Histone demethylating agents as potential S-adenosyl-l-methionine-competitors

Article abstract of DOI:10.1039/c6md00170j

Histone H3 methylation on K9 and/or K27 depends on histone lysine methyltransferases (KMTs). EZH2, one of the components of the PRC2 complex, catalyzes the trimethylation of histone H3K27, which is associated with transcriptional repression and tumor development. H3K9me3 mediated gene silencing may result from other KDMs such as G9a/GLP, SUV39H1-2, SETDB1, CCLD8 and RIZ1. Their disturbance leads to defective cell mitosis. It is therefore desirable to find small molecules that are able to decrease H3K9 and K27 tagging to reinitiate gene transcription. Most KDM inhibitors are still based on SAM co-factor competition/modulation. Herein, functional screening of a diversity library proved to be a useful tool for finding new specific KDM inhibitors; the use of SAM-based pharmacophoric models facilitated the understanding of their possible mechanism of action.

Full text of DOI:10.1039/c6md00170j

MedChemComm
View Article Online
View Journal
RESEARCH ARTICLE
Histone demethylating agents as potential
S-adenosyl-L-methionine-competitors†‡
Cite this: DOI: 10.1039/c6md00170j
a
b
c
a
d
R. Navakauskienė,§ M. Mori,§ M. S. Christodoulou,§ A. Zentelytė, B. Botta,
e
f
f
c
g
g
L. Dalla Via, F. Ricci, G. Damia, D. Passarella, C. Zilio and N. Martinet*
Histone H3 methylation on K9 and/or K27 depends on histone lysine methyltransferases (KMTs). EZH2, one
of the components of the PRC2 complex, catalyzes the trimethylation of histone H3K27, which is associ-
ated with transcriptional repression and tumor development. H3K9me3 mediated gene silencing may result
from other KDMs such as G9a/GLP, SUV39H1-2, SETDB1, CCLD8 and RIZ1. Their disturbance leads to de-
fective cell mitosis. It is therefore desirable to find small molecules that are able to decrease H3K9 and K27
tagging to reinitiate gene transcription. Most KDM inhibitors are still based on SAM co-factor competition/
modulation. Herein, functional screening of a diversity library proved to be a useful tool for finding new
specific KDM inhibitors; the use of SAM-based pharmacophoric models facilitated the understanding of
their possible mechanism of action.
Received 23rd March 2016,
Accepted 4th May 2016
DOI: 10.1039/c6md00170j
www.rsc.org/medchemcomm
(
K) residues in a non-histone protein, such p53, are also
Introduction
5
methylated by KDMs. Both the site (K) and the methylation
(from me1 to me3) conditions further facilitate the recogni-
tion of these histone tags by reader proteins. EZH2 is respon-
sible for the H3K27me3 found at transcriptionally repressed
promoters. Silent pericentric heterochromatin is marked by
H3K9me3, which can be deposited by several KMTs, includ-
ing G9a and its homologue GLP (together they form hetero-
dimeric complexes), SETDB1, RIZ1, CLLD8 and mainly
SUV39H1 and its homologue SUV39H2; all of which silence
gene expression, thereby promoting aberrant cancer cell pro-
1
,2
Lysine methylation was first observed in histones in 1965.
Described in 2000, SUV39H1 was the first SAM dependent en-
zyme, known as a KMT. For about 50 human KMTs, the cata-
6
3
7
lytic domain is composed of a core SET domain that is highly
conserved. More variable regions form the peptide substrate-
binding groove, and also bring the SAM co-factor binding
pocket at the core of the SET structure. Such KMTs methylate
the N-terminal tail of H3 at K4, K9, K27, and K36, and H4 at
4
K20. Methylation does not alter the charge of the histone
7
,8
tail, but influences the basicity and hydrophobicity of the his-
tone and thus, changes the histone's affinity for transcription
factors involved in genome transcription. Site-specific lysine
liferation, and also the repression of ESC pluripotency.
Targeting the SAM co-factor binding site of KDMs was the
first successful approach to attempt KMT inhibition.
Sinefungin A, isolated from Streptomyces, is a natural nucleo-
side analog of SAM and a non-selective inhibitor of KMTs.
Chaetocin, a fungal metabolite, was discovered next, as an in-
9
a
Department of Molecular Cell Biology, Institute of Biochemistry, Vilnius
University, Mokslininku st. 12, LT-08662 Vilnius, Lithuania
10
hibitor of Suv39H1. It demonstrated some dose-dependent
b
Center for Life Nano Science@Sapienza, Istituto Italiano di Tecnologia, viale
selectivity, when not additionally inhibiting EZH1-2 or SET7/
9, at concentrations below 90 μM. Chaetocin enantiomers
were also shown to have similar selective inhibitory activi-
ties. There are reports showing chaetocin to be a non-
specific and chemically reactive molecule, owing to its disul-
fide bridge, and it forms adducts with target enzymes, leading
Regina Elena 291, 00161 Rome, Italy
c
Dipartimento di Chimica, Università degli Studi di Milano, Via Golgi 19, 20133
Milano, Italy
d
11
Dipartimento di Chimica e Tecnologie del Farmaco, Università degli Studi di
Roma “La Sapienza”, Piazzale Aldo Moro 5, 00185 Roma, Italy
e
Dipartimento di Scienze del Farmaco, Università degli Studi di Padova, Via F.
Marzolo 5, 35131 Padova, Italy
1
2,13
f
to their irreversible inhibition.
Gliotoxin analogs isolated
Mario Negri Institute for Pharmacological Research, Via La Masa 19, 20156
Milano, Italy
from Penicillium, showed potent inhibitory activity, mainly
against G9a and Suv39H1, by dysregulation of the methyl/me-
thionine cycle to control intracellular SAM and SAH homeo-
stasis. Yuan et al. reported that BRD4770, a SAM mimetic,
reduced levels of di- and tri-methylated H3K9 in the pancre-
atic cancer cell line PANC-1. The DOT1L (a non-SET KDM)
g
Institute of Chemistry, UMR CNRS 7272, Parc Valrose, Nice 06108, Cedex 2,
France. E-mail: nadine.martinet@inserm.fr
†
‡
The authors declare no competing interests.
1
4
15
Electronic supplementary information (ESI) available. See DOI: 10.1039/
c6md00170j
Author that contributed equally to the study.
§
This journal is © The Royal Society of Chemistry 2016
Med. Chem. Commun.

View Article Online
Research Article
MedChemComm
inhibitor EPZ004777 is also a SAM competitor with a very
good selectivity for EZH2 and SETD7, killing selectively mixed
lineage leukemia cells. It is currently undergoing a phase I
trial in patients with advanced solid tumors and/or with re-
mL) and acetone (18 mL), butyl isocyanate (0.53 mL, 4.6 mM)
was added and the new solution was refluxed for 5 h. After
completion of the reaction, the solvent was evaporated under
vacuum; EtOAc was added and the organic layer was washed
with water and brine, dried with Na SO , filtered and evapo-
1
6
lapsed or refractory B-cell lymphoma.
2
4
These results support the potential for developing SAM
competitors/modulators as H3K9 and K27me3 inhibitors for
therapeutic purposes. This is actually the case for a potent
SAM hydrolase inhibitor, DZNep, which induces the accumu-
lation of SAH in cells to down regulate the KDM PRC2 cata-
rated. The residue was purified by flash chromatography
(hexane/EtOAc 1 : 1) to provide compounds 1 and 2 as amor-
phous solids (45% yield).
1
9
Compound 1. [a]
(Hex/EtOAc 1 : 1); H NMR (acetone-d ): δ = 10.2 (1H, br, s),
D
= −136.02 (c 0.28, MeOH); R
f
= 0.27
1
6
1
7
lytic activity, and thus, the H3K27me3 tag level. Unlike
DZNeP, GSK-343 and GSK-126 directly inhibit EZH2 by com-
peting with SAM, having nM potencies; they however, remain
7.51 (1H, d, J 7.8 Hz), 7.39 (1H, d, J 7.8 Hz), 7.12 (1H, t, J 8.0
Hz), 7.05 (1H, t, J 8.0 Hz), 5.04 (1H, d, J 16 Hz), 4.43 (1H, d, J
16 Hz), 4.32 (1H, dd, J 11 Hz J 5.1 Hz), 3.51 (2H, t, J 6.9 Hz),
3.29 (1H, dd, J 16 Hz J 6.0 Hz), 2.78 (1H, t, J 16 Hz), 1.61 (2H,
m), 1.34 (2H, m), 0.932 (3H, t, J 7.6 Hz); MS: 320.38 (M +
1
8–20
selective.
Curry and coauthors have shown that the com-
bined inhibition of KMTs (EHMT1/2 and EZH2) by GSK343
and UNC0638 increases growth inhibition in tumor cells,
more than the inhibition of only one KMT, and results in the
+
+
Na ) ; anal. calcd for C H N O : C, 68.67; H, 6.44; N, 14.13.
1
7
19 3 2
Found: C, 67.97; H, 6.26; N, 13.53.
2
1
19
re-expression of silenced genes.
Compound 2. [a]
D
= −209.91 (c 0.51, MeOH); R
f
= 0.16
1
Non-SAM interfering inhibitors were also developed. BIX-
(Hex/EtOAc 1 : 1); H NMR (acetone-d ): δ = 7.50 (2H, d, J 8.4
6
0
1294, a diazepine–quinazoline–amine derivative was con-
Hz), 7.19 (1H, t, J 8.6 Hz), 7.09 (1H, t, J 8.6 Hz), 5.61 (2H, d, J
17 Hz), 5.49 (1H, t, J 6.4 Hz), 5.16 (1H, d, J 16 Hz), 4.50 (1H,
d, J 16 Hz), 4.27 (1H, dd, J 11 Hz J 5.1 Hz), 3.48 (2H, t, J 6.9
Hz), 3.26 (1H, dd, J 16 Hz J 6.0 Hz), 2.68 (1H, t, J 16 Hz), 1.61
(2H, m), 1.34 (2H, m), 0.933 (3H, t, J 7.6 Hz); MS: 350.25 (M +
ceived to bind to the G9a/GLP channel, accommodating the
KDM protein substrate. It is a highly selective but transient
inhibitor, lowering H3K9 me2 levels in mouse ESC and fibro-
blasts; however, histone mono- and tri-methyl stages
2
2
+ +
remained unaffected.
More potent and more selective
Na ) ; anal. calcd for C H N O : C, 66.04; H, 6.47; N, 12.84.
1
8
21 3 3
second-generation compounds have been developed, based
on this mechanism and this initial scaffold; these include
E72, UNC321, UNC0638, and UNC0646.
Herein, we report our contribution to the KDM field with
the extensive study of 7 compounds, identified by functional
diversity screening, measuring the decrease in H3K9 and
K27me3 in MCF7 cells.
Found: C, 65.37; H, 6.29; N, 12.28.
Synthesis of compounds 3 and 4. According to the general
procedure above, compounds 3 and 4 were obtained as amor-
phous solids (48% yield) from compounds 10 and 11, after
purification by flash chromatography (Hex/EtOAc 1 : 1) and
evaporation of the solvent,.
2
3
2
0
Compound 3. [a]
D
= −127.95 (c 0.28, MeOH); MS: 320.36
M + Na ) ; anal. calcd for C17 : C, 68.67; H, 6.44; N,
4.13. Found: C, 67.99; H, 6.24; N, 13.52.
+
+
(
1
19 3 2
H N O
Materials and methods
Chemistry general procedures
1
9
Compound 4. [a]
D
= 206.54 (c 0.50, MeOH); MS: 350.28 (M
+ +
+
Na ) ; anal. calcd for C18 : C, 66.04; H, 6.47; N,
21 3 3
H N O
All reactions were carried out in oven-dried glassware and dry
solvents under nitrogen atmosphere, unless otherwise stated;
all solvents were purchased from Sigma-Aldrich, Milano, Italy
and used without further purification. Substrates and re-
agents were purchased from Sigma-Aldrich and used as re-
ceived. Thin layer chromatography was performed on Merck
12.84. Found: C, 65.35; H, 6.30; N, 12.30. For details see ref.
24 and 25.
Cell culture conditions
All reagents and culture media came from Sigma-Aldrich un-
less otherwise specified. All cell lines came from the Ameri-
can type collection, USA. MCF-7 cells (human mammary
breast adenocarcinoma) were grown in Eagle's minimum es-
sential medium. MSTO-211H cells (human biphasic mesothe-
lioma) and H460 cells (large cell lung carcinoma) were cul-
tured in Roswell Park Memorial Institute 1640 medium.
(Milano, Italy) precoated 60F254 plates. Reactions were moni-
tored by TLC on silica gel, with detection by UV light (254
nm). Flash chromatography was performed using silica gel
(240–400 mesh, from Merck). All tested compounds had a pu-
rity of >98%, confirmed via elemental analyses (CHN) using
1
a Perkin Elmer 2400 instrument. H-NMR spectra were
−
1
−1
recorded on a Bruker DRX-300 instrument, and reported rela-
tive to residual acetone-d6. Chemical shifts (δ) for proton and
carbon resonances are quoted in ppm, relative to tetra-
methylsilane, which was used as an internal standard. MS
spectra were recorded using a Waters Micromass Q-Tof micro
mass spectrometer.
Media were supplemented with 2.38 g L HEPES, 0.11 g L
−
1
sodium pyruvate, 2.5 g L glucose, 10% v/v heat-inactivated
fetal calf serum (Invitrogen, USA); 100 U mL penicillin, 100
μg mL streptomycin, and 0.25 μg mL amphotericin B
−
1
−
1
−1
(Gibco, Grand Island, NY, USA). Cells were seeded into each
4
well of a 24 well plate, at a density per well of 3 × 10 for both
4
General procedure for the synthesis of compounds 1 and
H460 and MSTO-211H, and at 6 × 10 for MCF7. Cultures
2
. To a solution of compounds 8 and 9 (1.0 g) in DMSO (8.0
were kept at 37 °C in a moist atmosphere of 5% carbon
Med. Chem. Commun.
This journal is © The Royal Society of Chemistry 2016

View Article Online
MedChemComm
Research Article
dioxide in air. After 24 h, various concentrations of com-
pounds were added in triplicate wells, and the cultures lasted
up to 72 h. A trypan blue assay was then performed to deter-
mine cell viability. Results were finally expressed as IC50
values; i.e., the concentration of the compound inducing a
0.18% v/v Tween 20, washed in PBS/Tween 20 v/v solution
without BSA, and finally probed with the primary antibody,
in accordance with the manufacturer's instructions. The
membrane was subsequently washed four times with the
PBS/Tween 20 solution and then incubated with the HPR
secondary antibody for 1 h at room temperature. Immuno-
reactive bands were detected by enhanced chemilumines-
cence (GE Healthcare Amersham). The antibodies used were
as follows: monoclonal anti-GAPDH (ab8245, from Abcam,
Cambridge, UK; polyclonal anti-me3 H3K27 (No. 07-449)
and the anti-me3 H3K9 (No. 07-442, from Millipore), and
the goat anti-rabbit (or anti-mouse) HPR secondary anti-
bodies from Dako Cytomation A/S (Copenhagen, Denmark);
GAPDH served as the protein loading control. The band
intensity of western blots was quantified by densitometric
analysis, using ImageJ 1.45 S software. Changes in the his-
tone modification level were expressed by comparison of
densitometry obtained for treated and untreated control
cells.
5
0% reduction in cell growth by comparison with untreated
control cultures.
For cell cycle and apoptosis analysis, MCF7 cells were
4
seeded into 6 well plates at 45 × 10 cells per well. For KDM
2
activity evaluation, MCF7 cells were seeded in 25 cm culture
4
flasks at 100 × 10 cells per flask. Aliquots of these last cul-
tures were also processed for DNA purification by classical
phenol/chloroform extraction. These DNA were further used
in methylation specific PCR to study the DNA methylation
state of the second exon of the retinoic acid receptor β, and
also to measure by ELISA, the global level of DNA 5 methyl
cytosine.
Cancer stem cell-like cultures BIX-01294
Two tumor initiating, enriched cell cultures, LA #83 and LA
Cell cycle analysis
2
6,27
#
100, were recently isolated,
spheroids in low adherence flasks (Corning, USA) under
stem cell-like conditions”; i.e., serum-free DMEM/F12
and were grown to form
Cells were treated with 15 μM of each compound for 24, 48
and 72 h. The cells were then collected and washed twice by
centrifugation (1000 × g for 5 min, at 4 °C) in PBS, then re-
suspended in 1 ml of ice cold PBS. Next, they were gently
vortexed, slowly adding 3.3 ml of cold absolute ethanol
dropwise, and stored at −20 °C overnight. Before staining,
the cells were centrifuged at 1000 × g for 10 min, at 4 °C.
They were washed twice with cold PBS, then re-suspended in
“
−
1
−1
supplemented with 5 μg mL insulin, 20 ng mL human re-
combinant epidermal growth factor (PeproTech, USA), 10 ng
−
1
mL basic fibroblastic growth factor (PeproTech) and B27
supplement (Gibco, USA). Differentiated cells DC #83 and DC
#
100 were obtained, respectively, from LA #83 and LA #110
by culturing them in specific “stem cell-like differentiating
conditions”; i.e., DMEM/F12 medium with 10% fetal bovine
serum for one week. These differentiated cells, and those de-
rived from the stem-like spheroids, were seeded in 384 well
−
1
5
00 μl PBS containing 50 μg ml protease inhibitor cocktail
−
1
and 50 μg ml RNAse A, and further incubated for 30 min in
the dark at 37 °C. Treated cells were analyzed with a
FACSCanto II apparatus, using the FACSDiva Software 6.0
4
plates at a concentration of 1.25 × 10 cells per mL. After 96
(
BD Biosciences, CA, USA). Ten thousand events were col-
h, the cells were treated (in triplicate) with 20 μM of each
compound, and cell survival was assessed after 72 h by the
MTT Cell Growth Assay (Millipore, USA), according to the
manufacturer's instructions. Results are an average of the
data obtained in triplicate, and are expressed as a percentage
of the value obtained for untreated cells; standard deviation
values were <5%.
lected for each sample. Non-specific fluorescence was moni-
tored using the untreated control cells.
Apoptosis analysis
Cells were treated with the compounds at 15 μM concentra-
tion, for either 48 or 72 h. The cells were harvested by centri-
fugation (1000 × g for 5 min and at room temperature), and
further washed twice with PBS. For apoptosis evaluation, the
ApoFlowEx® FITC Kit (EXBIO, Prague, Czech Republic) was
used, while annexin V staining was performed, according to
the manufacturer's instructions. The cells were finally ana-
lyzed with the FACSCanto II apparatus, using the FACSDiva
Software 6.0. Ten thousand events were collected for each
sample. Non-specific fluorescence was monitored using
annexin V and untreated control cells.
Protein isolation and western blot analysis
All reagents were obtained from Sigma-Aldrich, unless oth-
erwise specified. MCF7 cells were washed with ice cold PBS,
and lysed in 62.5 mM Tris, at pH of 6.8, 100 mM DTT, 2%
SDS w/v and 10% v/v glycerol, where benzonase from Merck-
−
1
Serono was added at a final concentration of 2.5 units ml .
The cellular lysates were homogenized on ice, through no.
2
1
1 gauged needles, and then centrifuged at 20 000 × g for
0 min at 4 °C. The supernatants were immediately ana-
KDM activities
lyzed, or frozen at −76 °C. For western blotting, supernatant
aliquots were spotted onto a PVDF membrane, Immobilon P
Cells were treated with 15 μM of each compound during 48
h. The cells were then harvested by centrifugation (at 1000 ×
g for 5 min at room temperature), and washed with PBS. We
used the EpiQuik™ Nuclear Extraction Kit (Epigentek, NY,
(from Millipore GmbH, Eschborn, Germany), placed in the
Spot-Blot system (GE Healthcare Amersham, Buckingham-
shire, UK), blocked with PBS containing 5% w/v of BSA and
This journal is © The Royal Society of Chemistry 2016
Med. Chem. Commun.

View Article Online
Research Article
MedChemComm
USA), according to the manufacturer's instructions, to obtain Results
nuclear extracts with which KDM activities were evaluated,
Compound synthesis
using the EpiQuik™ histone methyltransferase activity/inhi-
bition assay kits (for H3K9 and for H3K27, Epigentek, NY,
USA) according to the manufacturer's instructions. The
resulting absorbance was measured at 450 nm using the Infi-
nite M200 PRO reader (Tecan, Mannedorf, Switzerland).
The active assayed compounds are presented in Fig. 1. The
synthesis of compounds 5, 6 and 7 has been previously
2
4,25
reported.
The preparation of compounds 1, 2, 3 and 4 is
presented is Scheme 1. The reaction of L-tryptophan with
formaldehyde provided an inseparable mixture of com-
pounds 8 and 9, which was used in the next step without fur-
ther purification. Subsequently, treatment of 8 and 9 with bu-
tyl isocyanate, furnished the final compounds 1 and 2.
DNA methylation evaluation
DNA was treated with bisulfate, using the EZ DNA Methyla-
tion™ Kit according to the manufacturer's instructions. It
was further analyzed using methylation specific PCR, to as-
sess the methylation status of the retinoic acid receptor β sec-
ond exon, using amplicon melt curves. Global
Screening the compound library by western blotting
Western blotting is not a high throughput assay. Therefore,
only a subset of 400 compounds of our COST CM1106 com-
pound library (“Chemical Approaches to Targeting Drug Re-
sistance in Cancer Stem Cells”, at www.stemchem.org) was
screened for their capacity to decrease H3me3 levels for both
K9 and K27. As shown in Fig. 2, all the compounds decreased
H3me3 levels at both K9 and K27, except for compounds 5
and 1 (an enantiomer of compound 3), which did not demon-
strate much activity. Compound 2, an enantiomer of com-
pound 4, had similar inhibitory activity on both H3K9me3
5
-methylcytosine levels in the same specimens were also eval-
uated directly with the 5-mC DNA ELISA,™ according to the
manufacturer's instructions. DNA bisulfitation and ELISA kits
were from ZYMO research, USA.
Molecular modelling
SAM modelling. Pharmacophoric similarity was analyzed
by the following: i) canonical modelling with LigandScout 4.0
(
50%) and H3K27me3 (40%), as did compound 3, which had
2
8,29
from Inte:Ligand,
from Molecular Discovery
similarity search with ROCS software, version 3.2.0.4 from
ii) fingerprint-based analysis with FLAP
a higher activity for the H3K27me3 tag (60%) than for the
H3K9me3 tag (50%). Compound 4 had a low activity (20%)
for H3K27me3 but had higher activity on H3K9me3 (40%)
while compound 6 was very active on the H3K9me3 tag
3
0,31
and iii) shape and chemical
3
2,33
OpenEye.
During pharmacophore or ROCS queries, all
parameters were kept at their default values.
(
100%) and on the H3K27me3 tag (60%). Compound 7 was
Compounds docking into G9a. Docking was performed
active on both tags at K9 (70%) and K27 (30%). By compari-
son, DZNep was quite specific for H3K9me3, while BIX-01294
was more specific for the H3K9me3 level (80%), than for the
H3K27me3 level (25%). Since it is known that IC50 for BIX-
3
4,35
with the FRED docking program from OpenEye,
using
the standard resolution during the preliminary attempt and
the “high” docking resolution during the final run. The crys-
tallographic structure of G9a in the complex with S-adenosyl-
01294 and DNZep varies from 3.3 μM up to 11 μM for the for-
L-homocysteine (SAH), and
a small molecule inhibitor
mer, and from 0.08 μM up to 4.8 μM for the latter, depending
encoded by PDB ID:3K5K was downloaded from the Protein
Data Bank (PDB, www.rcsb.org/pdb) and was used as a rigid
39,40
on the used cell line,
both.
we used adjusted concentrations for
3
6
receptor in molecular docking. Ligand conformational anal-
ysis was performed with Omega2, using default parame-
3
7,38
ters,
and allowing the generation of up to 400 conformers
Cell growth inhibition
for each compound. A self-docking experiment was prelimi-
narily performed with SAH, designed de novo, to prove the
conformational sampling and docking efficacy of selected
programs (ESI‡).
The obtained results are shown in Table 1. All compounds,
except 6 and 7, have IC50 values of about 30–35 μM. Cell
growth inhibition was excellent for compound 6, which had
an IC50 of 5.44 ± 3.2 μM for MCF7 cells.
Fig. 1 Structures of the tested compounds: compound
3 is the
enantiomer of compound 1, and compound 4 is the enantiomer of
compound 2.
Scheme 1 Reagents and conditions: (a) HCHO, H
(b) butyl isocyanate, acetone, DMSO and reflux for 5 h.
2 4 2
SO , H O, r.t., 2 h;
Med. Chem. Commun.
This journal is © The Royal Society of Chemistry 2016

View Article Online
MedChemComm
Research Article
Fig.
3
Cell cycle changes after treatment with each compound.
Compound activity on cell cycle distribution was analysed by flow
cytometry, as detailed in the Materials and methods section. MCF7
were treated with each compound (15 μM) for 24, 48 and 72 h.
Control cells were untreated; standard variations are in grey.
cell accumulation in the G0/G1 phase when cells were treated
with BIX-01294 (ref. 41) and DZNep.
Fig. 2 Analysis of the compounds' effects on histone H3K9 (A) and
H3K27 (B) trimethylation by western blotting. MCF-7 cells were treated
for 48 h with the different compounds at a concentration of 15 μM, ex-
cept for DZNep (0.5 and 1 μM) and for BIX-01294 (3 μM). Western blot
analysis of isolated proteins was performed as specified above, using
specific antibodies against H3K9me3 and H3K27me3. A relative quanti-
fication was achieved by densitometry measurement of immune
detected protein bands, and results are expressed relative to the
untreated control cells (0%).
Compound induced cell apoptosis
Cells were incubated (or not) with 15 μM of each compound
for 48 and 72 h, and their fate was analyzed by flow cytome-
try. Results are presented in Table 2. All compounds seemed
to induce cell apoptosis. The strongest effect was observed
with compound 6, for which late apoptosis was increased of
about 20%. For the other compounds, late apoptosis increase
was about 9%. Necrosis was not observed.
Changes in cell cycle after MCF7 treatment with each
compound
Histone KDM activity is modulated in MCF-7 treated cells
The effect of each compound on cell cycle was analysed by
flow cytometry. MCF7 cells were treated (or not) with 15 μM
of each compound for 24, 48 and 72 h. The results, which are
shown in Fig. 3, indicate that the strongest effect on cell cycle
was induced by compounds 5, 6 and 7, for which the cells
were arrested in G2/M, with earlier and steadier activity for
compound 6; after 72 h, it was about 39.3%, when compared
to untreated cells. Compound 1 showed a mild tendency for
the increased arrest of cells in the S phase; 10.8% when com-
pared to control cells. Compounds 2, 3 and 4 demonstrated a
slight tendency to arrest cells in G0/G1; after 24 h it was
about 3%. In our previous studies, we also determined cancer
KDM activities were determined using nuclear extracts of
MCF7 treated with 15 μM of each compound for 48 h (Fig. 4).
All inhibited H3K9 KDM activity (Fig. 4A), with decreasing ef-
fects from compounds 5, 3, 6, 1, 7 and 2 (the kit reference en-
zyme is G9a). The results obtained for DZNep and BIX-01294
were similar and in almost the same range as for compound
4. However, the resulting activity for the H3K27 assay was not
convincing. We believe that the reference enzyme in this kit
is not relevant for estimating a global H3K27 KDM enzymatic
activity (Fig. 4B). A small activity increase was observed with
compounds 3, 5 and 6; the remaining compounds inhibited
Table 1 Cell growth inhibition for the 7 compounds. The anti-proliferative activities were tested using the MCF-7 cell line after 48 h of treatment with
each compound, using the MTT assay as described above
IC50 (μM) in cell lines
Compounds
MCF7
H460
MSTO-211H
See ref. 12, 33 and 34
Chaetocin
DZNep
BIX-01294
82–153 nM
0.8–0.24 μM
3.3–11 μM
1
2
3
4
5
6
7
37.5 ± 4
>20
>20
>20
>20
>20
0.26 ± 0.07
9.05 ± 1.2
>20
>20
>20
>20
>20
0.54 ± 0.12
6.6 ± 1.8
38.7 ± 7.6
32.3 ± 7.5
30.7 ± 5.6
30.4 ± 3.9
5.4 ± 3.2
34.8 ± 5.9
This journal is © The Royal Society of Chemistry 2016
Med. Chem. Commun.

View Article Online
Research Article
MedChemComm
Table 2 Compound induced cell apoptosis. The pathway of cell death induced by each compound was analysed by flow cytometry. Cells were incu-
bated with each compound at a 15 μM concentration for 48 and 72 h. Control cells were untreated
Compound
Control
Treatment time, h
48
Viable cells, %
Early apoptosis, %
Late apoptosis, %
Necrotic cells, %
74.4 ± 9.1
60.5 ± 6
69 ± 3.5
7.8 ± 5.3
20.1 ± 7.3
13 ± 3.9
27.9 ± 11.7
13.1 ± 2.5
20.6 ± 11.9
12.2 ± 3.6
19.5 ± 7.2
8.3 ± 2.5
7.6 ± 5.9
11.8 ± 2.2
13.8 ± 1.0
17.7 ± 4.6
13.5 ± 1.8
13.7 ± 0.4
11 ± 1.7
12.1 ± 2.8
12.1 ± 2.6
15.2 ± 3.5
11.8 ± 5
10.2 ± 1.9
7.6 ± 1
7
2
1
48
4.2 ± 2.5
6.7 ± 1.2
4.1 ± 2.2
7.1 ± 0.6
6.4 ± 0.7
7.2 ± 1.3
8.5 ± 2.6
7.2 ± 1.6
8.1 ± 4.1
4.8 ± 2.1
9 ± 1.1
8.1 ± 1.7
9.6 ± 2.5
6.2 ± 2.6
6.3 ± 0.9
6.3 ± 1.5
7.8 ± 1.2
5.9 ± 0.8
7
2
47.7 ± 8.3
69.4 ± 3.4
58.6 ± 12.8
70.4 ± 3.5
61.2 ± 5.6
71.1 ± 3.4
52.1 ± 15.7
61.4 ± 5.1
45.2 ± 6.8
51.5 ± 5.5
41.9 ± 10.9
64.6 ± 10.6
54.7 ± 14.2
67.2 ± 4.1
65.8 ± 8.6
32.2 ± 3.2
52.6 ± 7.3
2
48
7
2
3
48
7
2
4
48
7
2
25.5 ± 13.8
18.7 ± 7.6
29 ± 9.3
5
48
7
2
21 ± 4.7
6
48
11.9 ± 9
27.6 ± 4.2
33.2 ± 1.7
12.2 ± 6.1
15.8 ± 3.7
13.8 ± 5.6
14.3 ± 2.7
16.1 ± 4.5
24.1 ± 3.2
7
2
16.8 ± 10.9
13.6 ± 7.1
23.3 ± 13.1
12.6 ± 6.2
13.5 ± 7.1
12.8 ± 6.3
17.4 ± 5.1
7
48
7
2
BIX01294
DZNep
48
7
2
48
7
2
H3K27 KDM activity in the following order: 2, 7, BIX-01294,
DZNep and 4 (Fig. 4B).
sults, expressed as % of control survival, meaning that the
lower the %, the higher is the bio-activity. As can be ob-
served, compound 6 is equally active at the dose of 20 μM in
both the stem cell-like and differentiated cells, with a more
important effect on the stem cell-like state. It is more active
than any other compound, except for kinase inhibitors, and
in particular, it is better than the reference compound BIX-
Screening ovary cancer-like stem cells
The antiproliferative activity of the compounds was also
tested in two ovarian cancer stem cell-like lines (LA #83 and
LA #110), and the differentiated cells derived from them, as
described in Materials and methods. Table 3 reports the re-
01294. Chaetocin was repeatedly found to be too toxic in the
assay to draw useful conclusions.
DNA methylation studies
Using 50 ng of bisulfited DNA, amplicons were generated
with the SYBR green kit from Clinisciences (Paris, France),
according to the manufacturer's instructions. Primer pairing
was at 58 °C for 40 cycles with RAR β unmethylated
sense: TTGGGATGTTGAGAATGTGAGT; RAR β unmethylated
antisense: CTTACTCAACCAATCCAACCAAAACAA; RAR
β
Fig. 4 KDM activities. KDM activities were determined after MCF7
treatment with 15 μM compounds for 48, using treated nuclear
extracts. The values of the resulting enzymatic activity (A – H3K9
MTase activity,
B
–
H3K27 MTase activity) are presented as the
Fig. 5 RAR β methylation status by methylation specific qPCR for DNA
treated with compound 6 gives amplicons with a Tm 75.5 °C, while
control methylated amplicons have a Tm of 77 °C.
difference computed, comparing activities of the treated and
untreated control cells.
Med. Chem. Commun.
This journal is © The Royal Society of Chemistry 2016

View Article Online
MedChemComm
Research Article
Table 3 Proliferation of ovary cell lines no. 83 and 110 with their stem cell-like and differentiated states. The cells were grown as described in Materials
and methods. The average standard deviation was <5%
Compounds
LA #83
DC #83
LA #110
DC #110
Function and class
Tranylcypromine
Pargyline
BIX-01294
Chaetocin
DNZep
1
2
3
4
5
6
7
77.46
77.71
57.0
101.59
96.69
86.15
67.80
63.48
93.27
96.02
91.19
100.77
90.29
65.47
84.25
73.74
94.53
73.13
60.62
85.05
50.2
93.58
59.12
73.0
Histone demethylase inhibitor
Histone demethylase inhibitor
G9a inhibitor
Su (var) 3–9 inhibitor
EZH2 inhibitor
Indole analog
Indole analog
Indole analog
Indole analog
Pifithrin analog
Evodiamine S
Evodiamine derivative
Tyrosine kinase inhibitor
Tyrosine kinase inhibitor
Tyrosine kinase inhibitor
−3.25
42.45
81.93
91.79
64.73
67.81
70.21
53.22
77.26
9.8
−2.63
30.38
66.71
71.84
77.95
55.96
77.14
26.75
56.69
2.6
−4.22
13.64
80.78
93.81
86.39
82.20
80.12
43.09
87.74
54.2
Sunitinib
Erlotinib
Dasatinib
24.7
38.3
27.7
20.5
75.3
58.7
methylated sense: TGTCGAGAACGCGAGCGATTC; RAR β meth-
ylated antisense: CGACCAATCCAACCGAAACGA. Chaetocin and
compounds 6 (Fig. 5) and 7 and 2, 3 and 4 strongly demethyl-
ated DNA. Compound 5 and 1 did not demethylate DNA.
With the ELISA assay (Fig. 6), the most DNA demethyl-
ating compounds were in the following order: 2, 4 and 3; 6, 7
and 1, while for compound 5, there was no demethylation of
DNA, like untreated DNA, labelled here as number 8.
was extracted from all SAM/protein complexes available in
the PDB. The resulting 32 SAM conformations (ESI‡) were
studied by different in silico tools, namely Ligandscout, FLAP
2
8–33,42,43
and ROCS.
In Ligandscout, SAM conformers were
aligned and a ligand-based common features pharmacophore
model, composed of 8 H-bond acceptor features (HBA), 4
H-bond donor features (HBD), 2 aromatic ring features (AR),
2 positive ionisable features (PI), a negative ionisable feature
(NI) and excluded volumes, was generated (Fig. 7A). More-
over, a structure-based pharmacophore model was generated
by merging individual pharmacophores created by
Ligandscout on crystallographic SAM/protein complexes.
Pharmacophoric screening of 1–7 showed that all molecules,
with the only exception being 5, have a good pharmacophoric
similarity to SAM (Table 4).
The generation of the pharmacophore hypothesis in FLAP
was carried out by using the default parameters. All SAM con-
formers were aligned by the program, and a pharmacophore
composed of a combination of molecular interaction fields
generated with the following probes was obtained, and is
shown in Fig. 7B: H, DRY, N1, O, Acc, Don, and Hyd. Results
of the pharmacophoric screening of 1–7 performed with FLAP
The global DNA ELISA suggests that it was the
5
-methylcytosine levels that were decreased by the treatment
of the cells with the tested compounds.
Pharmacophoric analysis
Preliminary docking calculations, performed towards the
crystallographic structure of the G9a, highlighted that com-
pounds 1–7 have a clear preference for the SAM binding site,
rather than for the histone binding site (see below). Accord-
ingly, to depict the possible mechanism of action of 1–7, as
well as to drive further chemical optimization, we were
prompted to investigate their pharmacophoric similarity to
SAM. For this aim, the crystallographic binding mode of SAM
(
Table 4) are consistent with those provided by LigandScout,
showing that 1–7 have a noticeable pharmacophoric similar-
ity with SAM with the only exception of 5.
Finally, chemical similarity analysis was performed with
ROCS. Each SAM conformation was used to generate a ROCS
query (an example is shown in Fig. 7C), which was subse-
quently used to screen the multi-conformers database of 1–7
3
7,38
generated with Omega2.
Results were then averaged
(
Table 3). According to ROCS, 1–7 have shape and chemical
similarity with SAM, but there is a lower discrepancy between
molecules and a lower similarity with SAM was observed, in
comparison with Ligandscout and FLAP. Overall, results of
Table 4 unequivocally show that all compounds have a good
pharmacophoric similarity with SAM, with compound 5 being
the worst in ranking. These in silico data are in good corre-
spondence with results obtained in cellulo, and reinforce the
Fig. 6 Global DNA methylation in MCF-7 cells treated with different
compounds. 5-methylcytosine levels assessed by ELISA as described in
the section “Materials and methods”. Numbers 1–7 represent the cor-
responding compounds, the control sample is labelled number 8. The
SD is in grey.
This journal is © The Royal Society of Chemistry 2016
Med. Chem. Commun.

View Article Online
Research Article
MedChemComm
mechanism of action was consistent with SAM competition,
also from a structure-based perspective, the possible binding
mode of 1–7 towards the crystallographic structure of G9a
was investigated by molecular docking. Compound 6 was
endowed with the highest theoretical affinity for the SAM
binding site of G9a (Table 5). The binding mode of 6 is
shown in Fig. 7. Overall, 6 has a clear shape overlap with
SAM, and has a H-bond interaction with the side chain of
Tyr1154, whereas the highly lipophilic core of 6 interacts with
the hydrophobic residues Trp1159, Phe1166, Met1048 and
Tyr1085 from the SAM binding site in G9a (Fig. 8). Finally,
the carbonyl group of 6 is projected towards the solvent and
may have H-bond interactions with some of the water mole-
cules that are co-crystallized with G9a within the SAM bind-
3
6
ing site.
Discussion
A subset of 400 compounds of our COST CM1106 compound
library (“Chemical Approaches to Targeting Drug Resistance
in Cancer Stem Cells”) was screened by western blotting to
evince the resulting H3K9/K27me3 tag decreased levels. The
seven most active compounds were studied in detail. Studies
included the effects on MCF7 proliferation/apoptosis and cell
cycle. Proliferation was also measured on cancer stem cell-
like ovary tumor cells and their more differentiated stages.
Moreover, the KDM enzymatic activities contained in “com-
pound treated MCF7” was evaluated in vitro using nuclear ex-
tracts of compound treated cells. Only (S)-evodiamine and to
a much lesser extent, compound 7, showed significant KDM
inhibition activities, while (R)-evodiamine and derivatives at
position 5 in both the (R)- and (S)-evodiamine skeletons (14
derivatives) were inactive. As for compound 5, more than 15
analogues bearing a polar group, instead of a methyl group,
at the para position of the right phenyl were tested, but they
were inactive. Regarding the indole analogs (1–4), the pres-
ence of the indole moiety plays a significant role for KDM ac-
tivity, as is the case with chaetocin. This was also observed in
our most active molecule, (S)-evodiamine. Further analogues
are being synthesized in order to understand their structure–
activity relationship. At 200 μM, (S)-evodiamine was already
Fig.
7 Pharmacophoric models and queries. A) Common feature
pharmacophore model generated with Ligandscout; SAM is shown as
sticks, HBA is featured as red spheres, HDB is featured as blue spheres,
PI and NI as blue and red stars, respectively, whereas RE features are
shown as blue circles. Excluded volumes are grey spheres surrounding
other pharmacophore features. Figures were generated with
LigandScout
4.0
(http://www.inteligand.com/ligandscout).
B)
Pharmacophoric model generated by FLAP. A representative SAM
molecule is shown in sticks. H is shown as green mesh, DRY as gold
surface, N1 as blue surface, O as red surface, Acc as pink surface, Don
as violet surface, and Hyd as yellow surface. C) ROCS query. SAM is
shown as sticks. The query's shape is shown as a grey transparent
surface; HBA features are shown as red meshed spheres, HBD as blue
meshed spheres, anions as red spheres and cations as blue spheres.
hypothesis that tested compounds 1–7 may behave as SAM
competitors.
Molecular docking
Tested molecules were found to decrease the histone H3 tri-
methylation on K9 and K27 in cellulo. To check whether this
Table 4 Chemical or pharmacophoric similarity between SAM and compounds 1–7, measured by means of four different methods in silico
Relative pharmacophore-fit score
“ligand-based” (LigandScout)
Relative pharmacophore-fit score
“structure-based” (Ligandscout)
Glob-sum
(FLAP)
Average tanimoto
combo (ROCS)
Molecules
DZNep
BIX-01294
1
2
3
4
5
6
0.78
0.32
0.76
0
0.601
0.483
0.658
0.680
0.722
0.798
0.583
0.625
0.696a
1.203
1.115
0.629
0.669
0.711
0.708
0.713
0.621
0.684
0.648
1.887
0.566
0.548
0.566
0.729
0.370
0.508
0.561
0.997
0.510
0.504
0.499
0.651
0
0.519
0.478
0.996
7
SAM
a
Average of the Glob-sum score of all crystallographic SAM conformers tested against the pharmacophore.
Med. Chem. Commun.
This journal is © The Royal Society of Chemistry 2016

View Article Online
MedChemComm
Research Article
Table 5 Docking-based predicted affinity of 1–7 for the SAM binding site
of G9a
lar level by means of pharmacophore and chemical similarity
analyses, performed on available crystallographic binding
conformations of SAM. One may note that a pharmacophoric
model generated on the SAM binding modes, or a query built
on SAM, would lead to the selection of totally unspecific in-
hibitors of KMTs. In this respect, it is worth mentioning that
we used SAM binding conformations to proteins, which are
chiral molecules, and that our models (particularly those gen-
erated with Ligandscout and FLAP) were enriched by ex-
cluded volumes or shape features that mimic the shape and
steric restrictions of the SAM binding site. Accordingly, this
would allow the selection of small molecules endowed with a
certain degree of specificity for a number of KMTs, which
may require further studies as potential anticancer agents.
Concomitant inhibition of DNA methylation and histone
deacetylation are overcoming tumor resistance to chemother-
Molecule
Chemgauss4 score
1
2
3
4
5
6
7
−9.043
−8.256
−8.403
−8.686
−7.349
−10.253
−9.006
4
5
apy by “epi-resensitization”. Indeed, the failure of targeted
molecules to significantly increase a cancer patient's survival
overall, supports an evidence-based paradigm shift, away
from the systematic avoidance of unspecific molecules like
ours, towards their use for re-sensitization of tumors to clas-
sical therapies, because combined inhibitors of EZH2 and of
DNA methylation exhibit remarkable synergistic anti-
4
6
neoplastic activity in preclinical studies.
Conclusions
Methods
Screening by western blotting was used to identify small mol-
ecules able to decrease H3K9me3 and/or K27me3 levels in
MCF7 treated cells. The activity of positive hits was further
profiled for interference with MCF7 proliferation and cell cy-
cle. KMT enzymatic inhibition, cytosine demethylation and
effects on ovary CSC proliferation were also measured. Exten-
sive molecular modelling studies shed light on the possible
mechanism of action of the active compounds.
Fig.
8
Predicted binding mode of
6
endowed with the highest
predicted affinity to G9a. The molecule is shown as cyan sticks; G9a is
drawn as a cartoon, residues within 5 Å from the ligand are shown as
green lines. The H-bond interaction between 6 and Tyr1154 is shown
as a yellow dashed line, the crystallographic binding mode of SAH is
kept for comparison, and is shown as yellow sticks.
Results
Out of 400 compounds, seven were identified as active (four
chaetocin-like, one pifithrin analog and two evodiamine de-
rivatives). The cellular levels of H3K9me3 and K27me3 were
most noticeably reduced by evodiamine S (compound 6),
which also demetyhlated DNA, suggesting a SAM competition
that was somehow corroborated by extensive in silico studies.
In addition, compound 6 arrested MCF7 cells in the G2/M
phase, had a global cellular pro-apoptotic activity and re-
duced cancer cell proliferation (with an IC50 of 5.44 ± 3.2
μM), like the reference compound BIX-01294. A certain
targeting of ovary cancer stem cells was also seen; in vitro,
G9a was unselectively inhibited.
2
4
reported by us to also be an unspecific sirtuin inhibitor. In
Fig. 3 we observed that compound 6 is very active in blocking
cell cycles and induces early and steady apoptosis, features
4
4
that have also been described for other G9a inhibitors.
Using assays for individual KMTs is laborious and expensive
and therefore, we used a more global enzymatic assay; how-
ever, we observed that the kit used for K27 was not convinc-
ing. It demonstrates that without KMT functional testing
(
like we did), it would have been very unlikely that we would
have seen any KDM inhibition. Since most KDMs function as
complexes of three or more proteins, we believe that in cellulo
conditions are the best for screening.
An in depth molecular modeling study was conducted to
depict the possible mechanism of action of these molecules.
Particularly, preliminary docking simulations suggested that
molecules 1–7 may behave as SAM competitors. This possible
mechanism of action was further investigated at the molecu-
Conclusions
This work highlights the value of functional screening to
identify KDM inhibitors. DNA methyltransferases, KDMs and
other enzymes that use SAM as a co-factor could all be
This journal is © The Royal Society of Chemistry 2016
Med. Chem. Commun.

View Article Online
Research Article
MedChemComm
affected by such molecules. In the context of cancer stem cell
targeting by “epi-resensitization”, such wide effects could be
suitable.
3 R. Firestein, X. Cui, P. Huie and M. L. Cleary, Mol. Cell.
Biol., 2000, 20, 4900–4909.
4 T. Kouzarides, Curr. Opin. Genet. Dev., 2002, 12, 198–209.
5
S. Chuikov, J. K. Kurash, J. R. Wilson, B. Xiao, N. Justin, G. S.
Ivanov, K. McKinney, P. Tempst, C. Prives, S. J. Gamblin,
N. A. Barlev and D. Reinberg, Nature, 2004, 432, 353–360.
S. D. Taverna, H. Li, A. J. Ruthenburg, C. D. Allis and D. J.
Patel, Nat. Struct. Mol. Biol., 2007, 14, 1025–1040.
J. Z. Tan, Y. Yan, X. X. Wang, Y. Jiang and H. E. Xu, Acta
Pharmacol. Sin., 2014, 35, 161–174.
J. Liao, R. Karnik, H. Gu, M. J. Ziller, K. Clement, A. M.
Tsankov, V. Akopian, C. A. Gifford, J. Donaghey, C.
Galonska, R. Pop, D. Reyon, S. Q. Tsai, W. Mallard, J. K.
Joung, J. L. Rinn, A. Gnirke and A. Meissner, Nat. Genet.,
Abbreviations
BSA
CCLD8
CSC
Bovine serum albumin
KMT1F
Cancer stem cell
6
7
8
DMSO
DOTL1
DTT
Dimethylsulfoxide
Disruptor of telomeric silencing 1-like H3K79
Dithiotrithol
DZNep
ESC
EtOAc
EZH2
FLAP
G9a
3-Deazaneplanocin A
Embryonic stem cell
Ethyl acetate
Enhancer of zeste
Fingerprints for ligands and proteins
KMT1C
2
015, 47, 469–478.
9
C. Barbés, J. Sánchez, M. J. Yebra, M. Robert-Geró and C.
Hardisson, FEMS Microbiol. Lett., 1990, 57, 239–243.
1
0 E. Iwasa, Y. Hamashima, S. Fujishiro, E. Higuchi, A. Ito, M.
GLP
H
K
KMT
MS
G9a-like protein or KMT1C
Histone
Lysine
Histone lysine methyl transferase
Mass spectrometry
Methanol
Yoshida and M. Sodeoka, J. Am. Chem. Soc., 2010, 132,
4
078–4089.
1
1 E. Iwasa, Y. Hamashima, S. Fujishiro, D. Hashizume and M.
Sodeoka, Tetrahedron, 2011, 67, 6587–6599.
2 F. L. Cherblanc, K. L. Chapman, J. Reid, A. J. Borg, S.
Sundriyal, L. Alcazar-Fuoli, E. Bignell, M. Demetriades, C. J.
Schofield, P. A. Jr DiMaggio, R. Brown and M. J. Fuchter,
J. Med. Chem., 2013, 56, 8616–8625.
1
MeOH
PBS
Phosphate buffered saline
Parts per million
PPM
PRC2
RIZ1
RAR β
ROCS
SAH
Polycomb repressive complex 2
Retinoblastoma-interacting zinc-finger protein 1
Retinoic acid receptor beta
Rapid overlay of chemical structures
S-Adenosyl-homocysteine
S-Adenosyl-methionine
Suppressor of variegation, enhancer of zeste,
trithorax
1
3 F. L. Cherblanc, K. L. Chapman, R. Brown and M. J. Fuchter,
Nat. Chem. Biol., 2013, 9, 136–137.
1
4 M. Takahashi, Y. Takemoto, T. Shimazu, H. Kawasaki, M.
Tachibana, Y. Shinkai, M. Takagi, K. Shin-ya, Y. Igarashi, A.
Ito and M. Yoshida, J. Antibiot., 2012, 65, 263–265.
5 Y. Yuan, Q. Wang, J. Paulk, S. Kubicek, M. M. Kemp, D. J.
Adams, A. F. Shamji, B. K. Wagner and S. L. Schreiber, ACS
Chem. Biol., 2012, 7, 1152–1157.
SAM
SET
1
1
1
1
1
SETDB1
SD
SDS
KMT1E
Standard deviation
Sodium dodecyl sulfate
6 Y. Yao, P. Chen, J. Diao, G. Cheng, L. Deng, J. L. Anglin,
B. V. Prasad and Y. Song, J. Am. Chem. Soc., 2011, 133,
1
6746–16759.
SUV39H1-2 Suppressor of variegation 3–9 homolog 1–2 or
7 J. Tan, X. Yang, L. Zhuang, X. Jiang, W. Chen, P. L. Lee,
KMT1A and B.
R. K. Karuturi, P. B. Tan, E. T. Liu and Q. Yu, Genes Dev.,
2
007, 21, 1050–1063.
Acknowledgements
8 M. D. Amatangelo, A. Garipov, H. Li, J. R. Conejo-Garcia,
D. W. Speicher and R. Zhang, Cell Cycle, 2013, 12,
We thank the COST organization for partial funding of this
work and the Research Council of Lithuania (Project No. SEN-
2
113–2119.
9 M. T. McCabe, H. M. Ott, G. Ganji, S. Korenchuk, C.
Thompson, G. S. Van Aller, Y. Liu, A. P. Graves, A. Della
Pietra 3rd, E. Diaz, L. V. LaFrance, M. Mellinger, C.
Duquenne, X. Tian, R. G. Kruger, C. F. McHugh, M. Brandt,
W. H. Miller, D. Dhanak, S. K. Verma, P. J. Tummino and
C. L. Creasy, Nature, 2012, 492, 108–112.
1
2/2015). Caterine Zilio was supported by the Région PACA
(DNAfix-2014-02862 and 2014-07199). We thank Molecular Dis-
covery for free access to the software FLAP and the OpenEye
Free Academic Licensing Program for providing us an aca-
demic license for molecular modeling and chemoinformatics
software, and the Center for life Nano Science@Sapienza of
the Instituto Italiano di Tecnologia for financial support.
2
0 R. L. Momparler and S. Côté, Expert Opin. Invest. Drugs,
2
015, 24, 1031–1043.
2
1 E. Curry, I. Green, N. Chapman-Rothe, E. Shamsaei, S.
Kandil, F. L. Cherblanc, L. Payne, E. Bell, T. Ganesh, N.
Srimongkolpithak, J. Caron, F. Li, A. G. Uren, J. P. Snyder,
M. Vedadi, M. J. Fuchter and R. Brown, Clin. Epigenet.,
2015, 84, 1–12.
Notes and references
1
2
K. Murray, Biochemistry, 1964, 3, 10–15.
V. G. Allfrey, R. Faulkner and A. E. Mirsky, Proc. Natl. Acad.
Sci. U. S. A., 1964, 51, 786–794.
Med. Chem. Commun.
This journal is © The Royal Society of Chemistry 2016

View Article Online
MedChemComm
Research Article
2
2 S. Kubicek, R. J. O'Sullivan, E. M. August, E. R. Hickey, Q.
Zhang, M. L. Teodoro, S. Rea, K. Mechtler, J. A. Kowalski,
C. A. Homon, T. A. Kelly and T. Jenuwein, Mol. Cell,
31 S. Cross, M. Baroni, E. Carosati, P. Benedetti and S.
Clementi, J. Chem. Inf. Model., 2010, 50, 1442–1450.
32 ROCS 3.2.1.4: OpenEye Scientific Software, Santa Fe, NM,
http://www.eyesopen.com.
2
007, 25, 473–481.
2
3 M. Vedadi, D. Barsyte-Lovejoy, F. Liu, S. Rival-Gervier, A.
Allali-Hassani, V. Labrie, T. J. Wigle, P. A. Dimaggio, G. A.
Wasney, A. Siarheyeva, A. Dong, W. Tempel, S. C. Wang, X.
Chen, I. Chau, T. J. Mangano, X. P. Huang, S. D. Simpson,
S. G. Pattenden, L. J. Norris, D. B. Kireev, A. Tripathy, A.
Edwards, B. L. Roth, W. P. Janzen, B. A. Garcia, A. Petronis,
J. Ellis, P. J. Brown, S. V. Frye, C. H. Arrowsmith and J. Jin,
Nat. Chem. Biol., 2011, 7, 566–574.
33 P. C. D. Hawkins, A. G. Skillman and A. Nicholls, J. Med.
Chem., 2007, 50, 74–82.
34 FRED 3.0.1: OpenEye Scientific Software, Santa Fe, NM, http://
www.eyesopen.com.
35 M. McGann, J. Chem. Inf. Model., 2011, 51, 578–596.
36 F. Liu, X. Chen, A. Allali-Hassani, A. M. Quinn, G. A.
Wasney, A. Dong, D. Barsyte, I. Kozieradzki, G. Senisterra, I.
Chau, A. Siarheyeva, D. B. Kireev, A. Jadhav, J. M. Herold,
S. V. Frye, C. H. Arrowsmith, P. J. Brown, A. Simeonov, M.
Vedadi and J. Jin, J. Med. Chem., 2009, 52, 7950–7953.
37 OMEGA 2.5.1.4: OpenEye Scientific Software, Santa Fe, NM,
http://www.eyesopen.com.
38 P. C. D. Hawkins, A. G. Skillman, G. L. Warren, B. A. Ellingson
and M. T. Stahl, J. Chem. Inf. Model., 2010, 50, 572–584.
39 N. A. Malmquist, T. A. Moss, S. Mecheri, A. Scherf and M. J.
Fuchter, Proc. Natl. Acad. Sci. U. S. A., 2012, 109, 16708–16713.
40 J. Zhou, C. Bi, L. L. Cheong, S. Mahara, S. C. Liu, K. G. Tay,
T. L. Koh, Q. Yu and W. J. Chng, Blood, 2011, 118, 2830–2839.
41 J. Savickiene, G. Treigyte, I. Stirblyte, G. Valiuliene and R.
Navakauskiene, Leuk. Res., 2014, 38, 822–829.
2
2
4 M. S. Christodoulou, A. Sacchetti, V. Ronchetti, S. Caufin, A.
Silvani, G. Lesma, G. Fontana, F. Minicone, B. Riva, M.
Ventura, M. Lahtela-Kakkonen, E. Jarho, V. Zuco, F. Zunino,
N. Martinet, F. Dapiaggi, S. Pieraccini, M. Sironi, L. Dalla
Via, O. M. Gia and D. Passarella, Bioorg. Med. Chem.,
2
013, 21, 6920–6928.
5 M. S. Christodoulou, F. Calogero, M. Baumann, L. Dalla Via,
O. M. Gia, S. Pieraccini, M. Sironi, F. Dapiaggi, R. Bucci, G.
Broggini, S. Gazzola, S. Liekens, A. Silvani, M. Lahtela-
Kakkonen, N. Martinet, A. Nonel-Canals, I. Baxendale and D.
Passarella, Eur. J. Med. Chem., 2015, 92, 766–775.
2
2
6 F. Ricci, F. Bizzaro, M. Cesca, F. Guffanti, M. Ganzinelli, A.
Decio, C. Ghilardi, P. Perego, R. Fruscio, A. Buda, R. Milani,
P. Ostano, G. Chiorino, M. R. Bani, G. Damia and R.
Giavazzi, Cancer Res., 2014, 74, 6980–6990.
7 F. Ricci, S. Bernasconi, P. Perego, M. Ganzinelli, G. Russo, F.
Bono, C. Mangioni, R. Fruscio, M. Signorelli, M. Broggini
and G. Damia, Cell Cycle, 2012, 11, 1966–1976.
42 M. Mori, Y. Cau, G. Vignaroli, I. Laurenzana, A. Caivano, D.
Vullo, C. T. Supuran and M. Botta, ACS Chem. Biol.,
2015, 10, 1964–1969.
43 T. M. Steindl, D. Schuster, C. Laggner, K. Chuang, R. D.
Hoffmann and T. Langer, J. Chem. Inf. Model., 2007, 47,
563–571.
2
2
3
8 G. Wolber, A. A. Dornhofer and T. Langer, J. Comput.-Aided
Mol. Des., 2006, 20, 773–788.
9 G. Wolber and T. Langer, J. Chem. Inf. Model., 2005, 45,
44 Q. Yang, Z. Lu, D. Singh and J. U. Raj, Cell Proliferation,
2012, 45, 335–344.
45 B. Oronsky, N. Oronsky, J. Scicinski, G. Fanger, M. Lybeck
and T. Reid, Transl. Oncol., 2014, 7, 626–631.
46 R. L. Momparler and S. Côté, Expert Opin. Invest. Drugs,
2015, 24, 1031–1043.
1
60–169.
0 M. Baroni, G. Cruciani, S. Sciabola, F. Perruccio and J. S.
Mason, J. Chem. Inf. Model., 2007, 47, 279–294.
This journal is © The Royal Society of Chemistry 2016
Med. Chem. Commun.