Discovery of decamidine as a new and potent PRMT1 inhibitor

Article abstract of DOI:10.1039/c6md00573j

Protein arginine methyltransferase 1 (PRMT1) is a key player for the dynamic regulation of arginine methylation. Its dysregulation and aberrant expression are implicated in various pathological conditions, and a plethora of evidence suggests that PRMT1 inhibition is of significant therapeutic value. Herein, we reported the modification of a series of diamidine compounds with varied lengths in the middle alkyl linker for PRMT1 inhibition. Decamidine (2j), which possesses the longest linker in the series, displayed 2- and 4-fold increase in PRMT1 inhibition (IC₅₀ = 13 μM), compared with furamidine and stilbamidine. The inhibitory activity toward PRMT1 was validated by secondary orthogonal assays. Docking studies showed that the increased activity is due to the extra interaction of the amidine group with the SAM binding pocket, which is absent when the linker is not long enough. These results provide structural insights into developing the amidine type of PRMT1 inhibitors.

Full text of DOI:10.1039/c6md00573j

MedChemComm
View Article Online
View Journal
RESEARCH ARTICLE
Discovery of decamidine as a new and potent
PRMT1 inhibitor†‡
Cite this: DOI: 10.1039/c6md00573j
Jing Zhang,§a Kun Qian,§a Chunli Yan,b Maomao He,a Brenson A. Jassim,a
b
a
*
Ivaylo Ivanov and Yujun George Zheng
Protein arginine methyltransferase 1 (PRMT1) is a key player for the dynamic regulation of arginine methyla-
tion. Its dysregulation and aberrant expression are implicated in various pathological conditions, and a
plethora of evidence suggests that PRMT1 inhibition is of significant therapeutic value. Herein, we reported
the modification of a series of diamidine compounds with varied lengths in the middle alkyl linker for
PRMT1 inhibition. Decamidine (2j), which possesses the longest linker in the series, displayed 2- and 4-fold
increase in PRMT1 inhibition (IC50 = 13 μM), compared with furamidine and stilbamidine. The inhibitory ac-
tivity toward PRMT1 was validated by secondary orthogonal assays. Docking studies showed that the in-
creased activity is due to the extra interaction of the amidine group with the SAM binding pocket, which is
absent when the linker is not long enough. These results provide structural insights into developing the
amidine type of PRMT1 inhibitors.
Received 13th October 2016,
Accepted 30th December 2016
DOI: 10.1039/c6md00573j
www.rsc.org/medchemcomm
of the best studied PRMT1 inhibitors. The compound
stilbamidine (Fig. 1),15 which was discovered by the Jung
1. Introduction
Epigenetic regulation, mainly composed of DNA methylation
and histone modifications such as methylation and acetyla-
tion, plays an essential role in eukaryotic life and contributes
to a plethora of pathological conditions such as cancer. The
concept of epigenetic alternations being an important cause
of tumor initiation, progression and metastasis has prompted
intense study on the development of epigenetic drugs which
is best represented by the approval of two histone deacetylase
(HDAC) inhibitors (vorinostat and romidepsin) for the treat-
ment of cutaneous T-cell lymphoma.1–3 Besides acetylation,
histone arginine methylation is another significant epigenetic
player in this anti-cancer arena. Overexpression of protein ar-
ginine methyltransferase 1 (PRMT1), which is the major en-
zyme catalyzing monomethylation and asymmetric dimethyl-
ation of arginine residues of protein substrates, has been
linked with various cancer phenotypes, including leukemia,
lung, prostate, and breast cancers.4 Application of PRMT1 in-
hibitors or inactivation of PRMT1 by siRNA, has great poten-
tial to significantly hamper tumor cell growth.5–8
group, is the first diamidine compound with potent PRMT1
inhibitory activity. As a result of PRMT1 inhibition, histone
hypomethylation was induced while estrogen receptor activa-
tion was reduced by this compound in functional assays.15
Focused library screening by our group identified the com-
pound furamidine (DB75) and its analogs as low micromolar
PRMT1 inhibitors. The compound furamidine possessed a
two-fold lower IC50 value compared with stilbamidine, and it
showed >10-fold selectivity over the other PRMTs including
PRMT5. Supported by structural modelling, the amidine
groups in both stilbamidine and furamidine most likely
mimic the guanidino moiety of the arginine residue, and this
concept is also confirmed by the mechanistic study showing
that the compound furamidine is partially competitive to-
wards the substrate and noncompetitive towards the
So far, a handful of PRMT1 inhibitors have been discov-
ered or designed.7,9–15 Diamidine compounds represent one
a Department of Pharmaceutical and Biomedical Sciences, College of Pharmacy,
The University of Georgia, Athens, Georgia 30602, USA. E-mail: yzheng@uga.edu
b Department of Chemistry, Center for Diagnostics and Therapeutics, Georgia
State University, Atlanta, Georgia 30302, USA
† The authors declare no competing interests.
‡ Electronic supplementary information (ESI) available. See DOI: 10.1039/
c6md00573j
Fig.
1 Reported PRMT1 inhibitors which contain the amidine
§ Jing Zhang and Kun Qian contributed equally.
functionality.
This journal is © The Royal Society of Chemistry 2017
Med. Chem. Commun.

View Article Online
Research Article
MedChemComm
cofactor.7 Recently the SAR study of amidine compounds by
Yu et al. confirmed the importance of the amidine group,
showing that its deletion ablates PRMT1 inhibition.11 In that
study, the researchers identified compound 6d (Fig. 1) as a
potent PRMT1 inhibitor with an IC50 of 2.0 μM.
precipitation from H2O, or by preparative HPLC if needed.
The purity and identity of the compounds were validated by
NMR and mass spectrometry (see the ESI‡).
The PRMT1 inhibitory activities of compounds 2a–2j were
evaluated by using our previously established scintillation
proximity assay (SPA),19 which measured the amount of
methyl transfer from [3H]-SAM to biotinylated histone H4
peptide (ac-SGRGKGGKGLGKGGAKRHRKIJBiotin)-NH2, abbre-
viated as H4-20-Biotin). Furamidine was also used as a posi-
tive control. As shown in Fig. 2, there was no significant dif-
ference in PRMT1 inhibition among compounds 2a–2g, with
the residual PRMT1 activity ranging from 80% to 57%. There-
fore, the middle methylene length in the range of 0–4 (refer-
ring to the numbers of methylene units) was not particularly
sensitive to PRMT1 binding under the current assay condi-
tions. However, when the carbon length increased to above 5
methylene units, the PRMT1 inhibition difference was dis-
tinct enough to be observed following the ranking order of 2j
In our continuing studies on structural modifications of
diamidine compounds, we found that slightly varying the
linkers in stilbamidine can be tolerated (data not shown).
This inspired us to explore other diamidine compounds with
varied lengths of the middle linker such as pentamidine,
which was reported to be a DNA binder and has been clini-
cally used as an antiparasitic drug in a similar manner to
furamidine.16,17 In this study, we set up a facile synthesis of
amidine analogs with different hydrocarbon linkers between
the two benzamidine moieties. The activities of these com-
pounds on PRMT1 inhibition were studied with biochemical
assays. The most active compound 2j was docked into the ac-
tive site of PRMT1 to decipher the structural basis for the
linker length influence on the potencies of the inhibitors.
(decamidine)
> 2i (hexamidine) > 2h (pentamidine),
suggesting that the potency for PRMT1 inhibition increases
as the middle linker elongates to above 5.
2. Results and discussion
To determine whether the tested diamidine compounds
were specific for PRMT1, we also screened the inhibitory ac-
tivities of these compounds against PRMT5, the major type II
PRMT enzyme. As shown in Fig. 3, the majority of diamidine
compounds did not show significant PRMT5 inhibition, ex-
cept for compounds 2b, 2j and furamidine, which showed ca.
50% inhibition at 100 μM.
Then, the IC50 values in PRMT1 inhibition were deter-
mined for compounds 2h–2j, furamidine and stilbamidine
with dose-dependent assays. A typical inhibition curve is
shown in Fig. 4. As displayed in Table 1, consistent with the
single-dose inhibition profile, the IC50 values of 2h, 2i and 2j
showed a decreasing trend with 81 μM, 52 μM and 13 μM, re-
spectively. It should be noted that compound 2j is more po-
tent than furamidine, which showed an IC50 value of 22 μM
under these experimental conditions. Interestingly, compared
with furamidine, compound 2j displayed a reduced selectivity
over PRMT5 with an IC50 value of 43 μM.
The compounds phenamidine (2a) and pentamidine (2h)
were obtained commercially. The other compounds were syn-
thesized using the synthetic route illustrated in Scheme 1.
First, commercially available 4-cyanophenol was reacted with
an excess of dibromo- or dichloro-alkane under basic condi-
tions to yield the alkoxylbenzonitrile compounds 1. Classi-
cally, the cyano groups in these compounds could be
converted to amidine groups via the Pinner reaction through
an imino ether intermediate followed by the addition of am-
monia in ethanol.18 However, we noticed incomplete conver-
sion and a low yield by this method (data not shown). There-
fore, an easier method under basic conditions was
investigated. In this method, the intermediate 1 was treated
in LiHMDS (lithium bisIJtrimethylsilyl)amide, 1 N in tetrahy-
drofuran) at room temperature overnight, followed by
quenching with 2 N HCl solution. This method allowed com-
plete conversion of the cyano group to amidine, and there-
fore led to high yields. Meanwhile, the reaction is not sensi-
tive to moisture, so it is suitable for small-scale reaction and
can be easily handled. The crude product can be purified by
Fig. 2 PRMT1 inhibition of compounds 2a–2j and furamidine at 100
μM and 10 μM concentrations. The reaction buffer contains 50 mM
HEPES, 10 mM NaCl, 0.5 mM EDTA and 0.5 mM DTT, and has pH 8.0.
The concentrations of PRMT1, [3H]-SAM, and H4-20-Biotin are 0.01,
0.5, and 0.5 μM, respectively. The mixtures were incubated at room
temperature for 15 min.
Scheme 1 Synthetic route to diamidine compounds.
Med. Chem. Commun.
This journal is © The Royal Society of Chemistry 2017

View Article Online
MedChemComm
Research Article
Fig. 3 PRMT5 inhibition of compounds 2a–2j and furamidine at 100
μM and 10 μM concentrations. The reaction conditions were similar to
those Fig. 2, except that the reaction time was 45 min.
Fig. 4 The dose-dependent PRMT1 inhibition curve by compound 2j.
Fig. 5 The comparison of PRMT1 inhibition by compounds 2j and
furamidine. (a) Mass spectra of the reaction mixture with different
concentrations of compounds 2j and furamidine. From spectra I to V:
DMSO control, compound 2j (0.2 mM), compound 2j (0.5 mM),
furamidine (0.2 mM), furamidine (0.5 mM). The concentrations of
PRMT1, SAM, and H4-20 were 1, 200, and 50 μM, respectively. The
mixtures were incubated at 30 °C for 3.5 h. (b) Radiometric gel assay
and its quantification. The concentrations of PRMT1, [14C]SAM, and
H4-20 are 0.1, 20, and 100 μM, respectively. The reaction time was 30
min at 30 °C before it was quenched with 5 × SDS-loading buffer, sep-
arated by 15% SDS-PAGE, dried under vacuum and visualized by stor-
age phosphorimaging. The quantification was based on two
experiments.
Table 1 IC50 values of selected compounds for PRMT1 and PRMT5
IC50, μM
Compounds
PRMT1
PRMT5
Stilbamidine
Furamidine
Pentamidine (2h)
Hexamidine (2i)
Decamidine (2j)
52
22
81
52
13
2
2
3
3
1
712 161
256
1
582 100
357 155
43
1
To further validate the PRMT1 inhibition by 2j, mass
spectrometric analysis of the methylation reaction mixture
using regular SAM was conducted. As shown in the spectrum
I in Fig. 5a, the substrate histone 4 (1–20) with N-terminal
acetyl group (abbreviated as H4-20) was completely converted
to dimethyl H4-20 in the absence of any inhibitor. However,
when compound 2j was added at 0.2 mM or 0.5 mM concen-
trations (spectra II and III in Fig. 5a), this methyl transfer re-
action was completely inhibited, showing that only H4-20
was detected in the reaction mixture. In comparison, when
furamidine was used for inhibition, the MALDI result showed
that both H4-20 and dimethyl H4-20 were detected at 0.2 mM
concentration (spectra IV in Fig. 5a), suggesting an incom-
plete inhibition. We also examined the inhibition effect with
the radiometric gel assay. In this assay, different concentra-
tions of inhibitors were incubated with PRMT1, H4-20 and
[14C]SAM. The reaction mixture was resolved on a 15% SDS-
PAGE gel. The gel was vacuum dried and then subjected to
phosphorimaging detection. As shown in Fig. 5b, no methyla-
tion was observed when compound 2j was at 0.2 mM.
However, methylation was still observed at the same concen-
tration of furamidine. Quantitation analysis revealed that the
extent of methylation with compound 2j at 0.1 and 0.05 mM
was very similar to that with furamidine at 0.2 and 0.1 mM,
consistent with the IC50 measurement in which compound 2j
was two-fold more potent than furamidine.
Evidently, elongating the middle linker could significantly
enhance the potency of diamidine compounds for PRMT1 in-
hibition, resulting in compound 2j being more potent than
furamidine. To probe possible structural differences in the
PRMT1-inhibitor interaction at different linker lengths, com-
pounds 2i and 2j were docked into a homology model of
PRMT1 using Autodock4.2 as described in our previous
work.7 The docking scores for compounds 2i and 2j were
−10.1 and −11.0 kcal mol−1, respectively. Therefore, com-
pound 2j can form an energetically more favorable
This journal is © The Royal Society of Chemistry 2017
Med. Chem. Commun.

View Article Online
Research Article
MedChemComm
2j also form hydrophobic interactions with the side chain of
F36 and Y39, including π–π interactions (T-shape geometry)
from the N-terminal αX helix, which forms the roof of the
pocket above the ligand. These interactions thus tightly an-
chor 2i and 2j in the PRMT1 active site. Docking of 2i and 2j
also suggests that both compounds are nestled to allow good
shape complementarity to the cavity formed by the adjacent
substrate/cofactor sites in PRMT1 (Fig. 6c and d).
To probe the selectivity difference between compounds 2i
and 2j, we also docked both compounds with PRMT5. As can
be seen in Fig. S1,‡ both compounds only bind to the cofac-
tor site of PRMT5. For compound 2i, it spans both the cofac-
tor and arginine sites for PRMT1, but only the cofactor site
for PRMT5. For compound 2j, it binds to PRMT1 and PRMT5
similarly. This may partially explain the low selectivity of this
compound.
Fig. 6 Predicted conformation of 2i and 2j bound to PRMT1 from
docking. (a) Detailed view of 2i displaying the key PRMT1 residues. (b)
Detailed view of 2j displaying the key PRMT1 residues. The PRMT1
homology structure was constructed by using the known structure of
R. norvegicus PRMT3 (PDB code: 1F3L20) and H. sapiens PRMT3 (PDB
code: 3SMQ21). Compounds 2i and 2j are shown in a magenta ball and
stick representation. The PRMT1 residues involved in binding to 2i and
2j are shown as green sticks and labeled. The red dashed lines
represent hydrogen bonds. (c) Shape of the binding cavity of PRMT1
(grey) with 2i (magenta sphere). (d) Shape of the binding cavity of
PRMT1 (grey) with 2j (magenta sphere). The surface in panels (c) and
(d) is visualized with VMD.22
3. Conclusions
The results of the present study demonstrate that diamidine
compounds are capable of inhibiting the methyl transfer ac-
tivity of PRMT1 with varied potencies. Among the analogs
with different lengths of linkers (C0–C6, C10), the potency of
inhibition increased as the linker length increased, and the
most potent inhibitor was compound 2j (decamidine) with a
C10 linker. Compared with stilbamidine and furamidine, this
compound showed moderately increased PRMT1 inhibition,
but slightly decreased selectivity against PRMT5. Besides the
potency characterization using the SPA measurement, the in-
hibitory activity of 2j was also confirmed by mass spectrome-
try analysis of the reaction mixture and gel phosphorimaging
assay. The increased PRMT1 inhibition could be attributed to
the occupation by the amidine group of the SAM binding site
which resulted from the extension of the middle linker. In or-
der to reach this extra interaction, the linker needs to be long
enough to stretch out, and a linker length of over five methy-
lene groups appears to be the threshold for transitioning
from one binding mode to the other. Future work is
warranted for detailed investigation of diamidine compounds
in cellular PRMT1 activity inhibition as well as correlating
biochemical potencies with potential anti-cancer activities in
PRMT1 overexpressed tumor models. Compound 2j with a
C10 linker could be an interesting hit for further optimiza-
tion by finely tuning the length of the hydrocarbon linker
and the substitutions of the phenyl rings. Our effort towards
these directions will be reported in due time.
interaction with PRMT1, which is in accordance with biologi-
cal activity results. As shown in Fig. 6a and c, the diamidine
moiety of 2i inserts into both the SAM adenosine binding site
and the Arg substrate site analogous to the previously
reported binding mode of furamidine to PRMT1.7 Moreover,
the docking analysis shows that one of the amidine groups of
2i forms three favorable hydrogen bonds: one with residue
P24 and two with the backbone carbonyl groups of residues
K23 and E27. The second amidine group of 2i is stabilized
via hydrogen bonds to the side chains of E144 and E153 (two
important residues in the double-E loop). For compound 2j,
the first amidine group occupies the same SAM adenosine
site as in the 2i complex (via two hydrogen bonds to the
backbone carbonyl group of P24 and A26, a hydrogen bond
with the side chain of E129). However, the binding pattern
for the second amidine group of 2j is completely different. In
the case of 2j, the flexible 10-carbon linker can stretch and
position this amidine group into the cofactor methionine
binding site. In this orientation, the amidine interacts with
the Gly-rich loop of PRMT1 (residues G78 and G80) through
hydrogen bonds with the side chain of D76 and the backbone
carbonyl group of S79. The deep placement of the amidine in
the cofactor pocket with multiple interactions formed ac-
counts for the stronger binding of 2j with PRMT1 (Fig. 6d).
The different binding mode for the second amidine group
could explain the origin of the differences in the observed ac-
tivity of 2i and 2j. The phenyl rings and the linker of 2i and
Acknowledgements
This work was supported by the National Institute of Health
Grant R01GM086717 to Y. G. Z. and R01GM110387 to I. I. Jing
Zhang was supported by American Heart Association Postdoc-
toral Fellowship # 16POST27260147. The authors acknowl-
edge the Proteomics and Mass Spectrometry (PAMS) Facility
at UGA for the MS support, and Dr. D. Cui of the Chemistry
department at UGA for NMR assistance.
Med. Chem. Commun.
This journal is © The Royal Society of Chemistry 2017

View Article Online
MedChemComm
Research Article
References
12 A. Spannhoff, R. Machmur, R. Heinke, P. Trojer, I. Bauer, G.
Brosch, R. Schule, W. Hanefeld, W. Sippl and M. Jung,
Bioorg. Med. Chem. Lett., 2007, 17, 4150–4153.
13 N. Fontan, P. Garcia-Dominguez, R. Alvarez and A. R. de
Lera, Bioorg. Med. Chem., 2013, 21, 2056–2067.
14 M. S. Eram, Y. Shen, M. M. Szewczyk, H. Wu, G. Senisterra,
F. Li, K. V. Butler, H. U. Kaniskan, B. A. Speed, C. dela Sena,
A. Dong, H. Zeng, M. Schapira, P. J. Brown, C. H.
Arrowsmith, D. Barsyte-Lovejoy, J. Liu, M. Vedadi and J. Jin,
ACS Chem. Biol., 2016, 11, 772–781.
15 A. Spannhoff, R. Heinke, I. Bauer, P. Trojer, E. Metzger, R.
Gust, R. Schule, G. Brosch, W. Sippl and M. Jung, J. Med.
Chem., 2007, 50, 2319–2325.
16 F. C. Bocobo, A. C. Curtis and E. R. Harrell, J. Invest.
Dermatol., 1953, 21, 149–156.
17 V. Jean-Moreno, R. Rojas, D. Goyeneche, G. H. Coombs and
J. Walker, Exp. Parasitol., 2006, 112, 21–30.
18 A. Pinner and F. Klein, Ber. Dtsch. Chem. Ges., 1877, 10,
1889–1897.
1 S. Rangwala, C. Zhang and M. Duvic, Future Med. Chem.,
2012, 4, 471–486.
2 O. Khan and N. B. La Thangue, Immunol. Cell Biol., 2012, 90,
85–94.
3 S. Jain, J. Zain and O. O'Connor, J. Hematol. Oncol., 2012, 5, 24.
4 Y. Yang and M. T. Bedford, Nat. Rev. Cancer, 2013, 13, 37–50.
5 M. Yoshimatsu, G. Toyokawa, S. Hayami, M. Unoki, T.
Tsunoda, H. I. Field, J. D. Kelly, D. E. Neal, Y. Maehara, B. A.
Ponder, Y. Nakamura and R. Hamamoto, Int. J. Cancer,
2011, 128, 562–573.
6 J. Wang, L. Chen, S. H. Sinha, Z. Liang, H. Chai, S. Muniyan,
Y. W. Chou, C. Yang, L. Yan, Y. Feng, K. K. Li, M. F. Lin, H.
Jiang, Y. G. Zheng and C. Luo, J. Med. Chem., 2012, 55,
7978–7987.
7 L. Yan, C. Yan, K. Qian, H. Su, S. A. Kofsky-Wofford, W. C.
Lee, X. Zhao, M. C. Ho, I. Ivanov and Y. G. Zheng, J. Med.
Chem., 2014, 57, 2611–2622.
8 N. Cheung, T. K. Fung, B. B. Zeisig, K. Holmes, J. K. Rane,
K. A. Mowen, M. G. Finn, B. Lenhard, L. C. Chan and C. W.
So, Cancer Cell, 2016, 29, 32–48.
9 H. Hu, K. Qian, M. C. Ho and Y. G. Zheng, Expert Opin.
Invest. Drugs, 2016, 25, 335–358.
10 E. M. Bissinger, R. Heinke, A. Spannhoff, A. Eberlin, E.
Metzger, V. Cura, P. Hassenboehler, J. Cavarelli, R. Schule,
M. T. Bedford, W. Sippl and M. Jung, Bioorg. Med. Chem.,
2011, 19, 3717–3731.
11 X. R. Yu, Y. Tang, W. J. Wang, S. Ji, S. Ma, L. Zhong, C. H.
Zhang, J. Yang, X. A. Wu, Z. Y. Fu, L. L. Li and S. Y. Yang,
Bioorg. Med. Chem. Lett., 2015, 25, 5449–5453.
19 J. Wu, N. Xie, Y. Feng and Y. G. Zheng, J. Biomol. Screening,
2012, 17, 237–244.
20 X. Zhang, L. Zhou and X. Cheng, EMBO J., 2000, 19,
3509–3519.
21 A. Siarheyeva, G. Senisterra, A. Allali-Hassani, A. Dong, E.
Dobrovetsky, G. A. Wasney, I. Chau, R. Marcellus, T. Hajian,
F. Liu, I. Korboukh, D. Smil, Y. Bolshan, J. Min, H. Wu, H.
Zeng, P. Loppnau, G. Poda, C. Griffin, A. Aman, P. J. Brown,
J. Jin, R. Al-Awar, C. H. Arrowsmith, M. Schapira and M.
Vedadi, Structure, 2012, 20, 1425–1435.
22 W. Humphrey, A. Dalke and K. Schulten, J. Mol. Graphics,
1996, 14, 33–38.
This journal is © The Royal Society of Chemistry 2017
Med. Chem. Commun.