1027 in mDnmt3A described by Yoo et al.,[14] who used
a crystal structure lacking the cytidine substrate in the resolu-
tion. Still, we found an interaction of part A with the enzyme,
ꢁ 2014 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 590 – 601 591

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
with Arg165 (corresponding to Arg788 in DNMT3A) (Figure 1,
left-hand zoom). Based on these findings, we designed new
derivatives of SGI-1027 to explore the substrate binding pock-
ets guided by our docking model. More specifically, since few
interactions were modelized for the quinoline moiety (part A,
Figure 1), we began by modifying this part of the molecule
aiming at improving the affinity of the compounds for the
deoxycytidine binding pocket of DNMTs.
Exploring the deoxycytidine pocket
According to the above molecular docking studies, the quino-
line moiety (Figure 1, ring A) is superimposable on the flipped-
out deoxycytidine but lacks strong interactions in the
pocket. Thus, its modifications can improve the affinity for the
cytidine binding pocket. In a first series of compounds, this
substituent was replaced with aromatic and heteroaromatic
rings. The synthesis of compound 9 followed a published
route (Scheme 1).[17] The desired derivatives were then ob-
tained by direct reaction with miscellaneous aryl chlorides in
the presence of catalytic amounts of acid (Scheme 2).
Scheme 2. Synthesis of aryl derivatives. Reagents and conditions: a) ArCl, HCl
(cat.), MeOH/H2O/EtOH (1:1:1), reflux, 22 h.
pyrimidine and methylated aminopyrimidine, respectively,
were slightly less active. Dimethoxy triazine 13 lost most of
the inhibitory activity. In agreement with our model, the re-
placement of the quinoline moiety by bicyclic or tricyclic moi-
eties was in general detrimental on the inhibition activity com-
pared with the monocyclic substitution (14–18; Figure 2). In
fact, the introduction of an adenine as ring A induced a strong
decrease in the activity (analogue 17), comforting the hypoth-
esis that ring A resides in the cytidine pocket and not in the
AdoMet pocket. In contrast, the addition of a phenyl group on
the quinoline, analogue 16, kept the activity at a good level,
suggesting the presence of a hydrophobic pocket that might
be worth exploring in future optimizations.
Scheme 1. Synthesis of compound 9. Reagents and conditions: a) HCl, ethox-
yethanol, reflux, 3 h; b) Fe, AcOH, EtOH/H2O (2:1), reflux, 15 h; c) 4-nitroben-
zoyl chloride, Et3N, dioxane, 608C, 1 h.
In a second series of derivatives, we studied the influence of
the distance between rings A and B on the activity of the com-
pounds. Preparation of compounds comprising an intercalated
carbonyl group was easily achieved by acylation, albeit with
only moderate yields (19–22; Scheme 3). Not only was the
linker length changed in these compounds but also p-staking
interactions and hydrogen bond counts were modified.
Yields were generally good, except upon use of chloropur-
ine, because of its poor reactivity and the poor solubility of the
final product (17) giving rise to material loss during purifica-
tion. 4-Chloro-7-fluoroquinazoline—used to obtain 15—was
synthesized by reaction of N,N-dimethylformamide with 4-
fluoro-2-aminobenzoic acid followed by intramolecular cycliza-
tion and chloration with thionyl chloride.[18]
Next, methylene groups were introduced. In the absence of
a
reaction with the corresponding chloromethylpyridines,
methylene group introduction was performed by reductive
amination (derivatives 23 and 24; Scheme 4). Preparation of
pyridine-3- or -4-carboxaldehyde from the corresponding alco-
hols could be done either with pyridinium chromate or man-
ganese dioxide, but the latter was preferred because traces of
oxidants led to formation of unexpected compound 25 that
was used as a control (presumably via formaldehyde forma-
tion). All these modifications led to complete loss of activity
(compounds 19–25; Figure 2 and Table 1).
The compounds were assayed for their ability to inhibit the
catalytic activity of human DNMT3A at 32 and 10 mm (Table 1),
with parent compound SGI-1027 as a reference. Compound
10, where a pyridine replaces the quinoline of SGI-1027, inhib-
ited 91% of the enzyme activity at 32 mm, but this inhibition
dramatically dropped at 10 mm (EC50 value of 13 mmꢀ1 was
calculated). Compounds 11 and 12, bearing an unsubstituted
ꢁ 2014 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 590 – 601 592

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
Table 1. Screening results for all compounds against DNMT3A catalytic activity.
Compd
Inhibition [%][a]
Compd
Inhibition [%][a]
R
R’
32 mm
10 mm
R
R’
32 mm
10 mm
SGI-1027 (1)
A
76ꢀ5
1[b]
74ꢀ5
20
21
22
A
0[b]
0
8
9
A
A
n.d.
n.d.
A
A
0[b]
0
0[b]
8
29
10
11
A
A
91ꢀ4
65ꢀ3
32
17
23
24
A
A
0[b]
0
0
0[b]
12
13
A
A
69ꢀ1
82ꢀ1
25
30
A
B
4
0
35
36
47ꢀ1
56ꢀ11
14
15
A
A
19
6
25
19
31
32
B
B
75ꢀ8
73ꢀ3
53ꢀ8
53ꢀ13
16
17
A
A
55ꢀ10
81ꢀ5
33
36
B
C
46
29
18
3[b]
8
54ꢀ5
18
19
A
A
53ꢀ2
11
5
37
38
B
29
39
0[b]
D
62ꢀ22
48ꢀ8
[a] When inhibition was >50%, experiments were performed in triplicate/duplicate, and these data are reported as the meanꢀSEM. All other values are
single determinations. Not determined (n.d.). [b] Inhibition was measured at 20 mm.
ꢁ 2014 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 590 – 601 593

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
Exploring the AdoMet pocket
A third series of products was
devised to investigate the im-
portance of the aminopyrimidine
moiety (ring C in Figure 1). The
synthetic pathway was similar to
the one described above, but
used 4-chloroquinoline instead
of the chloropyrimidine as the
starting materials.
After coupling with 4-nitro-
aniline, the resulting nitro com-
pound (26) was reduced to
amino derivative 27 with iron in
acetic acid. Amidation with 4-
nitro-benzoyl chloride yielded
nitro compound 28, which was
similarly reduced to amine 29.
As before, compounds 30–33
were prepared by reaction of
the corresponding aryl chloride
in the presence of a catalytic
amount of hydrochloride acid
(Scheme 5). Compound 30 is an
Figure 2. Inhibition of DNMT3A by parent compound SGI-1027 (&) and synthesized compounds exploring the
deoxycytidine pocket (&), with varying tether length between the quinoline A and part B (&), addressing the
AdoMet pocket (&), and with different moiety sizes addressing the two substrates pockets (&). Percent inhibition
was determined at a concentration of 10 mm.
isomer of SGI-1027 with inversion of the central amide bond of
part B. This compound was recently described by Gamage
et al. and showed lower efficiency on DNMT1 degradation
than SGI-1027.[19] In agreement with these previous findings on
DNMT1,[17,19] methyl-amino pyrimidine 30 was found to be less
potent than SGI-1027 against DNMT3A, and its level of activity
was comparable to that of diaminopyrimidine 32 (Figure 2 and
Table 1). Unexpectedly, compound 32 was difficult to analyze
because it did not present a dose–reponse curve but rather
a plateau at 50% inhibition between 32 and 1 mm (data not
shown). This could be due to chemical-physical features of the
Scheme 3. Synthesis of acyl derivatives. Reagents and conditions: a) Ar-C(O)-
Cl, pyr., 08C!RT, o/n.
Scheme 5. Synthesis of quinoleine derivatives. Reagents and conditions: a) 4-
nitroaniline, HCl (cat), ethoxyethanol, reflux, 3 h; b) Fe, AcOH, EtOH/H2O
(2:1), reflux, 15 h; c) 4-nitrobenzoyl chloride, Et3N, dioxane, 608C, 1 h;
d) ArCl, HCl (cat), EtOH/H2O/MeOH (1:1:1), reflux, o/n.
Scheme 4. Synthesis of methylene derivatives. Reagents and conditions:
a) PCC, CH2Cl2, RT, 90 min; b) 1. 9, MeOH, Et3N, RT, 3.5 h; 2. AcOH, NaBH3CN,
RT, o/n; c) MnO2, CH2Cl2, RT, o/n.
ꢁ 2014 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 590 – 601 594

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
compound that we did not investigate further. Interestingly,
compound 31, containing the quinoline in C and a simplified
pyridine cycle in A, was the most potent compound at both 32
and 10 mm, with similar potency than SGI-1027 (Table 1).
Probing the size of cofactor and substrate pockets
Finally, in order to investigate the size of the cofactor and sub-
strate pockets, we introduced variations to parts A and C, ex-
ploring the effect of terminating the molecule with a single
ring on both ends (product 36), with a bicycle (derivatives 33
and 37) or a tricyclic system (Scheme 5).
Symmetrical compound 36, terminated with a pyridine
moiety, was synthesized according to a simplified synthetic
scheme. The central amide part B was easily prepared by reac-
tion between 4-nitrobenzoyl chloride and nitroaniline, followed
by simultaneous reduction of the nitro groups to afford com-
pound 35 (Scheme 6). A double arylation with 4-chloropyridine
led to 36 in three steps. This simplified pathway offered bis-
quinoline 37 (2+2 combination) and bis-acridine 38 (3+3
combination) derivatives.
Figure 3. hDNMT3A inhibition curves for SGI-1027 (&), and compounds 12
(~), 16 (!), 31 (^) and 32 (*). Data are the mean of triplicate determina-
tions.
The compound gave an EC50 value of 15ꢀ3 mm against
hDNMT1, comparable to SGI-1027 with an EC50 value of
10ꢀ1 mm, and a weak inhibition activity on G9A with an IC50
value of 28ꢀ4 mm, comparable to that of SGI-1027 (IC50 =59ꢀ
4 mm).
The compounds were tested for their ability to reactivate
gene expression by use of an integrated luciferase reporter
system under the control of a methylated cytomegalovirus
(CMV) promoter in leukemia KG-1 cells (CMV-luc assay reported
in Table 2), which inhibits its expression. In parallel, the antipro-
liferative activity of the compounds was tested in KG-1 leuke-
mia cells. In particular, the cytotoxicity in the KG-1 cells of the
compounds can interfere with the CMV-luc assay.
The fold induction of the luciferase signal was measured in
cells incubated for 24 hours in the presence of the test com-
pounds at the indicated concentrations and normalized to the
untreated control cells. Negative control compound 19, which
did not inhibit the enzyme, was also unable to reactivate luci-
ferase expression. Parent compound SGI-1027 gave a 16-fold
induction at 5 mm, but at 10 mm, a decrease in the luciferase
signal was observed because of its cytotoxicity in the KG-1 cell
line (EC50 =4.4 mm; Table 2). The most potent inhibitor of our
series, compound 31, reactivated luciferase expression starting
from 1 mm; however, cytotoxicity in the KG-1 cell line (EC50 =
1 mm; Table 2) hindered the measurement of the luciferase
signal at higher concentrations. Active bismonocyclic 12 gave
an 8.6-fold induction at 10 mm, a dose that did not affect cell
proliferation (Table 2). Compound 16 presented a similar activi-
ty profile as compound 31. Even if the cytotoxicity clearly in-
terefered with the measurement of luciferase induction, we
can conclude that all inhibitors induced re-expression of the
luciferase gene.
Scheme 6. Synthesis of symmetrical compounds. Reagents and conditions:
a) Et3N, dioxane, 608C, 1 h; b) Fe, AcOH, EtOH/H2O (2:1), reflux, 15 h; c) ArCl,
HCl (cat), EtOH/H2O/MeOH (1:1:1), reflux, o/n.
The inhibitory activity against DNMT3A at 32 and 10 mm con-
firmed that the best inhibitors remain derivative 31 terminated
by a single ring on one side and a two-ring system on the
other (Figure 2 and Table 1).
Biological activities
The most active compounds against DNMT3A, 12, 16 and 31
and inactive compound 19 (Figure 2) were further assayed to
determine the concentration at which 50% of efficacy of inhib-
ition is obtained (EC50) (Figure 3 and Table 2). We did not fur-
ther investigate active compound 32 because of its behavior
in the dose–reponse experiment discussed above. Compound
31, the most potent and efficient on hDNMT3A inhibition
(EC50 =0.9 mm, efficacy 90%), was tested on DNMT1 and on his-
tone lysine methyltransferases G9a inhibition for selectivity.
Conclusions
Three new derivatives of SGI-1027, compounds 12, 16 and 31,
showed similar DNMT inhibition activity and gene expression
induction as the parent compound, associated with prolifera-
tion inhibition on KG-1 cells in the micromolar range. Struc-
ture–activity relationship studies carried out on three parts of
ꢁ 2014 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 590 – 601 595

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
Table 2. Biochemical and biological activity of the most active derivatives. Inhibition (EC50) of human DNMT3A catalytic activity, cytotoxicity (EC50) against
leukemia KG-1 cells, and reactivation of luciferase expression of the CMV-luc construct in KG-1 cells.
Compd
EC50 [mm]
Fold reactivation[c]
hDNMT3A[a]
KG-1[b]
10 mm
5 mm
3.2 mm
1 mm
0.5 mm
0.32 mm
0.1 mm
SGI-1027 (1)
0.9 (0.7–1.0)
5.0 (1.9–13)
1.6 (1.0–2.5)
inactive
4.4 (3.2–6.2)
>10
7.6ꢀ1.3
8.6ꢀ0.9
6.1ꢀ2.7
1.0ꢀ0.1
3.2ꢀ0.5
15.7ꢀ0.2
2.6ꢀ0.1
6.0ꢀ2.0
1.0ꢀ0.1
3.9ꢀ0.1
7.1ꢀ1.2
1.8ꢀ0.1
1.4ꢀ0.1
1.2ꢀ0.1
1.1ꢀ0.1
1.1ꢀ0.1
1.2.ꢀ0.1
1.0ꢀ0.1
1.6ꢀ0.3
1.0ꢀ0.1
1.1ꢀ0.1
n.d.
1.0ꢀ0.1
1.3ꢀ0.2
0.9ꢀ0.1
1.0ꢀ0.1
1.0ꢀ0.1
1.0ꢀ0.1
1.1ꢀ0.1
12
16
19
31
1.3 (1.0–3.2)
1.3 (0.9–1.7)
0.96 (0.9–1.0)
5.5ꢀ1.3
1.2ꢀ0.1
3.9ꢀ0.3
1.3ꢀ0.1
1.0ꢀ0.1
2.0ꢀ0.2
0.9 (0.7–1.0)
[a] Concentration at which 50% inhibition of enzyme activity was observed; data represent the mean of 3–5 experiments; 95% confidence intervals are
given in parentheses. [b] Concentration at which 50% inhibition of cellular proliferation was observed; data represent the mean of 3–5 experiments; 95%
confidence intervals are given in parentheses. [c] Fold-induction (reactivation) of the luciferase gene controlled by methylated CMV compared with
untreated cells after 24 h of treatment. Data represent the meanꢀSEM of 2–3 experiments.
N4-(4-Nitrophenyl)-6-methylpyrimidine-2,4-diamine (6): 4-Nitroa-
the molecule indicated that the presence of a methylene or
niline 4 (6.00 g, 43.44 mmol) and 2-amino-4-chloro-6-methylpyrimi-
dine 5 (6.24 g, 43.44 mmol) were dissolved in 2-ethoxyethanol
(330 mL). A few drops of HCl (35%) were added, and the solution
was heated at reflux for 3 h, and then allowed to cool to RT over-
night. The reaction mixture was filtered, and the resulting solid
was rinsed with Et2O and dried in vacuo to afford compound 6 as
a pale yellow solid (8.53 g, 80%): 1H NMR (500 MHz, CD3OD): d=
2.22 (s, 3H), 6.04 (d, J=0.6 Hz, 1H), 7.99 (d, J=9.2 Hz, 2H),
8.18 ppm (d, J=9.4 Hz, 2H); 13C NMR (126 MHz, CD3OD): d=23.3,
97.8, 119.6, 125.8, 148.4, 162.9, 167.3 ppm; MS (ESI): m/z=246
[M+H]+.
carbonyl group as a linker to the quinoline moiety dramatically
decreased activity. Second, the effect of the size and nature of
the aromatic or heterocycle substituents at the ends of the
molecule played an important effect on inhibition: three-ring
structures decreased the activity, while two-ring substituents
were well tolerated, the best being the combination of a two-
ring on one side and a one-ring system on the other. Finally
the orientation of the central amide bond was found to have
little effect on the activity. These findings bring new insights
into the substituents that contribute to the DNMT inhibition
activity of this family of compounds for further design of new
more potent inhibitors.
N4-(4-Aminophenyl)-6-methylpyrimidine-2,4-diamine (7):
A re-
fluxing suspension of amine 6 (7.08 g, 28.9 mmol) in EtOH/H2O
(86 mL/43 mL) was treated with Fe powder (6.44 g, 115.5 mmol)
and AcOH (2.6 mL). The resulting dark brown suspension was
heated at reflux for 15 h, after which time, the hot mixture was fil-
tered through a pad of Celite. The filtrate was concentrated in va-
cuo and purified by flash chromatography (CH2Cl2/MeOH, 100:0!
80:20) to afford amine 7 as a brown solid (5.71 g, 92%): 1H NMR
(500 MHz, CD3OD): d=2.13 (s, 3H), 5.82 (s, 1H), 6.73 (d, J=8.6 Hz,
2H), 7.17 ppm (d, J=8.6 Hz, 2H); MS (ESI): m/z=216 [M+H]+.
Experimental Section
Chemistry
Reagents, chemicals and anhydrous solvents were purchased from
commercial suppliers and used without purification. Sensitive reac-
tions were performed under an argon atmosphere. NMR spectra
N-[4-(2-Amino-6-methylpyrimidin-4-ylamino)phenyl]-4-nitroben-
zamide (8): 4-Nitrobenzoyl chloride (1.21 g, 6.5 mmol) and Et3N
(0.99 g, 9.8 mmol) were sequentially added to a solution of amine
7 (1.40 g, 6.5 mmol) in dry dioxane (90 mL). The solution was
heated at 608C for 1 h, and the reaction mixture was then allowed
to cool to RT. Solvent was removed in vacuo, and the residue was
purified by flash chromatography on silica gel (CH2Cl2/MeOH,
were recorded on a Bruker Avance II spectrometer equipped with
1
a
13C cryoprobe at 500 MHz for H and 126 MHz for 13C; 2D experi-
ments were performed using standard Bruker programs. High-reso-
lution mass spectrometry (HRMS) was performed on a Bruker Mi-
croTOF spectrometer using electrospray ionization (ESI). Thin-layer
chromatography (TLC) was carried out on precoated silica gel
60F254 TLC plates (Merck), and spots were visualized by heating
after diping in KMnO4 solution. Column chromatography was car-
ried out on a Puriflash 430 apparatus (Interchim) equipped with
30 mm porated silica column.
100:0!70:30) to give compound
8
(1.63 g, 70%): 1H NMR
(500 MHz, [D6]DMSO): d=2.08 (s, 3H), 5.86 (s, 1H), 6.13 (s, 2H),
7.66–7.73 (m, 4H), 8.18 (d, J=8.9 Hz, 2H), 8.37 (d, J=8.9 Hz, 2H),
8.99 (s, 1H), 10.47 ppm (s, 1H); 13C NMR (126 MHz, [D6]DMSO): d=
23.5, 94.7, 119.5, 121.0, 123.6, 129.2, 132.5, 132.5, 137.3, 140.8,
149.1, 161.3, 162.9, 163.4, 165.0 ppm; MS (ESI): m/z=365 [M+H]+.
Analytical HPLC was performed on a VWR-Hitachi apparatus ELITE
LACHROM. A prepacked C18 reverse-phase column (Waters X-
Bridge RP-18, 4.6ꢂ100 mm, 5 mm) was used for analytical HPLC
with a binary gradient elution (solvent A: H2O; solvent B: MeCN)
and a flow rate of 1mLminꢁ1. Preparative HPLC was performed on
an apparatus equipped with a VWR International LaPrep pump
P110, a VWR LaPrep P314 dual l absorbance detector, and EZ-
Chrom software. A Waters Xbridge RP-18 column (19ꢂ250 mm,
10 mm) was used for preparative HPLC with a binary gradient elu-
tion (solvent A: H2O; solvent B: CH3CN) and a flow rate of
25 mLminꢁ1, and the UV absorbance was monitored at 250 and
320 nm.
N-[4-(2-Amino-6-methylpyrimidin-4-ylamino)phenyl]-4-amino-
benzamide (9): A refluxing suspension of 8 (1.07 g, 2.9 mmol) in
EtOH/H2O (13 mL/6.5 mL) was treated with Fe powder (654 mg,
11.7 mmol) and AcOH (1.3 mL). The resulting dark brown suspen-
sion was heated at reflux for 15 h, after which time, the resulting
mixture was filtered through a pad of silica using 7n NH3 in
MeOH. Solvent was removed in vacuo and purification by flash
chromatography (CH2Cl2/MeOH, 100:0!70:30) afforded amine 9 as
a brown solid (638 mg, 65%): 1H NMR (500 MHz, [D6]DMSO): d=
2.07 (s, 3H), 5.73 (s, 2H), 5.84 (s, 1H), 6.09 (s, 2H), 6.59 (d, J=
ꢁ 2014 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 590 – 601 596

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
8.6 Hz, 2H), 7.59–7.63 (m, 4H), 7.69 (d, J=8.6 Hz, 2H), 8.89 (s, 1H),
9.67 ppm (s, 1H); 13C NMR (126 MHz, [D6]DMSO): d=23.5, 94.7,
112.7, 119.9, 120.8, 121.4, 129.4, 133.8, 136.1, 152.0, 161.5, 163.2,
165.1 ppm; MS (ESI): m/z=335 [M+H]+.
120.8, 128.4, 129.0, 133.1, 136.6, 141.9, 161.4, 162.9, 164.6, 164.9,
166.0, 172.0 ppm; HRMS (ESI): m/z [M+H]+ calcd for C23H24N9O3:
474.2002, found: 474.1996.
N-[4-(2-Amino-6-methylpyrimidin-4-ylamino)phenyl]-4-(quinazo-
lin-4-ylamino)benzamide (14): Compound 14 was synthesized ac-
cording to the procedure described for 10, using amine 9 (30 mg,
0.09 mmol) and 4-chloroquinazoline (15 mg, 0.09 mmol). Purifica-
tion by preparative HPLC (H2O/CH3CN, 90:10!0:100) gave com-
pound 14 as a white solid (25 mg, 73%): 1H NMR (500 MHz,
[D6]DMSO): d=2.08 (s, 3H), 5.85 (s, 1H), 6.13 (br s, 1H), 7.67–7.72
(m, 4H), 7.83 (d, J=8.1 Hz, 1H), 7.90 (td, J=7.1, 1.3 Hz, 1H), 8.00
(d, J=8.7 Hz, 2H), 8.08 (d, J=8.7 Hz, 2H), 8.61 (d, J=8.4 Hz, 1H),
8.69 (s, 1H), 8.96 (s, 1H), 10.03 (br s, 1H), 10.10 ppm (s, 1H);
13C NMR (126 MHz, [D6]DMSO): d=23.6, 94.6, 115.4, 119.7, 120.9,
121.2, 123.1, 126.6, 128.0, 128.2, 129.6, 133.2, 133.3, 136.7, 142.4,
149.8, 154.4, 157.6, 161.4, 163.0, 164.7, 165.0 ppm; HRMS (ESI): m/z
[M+H]+ calcd for C26H23N8O: 463.1995, found: 463.1970.
N-[4-(2-Amino-6-methylpyrimidin-4-ylamino)phenyl]-4-(pyridin-
4-ylamino)benzamide (10): A refluxing suspension of amine 9
(100 mg, 0.3 mmol) in MeOH/H2O/EtOH (2 mL/2 mL/2 mL) was
treated with chloropyridine hydrochloride (34 mg, 0.3 mmol) and
one drop of HCl (35%). The resulting yellow solution was heated
at reflux for 22 h, and then was cooled to RT and concentrated in
vacuo. Purification by flash chromatography on silica gel (CH2Cl2/
MeOH, 100:0!70:30) gave compound 10 as a yellow solid (98 mg,
1
80%): H NMR (500 MHz, [D6]DMSO): d=2.28 (s, 3H), 6.18 (s, 1H),
7.29 (d, J=6.6 Hz, 2H), 7.50 (d, J=8.7 Hz, 2H), 7.78–7.83 (m, 4H),
8.10 (d, J=8.6 Hz, 2H), 8.36 (d, J=7.4 Hz, 2H), 10.40 (s, 1H), 10.68
(s, 1H), 11.01 ppm (s, 1H); 13C NMR (126 MHz, [D6]DMSO): d=18.4,
97.4, 109.5, 120.9, 121.7, 122.0, 129.5, 131.8, 134.0, 135.8, 140.4,
141.0, 152.3, 155.6, 155.7, 161.4, 164.6 ppm; HRMS (ESI): m/z
[M+H]+ calcd for C23H22N7O: 412.1886, found: 412.1873.
N-[4-(2-Amino-6-methylpyrimidin-4-ylamino)phenyl]-4-(7-fluoro-
quinazolin-4-ylamino)benzamide (15): Compound 15 was synthe-
sized according to the procedure described for 10, using amine 9
(50 mg, 0.15 mmol) and 4-chloro-7-fluoroquinazoline (25 mg,
0.15 mmol). Purification by preparative HPLC (H2O/CH3CN, 90:10!
N-[4-(2-Amino-6-methylpyrimidin-4-ylamino)-phenyl]-4-(pyrimi-
din-2-ylamino)benzamide (11): Compound 11 was synthesized ac-
cording to the procedure described for 10, using amine 9 (40 mg
(0.12 mmol) and 2-chloropyrimidine (14 mg, 0.12 mmol). Purifica-
tion by preparative HPLC (H2O/CH3CN, 90:10!0:100) gave com-
pound 11 as a yellow solid (20 mg, 41%): 1H NMR (500 MHz,
[D6]DMSO): d=2.08 (s, 3H), 5.85 (s, 1H), 6.11 (br s, 2H), 6.93 (t, J=
4.8 Hz, 1H), 7.65 (s, 4H), 7.91 (s, 4H), 8.55 (d, J=4.7 Hz, 2H), 8.94
(s, 1H), 9.98 ppm (d, J=9.3 Hz, 2H); 13C NMR (126 MHz, [D6]DMSO):
d=23.9, 95.0, 113.6, 118.0, 120.1, 121.3, 127.7, 128.8, 133.7, 136.9,
143.9, 158.6, 160.1, 161.8, 163.3, 165.2, 165.3 ppm; HRMS (ESI): m/z
[M+H]+ calcd for C22H21N8O: 413.1838, found: 413.1825.
1
0:100) gave compound 15 as a white solid (12 mg, 18%): H NMR
(500 MHz, [D6]DMSO): d=2.08 (s, 3H), 5.85 (s, 1H), 6.13 (s, 2H),
7.58–7.62 (m, 2H), 7.67 (s, 4H), 8.03 (d, J=15.5 Hz, 4H), 8.68–8.72
(m, 2H), 8.96 (s, 1H), 10.11 ppm (s, 1H); 13C NMR (126 MHz,
[D6]DMSO): d=23.6, 94.6, 111.9 (d, JC–F =20.3 Hz), 112.5, 116.1 (d, JC–
F =24.4 Hz), 119.7, 120.9, 121.3, 126.6 (d, JC–F =10.6 Hz), 128.2,
129.8, 133.2, 136.7, 142.1, 151.9 (d, JC–F =13.5 Hz), 155.6, 157.5,
161.4, 162.9, 164.7, 164.9, 164.8 ppm (d, JC–F =251.8 Hz); HRMS
(ESI): m/z [M+H]+ calcd for C26H22FN8O: 481.1901, found:
481.1884.
N-[4-(2-Amino-6-methylpyrimidin-4-ylamino)phenyl]-N-4-(2-
amino-6-methylpyrimidin-4-ylamino)benzamide (12): Compound
12 was synthesized according to the procedure described for 10,
using amine 9 (40 mg, 0.12 mmol) and 2-amino-4-chloro-6-methyl-
pyrimidine (17 mg, 0.12 mmol). Concentration in vacuo gave com-
pound 12 as a yellow amorphous solid (40 mg, 77%): 1H NMR
(500 MHz, [D6]DMSO): d= 2.08 (s, 3H), 2.11 (s, 3H), 5.87 (s, 1H),
5.98 (s, 1H), 6.10 (brs, 2H), 6.26 (brs, 2H), 7.66 (s, 4H), 7.90 (s, 4H),
9.00 (brs, 1H), 9.46 (brs, 1H), 10.00 ppm (brs, 1 H); 13C NMR
(126 MHz, [D6]DMSO): d=23.5, 23.6, 94.6, 95.5, 117.9, 119.6, 120.8,
126.9, 128.4, 133.3, 136.5, 144.1, 161.2, 161.4, 162.9, 162.9, 164.7,
164.9, 165.5 ppm. Filtration of a DMSO solution of 12 through
a pad of K2CO3 gave the free base of 12 as a pale yellow solid
N-[4-(2-Amino-6-methylpyrimidin-4-ylamino)phenyl]-4-(2-phen-
ylquinolin-4-ylamino)benzamide (16): Compound 16 was synthe-
sized according to the procedure described for 10, using amine 9
(30 mg, 0.09 mmol) and 4-chloro-2-phenylquinoline (22 mg,
0.09 mmol). Purification by preparative HPLC (H2O/CH3CN, 90:10!
1
0:100) gave compound 16 as a white solid (16 mg, 33%): H NMR
(500 MHz, [D6]DMSO): d=2.08 (s, 3H), 5.85 (s, 1H), 6.13 (br s, 2H),
7.45–7.48 (m, 1H), 7.50–7.53 (m, 2H), 7.56–7.60 (m, 3H), 7.65–7.69
(m, 4H), 7.72 (s, 1H), 7.69 (td, J=6.9, 1.3 Hz, 1H), 8.00 (dd, J=8.4,
1.0 Hz, 1H), 8.04 (d, J=8.7 Hz, 2H), 8.09–8.11 (m, 2H), 8.39 (d, J=
8.4 Hz, 1H), 8.95 (s, 1H), 9.34 (s, 1H), 10.10 ppm (s, 1H); 13C NMR
(126 MHz, [D6]DMSO): d=24.0, 95.0, 101.2, 120.0, 120.1, 120.1,
121.2, 122.6, 125.4, 127.5, 129.1, 129.2, 129.7, 129.7, 130.0, 130.4,
133.6, 137.0, 139.9, 147.7, 149.5, 157.1, 161.8, 163.4, 164.9,
165.3 ppm; HRMS (ESI): m/z [M+H]+ calcd for C33H28N7O:
538.2355, found: 538.2348.
1
quantitatively: H NMR (500 MHz, [D6]DMSO): d=2.08 (s, 3H), 5.85
(s, 1H), 6.11 (br s, 2H), 6.93 (t, J=4.8 Hz, 1H), 7.65 (s, 4H), 7.91 (s,
4H), 8.55 (d, J=4.7 Hz, 2H), 8.94 (s, 1H), 9.98 ppm (d, J=9.3 Hz,
2H); 13C NMR (126 MHz, [D6]DMSO): d=23.9, 95.0, 113.6, 118.0,
120.1, 121.3, 127.7, 128.8, 133.7, 136.9, 143.9, 158.6, 160.1, 161.8,
163.3, 165.2, 165.3 ppm; HRMS (ESI): m/z [M+H]+ calcd for
C22H21N8O: 413.1838, found: 413.1825.
N-[4-(2-Amino-6-methylpyrimidin-4-ylamino)phenyl]-4-(7H-
purin-6-ylamino)benzamide (17): Compound 17 was synthesized
according to the procedure described for 10, using amine 9
(50 mg, 0.15 mmol) and 6-chloropurine (23 mg, 0.15 mmol). Purifi-
cation by flash chromatography on silica gel (CH2Cl2/MeOH,
100:0!60:40) gave compound 17 as a white solid (3 mg, 5%):
1H NMR (500 MHz, [D6]DMSO): d=2.08 (s, 3H), 5.85 (s, 1H), 6.10 (s,
1H), 7.65 (d, J=2.8 Hz, 4H), 7.91 (d, J=8.8 Hz, 2H), 8.16 (d, J=
8.8 Hz, 2H), 8.35 (s, 1H), 8.92 (s, 1H), 9.96 ppm (s, 1H); MS (ESI):
m/z=453.2 [M+H]+.
N-[4-(2-Amino-6-methylpyrimidin-4-ylamino)phenyl]-4-(4,6-dime-
thoxy-1,3,5-triazin-2-ylamino)benzamide (13): Compound 13 was
synthesized according to the procedure described for 10, using
amine 9 (40 mg, 0.12 mmol) and 2-chloro-4,6-dimethoxy-1-3-5-tria-
zine (21 mg, 0.12 mmol). Purification by preparative HPLC (H2O/
CH3CN, 90:10!0:100) gave compound 13 as a white solid (30 mg,
1
54%): H NMR (500 MHz, [D6]DMSO): d=10.41 (s, 1H), 10.03 (s, 1H),
8.93 (s, 1H), 7.94 (d, J=8.8 Hz, 2H), 7.88 (d, J=8.9 Hz, 2H), 7.65 (s,
4H), 6.11 (br s, 2H), 5.85 (s, 1H), 3.94 (s, 6H), 2.07 ppm (s, 3H);
13C NMR (126 MHz, [D6]DMSO): d=23.5, 54.7, 94.6, 119.4, 119.6,
N-[4-(2-Amino-6-methylpyrimidin-4-ylamino)-phenyl]-4-(acrin-
din-9-ylamino)benzamide (18): Compound 18 was synthesized ac-
cording to the procedure described for 10, using amine 9 (60 mg,
ꢁ 2014 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 590 – 601 597

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
0.18 mmol) and 9-chloroacridine (38 mg, 0.18 mmol). Purification
by preparative HPLC (H2O/CH3CN, 90:10!0:100) gave compound
18 as a white solid (51 mg, 55%): 1H NMR (500 MHz, [D6]DMSO):
d=2.07 (s, 3H), 5.85 (s, 1H), 6.11 (br s, 2H), 6.88 (d, J=8.5 Hz, 2H),
6.90–6.99 (m, 2H), 7.35 (d, J=7.7 Hz, 2H), 7.50 (t, J=9.6 Hz, 1H),
7.70–7.60 (m, 5H), 7.85–7.79 (m, 2H), 7.97 (d, J=8.5 Hz, 2H), 8.15
(t, J=7.4 Hz, 2H), 8.92 (br s, 1H), 10.01 (br, 1H), 11.05 ppm (br s,
1H); 13C NMR (126 MHz, [D6]DMSO): d=23.5, 94.6, 114.8, 116.8,
117.7, 119.7, 120.8, 121.1, 124.3, 125.1, 127.2, 129.1, 129.6, 129.7,
130.4, 131.7, 133.4, 136.5, 150.8, 157.1, 161.4, 162.9, 164.7,
164.9 ppm; HRMS (ESI): m/z [M+H]+ calcd for C22H21N8O:
512.2193, found: 512.2198.
7.86 (td, J=6.7, 1.3 Hz, 1H), 7.63 (d, J=4.3 Hz, 1H), 7.77 (d, J=
4.3 Hz, 1H), 7.86 (td, J=6.9, 1.4 Hz, 1H), 7.93 (d, J=8.7 Hz, 2H),
8.00 (d, J=8.8 Hz, 2H) 8.03 (d, J=8.7 Hz, 2H), 8.15 (t, J=8.9 Hz,
2H), 8.66 (dd, J=8.6, 1.1 Hz, 1H), 9.06 (d, J=4.3 Hz, 1H), 9.25 (br s,
1H), 10.16 (s, 1H), 11.08 ppm (s, 1H); 13C NMR (126 MHz,
[D6]DMSO): d=22.5, 95.1, 119.3, 119.4, 120.0, 124.0, 125.2, 125.3,
126.5, 127.1, 128.7, 129.6, 130.2, 130.4, 133.7, 136.2, 141.6, 141.7,
148.0, 148.4, 150.4, 150.5, 161.5, 164.6, 165.7 ppm; MS (ESI): m/z=
490 [M+H]+.
N-[4-(2-Amino-6-methylpyrimidin-4-ylamino)phenyl]-4-(methyl-
amino)benzamide (23) and N-[4-(2-Amino-6-methylpyrimidin-4-
ylamino)phenyl]-4-(pyridin-3-ylmethylamino)benzamide
(24):
N-[4-(2-Amino-6-methylpyrimidin-4-ylamino)phenyl]-4-(isonicoti-
namide)benzamide (19): A suspension of isonicotinoyl chloride hy-
drochloride (50 mg, 0.28 mmol) in pyridine (0.5 mL) was cooled to
08C, and a solution of amine 9 (94 mg, 0.28 mmol) in pyridine
(1.5 mL) was added dropwise. The reaction was stirred at RT over-
night. The mixture was then filtered, and the isolated solid was pu-
rified by flash chromatography on silica gel (CH2Cl2/MeOH, 100:0!
70:30) to give compound 19 as a white solid (38 mg, 30%):
1H NMR (500 MHz, [D6]DMSO): d=2.07 (s, 3H), 5.88 (s, 1H), 6.11 (s,
1H), 7.68 (s, 4H), 7.99 (d, J=5.2 Hz, 6H), 8.77 (d, J=5.4 Hz, 2H),
9.14 (s, 1H), 10.30 ppm (s, 1H); HRMS (ESI): m/z [M+H]+ calcd for
C24H22N7O2: 440.1835, found: 440.1817.
Pyridinium chlorochromate (PCC; 395 mg, 1.83 mmol) was added
to a solution of 4-pyridinemethanol (100 mg, 0.92 mmol) in CH2Cl2
(5 mL), and the mixture was stirred for 90 min at RT. The resulting
mixture was diluted with CH2Cl2 (25 mL) and then washed with
water (2ꢂ25 mL) and brine (1ꢂ25 mL). The organic phase was
dried over Na2SO4, filtered and concentrated in vacuo. The residue
was dissolved in MeOH (2 mL), treated with Et3N (38 mg,
0.38 mmol) and amine 9 (40 mg, 0.12 mmol), and the solution was
stirred at RT for 3.5 h. AcOH (0.053 mL) and NaBH3CN (30 mg) were
then added, and the solution was stirred at RT overnight. The re-
sulting mixture was purified by preparative HPLC (H2O/CH3CN,
90:10!0:100) to give compound 23 as a light brown solid (9 mg,
17%) and compound 24 as a pale brown solid (1.1 mg, 3%). Com-
pound 23: 1H NMR (500 MHz, [D6]DMSO): d=2.07 (s,3H), 2.73 (d,
J=5.0 Hz, 3H), 5.83 (s, 1H), 6.09 (s, 2H), 6.31 (q, J=4.9 Hz, 1H),
6.67 (d, J=8.8 Hz, 2H), 7.62 (s, 4H), 8.60 (d, J=8.8 Hz,1H), 8.89 (s,
1H), 9.70 ppm (s, 1H); 13C NMR (126 MHz, [D6]DMSO): d=23.5,
29.3, 94.5, 110.4, 119.7, 120.6, 121.1, 133.7, 136.1, 152.5,161.4, 162.9,
169.2, 164.8, 165.0 ppm; HRMS (ESI): m/z [M+H]+ calcd for
C19H21N6O: 371.1771, found: 371.1589. Compound 24: 1H NMR
(500 MHz, [D6]DMSO): d=2.09 (s, 3H), 4.42 (d, J=6.3 Hz, 2H), 5.84
(s, 1H), 6.10 (s, 2H), 6.62 (d, J=8.8 Hz, 2H), 7.03 (t, J=6.2 Hz, 1H),
7.35 (d, J=5.8 Hz, 2H), 7.61 (s, 4H), 7.72 (d, J=8.8 Hz, 2H), 8.51 (d,
6.0 Hz, 2H), 8.91 (s, 1H), 9.71 ppm (s, 1H); 13C NMR (126 MHz,
[D6]DMSO): d=23.9, 45.3, 94.9, 111.7, 120.1, 121.1, 122.4, 122.7,
129.6, 134.0, 136.5, 149.5, 150.0, 151.4, 161.8, 163.8, 165.3,
165.4 ppm; HRMS (ESI): m/z [M+Na]+ calcd for C24H23N7ONa:
448.1862, found: 448.1844.
N-[4-(2-Amino-6-methylpyrimidin-4-ylamino)phenyl]-4-(nicotina-
mide)benzamide (20). Compound 20 was synthesized according
to the procedure described for 19, using nicotinoyl chloride hydro-
chloride (100 mg, 0.56 mmol) and amine 9 (188 mg, 0.56 mmol).
Purification by flash chromatography on silica gel (CH2Cl2/MeOH,
100:0!70:30) gave compound 20 as a light brown solid (40 mg,
1
30%): H NMR (500 MHz, [D6]DMSO): d=2.27 (s, 3H), 6.13 (s, 1H),
7.60 (dd, J=7.8.9, 4.9 Hz, 1H), 7.92–8.03 (m, 4H), 7.95 (d, J=8.8 Hz,
2H), 8.01 (d, 8.8 Hz, 2H), 8.34 (d, J=8.1 Hz, 1H), 8.79 (dd, J=4.7,
1.4 Hz, 1H), 9.14 (d, J=1.8 Hz, 1H), 10.27 (s, 1H), 10.52 (s, 1H),
10.74 (s, 1H), 12.49 ppm (s, 1H); 13C NMR (126 MHz, [D6]DMSO): d=
18.4, 97.1, 112.6, 119.5, 120.7, 123.6, 128.6, 129.8, 130.3, 133.7,
135.7, 141.9, 148.9, 152.3, 155.8, 161.4, 164.4, 164.8 ppm; HRMS
(ESI): m/z [M+H]+ calcd for C24H22N7O2: 440.1835, found: 440.1825.
N-[4-(2-Amino-6-methylpyrimidin-4-ylamino)phenyl]-4-(2-hy-
droxynicotinamide)benzamide (21): Compound 21 was synthe-
sized according to the procedure described for 19, upon use of 2-
hydroxy-3-nicotinic acid (90 mg, 0.65 mmol) and amine 9 (72 mg,
0.22 mmol). The resulting mixture was purified by flash chromatog-
raphy on silica gel (CH2Cl2/MeOH, 100:0!60:40) and by prepara-
tive HPLC (H2O/CH3CN, 90:10!0:100) to give compound 21 as
a light yellow solid (20 mg, 20%): 1H NMR (500 MHz, [D6]DMSO):
d=2.08 (s, 3H), 5.85 (s, 1H), 6.12 (s, 2H), 6.53 (dd, J=7.2, 5.9 Hz,
1H), 7.82 (d, J=8.8 Hz, 2H), 7.66 (s, 4H), 7.87 (dd, J=5.9, 2.3 Hz,
1H), 7.97 (d, J=8.8 Hz, 1H), 8.38 (dd, J=7.2, 2.3 Hz, 1H), 8.95 (s,
1H), 10.08 (s, 1H), 13.04 ppm (s, 1H); 13C NMR (126 MHz,
[D6]DMSO): d=23.6, 94.6, 107.1, 118.2, 118.9, 119.7, 120.9, 128.8,
129.5, 133.2, 136.6, 141.6, 143.2, 143.4, 161.4, 163.0, 164.5, 164.6,
164.9 ppm; HRMS (ESI): m/z [M+H]+ calcd for C24H22N7O3:
456.1784, found: 456.1753.
N-[4-(2-Amino-6-methylpyrimidin-4-ylamino)-phenyl]-4-(pyridin-
4-ylmethylamino)benzamide (25): MnO2 (2.10 g, 24.2 mmol) was
added to a solution of 3-pyridinemethanol (330 mg, 3.0 mmol) in
CH2Cl2 (5 mL), and the reaction mixture was stirred at RT overnight.
The mixture was filtered through a pad of celite, and solvent was
removed in vacuo. The residue was dissolved in MeOH (5 mL),
treated with Et3N (100 mg, 0.99 mmol) and amine 9 (80 mg,
0.24 mmol), and then stirred at RT overnight. AcOH (0.150 mL) and
NaBH3CN (145 mg) were then added, and the solution was stirred
at RT for 3 h. The resulting mixture was purified by preparative
HPLC (H2O/CH3CN, 90:10!0:100) to give compound 25 as a light
1
yellow solid (98 mg, 96%): H NMR (500 MHz, [D6]DMSO): d=2.06
(s, 3H), 4.39 (d, J=6.1 Hz, 2H), 5.84 (s, 1H), 6.10 (s, 2H), 6.65 (d, J=
8.8 Hz, 2H), 6.95 (t, J=6.2 Hz, 1H), 7.36 (dd, J=7.8, 4.8 Hz, 1H),
7.60 (br s, 4H), 7.72 (d, J=8.8 Hz, 2H), 7.75 (dt, J=7.8 Hz, 1.8 Hz,
1H), 8.45 (dd, J=4.8, 1.7 Hz, 1H), 8.60 (d, J=1.9 Hz, 1H), 9.71 (s,
1H), 8.90 ppm (s, 1H); 13C NMR (126 MHz, [D6]DMSO): d=23.6,
43.6, 94.5, 111.3, 119.7, 120.6, 121.9, 123.6, 129.2, 133.7, 135.1,
135.1, 136.2, 148.1, 149.0, 151.0, 161.4, 162.9, 164.9, 165.0 ppm;
HRMS (ESI): m/z [M+H]+ calcd for C24H24N7O1: 426.2042, found:
426.2037.
N-[4-(2-Amino-6-methylpyrimidin-4-ylamino)-phenyl]-4-(quino-
line-4-carboxylamide)benzamide (22). Compound 22 was synthe-
sized according to the procedure described for 19, using quinoline
carboxylic acid (85 mg, 0.49 mmol) and amine
9 (82 mg,
0.25 mmol). Purification by flash chromatography (CH2Cl2/MeOH,
100:0!70:30) gave compound 22 as a light yellow solid (27 mg,
1
23%): H NMR (500 MHz, [D6]DMSO): d=2.12 (s, 3H), 5.91 (s, 1H),
ꢁ 2014 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 590 – 601 598

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
N-(4-Nitrophenyl)quinolin-4-amine (26): Compound 26 was syn-
thesized according to the procedure described for 6, using 4-nitro-
aniline (4.13 g, 29.90 mmol) and 4-chloroquinoline (4.89 g,
29.90 mmol). The reaction mixture was filtered, and the resulting
solid was rinsed with Et2O and dried in vacuo to afford compound
HRMS (ESI): m/z [M+H]+ calcd for C27H24N7O: 462.2037, found:
462.2034.
N-[4-(Quinolin-4-ylamino)phenyl]-4-(pyridin-4-ylamino)benza-
mide (31): Compound 31 was synthesized according to the proce-
dure described for 10, using amine 29 (55 mg, 0.16 mmol) and 4-
chloropyridine hydrochloride (47 mg, 0.31 mmol). Purification by
preparative HPLC (H2O/CH3CN, 90:10!0:100) gave compound 31
26 as
a
pale brown solid (3.73 g, 47%): 1H NMR (500 MHz,
[D6]DMSO): d=7.26 (d, J=6.8 Hz, 1H), 7.80 (d, J=8.9 Hz, 2H), 7.87
(td, J=7.1, 1.3 Hz, 1H), 8.08 (td, J=7.1, 1.3 Hz, 1H), 8.15 (d, J=
8.5 Hz, 1H), 8.40 (d, J=9.1 Hz, 2H), 8.70 (d, J=6.8 Hz, 1H), 8.84 (d,
J=8.2 Hz, 1H), 11.23 (s, 1H), 15.03 ppm (s, 1H); 13C NMR (126 MHz,
[D6]DMSO): d=102.2, 118.6, 121.0, 124.4, 125.8, 128.0, 134.6, 139.0,
144.0, 144.6, 145.2, 154.3 ppm; HRMS (ESI): m/z [M+H]+ calcd for
C24H24N7O: 266.0924, found: 266.0928.
1
as a yellow solid (25 mg, 38%): H NMR (500 MHz, [D6]DMSO): d=
6.84 (d, J=5.5 Hz, 1H), 7.05 (d, J=6.4 Hz, 2H), 7.31 (d, J=8.8 Hz,
2H), 7.35 (d, J=7.0 Hz, 2H), 7.52 (td, J=7.0, 1.2 Hz, 1H), 7.69 (td,
J=6.9, 1.2 Hz, 1H), 7.84 (d, J=8.9 Hz, 2H), 7.86 (d, J=8.4 Hz, 1H),
7.97 (d, J=8.7 Hz, 2H), 8.28 (d, J=6.3 Hz, 2H), 8.39 (d, J=8.5 Hz,
1H), 8.43 (d, J=5.4 Hz, 1H), 8.93 (s, 1H), 9.18 (s, 1H), 10.17 ppm (s,
1H); 13C NMR (126 MHz, [D6]DMSO): d=100.8, 110.2, 117.6, 119.4,
121.2, 121.9, 123.3, 124.4, 127.6, 129.0, 129.1, 129.1, 135.5, 135.6,
143.8, 148.1, 148.8, 148.8, 150.2, 157.6, 164.6 ppm; HRMS (ESI): m/z
[M+H]+ calcd for C27H22N5O: 432.1824, found: 432.1815.
N-(Quinolin-4-yl)benzene-1,4-diamine (27): Compound 27 was
synthesized according to the procedure described for 7, using 26
(3.73 g, 14.07 mmol). The filtrate was concentrated in vacuo and
purified by flash chromatography (CH2Cl2/MeOH, 100:0!70:30) to
1
give amine 27 as a brown solid (2.93 g, 89%): H NMR (500 MHz,
N-[4-(Quinolin-4-ylamino)phenyl]-4-(2,6-diaminopyrimidin-4-yl-
amino)benzamide (32): Compound 32 was synthesized according
to the procedure described for 10, using amine 29 (80 mg,
[D6]DMSO): d=5.33 (br s, 2H), 6.56 (d, J=6.7 Hz, 1H), 6.69 (d, J=
8.6 Hz, 2H), 7.06 (d, J=8.6 Hz, 2H), 7.63–7.70 (m, 1H), 7.86–7.91
(m, 1H), 7.94 (d, J=8.3 Hz, 1H), 8.39 (d, J=6.6 Hz, 1H), 8.59 (d, J=
8.5 Hz, 1H), 10.11 ppm (br s, 1H); 13C NMR (126 MHz, [D6]DMSO):
d=99.5, 114.5, 117.4, 122.6, 122.9, 125.5, 126.1, 126.7, 132.5, 141.2,
144.6, 148.1, 154.1 ppm; MS (ESI): m/z=236 [M+H]+.
0.24 mmol)
and
2,6-diamino-4-chloropyrimidine
(32 mg,
0.24 mmol). Purification by preparative HPLC (H2O/CH3CN, 90:10!
1
0:100) gave compound 32 as a white solid (39 mg, 38%): H NMR
(500 MHz, [D6]DMSO): d=5.26 (s, 1H), 5.76 (br s, 2H), 5.92 (br s,
2H), 6.82 (d, J=5.3 Hz, 1H), 6.84 (d, J=5.3 Hz, 1H), 7.22 (d, J=
5.1 Hz, 1H), 7.34 (d, J=8.8 Hz, 1H), 7.52 (td, J=6.9, 1.4 Hz, 1H),
7.69 (td, J=6.9, 1.3 Hz, 1H), 7.79 (d, J=8.9 Hz, 1H), 7.88–7.82 (m,
5H), 8.39 (br d, J=8.6 Hz, 1H), 8.43 (d, J=5.3 Hz, 1H), 8.92 (br s,
2H), 10.06 ppm (br s, 1H); 13C NMR (126 MHz, [D6]DMSO): d=77.5,
100.9, 117.4, 117.7, 119.5, 121.3, 122.0, 123.5, 124.5, 125.6, 128.4,
129.2, 135.5, 135.9, 145.2, 148.2, 148.9, 150.7, 161.0, 162.9, 164.8,
165.0 ppm; HRMS (ESI): m/z [M+H]+ calcd for C26H23N8O1:
463.1989, found: 463.1987.
N-[4-(Quinolin-4-ylamino)phenyl]-4-nitrobenzamide (28). Com-
pound 28 was synthesized according to the procedure described
for 8, using 4-nitrobenzoyl chloride (1.37 g, 7.40 mmol) and amine
27 (1.74 g, 7.40 mmol). The filtrate was concentrated in vacuo and
purified by flash chromatography (CH2Cl2/MeOH, 100:0!70:30) to
give compound 28 as a yellow solid (2.60 g, 93%): 1H NMR
(500 MHz, [D6]DMSO): d=6.84 (d, J=6.3 Hz, 1H), 7.47 (d, J=8.7 Hz,
2H), 7.69–7.75 (m, 2H), 7.88–7.91 (m, 1H), 7.93–7.98 (m, 3H), 8.23
(d, J=8.8 Hz, 2H), 8.40 (d, J=8.8 Hz, 2H), 8.50 (d, J=6.3 Hz, 1H),
8.60 (d, J=8.5 Hz, 1H), 8.60 (d, J=8.5 Hz, 1H), 10.76 ppm (s, 1H);
MS (ESI): m/z=385 [M+H]+.
N-[4-(Quinolin-4-ylamino)phenyl]-4-(2-phenylquinolin-4-yl-
amino)benzamide (33): Compound 33 was synthesized according
to the procedure described for 10, using amine 29 (80 mg,
0.24 mmol) and 4-chloro-2-phenylquinoline (81 mg, 0.34 mmol).
Purification by preparative HPLC (H2O/CH3CN, 90:10!0:100) gave
N-[4-(Quinolin-4-ylamino)phenyl]-4-aminobenzamide (29): Com-
pound 29 was synthesized according to the procedure described
for 7, using 28 (1.40 g, 3.64 mmol). The filtrate was concentrated in
vacuo and purified by flash chromatography (CH2Cl2/MeOH,
100:0!80:20) to give amine 29 as a yellow solid (326 mg, 25%):
1H NMR (500 MHz, [D6]DMSO): d=5.77 (s, 1H), 6.61 (d, J=8.7 Hz,
2H), 6.78 (d, J=6.1 Hz, 1H), 7.36 (d, J=8.8 Hz, 2H), 7.65 (t, J=
8.9 Hz, 1H), 7.73 (d, J=8.7 Hz, 2H), 7.84–7.81 (m, 1H), 7.88 (d, J=
8.8 Hz, 1H), 7.91 (br d, J=8.3 Hz, 1H), 8.46 (d, J=6.1 Hz, 1H), 8.63
(br d, J=7.9 Hz, 1H), 9.89 ppm (br s, 1H); MS (ESI): m/z=355
[M+H]+.
1
compound 33 as a white solid (116 mg, 92%): H NMR (500 MHz,
[D6]DMSO): d=6.85 (d, J=5.2 Hz, 1H), 7.36 (d, J=8.8 Hz, 2H),
7.54–7.46 (m, 4H), 7.62–7.56 (m, 3H), 7.69 (td, J=6.9, 1.3 Hz, 1H),
7.37 (s, 1H), 7.77 (td, J=6.9, 1.3 Hz, 1H), 7.85–7.88 (m, 3H), 8.01
(dd, J=8.5, 0.9 Hz, 1H), 8.06 (d, J=8.7 Hz, 2H), 8.09–8.11 (m, 2H),
8.40 (br d, J=8.5 Hz, 2H), 8.44 (d, J=5.2 Hz, 1H), 8.94 (br s, 1H),
9.34 (br s, 1H), 10.15 ppm (br s, 1H); 13C NMR (126 MHz,
[D6]DMSO): d=101.4, 120.0, 120.1, 121.7, 122.5, 122.7, 123.9, 124.9,
125.5, 125.6, 127.5, 128.9, 129.2, 129.6, 129.7, 129.8, 130.0, 136.1,
136.2, 139.6, 139.9, 144.9, 147.7, 148.6, 149.3, 149.5, 151.1, 157.1,
165.2 ppm; HRMS (ESI): m/z [M+H]+ calcd for C37H28N5O:
558.2288, found: 558.2292.
N-[4-(Quinolin-4-ylamino)phenyl]-4-(2-amino-6-methylpyrimidin-
4-ylamino)benzamide (30): Compound 30 was synthesized ac-
cording to the procedure described for 10, using amine 29 (70 mg,
0.20 mmol) and 2-amino-4-chloro-6-methylpyrimidine (43 mg,
0.30 mmol). Purification by preparative HPLC (H2O/CH3CN, 90:10!
N-(4-Nitrophenyl)-4-nitrobenzamide (34): Compound 34 was syn-
thesized according to the procedure described for 8, using 4-nitro-
benzoyl chloride (6.72 mg, 3.62 mmol) and 4-nitroaniline (5.0 g,
3.62 mmol). Filtration of the precipitate and drying in vacuo gave
1
0:100) gave compound 30 as a white solid (58 mg, 64%): H NMR
(500 MHz, [D6]DMSO): d=5.95 (s, 1H), 6.27 (br s, 2H), 6.83 (d, J=
5.3 Hz, 1H), 7.34 (d, J=8.8 Hz, 1H), 7.52 (td, J=6.9, 1.2 Hz, 1H),
7.69 (td, J=6.8, 1.3 Hz, 1H), 7.83 (d, J=8.9 Hz, 2H), 7.86 (dd, J=
8.6, 1.0 Hz, 1H), 7.94–7.88 (m, 4H), 8.39 (br d, J=8.0 Hz, 1H), 8.43
(d, J=5.3 Hz, 1H), 8.92 (br s, 1H), 9.34 (br s, 1H), 10.12 ppm (br s,
1H); 13C NMR (126 MHz, [D6]DMSO): d=23.6, 95.5, 100.9, 117.9,
119.5, 121.3, 122.0, 123.5, 124.5, 126.7, 128.5, 129.2, 129.2, 135.6,
135.8, 144.2, 148.2, 148.9, 150.7, 161.1, 162.9, 164.9, 165.6 ppm;
compound 34 as
a
yellow solid (10.29 mg, 99%): 1H NMR
(500 MHz, [D6]DMSO): d=8.09 (d, J=8.3 Hz, 2H), 8.24 (d, J=8.9 Hz,
2H), 8.29 (d, J=9.2 Hz, 2H), 8.39 (d, J=8.8 Hz, 2H), 11.20 ppm (br
s, 1H); 13C NMR (126 MHz, [D6]DMSO): d=120.1, 123.6, 124.8, 129.6,
139.8, 142.8, 145.0, 149.4, 164.7 ppm.
N-(4-Aminophenyl)-4-aminobenzamide (35): Compound 35 was
synthesized according to the procedure described for 7, using
ꢁ 2014 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 590 – 601 599

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
compound 34 (5.10 g, 17.8 mmol). Filtration over a pad of silica
and concentration in vacuo gave diamine 35 as a brown solid
(1.08 g, 27%): 1H NMR (500 MHz, [D6]DMSO): d=5.72 (br s, 4H),
6.57 (d, J=8.6 Hz, 2H), 6.61 (d, J=8.7 Hz, 2H), 7.39 (d, J=8.8 Hz,
2H), 7.67 (d, J=8.6 Hz, 2H), 9.45 ppm (br s, 1H); 13C NMR
(126 MHz, [D6]DMSO): d=112.9, 116.5, 121.8, 122.3, 129.6, 131.9,
140.4, 152.2, 165.2 ppm; HRMS (ESI): m/z [M+Na]+ calcd for
C13H13N3ONa: 250.0951, found: 250.0951.
Biological evaluation
DNMT3A assay: DNMT3A enzyme inhibition was adapted from the
restriction-based fluorescence assay protocol described by Ceccaldi
et al.[20] Briefly, a 5’-labelled biotin oligonucleotide was hybridized
to its complementary strand labelled with 6-carboxyfluorescein at
the 3’-end and transferred into a 384-well microplate (black Opti-
plates; PerkinElmer) pre-coated with avidin. The duplex contains
a unique CpG site overlapping with a restriction site of a methyla-
tion-sensitive restriction enzyme. The human C-terminal DNMT3A
(a.a. 623-908), produced as described in Ref. [20], was added to
each well (200 ng/well) and mixed with the chemical compounds
at the desired concentration and freshly prepared AdoMet (20 mm
final concentration) to start the reaction in a total volume of 50 mL.
After 1 h incubation at 378C, each well was washed three times
with phosphate-buffered saline (PBS) containing 0.05% Tween-20
and NaCl (500 mm) and three more times with phosphate-buffered
saline Tween-20 (PBST). Specific fluorescence signals were detected
with the methylation-sensitive restriction enzyme HpyCH4IV (New
England Biolabs, Ipswich, MA, USA) as described, and measured on
a PerkinElmer Envision detector. The percent inhibition was calcu-
lated according to Equation (1), where X is the signal determined
in the absence of the inhibitor and Y is the signal obtained in the
presence of the inhibitor.
N-[4-(Pyridin-4-ylamino)phenyl]-4-(pyridin-4-ylamino)benzamide
(36): Compound 36 was synthesized according to the procedure
described for 10, using amine 35 (80 mg, 0.35 mmol) and chloro-
pyridine hydrochloride (110 mg, 0.74 mmol). Purification by HPLC
(H2O/CH3CN, 90:10!0:100) gave compound 36 as a white solid
1
(57 mg, 43%): H NMR (500 MHz, [D6]DMSO): d=6.84 (d, J=6.5 Hz,
2H), 7.04 (d, J=6.5 Hz, 2H), 7.18 (d, J=8.9 Hz, 2H), 7.30 (d, J=
8.6 Hz, 2H), 7.75 (d, J=8.9 Hz, 2H), 7.96 (d, J=8.6 Hz, 2H), 8.16 (d,
J=6.5 Hz, 2H), 8.28 (d, J=6.7 Hz, 2H), 8.72 (br s, 1H), 9.18 (br s,
1H), 10.10 ppm (br s, 1H); 13C NMR (126 MHz, [D6]DMSO): d=
108.8, 110.3, 117.8, 120.9, 121.4, 127.7, 129.2, 134.6, 135.8, 143.9,
148.9, 150.1, 150.4, 150.5, 164.6 ppm; HRMS (ESI): m/z [M+H]+
calcd for C23H20N5O: 382.1662, found: 382.1661.
N-[4-(Quinolin-4-ylamino)phenyl]-4-(quinolin-4-ylamino)benza-
mide (37): Compound 37 was synthesized according to the proce-
dure described for 10, using amine 35 (80 mg, 0.35 mmol) and 4-
chloroquinoline (173 mg, 1.06 mmol). Purification by HPLC (H2O/
CH3CN, 90:10!0:100) gave compound 37 as a white solid (86 mg,
51%): 1H NMR (500 MHz, [D6]DMSO): d=6.84 (d, J=5.3 Hz, 1H),
7.22 (d, J=5.1 Hz, 1H), 7.36 (d, J=8.8 Hz, 1H), 7.50 (d, J=8.5 Hz,
1H), 7.53 (td, J=7.1 Hz, 1.3 Hz, 1H), 7.58 (td, J=7.1 Hz, 1.0 Hz, 1H),
7.69 (td, J=6.8 Hz, 1.2 Hz, 1H), 7.74 (td, J=6.7 Hz, 1.3 Hz, 1H),
7.88–7.83 (m, 3H), 7.93 (d, J=8.3 Hz, 1H), 8.03 (d, J=8.7 Hz, 2H),
8.38 (d, J=8.7 Hz, 1H), 8.40 (d, J=8.5 Hz, 1H), 8.44 (d, J=5.2 Hz,
1H), 8.59 (d, J=5.1 Hz, 1H), 8.97 (br s, 1H), 9.27 (br s, 1H),
10.23 ppm (br s, 1H); 13C NMR (126 MHz, [D6]DMSO): d=101.0,
103.8, 119.5, 119.7, 120.5, 121.4, 122.4, 122.4, 123.5, 124.6, 125.1,
128.6, 129.1, 129.2, 129.3, 129.4, 129.6, 135.7, 135.7, 144.4, 146.4,
148.3, 148.8, 149.1, 150.6, 150.8, 164.8 ppm; HRMS (ESI): m/z
[M+H]+ calcd for C31H24N5O: 482.1975, found: 482.1974.
% Inhib: ¼ ½ðXꢁYÞ=Xꢂ ꢃ 100
ð1Þ
The ligand concentration at which 50% inhibition of enyme activi-
ty is observed (EC50) was determined by analysis of a concentration
range of the test compounds in triplicates. Nonlinear regression fit-
tings with sigmoidal dose–response (variable slope) were per-
formed with Prism 4.03 (GraphPad Software, Inc., La Jolla, CA,
USA).
DNMT1 and G9A assay: His-DNMT1 (182 kDa, human) was cloned,
expressed and purified as described by Lee et al.[21] The assays
were performed as described in the literature.[22] The reaction was
performed in a total reaction volume of 10 mL in low-volume non-
binding surface (NBSTM) 384-well microplates (Corning Inc.), con-
taining test compound (up to 1% DMSO), 1 mm of a S-adenosyl-l-
N-[4-(Acridin-9-ylamino)phenyl]-4-(acridin-9-ylamino)benzamide
(38): Compound 38 was synthesized according to the procedure
described for 10, using amine 35 (50 mg, 0.22 mmol) and 9-chlor-
oacridine (118 mg, 0.55 mmol). Purification by HPLC (H2O/CH3CN,
90:10!0:100) gave compound 38 as a white solid (40 mg, 31%):
1H NMR (500 MHz, [D6]DMSO): d=6.74 (d, J=8.8 Hz, 2H), 6.78 (d,
J=8.5 Hz, 2H), 6.97–6.92 (m, 4H), 7.51–7.41 (m, 8H), 7.71 (d, J=
8.7 Hz, 2H), 7.88 (d, J=8.5 Hz, 4H), 7.92 (d, J=8.3 Hz, 2H),
9.91 ppm (br s, 1H); 13C NMR (126 MHz, [D6]DMSO): d=117.4,
118.0, 118.5, 119.6, 120.3, 121.7, 121.8, 124.8, 126.9, 129.3, 130.4,
130.9, 133.5, 145.4, 151.5, 157.3, 164.9 ppm; HRMS (ESI): m/z
[M+H]+ calcd for C39H28N5O: 582.2288, found: 582.2281.
methionine (SAM)/[methyl-3H]SAM (3 TBqmmolꢁ1
, PerkinElmer)
mix in a ratio of 3:1 (isotopic dilution 1*:3), 0.3 mm of biotinylated
hemimethylated DNA duplex (5’-GATmCGCmCGATGmCGmCGAAT-
mCGmCGATmCGATGmCGAT-3’ and BIOT-5’-ATCGCATCGATCGCGATT-
CGCGCATCGGCGATC-3’), and 90 nm of DNMT1 in methylation
buffer (20 mm HEPES (pH 7.2), 1 mm EDTA, 50 mm KCl, 25 mgmLꢁ1
bovine serum albumin). The reaction was incubated at 378C for
2 h, then an aliquot (8 mL) was transferred into a streptavidin 96-
well scintillant-coated FlashPlate (PerkinElmer) containing 20 mm S-
adenosyl-l-homocysteine (SAH; 190 mL) in 50 mM Tris-HCl (pH 7.4).
The FlashPlate was agitated at RT for 1 h, washed three times with
200 mL of 0.05% TweenR-20 in 50 mm Tris-HCl (pH 7.4), and read in
200 mL of 50 mm Tris-HCl (pH 7.4) on TopCount NXT (PerkinElmer).
Docking analysis
CMV-luc reexpression: KG-1 cells, stably transfected with the firefly
luciferase (luc+ from pGL3 by Promega) reporter gene under the
control of the cytomegalovirus (CMV) promoter (from pEGFP-N1 by
Clontech Laboratories Inc.) partially methylated (50%), was seeded
at 20000 cells per well in a 96-well plate. After 24 h incubation in
the presence of the test compound or solvant (DMSO), the induc-
tion of the promoter was measured by quantification of luciferase
with the Brite-lite assay system (PerkinElmer) according to the
manufacturer’s protocol. The luminescence was measured on an
EnVision multilabel plate reader (PerkinElmer), and the data are ex-
The structure of Haemophilus haemolyticus cytosine-5 DNA methyl-
transferase (MHhaI C5 DNMT) was taken from the Protein Data
Bank (PDB: 2HR1[16]). Discovery Studio 3.0 (Accelrys Inc., San Diego,
CA, USA) to prepare the enzyme structures; alternate conforma-
tions were removed and incomplete chains with missing residues
and hydrogen atoms were added. The ligands were docked under
standard conditions using SYBYl-X 1.3 software (surflex-dock V2)
from Tripos L.P. (St. Louis, MO, USA). The images were prepared
with Benchware 3D Explorer, also from Tripos.
ꢁ 2014 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 590 – 601 600

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
Cancer Discovery 2011, 1, 598–607; c) L. Shen, H. Kantarjian, Y. Guo, E.
Lin, J. Shan, X. Huang, D. Berry, S. Ahmed, W. Zhu, S. Pierce, Y. Kondo, Y.
Oki, J. Jelinek, H. Saba, E. Estey, J. P. Issa, J. Clin. Oncol. 2010, 28, 605–
613.
pressed as the fold induction as compared with the DMSO control.
The mean of three experiments and the standard error is reported.
Antiproliferative activity: KG-1 human leukemia cells were obtained
from the American Type Culture Collection (ATCC, Manassas, VA,
USA) and cultivated in RPMI-1640 medium containing 4-(2-hydrox-
yethyl)-1-piperazineethanesulfonic acid (HEPES) and l-glutamine
(BE12-115F, Lonza, France) supplemented with 15% fetal calf
serum (Lonza, France), at 378C and under 5% CO2. Cells (2ꢂ104)
were seeded at day 0 in a 96-well plate. Solutions of test com-
pound, stored at ꢁ208C as a 10ꢁ2 m stock solution in 100% DMSO,
were freshly prepared by dilution on day 1 in RPMI-1640 medium.
Cells were treated with test compound solutions at a dose range
from 3.2 nm to 10 mm. Treatment was repeated on days 2 and 3,
and on day 4, cell viability was assessed using the ATPLite kit (AT-
Plite 1step Luminescence Assay System, ref: 3016739, PerkinElmer),
following the manufacturer’s instructions. Raw data were analyzed
with Prism 4.03 to generate EC50 values corresponding to the com-
pound concentration required to cause a 50% decrease in cell via-
bility as compared with untreated controls. The values presented
are the mean of at least two independent experiments.
[4] a) X. Cheng, R. M. Blumenthal, Structure 2008, 16, 341–350; b) R. Z. Ju-
rkowska, T. P. Jurkowski, A. Jeltsch, ChemBioChem 2011, 12, 206–222.
[5] a) J. Fahy, A. Jeltsch, P. B. Arimondo, Expert Opin. Ther. Pat. 2012, 22,
1427–1442; b) C. Gros, J. Fahy, L. Halby, I. Dufau, A. Erdmann, J. M. Gre-
goire, F. Ausseil, S. Vispe, P. B. Arimondo, Biochimie 2012, 94, 2280–
2296.
[6] J. M. Foulks, K. M. Parnell, R. N. Nix, S. Chau, K. Swierczek, M. Saunders,
K. Wright, T. F. Hendrickson, K. K. Ho, M. V. McCullar, S. B. Kanner, J.
Biomol. Screening 2012, 17, 2–17.
[7] J. Datta, K. Ghoshal, W. A. Denny, S. A. Gamage, D. G. Brooke, P. Phiasi-
vongsa, S. Redkar, S. T. Jacob, Cancer Res. 2009, 69, 4277–4285.
[8] G. J. Atwell, B. F. Cain, J. Med. Chem. 1973, 16, 673–678.
[9] D. Jia, R. Z. Jurkowska, X. Zhang, A. Jeltsch, X. Cheng, Nature 2007, 449,
248–251.
[10] J. Song, O. Rechkoblit, T. H. Bestor, D. J. Patel, Science 2011, 331, 1036–
1040.
[11] J. Song, M. Teplova, S. Ishibe-Murakami, D. J. Patel, Science 2012, 335,
709–712.
[12] a) J. L. Medina-Franco, T. Caulfield, Drug Discovery Today 2011, 16, 418–
425; b) J. L. Medina-Franco, F. Lopez-Vallejo, D. Kuck, F. Lyko, Mol. Diver-
sity 2011, 15, 293–304; c) J. L. Medina-Franco, J. Yoo, Mol. Diversity
2013, 17, 337–344; d) J. L. Medina-Franco, J. Yoo, Curr. Pharm. Des.
2013, 19, 2138–2147; e) J. Yoo, J.-L. Medina-Franco, Curr. Comput.-Aided
Drug Des. 2012, 8, 317–329; f) J. Yoo, J. L. Medina-Franco, J. Comput.-
Aided Mol. Des. 2011, 25, 555–567; g) J. Yoo, J. L. Medina-Franco, Curr.
Med. Chem. 2012, 19, 3475–3487.
Acknowledgements
The authors thank Dr. Jacques Fahy (USR3388 ETaC CNRS-Perre
Fabre) for initiating the program and Dr. Georges Massiot
(USR3388 ETaC CNRS-Perre Fabre) for proofreading the manu-
script. P.B.A. is a recipient of an ATIP award of the French Nation-
al Center for Scientific Research (CNRS), Equipe d’Excellence
Region Midi Pyrenꢀes (France) and the European Regional Devel-
opment Fund (FEDER) CNRS–Rꢀgion Midi Pyrenꢀes (France).
[13] J. Yoo, J. H. Kim, K. D. Robertson, J. L. Medina-Franco, Adv. Protein Chem.
Struct. Biol. 2012, 87, 219–247.
[14] J. Yoo, S. Choi, J. L. Medina-Franco, PLoS One 2013, 8, e62152.
[15] K. Takeshita, I. Suetake, E. Yamashita, M. Suga, H. Narita, A. Nakagawa, S.
Tajima, Proc. Natl. Acad. Sci. USA 2011, 108, 9055–9059.
[16] F. K. Shieh, B. Youngblood, N. O. Reich, J. Mol. Biol. 2006, 362, 516–527.
[17] P. Phiasivongsa, S. Redkar, S. A. Gamage, D. G. Brooke, W. A. Denny, D.
Bearss, H. Vankayalapati, (SuperGen Inc., Dublin, CA, USA), PCT Int. Pat.
Appl. WO 2009/049132 A1, 2009.
Keywords: cytotoxicity
·
DNA
methyltransferases
·
epigenetics · gene re-expression · inhibitors
[18] P. E. Zhichkin, L. H. Peterson, C. M. Beer, W. M. Rennells, J. Org. Chem.
2008, 73, 8954–8959.
[19] S. A. Gamage, D. G. Brooke, S. Redkar, J. Datta, S. T. Jacob, W. A. Denny,
Bioorg. Med. Chem. 2013, 21, 3147–3153.
[20] A. Ceccaldi, A. Rajavelu, C. Champion, C. Rampon, R. Jurkowska, G. Jan-
kevicius, C. Senamaud-Beaufort, L. Ponger, N. Gagey, H. D. Ali, J. Tost, S.
Vriz, S. Ros, D. Dauzonne, A. Jeltsch, D. Guianvarc’h, P. B. Arimondo,
ChemBioChem 2011, 12, 1337–1345.
[1] a) G. Auclair, M. Weber, Biochimie 2012, 94, 2202–2211; b) S. Sharma,
T. K. Kelly, P. A. Jones, Carcinogenesis 2010, 31, 27–36; c) S. B. Baylin,
P. A. Jones, Nat. Rev. Cancer 2011, 11, 726–734; d) A. Bird, Nature 2007,
447, 396–398; e) M. M. Suzuki, A. Bird, Nat. Rev. Genet. 2008, 9, 465–
476.
[2] a) S. L. Berger, T. Kouzarides, R. Shiekhattar, A. Shilatifard, Genes Dev.
2009, 23, 781–783; b) M. A. Dawson, T. Kouzarides, Cell 2012, 150, 12–
27.
[21] B. H. Lee, S. Yegnasubramanian, X. Lin, W. G. Nelson, J. Biol. Chem. 2005,
280, 40749–40756.
[22] C. Gros, L. Chauvigne, A. Poulet, Y. Menon, F. Ausseil, I. Dufau, P. B. Ari-
mondo, Nucleic Acids Res. 2013, 41, e185.
[3] a) M. Esteller, N. Engl. J. Med. 2008, 358, 1148–1159; b) R. A. Juergens, J.
Wrangle, F. P. Vendetti, S. C. Murphy, M. Zhao, B. Coleman, R. Sebree, K.
Rodgers, C. M. Hooker, N. Franco, B. Lee, S. Tsai, I. E. Delgado, M. A.
Rudek, S. A. Belinsky, J. G. Herman, S. B. Baylin, M. V. Brock, C. M. Rudin,
Published online on February 13, 2014
ꢁ 2014 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2014, 9, 590 – 601 601