Identifying novel selective non-nucleoside DNA methyltransferase 1 inhibitors through docking-based virtual screening

Article abstract of DOI:10.1021/jm501134e

The DNA methyltransferases (DNMTs) found in mammals include DNMT1, DNMT3A, and DNMT3B and are attractive targets in cancer chemotherapy. DNMT1 was the first among the DNMTs to be characterized, and it is responsible for maintaining DNA methylation patterns. A number of DNMT inhibitors have been reported, but most of them are nucleoside analogs that can lead to toxic side effects and lack specificity. By combining docking-based virtual screening with biochemical analyses, we identified a novel compound, DC-05. DC-05 is a non-nucleoside DNMT1 inhibitor with low micromolar IC₅₀ values and significant selectivity toward other AdoMet-dependent protein methyltransferases. Through a process of similarity-based analog searching, compounds DC-501 and DC-517 were found to be more potent than DC-05. These three potent compounds significantly inhibited cancer cell proliferation. The structure-activity relationship (SAR) and binding modes of these inhibitors were also analyzed to assist in the future development of more potent and more specific DNMT1 inhibitors.

Full text of DOI:10.1021/jm501134e

Subscriber access provided by CARNEGIE MELLON UNIV
Article
Identifying Novel Selective Non-Nucleoside DNA Methyltransferase
1 Inhibitors through Docking-Based Virtual Screening
Shijie Chen, Yulan Wang, Wen Zhou, Shanshan Li, Jianlong Peng, Zhe Shi,
Junchi Hu, Yu-Chih Liu, Hong Ding, Yijyun Lin, Linjuan Li, Sufang Cheng,
Jingqiu Liu, Tao Lu, Hualiang Jiang, Bo Liu, Mingyue Zheng, and Cheng Luo
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/jm501134e • Publication Date (Web): 21 Oct 2014
Downloaded from http://pubs.acs.org on October 26, 2014
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 43
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 43
Identifying Novel Selective Non-Nucleoside DNA Methyltransferase 1 Inhibitors
through Docking-Based Virtual Screening
1
2
3
4
5
6
7
Ⅱ
Shijie Chen#, ¶, Yulan Wang#, ¶, Wen Zhou , ¶, Shanshan Li‡, Jianlong Peng#, Zhe Shi, Junchi Hu#, YuꢀChih
8
9
Liu, Hong Ding#, Yijyun Lin, Linjuan Li#, §, Sufang Cheng#, Jingqiu Liu#, Tao Lu‡, Hualiang Jiang#, §, Bo
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Ⅱ
,
Liu *, Mingyue Zheng#,* and Cheng Luo#,*
#State Key Laboratory of Drug Research, Shanghai Institute of Materia Medica, Chinese Academy of
Sciences, 555 Zuchongzhi Road, Shanghai 201203, China
‡Laboratory of Molecular Design and Drug Discovery, School of Science, China Pharmaceutical University,
24 Tongjiaxiang, Nanjing 210009, China
In Vitro Biology, Shanghai ChemPartner LifeScience Co., Ltd., #5 Building,998 Halei Road, Shanghai
201203, China
Ⅱ
Guangdong Provincial Academy of Chinese Medical Sciences (The Second Affiliated Hospital of
Guangzhou University of Chinese Medicine), 55 Neihuanxi Road, Guangzhou 510006, China
§School of Life Science and Technology, Shanghai Tech University, Shanghai 200031, China
¶These authors contributed equally to this work.
*Correspondence:
Bo Liu: Telphone, 86ꢀ20ꢀ39318571. Email: doctliu@263.net.
Mingyue Zheng : Telphone, 86ꢀ21ꢀ50806600. Email: myzheng@simm.ac.cn.
Cheng Luo: Telphone, 86ꢀ20ꢀ50806600. Email: cluo@simm.ac.cn.
ACS Paragon Plus Environment

Page 3 of 43
Journal of Medicinal Chemistry
Abstract
1
2
3
4
5
6
7
8
9
The DNA methyltransferases (DNMTs) found in mammals include DNMT1, DNMT3A and DNMT3B, and
are attractive targets in cancer chemotherapy. DNMT1 was the first among the DNMTs to be characterized,
and it is responsible for maintaining DNA methylation patterns. A number of DNMT inhibitors have been
reported, but most of them are nucleoside analogs that can lead to toxic sideꢀeffects and lack specificity. By
combining dockingꢀbased virtual screening with biochemical analyses, we identified a novel compound,
DC_05. DC_05 is a nonꢀnucleoside DNMT1 inhibitor with low micromolar IC50 values and significant
selectivity towards other AdoMetꢀdependent protein methyltransferases. Through a process of
similarityꢀbased analog searching, compounds DC_501 and DC_517 were found to be more potent than
DC_05. These three potent compounds significantly inhibited cancer cell proliferation. The structureꢀactivity
relationship (SAR) and binding modes of these inhibitors were also analyzed to assist in the future
development of more potent and more specific DNMT1 inhibitors.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 43
Introduction
1
2
3
4
5
6
7
8
9
Epigenetic modifications such as DNA methylation, histone methylation and histone acetylation play
essential roles in all aspects of biology.1 Among all the epigenetic events, DNA methylation is most likely
the bestꢀknown epigenetic marker and has been shown to participate in gene expression control.2 In
mammals, DNA methylation occurs at the 5ꢀposition of cytosine, almost as in the context of CpG
dinucleotides that are clustered in CpG islands.1, 3 In cancer cells, alterations in DNA methylation can lead to
promoter hypermethylation at CpG islands and then silence tumor suppressor genes.4 Unlike genetic origins,
aberrations in DNA methylation are reversible, thus allowing cancer cells to revert to their normal state.5 To
date, DNMT1, DNMT3A and DNMT3B are the three different DNA methyltransferases that have been
identified in mammals.6, 7 DNMT1 is the most abundant among the three, and it serves as a substrate for
hemimethylated CpG dinucleotides and is responsible for maintaining DNA methylation patterns in
mammals.6, 8 DNMT3A and DNMT3B are associated with de novo methylation during embryonic
development.9 The inhibition of DNA methyltransferase activity can reactivate silenced tumor suppressor
genes (TSGs),10 and thus, DNMT inhibitors have become useful tools to revert cancer cells.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Two types of DNMT inhibitors have been discovered, namely nucleoside analogs and nonꢀnucleoside
analogs.1 Two nucleoside analogs, 5ꢀazacytidine (Vidaza) and 5ꢀazaꢀ2’ꢀdeoxycytidine (Decitabine), have
been approved by the FDA for treating myelodysplastic syndrome and leukemia.2 However, these drugs can
reportedly be incorporated into DNA and cause covalent trapping and the subsequent depletion of DNMTs.11
Moreover, these drugs are unstable, show low specificity and have significant toxic sideꢀeffects.12 Specific
interest has therefore been generated in nonꢀnucleosides. As shown in Chart 1, various nonꢀnucleoside
analogs have been reported, including the following: natural compounds such as genistein,13
(ꢀ)ꢀepigallocatechinꢀ3ꢀOꢀgallate (EGCG),14 and laccaic acid A;15 repurposed drugs such as hydralazine
(hypertension antagonist),16 procainamide and procaine (antiarrhythmic and anesthetic agents,
respectively);17, 18 novel inhibitors such as phthalimidoꢀLꢀtryptophan (RG108);4, 19 and the quinolone
ACS Paragon Plus Environment

Page 5 of 43
Journal of Medicinal Chemistry
derivative SGIꢀ1027.20 In comparison with nucleoside analogs, nonꢀnucleoside analogs are less likely to be
incorporated into DNA and hence offer a safer way to target DNA methylation. Unfortunately, the potency
of known nonꢀnucleoside DNMT inhibitors is much lower than that of nucleoside analogs, and most of
DNMT inhibitors are multiꢀtarget inhibitors with unknown mechanisms of action.12, 21 Moreover, as shown
in Chart 1, the demethylating activities of natural compounds and redirected drugs remain controversial.22, 23
There are two reported nonꢀnucleoside DNMT inhibitors called SGIꢀ1027 (IC50 = 35 ± 3 µM)24 and RG108
(IC50 = 390 ± 50 µM)25 that may cause demethylation by inhibiting DNMT directly, and their detailed
mechanisms of action remain unknown.19, 20 At present, there are no nonꢀnucleoside DNMT inhibitors
entering clinical trials, and it is therefore highly important to develop new DNMT inhibitors.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The virtual screening of compound databases has emerged as a powerful computational approach, and it is
also increasingly used in drug discovery projects.26ꢀ30 In addition to its convenience and low cost, this
technique can also help identify novel scaffolds, understand the binding mode of the active compound, and
provide some indications for compound optimization. In addition, virtual screening can also be
complemented with other ligandꢀbased approaches, such as similarity searching. To date, several successful
dockingꢀbased virtual screenings have been conducted with DNMT1.4, 31, 32
In this study, dockingꢀbased virtual screening was used to search for novel DNMT1 inhibitors from the
SPECS database. Highly ranked compounds were subjected to cluster analysis through virtual screening by
Glide. Fiftyꢀone compounds were selected and purchased for further DNMT1 inhibitor bioactivity testing.
Because it had a high inhibition rate, DC_05 was also the most potent compound in the group and showed
remarkable selectivity toward DNMT1. We therefore used this compound to obtain more structural analogs
within the SPECS database, and 19 analogs were selected to perform the bioactivity assay. DC_501 and
DC_517 both exhibited increased inhibitory potency in the enzymatic assays. Further experimental studies
demonstrated that these three potent compounds significantly inhibited cancer cell proliferation. Taken
together, this study indicates that DC_05 and its analogs are new potential DNMT1 inhibitors, and they can
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 43
provide us with new structural clues to develop more potent DNMT1 inhibitors.
1
2
3
4
5
Results and discussion
6
7
8
9
Docking-Based Virtual Screening. In the present study, a dockingꢀbased virtual screening strategy was
used to identify promising hits. The crystal structure of mouse DNMT1 bound to
SꢀadenosylꢀLꢀhomocysteine (SAH) and cytosine (PDB entry: 4DA4)33 was used for molecular docking. This
structure was selected because it is present in an active form and contains both the cofactor and DNA
substrate. The protein structure was prepared in the Protein Preparation Wizard Workflow of the
Schrödinger software package. The SPECS database (heep://www.specs.net) containing 198,745 compounds
was used as a ligand database. First, the compounds with unfavorable physicochemical properties were
filtered out using Pipeline Pilot 7.5. After removing the compounds with unfavorable properties, the
remaining 111,121 compounds were prepared with LigPrep to generate all stereoisomers and different
protonation states by Epik. The prepared ligands were docked in the SAH and cytosineꢀbinding pocket of
DNMT1 protein with the Glide module in Schrödinger. The docking procedure was validated by
reproducing the SAH binding mode with a rootꢀmeanꢀsquare deviation (RMSD) of 0.6 Å (Figure S1 in
supporting information (SI)). According to the Glide scores (GScores) and structural clustering, 51
structurally diverse compounds were selected and purchased for biochemical assay. The structures and
docking scores of the 51 compounds are provided in Table S1 in the SI.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
DNMT1 Inhibition Assays. The 51 candidate molecules selected by virtual screening were tested for
DNMT1 inhibition to validate their biochemical activities. From the ELISA DNMT1 activity assay, we used
the EpiQuik DNA Methyltransferase (DNMT) Activity/Inhibitor Assay Kit (Epigentek) and identified
DC_05 for its ability to inhibit DNMT1 activity by >60%, and this compound had a similar potency to that
of SAH against DNMT1 at a concentration of 200 ꢁM. In a more quantitative analysis, we measured the
methyltransferase activity of DNMT1 at a range of concentrations for this compound using the Hꢀ3ꢀlabeled
ACS Paragon Plus Environment

Page 7 of 43
Journal of Medicinal Chemistry
radioactive methylation assay and the IC50 value derived from the dose response curve, which was 10.3 ꢁM
(Figure 1). We also performed agarose gel electrophoresis to identify whether DC_05 binds to DNA.34
Doxorubicin was used as a positive control, and it has been shown to interact with DNA35. SGIꢀ1027 was
reported to bind slightly to DNA20, and it was also used as a positive control. SAH was used as a negative
control. As shown in Figure S2, doxorubicin and SGIꢀ1027 change the migration rate of DNA at higher
concentrations, and the decrease in emission intensity could be considered the result of EB (DNA bound)
replacement. By contrast, DC_05 and DC_517 have no such effect on the DNA and show similar results to
the negative SAH control and the DNA control without inhibitors. These results suggest that our compounds
do not act by binding to DNA. All the results suggest that DC_05 could be characterized as a structurally
novel DNMT1 inhibitor with a remarkable potency against DNMT1.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The Methyltransferase Enzymatic Selectivity of DC_05. In addition to DNMT1, there are many other
methyltransferases that can bind with SꢀadenosylꢀLꢀmethionine (SAM) to facilitate transmethylation
reactions. To investigate the selectivity of DC_05 for DNMT1, the compound was also evaluated for its
inhibitory activities against other important methyltransferases, including DNMT3A, DNMT3B and other
AdoMetꢀdependent enzymes, such as G9a (histone H3 lysine 9 methyltransferase),36 SUV39H1 (histone H3
lysine 9 methyltransferase),37 MLL1 (histone H3 lysine 4 methyltransferase),38, 39 SET7/9 (histone H3 lysine
4 methyltransferase)40 and PRMT1 (arginine methyltransferase).41 Under the tested conditions, DC_05
inhibited PRMT1 at a potency that was 3.6ꢀfold weaker than that of DNMT1. Moreover, it showed nearly no
inhibitory activities against DNMT3A, DNMT3B, MLL1, SET7/9, SUV39H1 and G9a (Table 1, Figure S3).
All these findings demonstrated that DC_05 is a DNMT1ꢀselective inhibitor. Here, the selectivity of
nonꢀnucleoside DNMT1 inhibitors is somewhat expected because these methyltransferases catalyze different
substrates and share a very low homology with DNMT1, even in their catalytic domain. The low
conservation of structure is revealed from the multiple sequence alignment analysis provided in Figure S4,
and among these comparisons, DNMT3A/3B and PRMT1 show relatively higher sequence identities with
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 43
DNMT1. To understand the basis of this selectivity, we superimposed the catalytic sites of DNMT1,
DNMT3A and PRMT1, and we briefly discuss the potential reasons for DC_05 selectivity in the supporting
information (Figure S5).
1
2
3
4
5
6
7
8
9
Similarity-Based Analog Searching and a Radioactive Methylation Assay Against DNMT1. In
accordance with the general notion that similar compounds have similar activity levels, similarityꢀbased
analog searching was conducted to select the compounds of interest. Because compound DC_05 has a
relatively high IC50 activity and good selectivity, an analog search using DC_05 as the query template was
performed to select more potential hit compounds. Finally, 19 compounds with similar scaffolds were
purchased from SPECS Corp, and their inhibitory activities against DNMT1 were assessed. Although most
of those compounds have chirality centers, the purchased samples were in the racemic form. Table 2 shows
the generic chemical structures of these compounds and their inhibitory activities against DNMT1. The
inhibition potencies are expressed as the inhibition rates at a concentration of 50 ꢁM. As shown both in
Table 2 and Figure 2, 8 of the 19 compounds showed more than 50% inhibition of the DNMT1 activity at 50
µM. We then tested these potent compounds within a range of concentrations by radioactive methylation
assay to compare with DC_05. DC_501 (IC50 = 2.5 µM) and DC_517 (IC50 = 1.7 µM) displayed 4.1ꢀ and
6.0ꢀfold higher activity than DC_05 against DNMT1, and DC_503, DC_504, DC_508, DC_514 and
DC_516 analogs were less potent, respectively. DC_512 showed a similar inhibitory activity as DC_05
(Table 2, Figure S6). This second round of compound searching and screening helped us to find many more
potential compounds that inhibited DNMT1 activity, and the results uncovered important information for hit
optimization and the study of structureꢀactivity relationships (SARs).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Surface Plasmon Resonance (SPR)-based Binding Assay. To more precisely validate these potential
DNMT1 inhibitors, we used the SPRꢀbased binding assay to measure direct interactions between DNMT1
and the compounds. As shown in Figure 3, the interactions were dose dependent and strong, further
confirming the activity of these potential compounds. The equilibrium dissociation constant (KD) between
ACS Paragon Plus Environment

Page 9 of 43
Journal of Medicinal Chemistry
DC_05 and DNMT1 is approximately 1.09 µM, and that of DC_517 is 0.91 µM. In comparison with
compound DC_05, DC_517 showed a higher affinity for DNMT1, which was consistent with the DNMT1
inhibition potency levels. However, because of the limited characterization of compound DC_501, we did
not obtain good SPR curves to calculate the KD between DC_501 and DNMT1.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Analyzing Chiral Enantiomers of DC_05 and DC_517. As shown in Table 2, DC_05 and its analogs
contain chiral centers, and the samples we purchased from SPECS are in racemic mixtures. DC_05 and
DC_517 enantiomers were synthesized to understand the effect of chirality on the activities of these
compounds. Radioactive methylation assays were performed to determine the activities of the DC_05 and
DC_517 enantiomers, in addition to their racemates. As shown in Figure 4B, (R)ꢀDC_05 (IC50 = 7.3 µM)
and (S)ꢀDC_05 (IC50 = 15.6 µM) show slightly different activities towards DNMT1, and the (R)ꢀ and
(S)ꢀenantiomers of DC_517 show very similar potencies (Figure 4E).
Molecular docking studies were also performed, and the putative binding modes of both enantiomers within
DNMT1 were carefully inspected (Figure 4C, F). The binding poses of (R)ꢀDC_05 and (S)ꢀDC_05 with
DNMT1 closely resemble one another (Figure 4C). Both the carbazolyls of (R)ꢀDC_05 and (S)ꢀDC_05 are
located in a large hydrophobic pocket formed by F1148, C1194, D1193 and M1172 in the SAM pocket and
establish a cation–π interaction with K1247, and the indolyls occupy the cytosine pocket and make hydrogen
bonds through the amino to E1269. Despite these similarities, their hydroxyls form hydrogen bonds with
different residues, in which the hydrogen bond in (R)ꢀDC_05 involves the E1171 residue, and it involves the
K1247 residue in (S)ꢀDC_05 (Figure 4C). E1171 is a conserved residue at the SAM pocket of DNMTs and
forms a hydrogen bond with SAH, but K1247 does not. The slightly different activities of (R)ꢀ and
(S)ꢀDC_05 may therefore be caused by the different hydrogen bonds with different residues. To account for
the improved inhibitory potency and the chiral effect of DC_517, we analyzed their interactions with
DNMT1. As shown in Figure 4F, one carbazolyl of DC_517 is located in the hydrophobic pocket of the
SAM pocket, which resembles DC_05 binding, and the other carbazolyl stretches outside the SAM pocket
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 43
and interacts with R1576, which has been shown to be an important residue that forms interactions with
DNMT1 inhibitors.3, 24, 25 Because R1576 is less than 4 Å away and has a carbazolyl group outside the SAM
pocket, the improved potency observed in DC_517 can therefore be ascribed to the gain of a cationꢀπ
interaction with R1576. In addition, the hydroxyl group of DC_517 forms hydrogen bonds with E1171, and
the imino group interacts with the primary chain oxygen of F1148. In comparing the putative binding poses
of two DC_517 enantiomers, we may find that they closely match with one another, which is consistent with
their activity assay results. The 2D ligand interaction diagrams of DC_05 and DC_517 enantiomers are
provided in Figures S7 and S8. In general, the binding modes of the enantiomers of these two compounds
within DNMT1 only show subtle differences that are consistent with their biochemical activities. These
results suggested that the chirality of the reported compound series may play a less important role in
determining their binding and activity toward DNMT1.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
SAR Analysis. The structureꢀactivity relationship of DC_05 and its analogs were identified and investigated
by similarity searching. As shown in Table 2, we found that the compounds with either scaffolds I or II
contain carbazolyl, and the compounds with scaffolds III and VI do not. Their DNMT1 inhibitory activities
suggest that the carbazolyl is essential for their activity. The putative binding poses of the compounds in
Table 2 were analyzed to understand the molecular basis for DNMT1 inhibition using the carbazolyl
molecule series. As shown in Figure 5A, the catalytic site of DNMT1 can be divided into two parts, namely
the cofactor binding site SAM pocket and the substrate binding site cytosine pocket, which are depicted as a
dashed green circle and a dashed orange circle, respectively. The carbazolyl groups of DC_05 is located in
the SAM pocket and establishes a cation–π interaction with K1247, and the indolyl group is located in the
cytosine pocket and forms a hydrogen bond via its imino group to E1269. To study the effects of the
substituents on the aliphatic chain (R1 and R3 in Table 2) in this activity, we compared the binding poses of
DC_05, DC_503, DC_504, DC_508 and DC_516. Figure 5B shows that they share a similar pattern, and it
shows that their filling of the cytosine pocket varies (Figure 5B). Generally, a better occupation of the
ACS Paragon Plus Environment

Page 11 of 43
Journal of Medicinal Chemistry
cytosine pocket corresponds to a higher inhibitory activity, and this result is consistent with a study by
Asgatay et al.25 For DC_05, the hydrogen bond formed with E1269 is essential to its binding because E1269
also forms a hydrogen bond with the cytosine and mediates the catalytic reaction under physiological
conditions.33, 42 In the SAM pocket, the carbazolyl of (R)ꢀ and (S)ꢀDC_05 are located in a large hydrophobic
region, and their hydroxyl groups form hydrogen bonds with E1171 and K1247, respectively (Figure 4C).
Similar interactions were also observed for the compounds displayed in Figure 5B, which suggests that the
hydrogen bonds are helpful for anchoring the carbazolyl group directly to the SAM pocket. The imino
groups (pKb≈3.1) of DC_05 and its analogs should mostly exist as a protonated aminium form at the given
pH (pH = 7.0±2.0), which was set by preparing ligand structures. There were ionic interactions between
these amino ions and E1171, which are barely affected by compound chirality. As mentioned above, most
hydroxyl groups in the compounds could form hydrogen bonds with E1171 or K1247, regardless of which
chirality they have. These interactions may explain why the chiral center does not have a major influence on
the binding affinity. Moreover, the chlorine (Cl) and bromine (Br) on the carbazolyl appear to be beneficial
to the inhibitory activity of these compounds because DC_501 (Figure 5C) is more potent than DC_05, and
DC_512 (Figure 5D) is more potent than DC_504 and DC_508. Because the halogen substituents on the
aromatic ring are typical hydrophobicityꢀenhancing groups, the increased activities of DC_501 and DC_512
can be partially ascribed to the reinforced hydrophobic interactions with the large hydrophobic region of the
SAM binding site. In addition, the lone pairs of the halogen substituents may also stabilize the cation–π
interaction of the carbazolyl with K1247. Another notable point is that DC_512 does not occupy the cytosine
pocket in comparison with DC_501, but the extra hydrogen bond formed with G1226 stabilizes its binding,
and the DC_512 activity is maintained (IC50 = 14.1 µM). To understand the improved potency of DC_512,
we compared the 2D ligand interaction diagrams of DC_512 with those of DC_504 and DC_508 and give a
brief discussion in Figure S9. As analyzed above, the improved potency of compound DC_517 can be
attributed to the cationꢀπ interaction between its outward carbazolyl and R1576, which has been highlighted
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 43
as an important residue that interacts with other nonꢀnucleoside DNMT inhibitors.3, 24, 25 However, from a
ligand efficiency (LE) perspective, DC_517 (LE = 0.21) is slightly inferior to DC_05 (LE = 0.24), which
suggests that the increased hydrophobic and cationꢀπ interactions are weak intermolecular forces and only
lead to a less significant gain in affinity. To achieve a more effective interaction with R1576, the carbazolyl
of DC_517 should be replaced by smaller groups, such as phenyl or other small negatively charged groups.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Cell-Based Activity. Collectively, compounds DC_05, DC_501 and DC_517 showed potential
DNMT1ꢀinhibiting activity in vitro, and we further explored whether these three most intriguing compounds
would affect the proliferation of cancer cells. We tested DC_05, DC_501 and DC_517 in HCT116 (human
colon cancer) and Capanꢀ1 (human pancreatic adenocarcinoma cells). An MTT assay was performed to
determine the compounds’ effects on cell proliferation and viability. In regular culture medium, DC_05,
DC_501 and DC_517 significantly inhibited cell proliferation at a low concentration. The doseꢀ and
timeꢀdependent inhibition profiles shown in Figure 6 clearly indicate that these compounds have remarkable
activities both in vitro and ex vivo (Figure 6A, B). As mentioned above, the enantiomers of DC_05 show
slightly different activities toward DNMT1, and the (R) and (S) enantiomers of DC_517 show very close
potencies in vitro. To understand the effect of the chirality on the proliferation of HCT116 and Capanꢀ1 cells,
we tested DC_05, (R)ꢀDC_05, (S)ꢀDC_05, DC_517, (R)ꢀDC_517 and (S)ꢀDC_517 on these two cell types.
Figure S10 shows that the enantiomers display the same level of antiꢀproliferative activities as their
corresponding racemates, which are also consistent with their DNMT1 inhibition potencies as determined by
radioactive methylation assay.
Notably, DC_517 displayed the highest antiꢀproliferative effects in the two types of cancer cells, and this
result was also fully consistent with the biochemical assay results. Moreover, in comparison with the
Capanꢀ1 cells, the HCT116 cells are more sensitive to all of the compounds, and thus, we chose compound
DC_517 for conducting apoptosis assays in HCT116 cells. The results demonstrated that DC_517 led to
doseꢀdependent apoptotic cell death in HCT116 cells (Figure 6C).
ACS Paragon Plus Environment

Page 13 of 43
Journal of Medicinal Chemistry
Conclusion
1
2
3
4
5
6
7
8
9
DNA methylation as catalyzed by DNMTs plays important roles in crucial cellular processes, such as
embryonic development or differentiation, and it is involved in cancer development. Most DNMT inhibitors
reported to date have low bioavailability and specificity. In the present study, we identified compound
DC_05 as a novel inhibitor of DNMT1 with an IC50 value of 10.3 µM by using the dockingꢀbased virtual
screening approach from a filtered small molecule SPECS compound database, which contains 198,745
compounds. DC_05 has shown remarkable selectivity towards other AdoMetꢀdependent protein
methyltransferases (DNMT3A, DNMT3B, G9a, SUV39H1, MLL1, SET7/9 and PRMT1). Subsequently,
similarityꢀbased analog searching was conducted to select more potent compounds. On the basis of the
structure of the selective inhibitor lead compound known as DC_05, 8 out of 19 DC_05 analogs were
identified as micromolar DNMT1 inhibitors. The radioactive methylation assay results identified compounds
DC_501 and DC_517 from this library of analogs as the two most potent DNMT1 inhibitors, with IC50
values of 2.5 µM and 1.7 µM, respectively. A Surface Plasmon Resonanceꢀbased binding assay precisely
validated these potential DNMT1 inhibitors by measuring direct interactions between DNMT1 and the
compounds. The equilibrium dissociation constant (KD) between DC_05 and DNMT1 is approximately 1.09
µM, and that of DC_517 is 0.91 µM, verifying the tight binding of these compounds with DNMT1. It should
be noted that these DC_05 analogs are chiral compounds, and our purchased samples are in racemic form.
To investigate how the chirality affects the activity of these compounds, asymmetric syntheses were
performed to obtain all DC_05 and DC_517 enantiomers, and their enzymatic assays revealed no obvious
activity differences between the enantiomers and their corresponding racemates. Our molecular docking
studies show that the binding poses of the two enantiomers with DNMT1 closely resemble one another,
which is consistent with their activity assay results. This finding suggests that the chiral structures of the
compounds do not influence its activities. Furthermore, when tested on HCT116 and Capanꢀ1 cells, DC_05,
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
DC_501 and DC_517 can significantly block the proliferation of cells at a low micromolar concentration.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 43
Given that the enantiomers show a similar potency against DNMT1 in biochemical assays, they also display
the same level of antiꢀproliferative activities in their corresponding racemates in cellꢀbased studies.
Moreover, the most potent inhibitor, which was compound DC_517, induced doseꢀdependent apoptotic cell
death in HCT116 cells.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Taken together, our experiments used virtual screening and similarity searching to find novel potent
nonꢀnucleoside DNMT1 inhibitors with considerable specificity for DNMT1. We identified the compounds
that are selective and highly potent inhibitors of DNMT1 via biochemical and cellular assays. Unlike the
nucleoside compounds, the DC_05ꢀinhibiting activity against DNMT1 is a result of binding to DNMT1
instead of incorporating into DNA. The SAR analysis and predicted binding mode studies of these analogs
show that a better occupation of the cytosine and SAM pockets is beneficial for the higher inhibitory activity
of inhibitors and that the formation of ionic interactions and/or hydrogen bonds in linking the two pockets is
necessary to form an interaction between the compounds and the DNMT1. In addition, the SAM pocket
shows a high affinity to hydrophobic groups. These results may provide meaningful clues for the future
structural optimization of these compounds and lay the foundation for the further development of DNMT1
inhibitors with greater potency and specificity for cancer therapy.
Experimental Section
Virtual Screening Protocol.
Preparing the Protein Structure. The crystal structure of murine DNMT1 that was resolved with the SAH
and DNA substrate was retrieved from the Protein Data Bank (PDB access number 4DA4). Specifically,
DNA fragments and the solvent molecules were deleted. The remaining protein structure was prepared using
the Protein Preparation Wizard module (Schrödinger, LLC: New York, NY, 2010) in the Maestro program
(Maestro version 9.1; Schrödinger, LLC: New York, NY, 2010) with standard Glide protocols.43 In brief, the
hydrogen atoms were properly added to the complexes, bond corrections were applied to the coꢀcrystallized
ACS Paragon Plus Environment

Page 15 of 43
Journal of Medicinal Chemistry
ligand, the hydrogen bond networks and flip orientations/tautomeric states of Gln, Asn, and His residues
were optimized to maximize hydrogen bond formation, and an exhaustive sampling was performed with
regards to hydrogen bond assignment. Finally, a restrained minimization on the ligandꢀprotein complexes
was performed with the OPLS_2005 force field and the default value for RMSD of 0.30 Å for nonꢀhydrogen
atoms was used.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Preparing the Ligand Database. A SPECS database containing 198,745 compounds was used as a ligand
database. The ligands were first filtered in Pipeline Pilot V7.5 (Pipeline Pilot, Accelrys Software Inc.: San
Diego, CA) to remove the molecules with unfavorable physicochemical properties, such as those containing
rare atoms (B, Si, Ni, Ti, Se and others), molecular weights larger than 450 and water solubility values (logS)
larger than ꢀ1 or smaller than ꢀ6. After the possibilities were filtered, there were 111,121 compounds
remaining. The threeꢀdimensional (3D) coordinates and all the stereoisomers of the ligands were generated
with LigPrep (LigPrep 2.4; Schrödinger, LLC: New York, NY, 2010), and their protonation states were
determined at a target pH 7.0 ± 2.0 with Epik in its default mode (Epik 2.1; Schrödinger, LLC: New York,
NY, 2010). The resulting structures were used as the starting point for all docking simulations.
Glide Docking Procedure. The Glide program was used to generate the grid file. The receptor grid was
defined as an enclosing box centered at the native ligand SAH to include the cofactor and substrate. Docking
was performed using Glide software (Glide 5.6, Schrödinger, LLC: New York, NY, 2010) with the standard
precision (SP) mode first, and then the top 10000 poses were reꢀdocked with the extra precision (XP) mode.
Topꢀranked compounds with GScores lower than ꢀ6.0 were then left for structural clustering with Pipeline
Pilot V7.5, in which the average number of molecules that each cluster contains was set to 20. Finally, the
compounds representing cluster centers were extracted and ranked according to their GScores, and then the
topꢀranked 51 structurally diverse compounds were selected for purchase from SPECS Corp. (The
Netherlands). The putative binding modes of DC_05 analogs were all generated using Glide in XP mode.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 43
Plasmid Construction, Protein Expression and Purification.
1
2
3
4
5
6
7
8
9
Sequence coding residues 731ꢀ1602 of mouse DNMT1 (mDNMT1) were cloned into a modified pET28a,
which encodes a SUMO tag after the Nꢀterminal His6 tag. The mDNMT1 protein was expressed in E. coli
BL21 (DE3) RIL cells. The cells were grown in L.B. (LuriaꢀBertani broth) at 37°C. When the O.D600
reached 0.6, the temperature was shifted to 15°C, and the cells were induced with 0.4 mM IPTG (isopropyl
βꢀDꢀ1ꢀthiogalactopyranoside) for 16 h. The protein was first purified through a HisTrap FF column (GE
Healthcare), and then, the His6ꢀSUMO tag was removed by ULP1 at 4°C overnight. After that, the
mDNMT1 protein was further purified through a Heparin HP column (GE Healthcare), followed by
gelꢀfiltration chromatography on a Superdex 200 10/300 column (GE Healthcare). The purified mDNMT1
protein was stored at ꢀ80°C in buffer containing 50 mM TrisꢀHCl (pH 8.0), 200 mM NaCl, 5 mM DTT, 5%
glycerol and 1 mM MgSO4.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
DNMT3A/DNMT3L, MLL1, DNMT3B, SUV39H1, SET7/9 and G9a were purchased from Shanghai
Chempartner Co., Ltd.
DNMT1 Inhibition Assays.
ELISA DNMT1 Activity Assay. All the compounds were first screened using an ELISA EpiQuik DNA
Methyltransferase (DNMT) Activity/Inhibitor Assay Kit (Epigentek). To measure the effects of the
compounds on mouse DNMT1 activity, 200 nM purified DNMT1 was incubated with 200 µM of the
different compounds and Sꢀadenosylmethionine (AdoMet) in the DNMT assay buffer in the assay plate at
37°C for 2 h.44 Next, every sample was incubated with the capture and detection antibody, followed by
incubation with developer solution for 10 mins at room temperature. The absorbance was measured at 450
nm using a POLARstar Omega microplate reader (BMG). SꢀAdenosylhomocysteine (AdoHcy) was used as
a positive control.
Radioactive Methylation Assay. The DNMT1 radioactive methylation inhibition assays were performed in
ACS Paragon Plus Environment

Page 17 of 43
Journal of Medicinal Chemistry
30 ꢁL reactions containing 0.1 µM adenosylꢀLꢀmethionine Sꢀ[methylꢀ3H] (3HꢀSAM,15 Ci/mmol, Perkin
Elmer), 0.25 µg/mL poly (dIꢀdC) · poly (dIꢀdC) (Sigma), 40 nM of DNMT1 in 50 mM TrisꢀHCl pH 8.0, 1
mM DTT, 5% glycerol and 100 ꢁg/mL BSA. The proteins were preꢀincubated with a range of compound
concentrations for 15 mins at room temperature before adding the substrate and [3H]ꢀSAM. After 60 mins of
incubating at 37°C, the reaction systems were transferred to a MultiScreen HTS Filter Plate (Millipore), and
the plate was washed 3 times with ddH2O via vacuum. The radioactivity was determined by liquid
scintillation counting (MicroBeta, Perkin Elmer). IC50 values were derived by fitting the data for the
inhibition percentage to a dose response curve by nonlinear regression in GraphPad Prism 5.0.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Methyltransferase Enzymatic Selectivity Assay
The DNMT3A/DNMT3L and DNMT3B/DNMT3L radioactive methylation inhibition assays were similar to
that of DNMT1, using 10 nM synthetic biotinylated DNA oligonucleotides (Life Technologies) as a
substrate. Sixty nM of DNMT3A/DNMT3L (BPS, 51106) or 25 nM DNMT3B/DNMT3L (BPS, 51104) was
preꢀincubated with various compound concentrations for 15 mins at room temperature before the substrate
and [3H]ꢀSAM were added. After 4 hours of incubation at room temperature, the reaction systems were
transferred to a 384ꢀwell streptavidinꢀcoated Flashplate Microplate (PerkinElmer) and then incubated for 1
hr at room temperature. The radioactivity was also determined by liquid scintillation counting (MicroBeta,
Perkin Elmer).
Methylation inhibition assays for MLL1, SUV39H1, SET7/9, and G9a were performed in modified Tris pH
9.0 buffer using AlphaLisa technology. Ten ꢁL of the reaction system contained a corresponding
concentration of SAM (Sigma) (MLL1 2 µM, SUV39H1 20 µM, SET7/9 0.22 µM and G9a 50 µM), which
was the Km value in each enzymatic reaction, plus 100 nM biotinylated peptide H3 (1ꢀ21) (synthesis by
GLChina) and the relevant enzyme concentration (2 nM MLL1, 3 nM SUV39H1, 0.2 nM SET7/9, and 0.03
nM G9a). Methylation inhibition assays for PRMT1 are similar to those of other methyltransferases in
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 43
modified Tris pH 8.0 buffer. The reaction system contained 0.8 µM SAM, 50 nM biotinylated peptide H4
(1ꢀ21) S1ac (synthesis by GLChina) and 0.1 nM PRMT1. The proteins were preꢀincubated with various
compound concentrations for 15 mins at room temperature before the substrate and SAM were added. After
60 mins of incubation at room temperature, acceptor and donor AlphaLisa beads were added according to
the manufacturer's recommendations. The signals were read in Alpha mode with an EnSpire Multimode
Plate Reader (PerkinElmer). IC50 values were derived by fitting the data for the inhibition percentage to a
doseꢀresponse curve by nonlinear regression in GraphPad Prism 5.0.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Surface Plasmon Resonance (SPR)-Based Binding Assays
The SPR binding assays were performed on a Biacore T200 instrument (GE Healthcare) at 25°C as
described.28 DNMT1 protein was covalently immobilized on a CM5 chip using a standard amineꢀcoupling
procedure in 10 mM sodium acetate (pH 5.0). The chip was first equilibrated with HBSꢀEP buffer (10 mM
HEPES (pH 7.4), 150 mM NaCl, 3 mM EDTA, 0.05% (v/v) surfactant P20, and 0.1% (v/v) DMSO)
overnight. The compounds were serially diluted with HBSꢀEP buffer and injected for 120 s (contact phase),
followed by 120 s (dissociation phase). The KD values of the tested compounds were determined by Biacore
T200 evaluation software (GE Healthcare).
Chemistry
Reagents were purchased from commercial sources and used as received. All anhydrous reactions were
performed under a nitrogen atmosphere. DMF was distilled from CaH2 prior to use. Melting points
(uncorrected) were determined on an XRCꢀ1 micro melting point apparatus. IR spectra were recorded on a
PerkinElmer Spectrum Two FTꢀIR. An HRMS was taken on a ThermoFisher LTQ Orbitrap XL instrument.
13
The 1H and C NMR experiments were performed on a Bruker AMꢀ400 spectrometer. The purities of all
tested compounds were determined by an HPLC (Agilent Technologies 1200 series) equipped with a Cꢀ18
boundedꢀphase column (Waters Symmetry C18, 4.6 mm × 250 mm, 5 um). A gradient elution was
ACS Paragon Plus Environment

Page 19 of 43
Journal of Medicinal Chemistry
performed with MeOH and water as a mobile phase, and the results were monitored at 254 nm. All tested
compounds were >95% pure.
1
2
3
4
5
6
7
Synthetic method for compounds (R)-DC_05
8
9
NH
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
H2N
NH
Cl
(R)
O
HN
NH
N
2
N
(R)
(S)
EtOH
KOH,DMF
O
HO
1
3
(S)-9-(oxiran-2-ylmethyl)-9H-carbazolyl (1). Powdered KOH (0.434 g, 6.85 mmol) was added to a
carbazole solution (1.000 g, 5.98 mmol) in DMF (6.0 mL) at ambient temperature and stirred for 30 min
until dissolved. (R)ꢀ(ꢀ)ꢀEpichlorohydrin (0.94 mL, 11.97 mmol) was added via syringe and the reaction was
stirred at room temperature overnight. Upon completion, the solution was partitioned between EtOAc and
H2O. The aqueous layer was extracted by EtOAc, and the combined organics were washed with saturated
aqueous NaCl, dried over Na2SO4, filtered, and concentrated in vacuo. The crude residue was recrystallized
from EtOAc/Hexane to yield compound 2 (475 mg, 35%).
Its characteristics were as follows: mp 75ꢀ77°C; IR (neat) 3051, 2995, 1594, 1483, 1452, 1324, 1217, and
1152 cmꢀ1; HRMSꢀESI: calcd for C15H14NO [M+H]+ 224.0997, found 224.1064; H NMR (CDCl3, 400
1
MHz) δ 8.12 (2H, s), 7.48 (4H, s), 7.29 (2H, s), 4.62ꢀ4.58 (1H, m), 4.39ꢀ4.35 (1H, m), 3.33 (1H, s),
2.79ꢀ2.78 (1H, d, J = 0.4), and 2.56 (1H, s), 13C NMR (CDCl3, 100 MHz) δ140.83 (2C), 125.99 (2C), 123.13
(2C), 120.44 (2C), 119.48 (2C), 108.90 (2C), 50.64, 45.39, and 44.68.
(R)-1-((2-(1H-indol-3-yl)ethyl)amino)-3-(9H-carbazol-9-yl)propan-2-ol (3). A solution of compound 2
(242 mg, 1.08 mmol) and tryptamine (720 mg, 4.34 mmol) in ethanol (10 mL) was stirred at room
temperature overnight. After the complete consumption of compound 2 (TLC), the solvent was evaporated
to dryness, and the residue was purified by chromatography (SiO2, 100% EtOAc) to produce compound 3 as
a white solid (177 mg, 44%).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 43
Its characteristics were as follows: mp 151ꢀ153°C; IR (neat) 3275, 3013, 2937, 1597, 1450, 1351, 1324,
1
2
3
4
1
1205, and 1111 cmꢀ1; HRMSꢀESI: calcd for C25H26N3O [M+H]+ 384.1998, found 384.2066. H NMR
5
6
7
8
(DMSOꢀd6, 400 MHz) δ 10.79 (1H, s), 8.13ꢀ8.11 (2H, d, J = 0.8), 7.63ꢀ7.61 (2H, d, J = 0.8), 7.53ꢀ7.51 (1H,
d, J = 0.8), 7.43ꢀ7.39 (2H, t, J = 0.8), 7.35ꢀ7.33 (1H, d, J = 0.8), 7.20ꢀ7.15 (3H, m), 7.08ꢀ7.05 (1H, t, J = 0.8),
6.99ꢀ6.95 (1H, t, J = 0.8), 5.04 (1H, brs), 4.50ꢀ4.45 (1H, m), 4.31ꢀ4.26 (1H, m), 4.03ꢀ4.01 (1H, t, J = 0.8),
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
13
2.85ꢀ2.82 (4H, m), 2.66ꢀ2.49 (2H, m), and 2.05 (1H, brs) C NMR (DMSOꢀd6, 100 MHz) δ140.60 (2C),
136.27, 127.30, 125.46 (2C), 122.56, 122.01, 120.80, 119.99(2C), 118.57 (2C), 118.30, 118.12, 112.57,
111.32, 109.71 (2C), 68.80, 53.04, 50.32, 47.06, and 25.55.
Synthetic method for compounds (S)-DC_05
NH
H2N
NH
Cl
(S)
O
HN
NH
N
4
N
(S)
KOH,DMF
(R)
EtOH
O
HO
1
5
(R)-9-(oxiran-2-ylmethyl)-9H-carbazole (4). Powdered KOH (0.434 g, 6.85 mmol) was added to a
carbazole solution (1.000 g, 5.98 mmol) in DMF (6.0 mL) at ambient temperature and stirred for 30 min
until dissolved. (S)ꢀ(+)ꢀEpichlorohydrin (0.94 mL, 11.97 mmol) was added via syringe and the reaction was
stirred at room temperature overnight. Upon completion, the solution was partitioned between EtOAc and
H2O. The aqueous layer was extracted by EtOAc, and the combined organics were washed with saturated
aqueous NaCl, dried over Na2SO4, filtered, and concentrated in vacuo. The residue was purified by
chromatography (SiO2, 4% EtOAc/Hexane) to yield compound 4 as a white solid (423 mg, 32%).
The compounds were as follows: mp 72ꢀ74°C; IR (neat) 3052, 2996, 1594, 1483, 1452, 1324, 1217, and
1153 cmꢀ1; HRMSꢀESI: calcd for C15H14NO [M+H]+ 224.0997, found 224.1066. 1H NMR (CDCl3, 400 MHz)
δ 8.08ꢀ8.06 (2H, d, J = 0.8), 7.47ꢀ7.41 (4H, m), 7.25ꢀ7.20 (2H, m), 4.59ꢀ4.54 (1H, m), 4.36ꢀ4.30 (1H, m),
13
3.30ꢀ3.27 (1H, m), 2.76ꢀ2.74(1H, t, J = 0.4), and 2.53ꢀ2.51 (1H, m), C NMR (CDCl3, 100 MHz) δ140.84
ACS Paragon Plus Environment

Page 21 of 43
Journal of Medicinal Chemistry
(2C), 126.01 (2C), 123.14 (2C), 120.46 (2C), 119.50 (2C), 108.90(2C), 50.66, 45.42, and 44.70.
1
2
3
4
5
6
7
8
9
(S)-1-((2-(1H-indol-3-yl)ethyl)amino)-3-(9H-carbazol-9-yl)propan-2-ol (5). A solution of compound 4
(223 mg, 1.00 mmol) and tryptamine (640 mg, 4.00 mmol) in ethanol (10 mL) was stirred at room
temperature overnight. After the complete consumption of compound 4 (TLC), the solvent was evaporated
to dryness, and the residue was purified by chromatography (SiO2, 100% EtOAc) to afford compound 5 as a
white solid (165 mg, 41%).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
It had the following characteristics: mp 158ꢀ160°C; IR 3275, 3013, 2908, 1597, 1450, 1351, 1324, 1205, and
1
1111 cmꢀ1; HRMSꢀESI: calcd for C25H26N3O [M+H]+ 384.1998, found 384.2066. H NMR (DMSOꢀd6, 400
MHz) δ 10.79 (1H, s), 8.13ꢀ8.11 (2H, d, J = 0.8), 7.63ꢀ7.61 (2H, d, J = 0.8), 7.53ꢀ7.51 (1H, d, J = 0.8),
7.43ꢀ7.39 (2H, t, J = 0.8), 7.35ꢀ7.33 (1H, d, J = 0.8), 7.20ꢀ7.15 (3H, m), 7.08ꢀ7.05 (1H, t, J = 0.8), 6.99ꢀ6.95
(1H, t, J = 0.8), 5.04 (1H, brs), 4.50ꢀ4.45 (1H, m), 4.31ꢀ4.26 (1H, m), 4.03ꢀ4.01 (1H, t, J = 0.8), 2.85ꢀ2.82
(4H, m), 2.66ꢀ2.49 (2H, m), and 2.05 (1H,brs) 13C NMR (DMSOꢀd6, 100 MHz) δ140.59 (2C), 136.26,
127.29, 125.45 (2C), 122.55, 122.00, 120.79, 119.98(2C), 118.57 (2C), 118.29, 118.11, 112.55, 111.31,
109.70 (2C), 68.79, 53.02, 50.31, 47.05, and 25.54.
Synthetic method for compounds (R)-DC_517
N
N
H2N
Cl
KOH,DMF
(R)
O
O
N
N
N
N
O
HO
OH
6
i-PrOH
(R)
(R)
HN
O
7
8
(R)-9,9'-(2-(oxiran-2-ylmethoxy)propane-1,3-diyl)bis(9H-carbazole) (7). To the solution of compound 6 (1
g, 2.56 mmol) in 20 mL of acetone, 2.4 g (25.6 mmol) of (R)ꢀ2ꢀ(chloromethyl)oxirane, 0.5 g (7.68 mmol) of
85% powdered KOH and 0.37 g (2.56 mmol) of anhydrous Na2SO4 were added. The mixture was stirred at
room temperature until compound 6 was completely consumed. The mixture was treated with EtOAc and
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 43
water. The organic layer was washed with distilled water until the wash water was neutral, and it was dried
over anhydrous Na2SO4 and filtered off. EtOAc was removed and the residue was crystallized with 40 mL of
petrol ether/EtOAc (10:1) to yield (R)ꢀ9,9'ꢀ(2ꢀ(oxiranꢀ2ꢀylmethoxy)propaneꢀ1,3ꢀdiyl)bis(9Hꢀcarbazole) (7)
as a white solid (910 mg, 79.7%).
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The compound had the following characteristics: mp 143ꢀ144°C; IR (neat) 3049, 2934, 2863, 1626, 1594,
1485, 1452, 1326, 1222, 1153, 1120, and 1061 cmꢀ1; HRMSꢀESI: calcd for C30H27N2O2 [M+H]+ 447.2073,
1
found 447.2060. H NMR (CDCl3, 400 MHz) δ8.09ꢀ8.04 (4H, m), 7.42ꢀ7.29 (6H, m); 7.24ꢀ7.18 (6H, m),
4.57ꢀ4.46 (2H, m), 4.37ꢀ4.29 (2H, m), 4.25ꢀ4.20 (1H, m), and 3.06ꢀ3.02 (1H, m); 2.98ꢀ2.94 (1H, m);
2.50ꢀ2.47 (1H, m), 2.33ꢀ2.31 (1H, t, J = 0.4); and 2.04ꢀ2.02 (1H, m). 13C NMR (CDCl3, 100 MHz) δ140.56,
140.53, 126.08 (2C), 125.99 (2C), 123.22, 123.13, 120.57 (2C), 120.49 (2C), 119.56 (2C), 119.48(2C),
109.00(8C), 78.59, 72.59, 50.39, 46.01(2C), and 44.32.
(R)-1-((1,3-di(9H-carbazol-9-yl)propan-2-yl)oxy)-3-(isopropylamino)propan-2-ol (8). A solution of
epoxide 7 (300 mg, 0.67 mmol) and isopropylamine (395 mg, 6.71 mmol, 10 equiv.) in isopropanol 4 mL
was heated to reflux for 1 h. After the complete consumption of the epoxide 7 (TLC), the solvent was
evaporated to dryness, and the crude product was purified by flash column chromatography on silica gel
(EtOAc/MeOH
20:1
as
eluent)
to
afford
(R)ꢀ1ꢀ((1,3ꢀdi(9Hꢀcarbazolꢀ9ꢀyl)propanꢀ2ꢀyl)oxy)ꢀ3ꢀ(isopropylamino)propanꢀ2ꢀol(8) as a white solid (161
mg, 47%).
The compound had the following characteristics: mp 108ꢀ110°C; IR(neat) 3627, 3303, 3049, 2930, 2868,
1626, 1595, 1484, 1452, 1325, 1223, 1153, 1120, and 1061 cmꢀ1; HRMSꢀESI: calcd for C33H36N3O2 [M+H]+
1
506.2808, found 506.2792. H NMR (CDCl3, 400 MHz) δ8.08ꢀ8.06 (4H, d, J = 0.8), 7.39ꢀ7.37 (4H, d, J =
0.8); 7.26ꢀ7.20 (8H, m), 4.55ꢀ4.50 (2H, m), 4.36ꢀ4.27 (2H, m), 3.22ꢀ3.18 (1H, m), 3.04ꢀ2.99 (2H, m);
2.47ꢀ2.41(1H, m), 2.12ꢀ1.98 (2H, m), 1.85 (2H, brs), and 0.87ꢀ0.85 (6H, d, J = 0.8). 13C NMR (CDCl3, 100
ACS Paragon Plus Environment

Page 23 of 43
Journal of Medicinal Chemistry
MHz) δ140.54 (2C), 126.06 (4C), 123.18 (2C), 120.65 (3C), 120.62 (3C), 119.56 (4C), 108.88 (3C), 108.82
(3C), 78.50, 74.35, 69.03, 44.81, 44.74, 45.75, 45.70, 22.90, and 22.83.
1
2
3
4
5
6
7
Synthetic method for compounds (S)-DC_517
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
N
N
H2N
Cl
KOH,DMF
(S)
O
O
N
N
N
N
O
HO
OH
6
i-PrOH
(S)
(S)
HN
O
9
10
(S)-9,9'-(2-(oxiran-2
-ylmethoxy)propane-1,3-diyl)bis(9H-carbazole) (9). To the solution of compound 6 (1 g, 2.56 mmol) in 20
mL of acetone 2.4 g (25.6 mmol) of (S)ꢀ2ꢀ(chloromethyl)oxirane, 0.5 g (7.68 mmol) of 85% powdered KOH
and 0.37 g (2.56 mmol) of anhydrous Na2SO4 were added. The mixture was stirred at room temperature until
compound 6 was completely consumed. The mixture was treated with EtOAc and water. The organic layer
was washed with distilled water until the wash water was neutral, and it was dried over anhydrous Na2SO4
and filtered off. The EtOAc was removed and the residue was crystallized with 40 mL of petrol ether/EtOAc
(10:1) to give (S)ꢀ9,9'ꢀ(2ꢀ(oxiranꢀ2ꢀylmethoxy)propaneꢀ1,3ꢀdiyl)bis(9Hꢀcarbazole) (9) as a white solid (840
mg, 73.6%).
The characteristics were as follows: mp 142ꢀ144°C; IR (neat) 3049, 2934, 2863, 1626, 1594, 1485, 1452,
1326, 1222, 1153, 1120, and 1061 cmꢀ1; HRMSꢀESI: calcd for C30H27N2O2 [M+H]+ 447.2072, found
447.2060. 1H NMR (CDCl3, 400 MHz) δ8.09ꢀ8.04 (4H,m), 7.42ꢀ7.29 (6H, m), 7.24ꢀ7.18 (6H, m), 4.57ꢀ4.46
(2H, m), 4.37ꢀ4.29 (2H, m), 4.25ꢀ4.20 (1H, m), 3.06ꢀ3.02 (1H, m); 2.98ꢀ2.94 (1H, m); 2.49ꢀ2.47 (1H, m),
2.33ꢀ2.31 (1H, t, J = 0.4); and 2.04ꢀ2.02 (1H, m). 13C NMR (CDCl3, 100 MHz) δ140.56 (2C), 126.08 (2C),
125.99 (2C), 123.21, 123.13, 120.57 (2C), 120.49 (2C), 119.56 (2C), 119.48 (2C),109.00 (8C), 78.58, 72.59,
50.39, 46.00 (2C), and 44.31.
(S)-1-((1,3-di(9H-carbazol-9-yl)propan-2-yl)oxy)-3-(isopropylamino)propan-2-ol (10). A solution of
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 43
epoxide 9 (300 mg, 0.67 mmol) and isopropylamine (395 mg, 6.71 mmol, 10 equiv.) in isopropanol 4 mL
was heated to reflux for 1 h. After the complete consumption of epoxide 9 (TLC), the solvent was
evaporated to dryness, and the crude product was purified by flash column chromatography on silica gel
1
2
3
4
5
6
7
8
9
(EtOAc/MeOH
20:1
as
eluent)
to
afford
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(S)ꢀ1ꢀ((1,3ꢀdi(9Hꢀcarbazolꢀ9ꢀyl)propanꢀ2ꢀyl)oxy)ꢀ3ꢀ(isopropylamino)propanꢀ2ꢀol(10) as a white solid (150
mg, 44%).
The characteristics were as follows: mp 91ꢀ93°C; IR (neat) 3627, 3303, 3049, 2930, 2868, 1626, 1595, 1484,
1452, 1325, 1223, 1153, 1120, and 1061 cmꢀ1. HRMSꢀESI: calcd for C33H36N3O2 [M+H] + 506.2807, found
506.2791. 1H NMR (CDCl3, 400 MHz) δ8.08ꢀ8.06 (4H, d, J = 0.8), 7.39ꢀ7.37 (4H, d, J = 0.8); 7.25ꢀ7.20 (8H,
m), 4.55ꢀ4.50 (2H, m), 4.37ꢀ4.28 (2H, m), 3.24ꢀ3.22 (1H, m), 3.07ꢀ2.99 (2H, m); 2.47ꢀ2.41(1H, m),
13
2.14ꢀ1.99 (4H, m); and 0.88ꢀ0.86 (6H, d, J = 0.8). C NMR (CDCl3, 100 MHz) δ140.54(2C), 126.07(4C),
123.18(2C), 120.65(3C), 120.62 (3C), 119.57 (4C), 108.89 (3C), 108.83 (3C), 78.49, 74.29, 68.87, 44.89,
44.78, 45.75, 45.69, 22.70, 22.64, 46.00 (2C), and 44.31.
Cell culture, cell proliferation assays and flow cytometry assays
HCT116 (colon cancer) and Capanꢀ1 (pancreatic cancer) cell lines were purchased from the American Type
Culture Collection (ATCC) and were maintained in McCoy’s 5A and DMEM (Life technologies)
supplemented with 10% fetal bovine serum (Life Technologies). The cells were cultured at 37°C in a
humidified atmosphere of 5% CO2 in a CO2 incubator. The compounds were dissolved in DMSO and stored
at ꢀ20°C. Control cells were treated with media containing an equal concentration of DMSO. For the cell
proliferation assay, cells were treated with different compound concentrations. HCT116 and Capanꢀ1 cell
viabilities were measured by MTT assay, and the absorbance was measured at 490 nm. The percent
proliferation was calculated by normalizing the absorbance to that of the control cells.
The cell apoptosis was determined by dual staining with 7ꢀaminoꢀactinomycin (7ꢀAAD) and annexin V
ACS Paragon Plus Environment

Page 25 of 43
Journal of Medicinal Chemistry
conjugated to phycoerythrin (PE) (BD Pharmingen). HCT116 cells were treated with 0, 0.75, 1.5 and 3 ꢁM
DC_517 for 24 hours, and then 1×106 cells were stained with annexin VꢀPE and 7ꢀAAD for 15 mins away
from the light and analyzed by flow cytometry using a LSR II cytometer (BD Pharmingen). Annexin V+
and/or 7ꢀAAD+ cells were identified as apoptotic, and cells excluding both annexin VꢀPE and 7ꢀAAD were
considered viable. The annexin Vꢀ/PIꢀ cells that were normalized to untreated controls are represented as a
percentage of live cells.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Associated content:
Supporting Information
A validation of the molecular docking; Agarose gel electrophoresis assay for compounds; The chemical
structure and Glide Score for 51 compounds selected from virtual screening; Biochemical assays for DC_05
against DNMT3A, DNMT3B, G9a, SUV39H1, MLL1, SET7/9 and PRMT1; A sequence alignment of the
catalytic site of DNMT1, DNMT3A, PRMT1, G9a, MLL1 and SET7/9; Superimposition of the catalytic
domains of PRMT1, DNMT3a and DNMT1 with (R)ꢀDC_05 and (S)ꢀDC_05 in cartoon mode;
Doseꢀresponse plots for selected compounds DC_501, DC_503, DC_504, DC_508, DC_512, DC_514,
DC_516 and DC_517 against DNMT1; The 2D ligand interaction diagram of enantiomers of DC_05 and
DC_517; The 2D ligand interaction diagram of DC_512, DC_504 and DC_508; and the cellular activity of
DC_05, (R)ꢀDC_05, (S)ꢀDC_05, DC_517, (R)ꢀDC_517 and (S)ꢀDC_517. This material is available free of
charge via the Internet at http://pubs.acs.org.
ACKNOWLEDGMENT
This work was supported by the HiꢀTech Research and Development Program of China (2012AA020302 to
CL), the National Natural Science Foundation of China (21210003 and 81230076 to HJ，81430084 to MZ,
81202398 to BL, 21472208 and 91229204 to CL) and the National Science and Technology Major Project
“Key New Drug Creation and Manufacturing Program” (2014ZX09507002ꢀ005ꢀ012 to MZ).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 43
ABBREVIATIONS USED
1
2
3
4
5
6
7
8
9
ꢁ, micro; ꢁM, micromole per liter; °C, degrees Celsius; 2D, two dimensional; 3D, three dimensional;
7ꢀAAD , 7ꢀaminoꢀactinomycin ; AdoHcy, SꢀAdenosylhomocysteine; AdoMet, Sꢀadenosylmethionine; ATCC,
American Type Culture Collection; DMEM, dulbecco's modified eagle medium; DMSO, dimethyl
sulfoxide; DNA, deoxyribonucleic acid; DNMT, DNA methyltransferase; DNMT1, DNA methyltransferase
1; DNMT3A, DNA methyltransferase 3A; DNMT3B, DNA methyltransferase 3B; DTT，dithiothreitol；
EDTA, ethylenediaminetetraacetic acid; ELISA , enzymeꢀlinked immunosorbent assay; FDA, Food and
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Drug
Administration;
GScores, Glide Scores;
h,
hour;
HEPES,
3
4ꢀ(2ꢀhydroxyethyl)ꢀ1ꢀpiperazineethanesulfonic acid; HꢀSAM, AdenosylꢀLꢀmethionine Sꢀ[methylꢀ3H]; IC50,
halfꢀmaximum inhibitory concentration; IPTG, Isopropyl βꢀDꢀ1ꢀthiogalactopyranoside; KD, equilibrium
dissociation constant; L.B., LuriaꢀBertani; mM, millimole per liter; min, minute;
MTT,
(3ꢀ(4,5ꢀDimethylthiazolꢀ2ꢀyl)ꢀ2,5ꢀdiphenyltetrazolium bromide; nM, nanomole per liter; nm, nanometer;
PDB,
Protein
Data
Bank;
RMSD, rootꢀmean square deviation; PRMT1,
protein
arginine
1;
Nꢀmethyltransferase
SAH, SꢀAdenosylꢀLꢀhomocysteine; SAM, SꢀadenosylꢀLꢀmethionine; SAR, structureꢀactivity relationship; S
P, standard precision; SPR, surface plasmon resonance; Tris, tris(hydroxymethyl)aminomethane; TSGS,
tumor suppressor genes; XP, extra precision;
Reference：
1. Yoo, C. B.; Jones, P. A. Epigenetic therapy of cancer: past, present and future. Nat Rev Drug Discov
2006, 5, 37ꢀ50.
2. Esteller, M. Epigenetics in cancer. N Engl J Med 2008, 358, 1148ꢀ1159.
3. Suzuki, T.; Tanaka, R.; Hamada, S.; Nakagawa, H.; Miyata, N. Design, synthesis, inhibitory activity,
and binding mode study of novel DNA methyltransferase 1 inhibitors. Bioorg Med Chem Lett 2010, 20,
1124ꢀ1127.
ACS Paragon Plus Environment

Page 27 of 43
Journal of Medicinal Chemistry
4. Siedlecki, P.; Garcia Boy, R.; Musch, T.; Brueckner, B.; Suhai, S.; Lyko, F.; Zielenkiewicz, P. Discovery
of two novel, smallꢀmolecule inhibitors of DNA methylation. J Med Chem 2006, 49, 678ꢀ683.
5. Jones, P. A.; Baylin, S. B. The epigenomics of cancer. Cell 2007, 128, 683–692.
6. Goll, M. G.; Bestor, T. H. Eukaryotic cytosine methyltransferases. Annu Rev Biochem 2005, 74,
481ꢀ514.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
7. Bestor, T. H. The DNA methyltransferases of mammals. Hum Mol Genet 2000, 9, 2395ꢀ2402.
8. Leonhardt, H.; Page, A. W.; Weier, H. U.; Bestor, T. H. A targeting sequence directs DNA
methyltransferase to sites of DNA replication in mammalian nuclei. Cell 1992, 71, 865ꢀ873.
9. Castellano, S.; Kuck, D.; Viviano, M.; Yoo, J.; LopezꢀVallejo, F.; Conti, P.; Tamborini, L.; Pinto, A.;
MedinaꢀFranco, J. L.; Sbardella, G. Synthesis and biochemical evaluation of delta(2)ꢀisoxazoline derivatives
as DNA methyltransferase 1 inhibitors. J Med Chem 2011, 54, 7663ꢀ7677.
10. Egger, G.; Liang, G.; Aparicio, A.; Jones, P. A. Epigenetics in human disease and prospects for
epigenetic therapy. Nature 2004, 429, 457ꢀ463.
11. Stresemann, C.; Lyko, F. Modes of action of the DNA methyltransferase inhibitors azacytidine and
decitabine. Int J Cancer 2008, 123, 8ꢀ13.
12. Gros, C.; Fahy, J.; Halby, L.; Dufau, I.; Erdmann, A.; Gregoire, J. M.; Ausseil, F.; Vispe, S.; Arimondo,
P. B. DNA methylation inhibitors in cancer: recent and future approaches. Biochimie 2012, 94, 2280ꢀ2296.
13. Fang, M. Z.; Chen, D.; Sun, Y.; Jin, Z.; Christman, J. K.; Yang, C. S. Reversal of hypermethylation and
reactivation of p16INK4a, RARbeta, and MGMT genes by genistein and other isoflavones from soy. Clin
Cancer Res 2005, 11, 7033ꢀ7041.
14. Fang, M. Z.; Wang, Y.; Ai, N.; Hou, Z.; Sun, Y.; Lu, H.; Welsh, W.; Yang, C. S. Tea polyphenol
(ꢀ)ꢀepigallocatechinꢀ3ꢀgallate inhibits DNA methyltransferase and reactivates methylationꢀsilenced genes in
cancer cell lines. Cancer Res 2003, 63, 7563ꢀ7570.
15. Fagan, R. L.; Cryderman, D. E.; Kopelovich, L.; Wallrath, L. L.; Brenner, C. Laccaic acid A is a direct,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 43
DNAꢀcompetitive inhibitor of DNA methyltransferase 1. J Biol Chem 2013, 288, 23858ꢀ23867.
1
2
3
4
5
6
7
8
9
16. Zambrano, P.; SeguraꢀPachero, B.; PerezꢀCardenas, E.; Cetina, L.; RevillaꢀVazquez, A. R.; TajaꢀChayeb,
L. A. T.; ChavezꢀBlanco, A.; Angeles, E.; Cabrera, G.; Sandoval, K.; TrejoꢀBecerril, C.; ChanonaꢀVilchis, J.;
DuenasꢀGonzalez, A. A phase I study of hydralazine to demethylate and reactivate the expression of tumor
suppressor genes. Bmc Cancer 2005, 5, 44ꢀ56.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
17. VillarꢀGarea, A.; Fraga, M. F.; Espada, J.; Esteller, M. Procaine is a DNAꢀdemethylating agent with
growthꢀinhibitory effects in human cancer cells. Cancer Res 2003, 63, 4984ꢀ4989.
18. Cornacchia, E.; Golbus, J.; Maybaum, J.; Strahler, J.; Hanash, S.; Richardson, B. Hydralazine and
procainamide inhibit Tꢀcell DNA methylation and induce autoreactivity. J immunol 1988, 140, 2197ꢀ2200.
19. Brueckner, B.; Garcia Boy, R.; Siedlecki, P.; Musch, T.; Kliem, H. C.; Zielenkiewicz, P.; Suhai, S.;
Wiessler, M.; Lyko, F. Epigenetic reactivation of tumor suppressor genes by a novel smallꢀmolecule
inhibitor of human DNA methyltransferases. Cancer Res 2005, 65, 6305ꢀ6311.
20. Datta, J.; Ghoshal, K.; Denny, W. A.; Gamage, S. A.; Brooke, D. G.; Phiasivongsa, P.; Redkar, S.; Jacob,
S. T. A new class of quinolineꢀbased DNA hypomethylating agents reactivates tumor suppressor genes by
blocking DNA methyltransferase 1 activity and inducing its degradation. Cancer Res 2009, 69, 4277ꢀ4285.
21. Seidel, C.; Florean, C.; Schnekenburger, M.; Dicato, M.; Diederich, M. Chromatinꢀmodifying agents in
antiꢀcancer therapy. Biochimie 2012, 94, 2264ꢀ2279.
22. Chuang, J. C.; Yoo, C. B.; Kwan, J. M.; Li, T. W. H.; Liang, G. N.; Yang, A. S.; Jones, P. A. Comparison
of biological effects of nonꢀnucleoside DNA methylation inhibitors versus 5ꢀazaꢀ2 'ꢀdeoxycytidine. Mol
Cancer Ther 2005, 4, 1515ꢀ1520.
23. Stresemann, C.; Brueckner, B.; Musch, T.; Stopper, H.; Lyko, F. Functional diversity of DNA
methyltransferase inhibitors in human cancer cell lines. Cancer Res 2006, 66, 2794ꢀ2800.
24. Valente, S.; Liu, Y. W.; Schnekenburger, M.; Zwergel, C.; Cosconati, S.; Gros, C.; Tardugno, M.;
Labella, D.; Florean, C.; Minden, S.; Hashimoto, H.; Chan, Y. Q.; Zhang, X.; Kirsch, G.; Novellino, E.;
ACS Paragon Plus Environment

Page 29 of 43
Journal of Medicinal Chemistry
Arimondo, P. B.; Miele, E.; Ferretti, E.; Gulino, A.; Diederich, M.; Cheng, X. D.; Mai, A. Selective
nonꢀnucleoside inhibitors of human DNA methyltransferases active in cancer Including in cancer stem cells.
J Med Chem 2014, 57, 701ꢀ713.
1
2
3
4
5
6
7
8
9
25. Asgatay, S.; Champion, C.; Marloie, G.; Drujon, T.; SenamaudꢀBeaufort, C.; Ceccaldi, A.; Erdmann, A.;
Rajavelu, A.; Schambel, P.; Jeltsch, A.; Lequin, O.; Karoyan, P.; Arimondo, P. B.; Guianvarc'h, D. Synthesis
and evaluation of analogues of NꢀPhthaloylꢀLꢀtryptophan (RG108) as inhibitors of DNA methyltransferase 1.
J Med Chem 2014, 57, 421ꢀ434.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
26. Liang, Z.; Ding, X.; Ai, J.; Kong, X.; Chen, L.; Chen, L.; Luo, C.; Geng, M.; Liu, H.; Chen, K.; Jiang, H.
Discovering potent inhibitors against cꢀMet kinase: molecular design, organic synthesis and bioassay. Org
Biomol Chem 2012, 10, 421ꢀ430.
27. Kong, X.; Qin, J.; Li, Z.; Vultur, A.; Tong, L.; Feng, E.; Rajan, G.; Liu, S.; Lu, J.; Liang, Z.; Zheng, M.;
Zhu, W.; Jiang, H.; Herlyn, M.; Liu, H.; Marmorstein, R.; Luo, C. Development of a novel class of
BꢀRaf(V600E)ꢀselective inhibitors through virtual screening and hierarchical hit optimization. Org Biomol
Chem 2012, 10, 7402ꢀ7417.
28. Chen, B.; Ye, F.; Yu, L.; Jia, G.; Huang, X.; Zhang, X.; Peng, S.; Chen, K.; Wang, M.; Gong, S.; Zhang,
R.; Yin, J.; Li, H.; Yang, Y.; Liu, H.; Zhang, J.; Zhang, H.; Zhang, A.; Jiang, H.; Luo, C.; Yang, C. G.
Development of cellꢀactive N6ꢀmethyladenosine RNA demethylase FTO inhibitor. J Am Chem Soc 2012,
134, 17963ꢀ17971.
29. Wang, J.; Chen, L.; Sinha, S. H.; Liang, Z.; Chai, H.; Muniyan, S.; Chou, Y. W.; Yang, C.; Yan, L.; Feng,
Y.; Li, K. K.; Lin, M. F.; Jiang, H.; Zheng, Y. G.; Luo, C. Pharmacophoreꢀbased virtual screening and
biological evaluation of small molecule inhibitors for protein arginine methylation. J Med Chem 2012, 55,
7978ꢀ7987.
30. Li, K. K.; Luo, C.; Wang, D.; Jiang, H.; Zheng, Y. G. Chemical and biochemical approaches in the study
of histone methylation and demethylation. Med Res Rev 2012, 32, 815ꢀ867.
ACS Paragon Plus Environment