Development of Potent Type i Protein Arginine Methyltransferase (PRMT) Inhibitors of Leukemia Cell Proliferation

Article abstract of DOI:10.1021/acs.jmedchem.7b01134

Protein Arginine Methyltransferases (PRMTs) are crucial players in diverse biological processes, and dysregulation of PRMTs has been linked to various human diseases, especially cancer. Therefore, small molecules targeting PRMTs have profound impact for both academic functional studies and clinical disease treatment. Here, we report the discovery of N¹-(2-((2-chlorophenyl)thio)benzyl)-N¹-methylethane-1,2-diamine (28d, DCPR049-12), a highly potent inhibitor of type I PRMTs that has good selectivity against a panel of other methyltransferases. Compound 28d effectively inhibits cell proliferation in several leukemia cell lines and reduces the cellular asymmetric arginine dimethylation levels. Serving as an effective inhibitor, 28d demonstrates the mechanism of cell killing in both cell cycle arrest and apoptotic effect as well as downregulation of the pivotal mixed lineage leukemia (MLL) fusion target genes such as HOXA9 and MEIS1, which reflects the critical roles of type I PRMTs in MLL leukemia. These studies present 28d as a valuable inhibitor to investigate the role of type I PRMTs in cancer and other diseases.

Full text of DOI:10.1021/acs.jmedchem.7b01134

Subscriber access provided by UNIVERSITY OF ADELAIDE LIBRARIES
Article
Development of Potent Type I Protein Arginine Methyltransferase
(PRMT) Inhibitors Inhibiting Leukemia Cells Proliferation
Chen Wang, Hao Jiang, Jia Jin, Yiqian Xie, Zhifeng Chen, Hao Zhang, Fulin
Lian, Yu-Chih Liu, Chenhua Zhang, Hong Ding, Shijie Chen, Naixia Zhang,
Yuanyuan Zhang, Hualiang Jiang, Kaixian Chen, Fei Ye, Zhiyi Yao, and Cheng Luo
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.7b01134 • Publication Date (Web): 11 Oct 2017
Downloaded from http://pubs.acs.org on October 12, 2017
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 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Development of Potent Type I Protein Arginine
Methyltransferase (PRMT) Inhibitors Inhibiting
Leukemia Cells Proliferation
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
a,b,c#
b,c#
a,#
b
d
b
Chen Wang
, Hao Jiang , Jia Jin , Yiqian Xie , Zhifeng Chen , Hao Zhang , Fulin
b
e
e
b
b
b
Lian , YuꢀChih Liu , Chenhua Zhang , Hong Ding , Shijie Chen , Naixia Zhang ,
b,
b
b,d
a,
f*
Yuanyuan Zhang *, Hualiang Jiang , Kaixian Chen , Fei Ye *, Zhiyi Yao and Cheng
b,
Luo *
a
College of Life Sciences, Zhejiang SciꢀTech University, Hangzhou, China
b
Drug Discovery and Design Center, CAS Key Laboratory of Receptor Research, State
Key Laboratory of Drug Research, Shanghai Institute of Materia Medica, Chinese
Academy of Sciences, 555 Zuchongzhi Road, Shanghai 201203, China
c
University of Chinese Academy of Sciences, 19 Yuquan Road, Beijing 100049, China
d
School of Life Science and Technology, ShanghaiTech University, 100 Haike Road,
Shanghai 201210, China
e
In Vitro Biology, Shanghai ChemPartner Life Science Co., Ltd., #5 Building, 998, Halei
Road, Shanghai 201203, China
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
f
College of Chemical and Environmental Engineering, Shanghai Institute of Technology,
Shanghai, 210032, China
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ACS Paragon Plus Environment

Page 3 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
ABSTRACT: Protein Arginine Methyltransferases (PRMTs) are crucial players in
diverse biological processes and dysregulation of PRMTs has been linked to various
human diseases, especially cancer. Therefore, small molecules targeting PRMTs have
profound impact for both academic functional studies and clinical disease treatment.
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
Here,
we
report
the
discovery
of
1
1
N ꢀ(2ꢀ((2ꢀchlorophenyl)thio)benzyl)ꢀN ꢀmethylethaneꢀ1,2ꢀdiamine (28d, DCPR049_12),
a highly potent inhibitor of type I PRMTs that has good selectivity against a panel of
other methyltransferases. 28d effectively inhibits cell proliferation in several leukemia
cells and reduces the cellular asymmetric arginine dimethylation levels. Serving as an
effective inhibitors, 28d underlies the mechanism of cell killing in both cell cycle arrest
and apoptotic effect as well as downregulation of the pivotal mixed lineage leukemia
(MLL) fusion target genes such as HOXA9 and MEIS1, which reflects the critical roles of
type I PRMTs in MLL leukemia These studies represent 28d as a valuable inhibitor to
investigate the role of type I PRMTs in cancer and other diseases.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
INTRODUCTION
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The involvement of PRMTs in human tumorigenesis has been broadly reported. PRMTs
expression is upregulated and associated with aggressive clinical features and poor
1, 2
survival in human malignancies. Overexpression of PRMT1 promotes the survival and
invasion of cancer cells while knockdown of PRMT1 has been shown to arrest the growth
3
ꢀ7
of cancer cells. In addition, accumulating evidence suggests the important roles of
8
PRMT1 in malignant hematopoiesis. Chromosomal translocation at t(8;21) generates a
fusion protein acute myelogenous leukemia 1ꢀRUNX1 translocation partner 1
(AML1ꢀETO), which acts as an oncogenic transcription factor and occurs in 15% of de
9
novo acute myeloid leukemia (AML) cases. PRMT1 can interact and methylate
AML1ꢀETO fusion protein, promoting its transcriptional activation and selfꢀrenewal
1
0
capability. PRMT1 can also cooperate with the MLL fusion proteins such as lysine
methyltransferase 2AꢀSH3 domain containing GRB2 like 1 (MLLꢀEEN) or lysine
11, 12
methyltransferase 2Aꢀgrowth arrest specific 7 (MLLꢀGAS7) in MLL leukemia.
Moreover, the enzymatic activity of PRMT1 has shown to be essential for MLLꢀmediated
transformation, as evidenced by the disability of the transforming capacity in the PRMT1
catalyticꢀdead mutant. In addition, PRMT4 (also known as coactivator associated
arginine methyltransferase 1 (CARM1)) and PRMT6 dysregulation have been reported to
ACS Paragon Plus Environment

Page 5 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2
, 13
be involved in diverse cancer types including colorectal, bladder and lung cancer.
These findings not only unveil the critical functions of PRMTs in oncogenesis but also
make PRMT family of enzymes promising therapeutic targets in drug discovery.
Therefore, PRMTs modulators are valuable compounds to both facilitate the dissection of
PRMTs biological function and speed up the development of candidate drug leads
targeting PRMTs.
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
So far nine PRMTs have been well reported in mammalian cells, and these enzymes
1
4
can be divided into three types. Type I PRMTs consisting of PRMT1, 2, 3, 4, 6 and 8
catalyze the transfer of methyl group(s) from SꢀadenosylꢀLꢀmethionine (SAM) onto
arginine residues to form monomethylation (MMA) and asymmetric dimethylation
(
ADMA) while type II PRMTs consisting of PRMT5 and 9 catalyze formation of monoꢀ
1
5, 16
and symmetric dimethylation (SDMA).
Type III PRMTs, which only contains
1
7
PRMT7, generates monomethylation of arginine. Among these enzymes, PRMT1 is the
predominant type I methyltransferase which accounts for ~ 90% of all cellular arginine
methylation events and has a wide range of substrate spectrum including histone H4
residue Arg 3 (H4R3), transcriptional factor forkhead box O1 (FOXO1) and abundant
1
8ꢀ20
proteins involving in RNA processing.
Through methylation of various substrates,
PRMT1 is reported to impact on diverse cellular processes including DNA damage
21ꢀ24
repair, RNA splicing process, signal transduction and gene transcription.
Besides,
application of the panꢀdimethyl arginine antibodies has benefited the functional study of
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
PRMTs. Researches with these antibodies have shown that loss of PRMT1 activity
switches the substrate arginine methylation status from asymmetry to symmetry in
2
5
cellular level. These validated antibodies are powerful tools in chemical biological
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
evaluations of PRMTs inhibition in cells.
26
To date, many PRMT inhibitors have been reported (Chart 1). Early studies often
used Sꢀadenosyl methionine (SAM) anologues (including Sꢀadenosylhomocysteine
2
7
28
(SAH),
5'ꢀdeoxyꢀ5'ꢀ(1,4ꢀdiaminoꢀ4ꢀcarboxybutyl)adenosine (1, sinefungin), and
2
9
methylthioadenosine (2, MTA) ) to explore PRMT1 biology in various cell models.
Since SAM is a common methyl group donor, these panꢀmethyltransferase inhibitors may
have the offꢀtarget effects through inhibiting other melthyltransferases, leading to
3
0
confounded phenotypes to interpret. Compound 3 (AMIꢀ1) was the first molecular that
specifically inhibited type I PRMTs (PRMT1, 3, 4 and 6) activity at low micromolar level
in vitro and has been widely used as a chemical tool under various biological conditions.
In recent years, many efforts have been made to seek more potent and selective inhibitors
31
32
33
toward type I PRMTs. Compounds like 4 (DCLX069), 5 (DB75), 6 (MHIꢀ21) show
improved activity and selectivity compared with compound 3, however, their biochemical
IC50 values trap in the micromolar range. In addition, PRMT4 (CARM1) inhibitor 7
3
4
(
CMPDꢀ1), with the potency at nanomolar range, has claimed good selectivity over
other PRMTs. Notably, while this work is in preparation, PRMT6 inhibitor 8
35
36
(
EPZ020411) and type I PRMT inhibitor 9 (MS023) have been identified. Our
ACS Paragon Plus Environment

Page 7 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
previous virtual screening work identified compound 4 as PRMT1 inhibitor which
3
1
theoretically occupied SAMꢀbinding site to exert the inhibitory effect. In this study, we
fineꢀtune our strategies and focus on the substrateꢀbinding site to seek compounds with
enhanced activity both in vitro and in 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
Chart 1. Chemical structures of the reported PRMT inhibitors.
Here, based on virtual screening and chemical optimizations, we identified two potent
type
I
PRMT
inhibitors,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
1
N ꢀ(2ꢀ((2ꢀchlorophenyl)thio)benzyl)ꢀN ꢀmethylethaneꢀ1,2ꢀdiamine (28d, DCPR049_12)
1
1
2
and N -(2ꢀ((2ꢀchlorophenyl)thio)benzyl)-N ,N -dimethylethaneꢀ1,2ꢀdiamine
(31d,
DCPR049_13) with biochemical IC50 values of 5.3 nM and 6.0 nM respectively.
Moreover, 28d and 31d exhibited potent inhibitory effect on leukemia cells and altered
the endogenous arginine methylation levels. They modulated key leukemogenic
homeobox genes, thus reflecting the critical role of type I PRMTs in human MLL
leukemia. Due to their high potency and selectivity against type I PRMTs, 28d and 31d
will be valuable compounds to dissect type I PRMT biology as well as facilitate
PRMTsꢀguided drug discovery progress.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
RESULTS
Discovery of a lead compound 13a (DCPR049) through virtual screening
In our previous study, we identified several inhibitors of PRMT1 based on virtual
31
screening targeting SAMꢀbinding site. For achieving inhibitors with better activity and
selectivity, an optimized structureꢀbased virtual screening strategy which focused on the
substrateꢀbinding pocket was adopted to identify smallꢀmolecule PRMT1 inhibitors with
novel scaffolds (Figure 1A). To date, the crystal structure of human PRMT1 (hPRMT1)
has not been determined. Given the highly conserved sequences shared among hPRMT1,
rat PRMT1 (rPRMT1) and human PRMT6 (hPRMT6) (Figure S1) as well as the
ACS Paragon Plus Environment

Page 9 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
conformation similarity of the Nꢀterminal helix αX domain which is important for
PRMT1 activity, we generated a homology model for hPRMT1 based on the Xꢀray
2
9
35
structure of rPRMT1 (PDB ID: 1OR8) and hPRMT6 (PDB ID: 4Y30) . The modeled
hPRMT1 structure was used as a target to perform molecular docking screening against
SPECS database (http://www.specs.net). To refine the database, we filtered it by
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
3
7ꢀ40
Lipinski’s rule of five and removed panꢀassay interference compounds (PAINS).
The
remaining 182,014 compounds were subsequently docked into the substrate arginine
binding site using energy scoring function of DOCK4.0. The topꢀranked 10500
candidates were evaluated and ranked by Glide SP and XP modes integrated in Maestro
41
9
.2, resulting a list of 600 compounds. To ensure the scaffold diversity in the hits, the
topꢀranked 600 molecules were clustered into 100 groups with Pipeline Pilot, version 7.5
(Pipeline Pilot; Accelrys Software Inc., San Diego, CA). Finally, 107 compounds were
purchased from Specs Company for further biochemical evaluations.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 1. Virtualꢀscreening procedures and PRMT1 activity assay results in vitro. A.
Flowchart of the integrated virtual screening for PRMT1 inhibitors. B. Inhibitory activity
of top 23 compounds from virtual screening determined by TRꢀFRET.
Inhibition of PRMT1 enzymatic activity and characterization in biophysical assay
The 107 candidate molecules selected by virtual screening were tested for PRMT1
inhibition to determine their biochemical activities. First, Timeꢀresolved Flourescence
Resonance Energy Transfer (TRꢀFRET) assay was carried out. The reactions were run
with Hisꢀtagged rPRMT1 (11ꢀ353) as the enzyme and biotinylated H4 peptide (1ꢀ21) as
the substrate. The reaction mixture was treated with europiumꢀlabeled histone H4
arginine 3 asymmetric dimethylation (H4R3me2a) antibody and Ulight streptavidin and
the final measurement was performed to detect methylation products. From the enzyme
ACS Paragon Plus Environment

Page 11 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
inhibition assay, we chose 7 compounds for their ability to inhibit PRMT1 activity by
>80% at the concentration of 100 ꢁM (Figure 1B and Figure S2). To further validate the
activities of the 7 compounds, we determined their IC50 values against PRMT1 by
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
3
Hꢀlabeled radioactive methylation assay (Figure 2AꢀB and Figure S3). Amongst them,
compound 13a (Figure 2A) exhibited the highest potency with an IC value of 3.0 ꢁM.
5
0
To confirm the binding between 13a and PRMT1, both CarrꢀPurcellꢀMeiboomꢀGill
(CPMG) and saturation transfer difference (STD) NMR experiments were performed. As
shown in Figure 2CꢀD, a strong signal in STD experiment and a significant decreased
signal in CPMG spectrum were detected under the condition of 5 ꢀM PRMT1 and 200
ꢀ
M compound 13a, indicating the mutual binding between PRMT1 and the compound.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Figure 2. Structures of compounds and biochemical or biophysical assay results. A.
Structure of 13a and IC50 determination for 13a against PRMT1 by radioisotope. B.
Structures of four compounds with poor activity against PRMT1 determined by
radioisotope. C. STD NMR experiments of 5 ꢁM PRMT1 in the presence of 200 ꢁM
compound 13a. D. Dose dependent CPMG spectra for 200 ꢁM compound 13a (red) in
the presence of 5 ꢁM PRMT1 (blue), 2 ꢁM PRMT1 (green) and 1 ꢁM PRMT1 (aqua).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Analysis of the binding mode and chemical modification of compound 13a
To gain a better understanding of the molecular basis for the inhibitory activity of
42
compound 13a toward PRMT1, a putative binding mode was docked by Glide with XP
4
3
mode . As shown in Figure 3A, compound 13a occupied the negativeꢀcharged binding
site of the substrate arginine. The ethylenediamine side chain of this compound formed
hydrophobic interactions with E144, Y148, E153 and S154, which were critical for
ACS Paragon Plus Environment

Page 13 of 71
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
Figure 3. Predicted binding mode and structural modification of compound 13a. A.
Superimposition of the binding modes of 13a and substrate arginine (PDB ID: 1OR8).
Compound 13a is shown as magenta sticks, and substrate arginine is shown as cyan
sticks, with the surface of PRMT1 depicted in vacuum electrostatics. B. Binding mode of
13a. It is shown as magenta sticks, and key residues are shown as gray sticks. Hydrogen
bonds are shown as a red dashed line. C. Schematic diagram showing putative
interactions between PRMT1 and 13a. Residues involved in the hydrophobic interactions
are shown as starbursts, and hydrogenꢀbonding interactions are denoted by dotted green
lines. D. Chemical modification of 13a, which leads to 28d and 31d with enhanced
activities. The bold bonds mimic side chain of substrate arginine.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
methylation catalysis. It has been reported that mutations of E144 and E153 could
29
significantly reduce or totally abolish the activity of PRMT1. Besides, a hydrogen bond
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
was found between the nitrogen in the ethylenediamine moiety and the OH group of
residue Y39, which was located at the Nꢀterminal helix αX (Figure 3B). There were also
some hydrophobic interactions between the naphthalene ring of 13a and residues in αX
and the adjacent αY (Figure 3C).
Table 1. Chemical Modification of compound 13a
No.
R1
R2
R3
H
R4
R5
IC (ꢁM)
50
2
2
3a
4a
> 200
34
H
H
H
2
3b
H
H
> 50
> 50
24b
ACS Paragon Plus Environment

Page 15 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2
5b
H
3.6
2
2
2
3c
4c
5c
H
> 200
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
H
H
H
45
0.71
> 50
54
2
2
3d
4d
H
H
H
H
H
H
25d
28d
31d
0.39
0.0053
0.006
H
H
H
To improve the molecular inhibitory activities against PRMT1, we synthesized a
series of phenyl sulfide derivatives for investigating the structure activity relationship
(
SAR) based on the predicted binding mode of lead compound (Figure 3D). We found
that the substitution of naphthalene ring with benzene ring resulting in 25b displayed
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
similar activity to 13a, with an IC value of 3.6 ꢁM (Table 1, potency (IC ) data shown
5
0
50
in table 1 were measured using radioisotope assay). However, introduction of carbonyl to
the ethylenediamine side chain (R2 position) or replacement of the methyl group with
hydrogen atom (R3 position) totally abolished or remarkably reduced the activity of
inhibitors (23a, 24a, 23b, 24b Table 1), indicating the essential role of the tertiary amine
in the middle of the ethylenediamine chain. Next, we explored introduction of halogen
atom to the benzene ring of 25b (Figure 3D). Halogen bond, a specific intermolecular
interaction between a halogen atom and an electronꢀrich partner (O, N, S, or π), is a
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4
4
distinct class of hydrogenꢀbondꢀlike interactions. Compounds with halogen atom
attached to the benzene ring may form halogen bonding with surrouding residues (e.g,
Y35, S38). As expected, the inhibitory activity of 25c, with orthoꢀsubstituted fluorine
atom on the benzene ring in R1 position, was improved by about 5ꢀfold. Introducing a
chlorine atom to benzene ring also increased activity (25d, IC = 0.39 ꢁM, Table 1). As
5
0
shown in Figure 3A, the ethylenediamine moiety of 13a mimics the conformation of
4
5
substrate arginine side chain.
It is known that in the typical S 2ꢀfavored catalytic
N
process of PRMT1, the first and second methyl are transferred onto the substrate arginine
4
5
one after another. We also noticed PRMT4 (CARM1) inhibitor CMPDꢀ1 features a
similar side chain which was directed towards the bottom of the arginineꢀbinding
34
cavity. Hence we wanted to mimic the substrate changes during the catalytic process.
We designed and synthesized compounds 28d and 31d, which replaced the terminal
tertiary amine with primary amine or secondary amine respectively (Figure 3D and
ACS Paragon Plus Environment

Page 17 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Scheme 1). The IC values of 28d and 31d were determined using radiolabel assay to be
5
0
5.3 nM and 6.0 nM, respectively (Table 1). Strikingly, the introduction of hydrogenꢀbond
donor significantly improved the activity of inhibitors. The molecular modeling of 28d
implied that the improved affinity is probably contributed by hydrogen bonds and
electrostatic attraction interactions between the tail group of the inhibitors and the
negativeꢀcharged binding pocket which is mainly formed by E144, M146, Y148, E153,
as well as halogen bond with Y35 located at the helix αX (Figure S4). While we were
performing the biological experiments, two compounds 8 and 9 with the similar side
chain were reported, which further confirmed the ethylenediamino group is an arginine
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
3
5, 36
mimetic and an excellent moiety for targeting type
I
PRMTs.
What’s more, the
optimized compounds 28d and 31d were again evaluated by both CPMG and STD
experiments. The results showed obvious binding signals between PRMT1 and both of
the compounds, confirming the inhibitory potency (Figure S5).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
Scheme 1. Reagents and conditions: (i) Cu, K2CO3, Aryl halides, DMF, reflux, 12 h. (ii)
Oxalyl dichloride, DMF, CH2Cl2, 0 ℃ to RT, h. (iii)
4
N1,N1ꢀdimethylethaneꢀ1,2ꢀdiamine, Ph, RT, 12 h. (iv) NaBH4, (C2H5)2O·BF3, THF, 1
h at RT and 3 h reflux. (v) HCOOH, 36% HCHO, 100℃, 3 h. (vi) Tertꢀbutyl
(
2ꢀ(methylamino)ethyl)carbamate, NaBH(OAc)3, toluene, 0℃ to RT, 4 h. (vii) NaBH4,
(
C2H5)2O·BF3, THF, 1 h at RT and 3 h reflux. (viii) HCl/dioxane, RT, 16 h. (ix) NaH,
ACS Paragon Plus Environment

Page 19 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
MeI, THF, 0℃ to RT, 17 h. (x) NaBH4, (C2H5)2O·BF3, THF, 1 h at RT and 3 h reflux.
(xi) HCl/dioxane, RT, 16 h.
Overall, based on combination of molecular modeling and chemical modifications, we
improved the potency of PRMT1 inhibitors by over 500ꢀfold, highlighting the efficiency
of structural optimizations.
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
Mechanism of action (MOA) studies and selectivity of 28d
To study MOA of 28d, we performed a kinetic experiment using radiolabel method to
determine the potency (IC ) of 28d using various concentrations of SAM or peptide. As
50
shown in Figure S6, 28d showed uncompetitive inhibition with respect to cofactor SAM
on PRMT1, and showed noncompetitive inhibition with respect to peptide on PRMT1.
The uncompetitive pattern of inhibition on SAM has also been observed in the study by
36
Eram et al. where they explored MOA of a similar inhibitor 9 against PRMT3.
However, in their study, compound 9 displayed noncompetitive pattern against PRMT1
with respect to SAM. The difference might be due to the lower concentrations of SAM
used in our assay. In the crystal structure (PDB ID: 1OR8), the cofactor SAM binding
site was around the peptide substrate binding site, we suggest that SAM may assist in
stabilizing peptide substrate binding pocket, which is benefit for the binding of 28d to
PRMT1. In other words, there is a preferred binding sequence for productive
PRMT1ꢀSAMꢀcompounds complex formation in which SAM binds before the
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
compounds. While at the low concentration of SAM, the protein may be flexible and
could hardly bind the compounds. Thus the inhibitory activity towards PRMT1 of 28d
was much weaker at low SAM concentrations. This possibility was also supported by the
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
4, 36
ITC experiments conditions of 7 and 9, in which SAH or SAM were added.
In the
34
study of compound 7, Sack et al. showed the ITC results in the absence of SAH,
demonstrating that 7 cannot bind to PRMT4 without SAH. Considering the conserved
structure among type I PRMTs, we suggest cofactor binding will induce structural
changes of PRMT1 that leads to appropriate formation of the arginineꢀbinding pocket.
Thus at very low concentrations of SAM will reduce the binding efficiency of 28d, which
results in the uncompetitive inhibition with respect to cofactor SAM.
Notably, we acquired the same result on the noncompetitive mode of inhibition with
respect to peptide towards PRMT1 like 9. However, in the previous studies like
3
4ꢀ36
compound 7, 8 and 9,
the crystal structures clearly showed the ethylenediamine
moiety mimics the conformation of substrate arginine side chain and occupies the peptide
substrate binding site so that we suggest 28d may share the similar MOA with them since
the ethylenediamine side chain of 28d is similar with them. And the possible explanation
about the contradictory results is there might be conformation change about the αꢀhelix at
the Nꢀterminus of the protein. As shown in crystal structures of type I PRMTs with
inhibitors, the αꢀhelix was completely floded and locked the substrate binding pocket,
which might prevent the binding of peptide. Thus the proteins once bind the compounds
ACS Paragon Plus Environment

Page 21 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
and form a closed state with compounds. Then peptide would not enter the binding cavity
by competing the compounds, even in a high concentration of peptide substrate.
In summary, although the kinetic experiment didn’t show competitive inhibition with
respect to peptide, we still can predict the binding model, in which 28d occupies the
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
6
, 10, 28
substrate binding site of PRMT1, according to the previous studies.
To assess the selectivity of 28d, we evaluated the inhibitory activities of 28d against
other PRMT family enzymes (PRMT3, PRMT4, PRMT5, PRMT6, PRMT8) and a panel
of other 10 methyltransferases and acetyltransferases at the concentrations of 50 ꢀM and
1
00 ꢀM (Figure 4). We found 28d exhibited obvious and diverse inhibitory activities
towards type I PRMTs, demonstrating its role as a pan inhibitor. However, 28d showed
minimal inhibitory effects on the other methyltransferases including 7 PKMTs (EZH2,
DOT1L, G9a, MLL1, PRDM9, SMYD2 and SUV39H1) as well as DNA
methyltransferases (DNMT1 and DNMT3A/3L) and histone acetyltransferase GCN5,
indicating its good selectivity against other methyltransferases (Figure 4G).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ACS Paragon Plus Environment

Page 23 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Figure 4. Determination of the selectivity of 28d in biochemical assays. AꢀF. IC50
determination of 28d against PRMT1, PRMT3, PRMT4, PRMT5, PRMT6 and PRMT8
using AlphaLISA method. G. Determination of 28d against GCN5, DOT1L, EZH2,
DNMT3A/3L, DNMT1, G9a, MLL1, PRDM9, SMYD2 and SUV39H1 at 50 ꢀM and
100 ꢀM.
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
Inhibitory effect on the proliferation of leukemia cells
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Figure 5. 28d decreased the cell viability in leukemia cells and knockdown of PRMT1
inhibited MV4ꢀ11 cell growth. A,B. Treatment with 28d for 96 h doseꢀdependently
reduced the proliferation of leukemia cells (A) but not normal cells (B). C. Knockdown
of PRMT1 was validated through western blotting. Both PRMT1 expression and arginine
dimethylation levels were reduced. GAPDH and Histone H4 were used as loading
controls. D. Genetic knockdown of PRMT1 reduced cell proliferation in MV4ꢀ11 cells.
The results are mean ± SD of three replicates. **P < 0.01, ***P < 0.001
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
To detect whether 28d and 31d were sufficient to inhibit cancer cell proliferation, we
screened eight leukemia cell lines (KOPNꢀ8, KMS11, MV4ꢀ11, THPꢀ1, RS4.11,
RCHꢀACV, REH and U937) using CellTiterꢀGlo assay. Results showed these two
compounds were potent at inhibiting cell proliferation in a doseꢀdependent manner, and
notably showed high potency toward MV4ꢀ11 cells (Figure 5A and Figure S7A). We also
found these two compounds exhibited minimal effects on cell proliferation of human
normal cells (human umbilical vein endothelial HUVꢀECꢀC cells, human normal lung
MRCꢀ5 cells and human renal cortical tubule epithelial RCTEC cells) (Figure 5B and
Figure S7B). The potent cellꢀkilling activity of type I PRMT inhibitors suggests PRMTs
might be a regulator of leukemia cell proliferation. Since the biological function of
PRMTs in leukemia is largely unknown, the effects of abrogation of the predominant
enzyme PRMT1 activity on leukemia cell proliferation were examined using validated
ACS Paragon Plus Environment

Page 25 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
PRMT1 shRNA. Immunoblot results showed both PRMT1 expression level and the
substrate asymmetric methylation levels were reduced (Figure 5C). Besides, depletion of
PRMT1 reduced cell proliferation of MV4ꢀ11 cells (Figure 5D). However, genetic
knockdown of PRMT1 showed a prolonged and rather small effect on cell proliferation
compared with the inhibitor treatment. Given that the inhibitors exhibited the role as pan
inhibitors, the cellꢀkilling effect may attribute to combined inhibition of type I PRMTs.
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
Alteration of the arginine methylation patterns in cellular levels
The most sensitive cell MV4ꢀ11 was chosen as the cell model to further characterize the
two compounds. We first performed a longꢀterm treatment of the two compounds against
MV4ꢀ11 cells and found both 28d and 31d greatly reduced
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 6. The antiꢀproliferative effect and arginine methylation alterations upon treatment
with 28d in MV4ꢀ11 cells. A. Longꢀterm treatment of MV4ꢀ11 cells with 28d inhibited
cell proliferation in a timeꢀdependent manner. B. Treatment with 28d for 48 h
doseꢀdependently inhibited the methyltransferase activity of PRMT1 in MV4ꢀ11 cells.
ADMA, SDMA) and H4R3me2a levels were detected by western blotting. C. The
negative compound 24b (DCPR049_4, shown in biochemical experiments) showed
minimal inhibition effects toward MV4ꢀ11 cells at 12.5 ꢁM and 6.25 ꢁM compared to
28d and 31d. Results shown are mean ± SD of three replicates.
ACS Paragon Plus Environment

Page 27 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
MV4ꢀ11 cell growth in a timeꢀdependent manner with a 12 day IC value of 0.81 ꢁM
5
0
and 1.08 ꢁM respectively (Figure 6A and Figure S8A). Using Western blot, we
confirmed these two compounds could alter the arginine methylation patterns in cellular
levels. After treating MV4ꢀ11 cells for 48 hours, they could potently block the global
levels of ADMA and concurrently increase the global levels of SDMA in a
concentrationꢀdependent manner (Figure 6B and Figure S8B). These effects were in line
with the PRMT1 knockout results in MEF cells, which revealed the competitive pattern
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
2
5
between the two arginine methylation types.
In addition, we observed a
doseꢀdependently decrease of H4R3me2a level. This result demonstrated the PRMT1
engagement effect of the compounds in cellular context since PRMT1 is the only known
46
contributor modifying this site. Also, upon treatment with 28d, the expression levels of
histone H3 arginine 2 asymmetric dimethylation (H3R2me2a) and histone H3 arginine 17
asymmetric dimethylation (H3R17me2a) decreased in a concentrationꢀdependent way,
indicating the inhibitory effect towards PRMT6 and CARM1 (PRMT4) respectively
(Figure S9). We next chose a structurally similar but far less active compound 24b to rule
out the possibility of the offꢀtarget effects of 28d and 31d. As expected, 24b exhibited
low potency against MV4ꢀ11 cells, and correspondingly showed little effects on the
arginine methylation patterns (Figure 6C and Figure S10), indicating minimal offꢀtarget
effects of our compounds.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 71
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Induction of cell cycle arrest and cell apoptosis and blockage of leukemogenic genes
expression
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 7. Treatment with 28d led to cell cycle arrest, apoptosis induction and
leukemogenic gene regulation in MV4ꢀ11 cells. A, B. Treatment of 28d arrested MV4ꢀ11
cells at G1 phase at 72 h (A) and induced apoptosis at 96 h (B). C. Quantitative RTꢀPCR
ACS Paragon Plus Environment

Page 29 of 71
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
analysis revealed 28d doseꢀdependently regulated the MLLꢀfusions target genes HOXA9,
HOXA10, MEIS1 and MNDA. Data shown are mean ± SD of three replicates.
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
To illustrate the antiproliferation mechanism, we examined the effects of the two
compounds on cell cycle and cell apoptosis through flow cytometric analysis. Results
showed a concentrationꢀdependent cell cycle arrest at G1 phase (Figure 7A and Figure
S11A) and an increased apoptosis (Figure 7B and Figure S11B) upon treatment with our
compounds for 72 hours and 96 hours respectively. These data supported the idea that
these compounds were not only cellꢀcycle arrestors, but also inducers of apoptosis.
Furthermore, we explored whether treatment with the two compounds could block
expression of leukemogenic genes in MV4ꢀ11 cells which originates from acute
monocytic leukemia and bears MLLꢀAF4 leukemic fusion gene. Overexpression of
homeobox genes including homeobox A9 (HOXA9), homeobox A10 (HOXA10) and the
HOX cofactor Meis homeobox 1 (MEIS1) has been considered as hallmarks of acute
4
7
leukemia carrying MLL rearrangements. MEIS1 was reported to synergize with
HOXA9 to facilitate transforming of myeloid progenitors and downꢀregulation of MEIS1
4
8, 49
and HOXA genes reduced proliferation in MLLꢀrearranged acute leukemia.
By using
quantitative realꢀtime PCR analysis, we examined the effects of the two compounds on
50
HOXA9, HOXA10 and MEIS1 mRNA levels. Treatment with the two compounds led to
profound doseꢀdependent decreases of all three transcripts in MV4ꢀ11 cells (Figure 7C
and Figure S11C). Meanwhile, the transcriptional level of myeloid cell nuclear
ACS Paragon Plus Environment