Mannich bases of 1,2,4-triazole-3-thione containing adamantane moiety: Synthesis, preliminary anticancer evaluation, and molecular modeling studies

Article abstract of DOI:10.1111/cbdd.12920

A series of 18 novel N-Mannich bases derived from 5-adamantyl-1,2,4-triazole-3-thione was synthesized and characterized using NMR spectroscopy and X-ray diffraction technique. All derivatives were evaluated for their anticancer potential against four human cancer cell lines. Several tested compounds exerted good cytotoxic activities on K562 and HL-60 cell lines, along with pronounced selectivity, showing lower cytotoxicity against normal fibroblasts MRC-5 compared to cancer cells. The effects of compounds 5b, 5e, and 5j on the cell cycle were investigated by flow cytometric analysis. It was found that these compounds cause the accumulation of cells in the subG1 and G1 phases of the cell cycle and induce caspase-dependent apoptosis, while the anti-angiogenic effects of 5b, 5e, and 5j have been confirmed in EA.hy926 cells using a tube formation assay. Further, the interaction of Bax protein with compound 5b was investigated by means of molecular modeling, applying the combined molecular docking/molecular dynamics approach.

Full text of DOI:10.1111/cbdd.12920

Received Date : 02-Sep-2016
Revised Date : 08-Nov-2016
Accepted Date : 21-Nov-2016
Article type : Research Article
Mannich Bases of 1,2,4-Triazole-3-thione Containing Adamantane Moiety:
Synthesis, Preliminary Anticancer Evaluation, and Molecular Modeling
Studies
Milorad Z. Milošev,1 Katarina Jakovljević,2 Milan D. Joksović,2 Tatjana Stanojković,3 Ivana Z.
Matić,3 Milka Perović,4 Vesna Tešić,4 Selma Kanazir,4 Milan Mladenović,2 Marko V. Rodić,5
Vukadin M. Leovac,5 Snežana Trifunović6 and Violeta Marković2,*
1Faculty of Medicinal Science, University of Kragujevac, S. Markovića 69, 34000 Kragujevac, Serbia
2Faculty of Science, Department of Chemistry, University of Kragujevac, R. Domanovića 12, 34000 Kragujevac,
Serbia
3Institute of Oncology and Radiology of Serbia, Pasterova 14, 11000 Belgrade, Serbia
4Institute for Biological Research “Siniša Stanković”, Department of Neurobiology, University of Belgrade, Bulevar
Despota Stefana 142, 11060 Belgrade, Serbia
5Faculty of Sciences, University of Novi Sad, Trg D. Obradovića 3, 21000 Novi Sad, Serbia
6Faculty of Chemistry, University of Belgrade, Studentski trg 12–16, 11000 Belgrade, Serbia
*Corresponding author: Violeta Marković, markovicvioleta@kg.ac.rs
Abstract
A series of 18 novel N-Mannich bases derived from 5-adamantyl-1,2,4-triazole-3-thione was synthesized
and characterized using NMR spectroscopy and X-ray diffraction technique. All derivatives were
evaluated for their anticancer potential against four human cancer cell lines. Several tested compounds
exerted good cytotoxic activities on K562 and HL-60 cell lines, along with pronounced selectivity,
showing lower cytotoxicity against normal fibroblasts MRC-5 compared to cancer cells. The effects of
compounds 5b, 5e and 5j on the cell cycle were investigated by flow cytometric analysis. It was found
that these compounds cause the accumulation of cells in the subG1 and G1 phases of the cell cycle and
induce caspase-dependent apoptosis, while the anti-angiogenic effects of 5b, 5e and 5j have been
This article has been accepted for publication and undergone full peer review but has not been
through the copyediting, typesetting, pagination and proofreading process, which may lead to
differences between this version and the Version of Record. Please cite this article as doi:
10.1111/cbdd.12920
This article is protected by copyright. All rights reserved.

confirmed in EA.hy926 cells using a tube formation assay. Further, the interaction of Bax protein with
compound 5b was investigated by means of molecular modeling, applying the combined molecular
docking/molecular dynamics approach.
Keywords: triazoles, Mannich bases, adamantane, anticancer activity, molecular modeling
Introduction
The chemistry of 1,2,4-triazoles and their fused heterocyclic derivatives has received
considerable attention owing to their synthetic and biological importance. Among them, thione-
substituted 1,2,4-triazole ring systems have been well studied due to a wide variety of their biological
activities, such as antibacterial,[1] antifungal,[2] anticancer,[3] anti-inflammatory,[4] and hypoglycemic[5]
properties. Additionally, 1,2,4-triazoles are important precursors for other heterocyclic compounds, such
as Mannich bases,[6] thioureas,[7] Schiff bases,[8] triazolothiadiazoles, triazolothiadiazines[9] etc. Some
derivatives of 4,5-substituted-1,2,4-triazole-5-thiones were synthesized and found to be good cytotoxic
agents against thymocytes.[10] Anticancer screening studies for a series of Mannich bases of 3-substituted
4-(5-nitro-2-furfurylidene)amino-1,2,4-triazole-5-thiones were performed against a panel of 60 cell lines
showing the GI50 values in the range of 10-100 μM.[3]
The adamantyl group is known to positively regulate the therapeutic index of many compounds,
and seven of its derivatives found use in current clinical practice, in treatment of viral infections,
neurodegenerative disorders, acne vulgaris and type 2 diabetes mellitus.[11] RARγ-selective adamantyl-
based compound induces G0/G1 cell cycle arrest and apoptosis in human breast carcinoma and human
leukemia cell lines in a p53-independent fashion.[12] It is also found to be a potent growth inhibitor of cell
lung cancer (NSCLC) lines with IC50 values in the range of <0.13 − 0.53 μM.[13]
The scientific literature regarding anticancer activity of compounds which contain both
adamantoyl and 1,2,4-triazole moieties is very scarce. One of such compounds, 4-(1-adamantyl)-5-[2-(3-
hydroxynapthyl)]-2H-1,2,4-triazole-3(4H)-thione was evaluated for its in vitro anticancer activity against
MCF-7 breast cancer cell line giving IC50 value of 8.15 μM.[14]
Most common chemotherapeutic drugs, doxorubicin, cisplatin and 5-fluorouracil, used for
treating a wide range of cancers, induce reduced cancer cell division and growth. Chemotherapy does not
distinguish between a cancer and normal cells and destroys not only the fast-growing cancer cells but also
healthy cells in the body. Therefore, it is of most importance to design and perform biological screening
of new pharmacophores, which could potentially enhance the selectivity of cytotoxic agents against
cancer cell lines and decrease their toxic effects towards normal cells. Numerous newly synthesized
compounds which possess excellent cytotoxic activity, display very low cancer selectivity. Having in
mind these facts, in this paper we examined cytotoxicity, cancer selectivity and plausible mechanism of
action of 18 novel N-Mannich bases of 5-adamantyl-1,2,4-triazole-3-thione, which displayed a
pronounced selectivity showing IC50 values for MRC-5 significantly higher in comparison to cancer cell
lines. In addition, molecular action mode of these compounds was investigated through western blot
analysis of key proteins involved in apoptosis, anti-angiogenic activity and molecular docking studies.
This article is protected by copyright. All rights reserved.

Methods and Materials
Physical measurements and methods
Melting points were determined on a Mel-Temp capillary melting points apparatus, model 1001
and are uncorrected. Elemental (C, H, N, S) analysis of the samples was carried out in the Center for
Instrumental Analysis, Faculty of Chemistry, Belgrade. IR spectra were recorded on a Perkin Elmer
1
Spectrum One FT-IR spectrometer with a KBr disc. H and 13C NMR spectra were recorded on a Varian
Gemini 200 MHz spectrometer. Crystal structures were determined by single-crystal X-ray
crystallography. Diffraction measurements were performed on an Oxford Diffraction Gemini S
diffractometer. Structures were solved by using the SHELXT,[15] and refined with the SHELXL.[16] Details
of preparation and structure determinations are given in the Supporting information.
In vitro cytotoxicity assay
Human cervical adenocarcinoma HeLa, human chronic myelogenous leukemia K562, human
acute promyelocytic leukemia HL-60, human lung carcinoma A549 and normal human lung fibroblasts
MRC-5 were grown in RPMI 1640 medium (Sigma Aldrich). Nutrient medium was supplemented with
10% fetal bovine serum, L-glutamine, and penicillin-streptomycin (Sigma Aldrich). Medium for HL-60
cells was supplemented with 20% fetal bovine serum. HeLa (2000 cells per well), K562, A549 and MRC-
5 (5000 cells per well) and HL-60 (7000 cells per well), were seeded in 96-well plates. The tested
compounds were added to the cells 20 h later (for adherent cell lines) or 2 h later (for cell lines that grow
in a suspension). The nutrient medium was added to control cells. Final concentrations of compounds
were 6.25 μM, 12.5 μM, 25 μM, 50 μM and 100 µM. Cisplatin was used as a positive control. After 72 h
treatment, the survival of cells was measured by MTT test, as described earlier.[17,18] More details of
determination of cell survival are given in the Supporting information.
Cell cycle analysis
Chronic myelogenous leukemia K562 cells were seeded into a 6-well plates (2.5x105 cells/well), and after
2 h treated with compounds, with the exception of control cells. Tested concentrations corresponded to
IC50 and 2IC50 values. After 24 h treatment, the cells were collected, and fixed as described elsewhere.[19]
Cell cycle distribution was analyzed after incubation of cells with RNase and propidium iodide. The
statistical significance of differences in percentages of cells at specific cell cycle phases between control
samples and samples exposed to tested compounds was evaluated using Repeated Measures ANOVA
with Tukey’s multiple comparison test. The p-values less than 0.05 were regarded as statistically
significant.
Morphological analysis of cell death mode
K562 cells were seeded into 6-well plates in 2 mL of complete nutrient medium (2x105 cells/well ). After
2 h, the cells were treated with 2IC50 concentrations of the investigated compounds. After 24 h treatment,
the cells were collected by centrifugation and stained with acridine orange/ethidium bromide mixture (3
μg/ml AO and 10 μg/ml EB in phosphate-buffered saline), and visualized under a fluorescence
microscope (Carl Zeiss PALM MicroBeam with Axio Observer.Z1 using AxioCam MRm).
This article is protected by copyright. All rights reserved.

Western blot analysis
Whole cell lysates of K562 cells (2×106) treated with compounds 5b, 5e and 5j (IC50
respectively) for 24h were prepared by homogenization of harvested cells in buffer containing 0.1%
Triton X-100 (Sigma Aldrich) and triple protease inhibitor (Roche Applied Science). Following
sonication (3x10 s) using UP 50H Ultraschallsrozessor (Hielscher GmbH) and centrifugation (30 min/20
000xg) of lysates, protein concentration of the supernatant was determined by BCA assay (Thermo
Scientific Pierce). Samples were separated on an SDS-PAGE and transferred to a PVDF membrane as
described previously.[20] To block non-specific binding, the membranes were incubated for 1h at room
temperature (RT) in 5% non-fat dry milk/TBST (150 mM NaCl, 50 mM Tris, pH 7.4, and 0.05%
Tween20). The following antibodies were used for overnight incubation at 4 °C in 2% blocking solution:
rabbit anti-caspase 3 (Santa Cruz Biotechnology), rabbit anti-cleaved caspase 3 (Cell Signaling), rabbit
anti-caspase 8 (Santa Cruz Biotechnology), mouse anti-caspase 9 (Santa Cruz Biotechnology) and mouse
anti-Bax (1:300, Santa Cruz Biotechnology). After several rinses in TBST, the membranes were
incubated for 1h in the appropriate Horse Radish Peroxidase (HRP)-conjugated secondary antibodies
(Santa Cruz Biotechnology). HRP-immunoreactive bands were visualized using enhanced
chemiluminescence (ECL, GE Healthcare) and film exposure (Kodak Biomax). Each blot was
subsequently re-probed with goat anti-GAPDH (glyceraldehyde 3-phosphate dehydrogenase) antibody
(Santa Cruz Biotechnology), as a control for protein load.
Tube formation assay
To examine the in vitro antiangiogenic activity of selected compounds 5b, 5e and 5j on human umbilical
vein endothelial EA.hy926 cells, the tube formation assay was performed. Briefly, 24-well plates were
coated with Corning® Matrigel® Basement Membrane Matrix, (Corning, 356234). EA.hy926 cells were
seeded onto the coated wells (4x104 cells per well), with or without the investigated compounds.
Concentrations of the compounds 5b, 5e and 5j corresponded to IC20 values (15 µM) previously
determined for this cell line for 24h treatment. After 24 h of incubation, the cells were photographed
under the inverted microscope.
Preparation of enzymes crystal structures
Molecular docking, homology modeling and molecular dynamics protocols were conducted using
experimentally available Bax crystal structure (PDB ID: 1F16),[21] Bax crystal structure in complex with
Bcl-XL (PDB ID: 3PL7)[22] and Bcl-2 (PDB ID: 2XA0),[23] or Bax with C terminus truncated in complex
with CHAPS (PDB ID: 4BH8),[24] retrieved from the Protein Data Bank. Details of molecular docking
studies are given in the Supporting information.
Results and Discussion
Chemistry
The synthetic pathway for the targeted novel N-Mannich bases of 5-adamantyl-1,2,4-triazole-3-
thione is presented in Scheme 1. The synthesis starts from a commercially available 1-
adamantanecarboxylic acid 1 which was converted into 1-adamantanecarbonyl chloride 2 in the reaction
with thionyl chloride in the presence of catalytic amounts of N,N-dimethylformamide (DMF). The
corresponding acylthiosemicarbazide 3 was obtained with very high purity by stirring acid chloride 2 with
This article is protected by copyright. All rights reserved.

thiosemicarbazide in dry tetrahydrofuran (THF) at room temperature. The intramolecular cyclization of
acylthiosemicarbazide 3 was performed by addition of 2M aqueous solution of NaOH and refluxing for 3
hours. After refluxing and cooling, the clear solution was acidified using 2M HCl until pH 1 was reached
followed with the formation of the precipitate of the 5-adamantyl-1,2,4-triazole-3-thione 4. The final
compounds 5a-r were prepared in a Mannich reaction which included stirring of the equimolar
concentrations of 5-adamantyl-1,2,4-triazole-3-thione 4, formalin and corresponding primary aromatic
amine in THF for two days. Then, the solvent was evaporated, and EtOH was added to the residue. The
compounds 5a-r were obtained with excellent purity excluding any need for further purification.
1
The structure of all synthesized compounds was determined by means of H and 13C NMR
spectroscopy (see Supporting information, Figures S1-S38).
Scheme 1: Reagents and conditions: a) SOCl2, DMF, CH2Cl2, 2 h, r.t.; b) NH2NHC(S)NH2, THF, 6 h, r.t.; c) NaOH,
3h, reflux; HCl; d) CH2O, Ar-NH2, THF, 48 h, r.t.
Theoretically, the alkylation of 5-substituted-1,2,4-triazole-3-thione can be performed at N2 and
N4 atom of triazole ring. Herein, we have obtained a suitable single crystal of 5e and by means of X-ray
crystallography unambiguously demonstrated that the aminomethylation occurs exclusively at N2 atom of
triazole moiety (Figure 1, CCDC 1465502–1465503). For more details see Supporting information (cif
file and checkcif report).
Figure 1: Molecular structure of 5e. Molecule A of the monoclinic polymorph is chosen as the representative.
Biology
In vitro cytotoxicity
The cytotoxic activity of 5-phenyl-1,2,4-triazole-3-thione was determined, proving this
compound to be inactive against K562 cell line showing IC50 value of 200.00 µM. In order to determine
its influence on the activity, the corresponding phenyl group was substituted with adamantyl group. The
cytotoxic activity of the synthesized 5-adamantyl-1,2,4-triazole-3-thione (4) was also determined against
K562 cells and the IC50 value for this compound was 101.38 µM. This increase in the activity, caused by
the incorporation of the adamantyl group, has encouraged us to perform the synthesis of the
corresponding Mannich bases starting from the intermediate 4 and to investigate their cytotoxic potential.
Cytotoxic activity of the newly synthesized triazoles was evaluated against four malignant cell lines
(HeLa, K562, HL-60 and A549) as well as against normal, non-transformed MRC-5 cells. The obtained
IC50 values are shown in Table 1. Most of the compounds that displayed the highest cytotoxic activities
against K562 cell line possess a substituent at ortho-position of the phenyl ring (5b, 5e, 5h, 5j). On the
other hand, all the tested compounds showed comparable activity against HL-60 cells, where compounds
containing different substituents at meta-position (5c, 5f, 5k) exerted the weakest activity. The electronic
effect of the substituent does not have a noticeable influence on the activity, since both electron-donating
and electron-withdrawing substituents showed similar effectiveness against HL-60 cancer cells. Some of
the active and highly selective compounds against K562 cell line 5b, 5e and 5j were chosen for further
examination of the mechanisms of their anticancer activity.
This article is protected by copyright. All rights reserved.

Table 1: The in vitro cytotoxic activity of investigated compounds
Cell cycle analysis and fluorescence microscopy
We examined the mechanisms of cytotoxic activity of the selected compounds 5b, 5e and 5j by
flow cytometric analysis, using propidium iodide to label DNA. Figure 2 shows cell cycle distribution of
K562 myelogenous leukemia cells incubated in the absence or presence of investigated compounds. The
obtained results demonstrate that after 24 h treatment with IC50 and 2IC50 concentrations of triazoles 5e,
and 5j the percentages of K562 cells in the G1 cell cycle phase were significantly higher compared with
the percentage of cells within G1 phase in the control cell sample. The G1 cell cycle phase accumulation
was also observed in K562 cells exposed to compound 5b, although this increase was not statistically
significant. The increase in the G1 phase cell population was accompanied by a decrease in the G2/M
phase cell populations. Our results are in accordance with literature data about different triazoles which
induce G1 cell cycle phase arrest.[25,26] In addition, treatment of K562 cells with 2IC50 concentrations of
5b, 5e, and 5j led to increase of cells in the subG1 phase in comparison to control cells.
Figure 2: Changes in the cell cycle phase distribution of human myelogenous leukemia K562 cells induced by the
compounds 5b, 5e and 5j after 24h treatment (tested concentrations corresponded to IC50 (A) and 2IC50 values (B)).
The results are presented as the mean S.D. from two independent experiments. Statistically significant differences
between control and treated cells are marked with * (p<0.05).
After 24 h treatment with triazoles 5b, 5e and 5j, K562 cells were stained with a mixture of
acridine orange and ethidium bromide and analyzed by fluorescence microscopy. As shown in Figure 3,
the morphological changes typical for apoptotic cell death were observed in K562 cells treated with 2IC50
concentrations of compounds. In the non-treated control the K562 cells had normal morphology. In
contrast, early apoptotic cells stained green with condensed chromatin in the shrunken nuclei or even
fragmented nuclei could be seen in the cell samples exposed to triazoles in addition to orange-red stained
late apoptotic cells. The obtained results indicate that 5b, 5e and 5j are able to induce apoptotic cell death
in myelogenous leukemia K562 cells.
Figure 3: Photomicrographs of acridine orange/ethidium bromide stained control K562 cells, and K562 cells treated
with 2IC50 concentrations of compounds 5b, 5e and 5j for 24 h.
Western blot analysis
Molecular mechanisms underlying cytotoxic activity of the compounds 5b, 5e and 5j were further
demonstrated by Western blot analysis (Figure 4). Activation of both principal apoptotic signal
pathways[27] results in the activation of cysteine aspartic acid-containing proteases (caspases). Caspase-8
and caspase-9 are usually involved in extrinsic or death receptor-mediated pathway and the intrinsic or
mitochondria-mediated pathway, respectively, while caspase-3 is downstream of both caspase-8 and
caspase-9. The caspases are activated by cleavage, which further activates the downstream substrate
molecules.[28]
This article is protected by copyright. All rights reserved.

Figure 4: Induction of caspase-dependent apoptosis by compounds 5b, 5e and 5j at IC50 concentrations in K562
myelogenous leukemia cells after 24h treatment. Immunoblotting was performed using anti-caspase 3, anti-cleaved
caspase 3, anti-caspase 8, anti-caspase 9 and anti-Bax antibodies. The anti-GAPDH antibody was used as a control
for protein load. Besides full-length caspases – procaspases, the fragments of activated caspase-3 (subunits p19 and
p17), fragments of activated caspase-8 (the cleaved intermediates p43/p41, the active fragments/subunits p18, p16
and p10) and fragments of activated caspase-9 (subunits p37 and p35) were detected.
The protein levels of full-length caspase-8 and caspase-3 in cells treated with compounds 5b, 5e
and 5j were increased in comparison to untreated, control cells. As shown in Figure 4, this increase was
accompanied by strong appearance of active fragments. The most intensive effect was recognized for the
compound 5j, while the effects declined for compounds 5b and 5e, respectively. Although an increase in
protein levels of full-length caspase-9 was also observed in treated cells, it was less prominent comparing
to other procaspases examined.
We further examined whether the protein expression levels of Bax were altered following the
treatment with compounds 5b, 5e and 5j. Bax is a member of the Bcl-2 protein family. As key regulators
of apoptosis, these proteins are involved in the control of mitochondrial permeability and, in particular,
the release of apoptogenic proteins from this organelle.[29] Western blot analysis showed that Bax
expression was substantially upregulated in K562 cells treated with all three compounds tested.
Furthermore, a smaller band of
̴18-kDa was detected in addition to a 21-kDa band in treated cells.
Detected cleavage fragment of Bax protein, p18, has been reported to be produced by calpain and to
accelerate apoptosis in numerous tumor cell lines following treatment with various chemotherapeutic or
biological agents.[30] In contrast, 18-kDa Bax was not detected in the control cells. We further
demonstrated that caspase-dependent apoptosis is involved in compounds 5b-, 5e- and 5j-induced
cytotoxicity in K562 cells, and that both intrinsic and extrinsic pathway seem to be involved in induction
of apoptosis. The intrinsic pathway is characterized by mitochondrial dysfunction with release of caspase
activators, followed by activation of caspase-9 and caspase-3. Mitochondrial-induced apoptosis further
requires involvement of the Bcl-2 family, including Bax. Caspase-8, on the other hand, is one of the
upstream mediators of cell death and its activation is associated with an increase in TNF-like cytokines,
such as TNF-α and Fas ligand (FasL; CD95).[31] Caspase-8 is inactivated in a variety of human cancers,
which may promote tumor progression as well as resistance to current treatment approaches.[32,33]
A
significant increase in its full length and cleaved fragments is consistent with the high cytotoxic efficacy
of selected compounds. In addition, augmentation of apoptosis in treated cells can also be attributed to up-
regulation of calpain-produced Bax fragment. Previous studies have demonstrated that 18-kDa Bax, a
characteristic feature of Bax activation, is a more potent inducer of apoptosis than 21-kDa Bax.[34] Further
research confirmed that interruption of Bax cleavage reduces drug-induced apoptosis, while depression of
18-kDa Bax degradation significantly augments drug-mediated apoptosis.[35]
Anti-angiogenic activity
Angiogenesis represents the formation of new blood vessels and has a fundamental role in
development, reproduction and repair of the tissue including tumors.[36,37] Dawson et al. have showed that
an adamantyl-substituted retinoid-derived molecule effectively reduces human microvascular endothelial
(HMVE) cell growth (IC50=0.3 μM) and suggested its potential in vivo antiangiogenic activity.[38]
Itraconazole, a compound containing 1,2,4-triazole, has been found to possess potent and selective
inhibitory activity against multiple key aspects of tumor-associated angiogenesis, both in vitro and in
vivo.[39] One of the most frequently used and best characterized permanent human vascular endothelial
cell line in angiogenesis research is EA.hy926.[40] To evaluate the anti-angiogenic potential of the
compounds 5b, 5e and 5j applied at subtoxic concentrations, a tube formation assay, also known as in
This article is protected by copyright. All rights reserved.

vitro angiogenesis assay was performed. As it could be seen in the Figure 5 reduced number of elongated
and connected EA.hy926 cells in addition to significantly inhibited formation of tubular and polygonal
structures was observed in the cell samples treated with compounds 5b, 5e and 5j compared with control
cells. The compound 5j exerted more pronounced anti-angiogenic effect in comparison to 5b and 5e. Our
results point out the significant anti-angiogenic potential of investigated triazoles 5b, 5e, 5j.
Figure 5: Effects on in vitro angiogenesis of EA.hy926 cells. Photomicrographs of control EA.hy926 cells and cells
exposed to subtoxic IC20 concentrations (15 µM) of compounds 5b, 5e and 5j after 20 h treatment.
The mechanism of 5b pro-apoptotic activity
The treatment of chronic myelogenous leukemia K562 cells with compound 5b elevated levels of
pro-apototic protein Bax, caspase-8 and main effector caspase-3, demonstrating that this triazole triggered
both intrinsic and extrinsic apoptotic pathways.[41] In modern apoptosis-based anticancer therapy, one of
the recognized approaches is to identify small molecules that activate pro-apoptotic Bcl-2 proteins[42] like
Bax. Consistently, 5b, as confirmed inducer of apoptosis, might exert its anticancer action as small-
molecule Bax agonist.[43]
Thus, 5b proapoptotic mode of action is exerted most likely via the displacement model
pathway[44] which involves: 1) the Bax activation, as an executioner protein, with agonist, 2) the
interaction of Bax-5b proapototic complex with anti-apoptotic enzymes Bcl-XL and/or Bcl-2, 3) the
extrusion of Bax-5b from anti-apoptotic enzymes with BH3 sensitizer proteins like Bad, Bik, Bmf, Bnip3,
Noxa or Hrk, 4) the Bax-5b penetration into the outer mitochondrial membrane (MOM) and dimerisation,
5) membrane disruption to release intermembrane space (IMS) proteins such as cytochrome c,
OMI/HTRA2, SMAC/DIABLO, and endonuclease G,[45] 6) the activation of caspase-3 as an effector of
apoptosis.[46] The displacement model steps 1), 2) (see Supporting information), and 4) (see Supporting
information) were considered in this report by means of molecular modelling to reveal the pharmacology
of 5b, applying the combined molecular docking/molecular dynamics approach.
Bax agonists target the C-terminal α9 helix.[42,43] To the best of our knowledge there is no
available crystal data regarding the binding modes of any Bax agonists, while there is limited structure-
based derived information about small-molecule Bax agonists positioning.[42,43] Therefore, the blind
docking procedure had been applied in order to outline the 5b bioactive conformation (Figure 6) using the
scoring function available in AutoDock4.2[47] program. During the Bax activation (step 1), 5b was
stabilized within the α9 helix[42] mainly due to the established advantageous hydrophobic interactions
with the helix residues. Hence, the adamantyl moiety was involved in the steric clash with Val180 side
chain. The o-tolyl scaffold was encircled by Ala183 and Ile187 where the hydrophobic interference is
constituted via the aromatic ring inasmuch as the o-methyl group was pointed away from the C-terminal
domain. The orientation of adamantyl and o-tolyl groups provided some more interesting observations.
Both of the functionalities were positioned in front of the Ser184, a key amino acid appointed as a target
for post-translational phosphorylation during the prevention of apoptotic process[42] by anti-apoptotic
proteins. If this post-translational modification is avoided, Ser184 provides the C-terminal domain
stabilization by means of C-terminus helix merging with the outer mitochondrial membrane.
Consequently, the pro-apototic protein Bax prevails in effort to induce apoptosis.[48]
Small-molecule Bax agonists induce conformational changes in Bax by blocking Ser184
phosphorylation, facilitating Bax insertion into mitochondrial membranes and forming Bax oligomers.[43]
The role of the 5b as the small-molecule Bax agonist in the induction of apoptosis will be discussed lately
This article is protected by copyright. All rights reserved.

in text but at this point it is crucial to accentuate that adamantyl and o-tolyl groups both served as a sort of
spatial gate in front of the Ser184 and therefore prevented the amino acid phosphorylation before the
actual interaction of Bax with the anti-apoptotic proteins.
The second level of Bax-5b complex stabilization was provided by 1,2,4-triazole-3(4H)-thione
ring where sulfanylidene group as hydrogen-bond (HB) acceptor was narrow to Ser101 forming a strong
hydrogen bond (dHB = 2.501 Å), while the HB donor nitrogen at position 4 of heterocyclic ring was in the
close proximity to Asp102 (dHB = 1.767 Å). At last, the imino nitrogen at position 1 was electrostatically
attracted by the α4 helix residue Asp98 and during the rigid docking simulation this interaction has not
disturbed the hydrogen bond between the Asp98 and Ser184.[21] Accordingly, 5b positioning within the
Bax C-terminal domain was characterized by the binding affinity of Kd = -9.34 kcal/mol and inhibition
constant of Ki = 0.76 μM.
Figure 6: The best docked pose of 5b (agonist colored in blue) as predicted by AutoDock4.2. The Bax α1 helix is
colored in red, α2 helix in forest green, α3 helix in conflower blue, α4 helix in magenta, α5 helix in gold, α6 helix in
steel blue, α7 helix in olive grab, α8 helix in siena, and α9 helix in orange. (For interpretation of the references to
colour in this figure legend, the reader is referred to the web version of this article.)
The general stability of Bax-5b was confirmed by molecular dynamics study (Figure S41,
Supporting information). Thus, the complex was stabile after 0.82 ns by means of RMSD (Figure S41A)
and RMSF values (ligand interactions in range from 0.597 to 1.638 Å; Figure S41B). The energy
decomposition procedure for Bax-5b complex outlined that calculated binding affinity of 5b (Table 2)
correctly predicted the possibility of agonist to interact with the Bax C terminus. The low value of
binding energy of 5b (ΔGbind = -51.78 kcal/mol) suggests that the agonist is easily bound to the α9 helix.
As expected, hydrophobic interactions were dominant for 5b binding (ΔEvdw = -54.13 kcal/mol) while the
significant contribution was perceived for electrostatic interactions, too (ΔEele = -33.52 kcal/mol). The
relatively low value for the solvation energy (ΔEsolv = 2.33 kcal/mol) showed that 5b was stable but no so
tightly bound to the C terminus. In that manner, 5b suffered conformational perturbation immediately
after the start of simulation (0.01 ns) in manner that hydrogen bonds between the sulfanylidene group and
Ser101 as well as between nitrogen at position 4 and Asp102 were broken and 1,2,4-triazole-3(4H)-thione
ring was rotated towards Asp98. This advantageous alignment of 1,2,4-triazole-3(4H)-thione was
supported by the establishment of new hydrogen bond between the nitrogen at position 4 and Asp98-
Met99 peptide bond carbonyl oxygen, which arouse after 0.149 ns and remained stabile until the end. The
disagreement between the pose obtained by molecular dynamics (5b-MD) and the best docked one (5b-
BD) however cannot be related to false docking outcome, but can be attributed to applied rigid docking
protocol where no flexibility for amino acids was allowed. The RMSD between the 5b-BD and 5b-MD
amounted 1.748 Å, suggesting that both 5b-BD and 5b-MD are actually borderline solutions of 5b
physiologically conformation and the active structure itself is in-between the 5b-MD and 5b-BD. As a
result of docking/dynamics coupled simulation, Asp98 was estranged from Ser184 and kept away from
latter residue due to hydrophobic repulsion between the 5b sulfanylidene group and Asp98 side chain
methylene group initially observed after 0.28 ns of simulation and occurred during the whole period of
stabilization. This predicted conformational change for protein is vital for the Bax pro-apoptotic efforts
inasmuch as it resulted is the degeneration of Asp98-Ser184 hydrogen bond[21] occurring in agonist-free
Bax structure. Particularly, if this hydrogen bond remains stabile, Bax may actually inhibit itself even
before the interaction with anti-apoptotic proteins since it becomes incapable to invade the mitochondrial
membrane via the C-terminus.[21,49] On the other hand, the elimination of Ser184-Asp98 hydrogen bond
promotes the dissociation of the Bax C-terminal helix from the hydrophobic groove through major
conformational change, making it more accessible to mitochondrial membrane.[21,49] In that manner, 5b
initially exerts pro-apoptotic potential by forcing the Asp98 away from Ser184 and facilitates Bax-
mediated apoptotic event.
This article is protected by copyright. All rights reserved.

Table 2: Binding free energies and individual energy terms of ATP-rTopIIα complex with various inhibitors
Utilizing the molecular modeling approach, it was decidedly proven that Bax-5b complex as
apoptosis initiator. The remaining steps of the displacement model pathway still remain poorly
understood by means of luck of experimental support, but for the purpose of 5b pro-apoptotic potential
understanding the contribution in the form of molecular modeling (Supporting information) was given for
steps 2) and 4), outlining the interaction of Bax-5b with anti-apoptotic proteins Bcl-XL and Bcl-2 and Bax-
5b facilitation of MOM disturbance-based apoptosis, respectively.
An additional discussion relating to molecular docking and molecular dynamics studies of Bax-
5b-Bcl-XL/Bax-5b-Bcl2 complexes formation is presented in Supporting information.
Conclusion
A novel class of 1,2,4-triazole N-Mannich bases containing adamantane moiety has been
synthesized. The confirmation of the structure of these compounds and determination of the exact
nitrogen atom at which Mannich reaction occurs were obtained by X-ray diffraction technique. The
investigated compounds showed moderate to good cytotoxic activities against four malignant cell lines,
with K562 and HL-60 cells being significantly more sensitive. The low toxicity of tested derivatives
towards normal MRC-5 cell line indicated their prominent selectivity in the anticancer action. The
selected active compounds 5b, 5e and 5j induced accumulation of K562 cells in subG1 and G1 phase.
Western blot analysis demonstrated that selected compounds exert their cytotoxic activity through the
caspase-dependent apoptosis and upregulation of Bax expression levels. Also, it was found, using a tube
formation assay, that tested derivatives exhibit antiangiogenic effects. Finally, the molecular modeling
study revealed, applying the combined molecular docking/molecular dynamics approach, that compound
5b has a role in protection of Bax Ser184 from phosphorylation in order to facilitate the proapoptotic
event.
Acknowledgements
The authors are grateful to the Ministry of Science and Technological Development of the Republic of
Serbia for financial support (Grant Nos 172016, 175011, III43004 and 173056).
Conflict of interest
The authors declare no conflict of interest.
References
[1]
G. L. Almajan, S. F. Barbuceanu, E. R. Almajan, C. Draghici, G. Saramet, Eur. J. Med. Chem. 2009, 44,
3083.
[2]
[3]
[4]
[5]
B. L. Wang, X. H. Liu, X. L. Zhang, J. F. Zhang, H. B. Song, Z. M. Li, Chem. Biol. Drug Des. 2011, 78, 42.
S. Holla, B. Veerendra, M. K. Shivananda, B. Poojary, Eur. J. Med. Chem. 2003, 38, 759.
E. Palaska, G. Sahin, P. Kelicen, N. T. Durlu, G. Altinok, Farmako 2002, 57, 101.
A. K. M. Iqbal, A. Y. Khan, M. B. Kalashetti, N. S. Belavagi, Y-D. Gong, I. A. M. Khazi, Eur. J. Med. Chem.
2012, 53, 308.
[6]
T. Plech, B. Kaproń, A. Paneth, U. Kosikowska, A. Malm, A. Strzelczyk, P. Stączek, Ł. Świątek, B. Rajtar,
M. Polz-Dacewicz, Eur. J. Med. Chem. 2015, 97, 94.
[7]
[8]
I. Küçükgüzel, S. G. Küçükgüzel, S. Rollas, M. Kiraz, Bioorg. Med. Chem. Lett. 2001, 11, 1703.
İ. Küçükgüzel, Ş. G. Küçükgüzel, S. Rollas, G. Otük-Saniş, O. Ozdemir, I. Bayrak, T. Altuğ, J. P. Stables,
Farmaco, 2004, 59, 893.
[9]
I. Khan, A. Ibrar, N. Abbas, Eur. J. Med. Chem. 2013, 63, 854.
[10] A. Ts. Mavrova, D. Wesselinova, Y. A. Tsenov, P. Denkova, Eur. J. Med. Chem. 2009, 44, 63.
[11] J. Liu, D. Obando, V. Liao, T. Lifa, R. Codd, Eur. J. Med. Chem. 2011, 46, 1949.
[12] Z. M. Shao, M. I. Dawson, X. S. Li, A. K. Rishi, M. S. Sheikh, Q. X. Han, J. V. Ordonez, B. Shroot, J. A.
Fontana, Oncogene, 1995, 11, 493.
This article is protected by copyright. All rights reserved.

[13] S. Y. Sun, P. Yue, M. I. Dawson, B. Shroot, S. Michel, W. W. Lamph, R. A. Heyman, M. Teng, R. A.
Chandraratna, K. Shudo, W. K. Hong, R. Lotan, Cancer Res. 1997, 57, 4931.
[14] Z. K. Genc, S. Tekin, S. Sandal, M. Genc, Res. Chem. Intermed. 2015, 41, 6229.
[15] G. M. Sheldrick, Acta Crystallogr. Sect. A 2015, 71, 3.
[16] G. M. Sheldrick, Acta Crystallogr. Sect. C 2015, 71, 3.
[17] T. Mosmann, J. Immunol. Methods 1983, 65, 55.
[18] M. Ohno, T. Abe, J. Immunol. Methods 1991, 145, 199.
[19] M. G. Ormerod (ed.), Flow Cytometry: A Practical Approach, 3rd ed., Oxford University Press, Oxford, 2000.
[20] V. Marković, N. Debeljak, T. Stanojković, B. Kolundžija, D. Sladić, M. Vujčić, B. Janović, N. Tanić, M.
Perović, V. Tešić, J. Antić, M. D. Joksović, Eur. J. Med. Chem. 2015, 89, 401.
[21] M. Suzuki, R. J. Youle, N. Tjandra, Cell 2000, 103, 645.
[22] P. E. Czabotar, E. F. Lee, G. V. Thompson, A. Z. Wardak, W. D. Fairlie, P. M. Colman, J. Biol. Chem. 2011,
286, 7123.
[23] B. Ku, C. Liang, J. U. Jung, B. H. Oh, Cell Res. 2011, 21, 627.
[24] P. E. Czabotar, D. Westphal, G. Dewson, S. Ma, C. Hockings, W. D. Fairlie, E. F. Lee, S. Yao, A. Y. Robin,
B. J. Smith, D. C. Huang, R. M. Kluck, J. M. Adams, P. M. Colman, Cell 2013, 152, 519.
[25] C. Y. Chen, P. H. Lee, Y. Y. Lin, W. T. Yu, W. P. Hu, C. C. Hsu, Y. T. Lin, L. S. Chang, C. T. Hsiao, J. J.
Wang, M. I. Chung, Bioorg. Med. Chem. Lett. 2013, 23(24), 6854.
[26] T. Srinivasa Reddy, H. Kulhari, V. Ganga Reddy, A. V. Subba Rao, V. Bansal, A. Kamal, R. Shukla, Org.
Biomol. Chem. 2015, 13(40), 10136.
[27] S. Fulda, K. M. Debatin, Oncogene 2006, 25(34), 4798.
[28] S. A. Cillessen, C. J. Meijer, M. Notoya, G. J. Ossenkoppele, J. J. Qudejans, J. Pathol. 2010, 220, 509.
[29] J. M. Adams, S. Cory, Trends Biochem. Sci. 2001, 26(1), 61.
[30] D. E. Wood, A. Thomas, L. A. Devi, Y. Berman, R. C. Beavis, J. C. Reed, E. W. Newcomb, Oncogene 1998,
17, 1069.
[31] J. M. Blander, Nat. Rev. Immunol. 2014, 14(9), 601.
[32] T. Decker, M. Oelsner, R. J. Kreitman, G. Salvatore, Q. C. Wang, I. Pastan, C. Peschel, T. Licht, Blood 2004,
103, 2718.
[33] N. Li, P. Lin, C. Cai, Z. Pan, N. Weisleder, J. Ma, Am. J. Physiol. Cell Physiol. 2009, 296, 267.
[34] H. Toyota, N. Yanase, T. Yoshimoto, M. Moriyama, T. Sudo, J. Mizuguchi, Cancer Lett. 2003, 189(2), 221.
[35] D. E. Wood, E. W. Newcomb, Exp. Cell Res. 2000, 256, 375.
[36] S. Liekens, E. de Clercq, J. Neyts, Biochem. Pharm. 2001, 61, 253.
[37] P. Carmeliet, R. K. Jain, Nature 2000, 407, 249.
[38] M. I. Dawson, Z. Xia, G. Liu, J. A. Fontana, L. Farhana, B. B. Patel, S. Arumugarajah, M. Bhuiyan, X.-K.
Zhang, Y.-H. Han, W. B. Stallcup, J.-I. Fukushi, T. Mustelin, L. Tautz, Y. Su, D. L. Harris, N. Waleh, P. D.
Hobbs, L. Jong, W.-R. Chao, L. J. Schiff, B. P. Sani, J. Med. Chem. 2007, 50, 2622.
[39] B. T. Aftab, I. Dobromilskaya, J. O. Liu, C. M. Rudin, Cancer Res. 2011, 71(21), 6764.
[40] D. Bouis, G. A. P. Hospers, C. Meijer, G. Molema, N. H. Mulder, Angiogenesis 2001, 4, 91.
[41] A. Aouacheria, F. Brunet, M. Gouy, Mol. Biol. Evol. 2005, 22, 2395.
[42] M. Xin, R. Li, M. Xie, D. Park, T. K. Owonikoko, G. L. Sica, P. E. Corsino, J. Zhou, C. Ding, M. A. White,
A. T. Magis, S. S. Ramalingam, W. J. Curran, F. R. Khuri, X. Deng, Nat. Commun. 2014, 5, 4935.
[43] E. Gavathiotis, D. E. Reyna, J. A. Bellairs, E. S. Leshchiner, L. D. Walensky, Nat. Chem. Biol. 2012, 8(7),
639.
[44] L. Chen, S. N. Willis, A. Wei, B. J. Smith, J. I. Fletcher, M. G. Hinds, P. M. Colman, C. L. Day, J. M.
Adams, D. C. Huang, Mol. Cell 2005, 17, 393.
[45] T. Kuwana, L. Bouchier-Hayes, J. E. Chipuk, C. Bonzon, B. A. Sullivan, D. R. Green, D. D. Newmeyer, Mol.
Cell 2005, 17, 525.
[46] A. Shamas-Din, J. Kale, B. Leber, D. W. Andrews, Cold Spring Harb. Perspect. Biol. 2013. 5, a008714.
[47] G. M. Morris, R. Huey, W. Lindstrom, M. F. Sanner, R. K. Belew, D. S. Goodsell, A. J. Olson, J. Comput.
Chem. 2009, 16, 2785.
[48] L. P. Billen, C. L. Kokoski, J. F. Lovell, B. Leber, D. W. Andrews, PLoS Biol. 2008, 6, e147.
[49] A. M. Petros, E. T. Olejniczak, S. W. Fesik, Biochim. Biophys. Acta 2004, 1644, 83.
This article is protected by copyright. All rights reserved.

Supporting Information
Additional Supporting Information may be found online in the supporting information tab for this
article:
Appendix S1. Procedure for the preparation of 5a-r, spectral data, crystal structure determination
details and structural data for 5e, details of biological experiments and molecular docking
studies.
Figure S1-S38. 1H and 13C NMR spectra of 4 and 5a-r.
Table S1-S4. Crystallographic data for 5e.
Figure S39 and S40. Hydrogen bonding in the orthorhombic and monoclinic form, respectively.
Figure S41-S45. Molecular docking and molecular dynamics studies of the interaction of 5b
with Bax.
Appendix S2. CIF and CheckCIF files containing crystallographic information and full structural check
regarding compound 5e.
Figure caption list:
Scheme 1: Reagents and conditions: a) SOCl2, DMF, CH2Cl2, 2 h, r.t.; b) NH2NHC(S)NH2,
THF, 6 h, r.t.; c) NaOH, 3h, reflux; HCl; d) CH2O, Ar-NH2, THF, 48 h, r.t.
Figure 1: Molecular structure of 5e. Molecule A of the monoclinic polymorph is chosen as the
representative.
Figure 2: Changes in the cell cycle phase distribution of human myelogenous leukemia K562
cells induced by the compounds 5b, 5e and 5j after 24h treatment (tested concentrations
corresponded to IC50 (A) and 2IC50 values (B)). The results are presented as the mean S.D.
from two independent experiments. Statistically significant differences between control and
treated cells are marked with * (p<0.05).
Figure 3: Photomicrographs of acridine orange/ethidium bromide stained control K562 cells,
and K562 cells treated with 2IC50 concentrations of compounds 5b, 5e and 5j for 24 h.
Figure 4: Induction of caspase-dependent apoptosis by compounds 5b, 5e and 5j at IC50
concentrations in K562 myelogenous leukemia cells after 24h treatment. Immunoblotting was
performed using anti-caspase 3, anti-cleaved caspase 3, anti-caspase 8, anti-caspase 9 and anti-
Bax antibodies. The anti-GAPDH antibody was used as a control for protein load. Besides full-
length caspases – procaspases, the fragments of activated caspase-3 (subunits p19 and p17),
fragments of activated caspase-8 (the cleaved intermediates p43/p41, the active
fragments/subunits p18, p16 and p10) and fragments of activated caspase-9 (subunits p37 and
p35) were detected.
Figure 5: Effects on in vitro angiogenesis of EA.hy926 cells. Photomicrographs of control
EA.hy926 cells and cells exposed to subtoxic IC20 concentrations (15 µM) of compounds 5b, 5e
and 5j after 20 h treatment.
Figure 6: The best docked pose of 5b (agonist colored in blue) as predicted by AutoDock4.2.
The Bax α1 helix is colored in red, α2 helix in forest green, α3 helix in conflower blue, α4 helix
This article is protected by copyright. All rights reserved.

in magenta, α5 helix in gold, α6 helix in steel blue, α7 helix in olive grab, α8 helix in siena, and
α9 helix in orange. (For interpretation of the references to colour in this figure legend, the reader
is referred to the web version of this article.)
Table 1: The in vitro cytotoxic activity of investigated compounds
IC50 SD [μM]
HeLa
K562
HL-60
A549
MRC-5
≈100
51.19 0.40
61.09 4.92
76.78 1.17
49.73 2.30
66.82 8.11
66.42 6.35
63.21 9.32
82.84 1.78
>100
64.01 2.09
54.60 0.84
78.71 7.61
77.25 0.37
71.18 1.10
42.01 1.30
74.17 0.41
54.98 6.79
58.68 3.62
2.24 0.26
35.19 1.71
25.42 0.89
40.35 0.57
26.89 1.21
27.71 2.46
30.62 2.85
30.48 0.72
24.26 0.17
63.84 2.08
27.05 5.35
32.98 5.14
41.68 1.84
32.40 0.10
36.98 2.54
29.20 4.88
36.70 0.54
33.94 3.08
35.13 0.46
5.45 0.18
22.74 2.70
21.91 1.47
30.30 1.67
25.16 3.27
21.02 2.76
34.09 2.51
24.26 2.27
22.66 3.48
24.73 0.23
23.76 2.63
28.10 3.85
26.26 4.58
23.32 3.93
26.87 2.12
25.82 0.45
25.76 1.44
23.32 1.84
23.56 0.72
3.37 0.06
75.24 10.32
72.90 1.95
84.46 1.71
85.77 0.82
94.04 0.13
92.90 3.78
≈100
5a
5b
5c
5d
5e
5f
5g
5h
5i
5j
5k
5l
5m
5n
5o
5p
5q
5r
≈100
≈100
93.38
>100
>100
>100
>100
>100
87.66
≈100
96.54 4.89
>100
87.98 3.10
97.94 2.92
>100
>100
>100
≈100
>100
90.05 12.86
55.49 1.36
83.28 10.82
83.36 5.34
96.25 2.14
14.42 0.70
55.11
≈100
≈100
>100
9.24 0.18
cisplatin
Table 2: Binding free energies and individual energy terms of ATP-rTopIIα complex with various
inhibitors
Complex
ΔEele
ΔEvdw
ΔGsolv
-TΔS
ΔGbind
(kcal/mol)
-33.52 0.31
-34.39 0.41
-34.26 0.46
(kcal/mol)
-54.13 0.46
-67.59 0.46
-67.41 0.61
(kcal/mol)
2.33 0.43
6.38 0.52
3.71 0.72
(kcal/mol)
-33.54 0.24
-34.32 0.32
(kcal/mol)
-51.78 0.34
-61.28 0.22
-60.42 0.25
Bax-5b
Bax-5b-Bcl-XL
Bax-5b-Bcl2
-37.54 0.15
aΔEele - electrostatic energy difference. ΔEvdw - van der Waals energy difference. ΔGsolv - solvation free
b
c
energy. dΔS - the entropy change. eΔGbind - predicted binding free energy.
This article is protected by copyright. All rights reserved.

This article is protected by copyright. All rights reserved.

This article is protected by copyright. All rights reserved.

This article is protected by copyright. All rights reserved.