Probing simple structural modification of α-mangostin on its cholinesterase inhibition and cytotoxicity

Article abstract of DOI:10.1002/ardp.202000156

α-Mangostin has been reported to possess a broad range of pharmacological effects including potent cholinesterase inhibition, but the development of α-mangostin as a potential lead compound is impeded by its toxicity. The present study investigated the impact of simple structural modification of α-mangostin on its cholinesterase inhibitory activities and toxicity toward neuroblastoma and liver cancer cells. The dialkylated derivatives retained good acetylcholinesterase (AChE) inhibitory activities with IC₅₀ values between 4.15 and 6.73 μM, but not butyrylcholinesterase (BChE) inhibitory activities, compared with α-mangostin, a dual inhibitor (IC₅₀: AChE, 2.48 μM; BChE, 5.87 μM). Dialkylation of α-mangostin produced AChE selective inhibitors that formed hydrophobic interactions at the active site of AChE. Interestingly, all four dialkylated derivatives of α-mangostin showed much lower cytotoxicity, being 6.4- to 9.0-fold and 3.8- to 5.5-fold less toxic than their parent compound on neuroblastoma and liver cancer cells, respectively. Likewise, their selectivity index was higher by 1.9- to 4.4-fold; in particular, A2 and A4 showed improved selectivity index compared with α-mangostin. Taken together, modification of the hydroxyl groups of α-mangostin at positions C-3 and C-6 greatly influenced its BChE inhibitory and cytotoxic but not its AChE inhibitory activities. These dialkylated derivatives are viable candidates for further structural modification and refinement, worthy in the search of new AChE inhibitors with higher safety margins.

Full text of DOI:10.1002/ardp.202000156

Received: 14 May 2020
Revised: 22 June 2020
Accepted: 11 July 2020
|
|
DOI: 10.1002/ardp.202000156
F U L L P A P E R
Probing simple structural modification of α‐mangostin on its
cholinesterase inhibition and cytotoxicity
1
,2
3
3
3
Kooi‐Yeong Khaw
| Pravin Kumar | Siti Rafidah Yusof | Surash Ramanathan
|
1
Vikneswaran Murugaiyah
1
Discipline of Pharmacology, School of
Pharmaceutical Sciences, Penang, Malaysia
Abstract
2
School of Pharmacy, Monash University
α‐Mangostin has been reported to possess a broad range of pharmacological effects
including potent cholinesterase inhibition, but the development of α‐mangostin as a
potential lead compound is impeded by its toxicity. The present study investigated
the impact of simple structural modification of α‐mangostin on its cholinesterase
inhibitory activities and toxicity toward neuroblastoma and liver cancer cells. The
dialkylated derivatives retained good acetylcholinesterase (AChE) inhibitory activ-
ities with IC50 values between 4.15 and 6.73 µM, but not butyrylcholinesterase
Malaysia, Bandar Sunway,
Selangor Darul Ehsan, Malaysia
3
Centre for Drug Research, Penang, Malaysia
Correspondence
Vikneswaran Murugaiyah, Discipline of
Pharmacology, School of Pharmaceutical
Sciences, Universiti Sains Malaysia, Penang
11800, Malaysia.
Email: vicky@usm.my
(
BChE) inhibitory activities, compared with α‐mangostin, a dual inhibitor (IC50:
Funding information
AChE, 2.48 µM; BChE, 5.87 µM). Dialkylation of α‐mangostin produced AChE se-
lective inhibitors that formed hydrophobic interactions at the active site of AChE.
Interestingly, all four dialkylated derivatives of α‐mangostin showed much lower
cytotoxicity, being 6.4‐ to 9.0‐fold and 3.8‐ to 5.5‐fold less toxic than their parent
compound on neuroblastoma and liver cancer cells, respectively. Likewise, their
selectivity index was higher by 1.9‐ to 4.4‐fold; in particular, A2 and A4 showed
improved selectivity index compared with α‐mangostin. Taken together, modifica-
tion of the hydroxyl groups of α‐mangostin at positions C‐3 and C‐6 greatly influ-
enced its BChE inhibitory and cytotoxic but not its AChE inhibitory activities. These
dialkylated derivatives are viable candidates for further structural modification and
refinement, worthy in the search of new AChE inhibitors with higher safety margins.
Ministry of Higher Education Malaysia and
Universiti Sains Malaysia,
Grant/Award Number: No. 1001/PFARMASI/
815093
K E Y W O R D S
acetylcholinesterase, cholinesterase inhibition, cytotoxicity, structural modification,
α‐mangostin
[
4,7,8]
1
| INTRODUCTION
the different substituents.
Naturally occurring xanthone are
classified into six groups depending on their chemical framework of
the tricyclic scaffold, including the prenylated xanthones, which are
Xanthones are low‐molecular‐weight polyphenolic compounds found
in plants. They have a nucleus with a dibenzo‐α‐pyrone scaffold,
which is symmetric and occurs as either fully aromatized or as di-
[
9]
restricted to the plant of the family Guttiferae. One of the well‐
studied prenylated xanthones is α‐mangostin, the major xanthone
found in Garcinia mangostana.
[
1–5]
hydro, tetrahydro, and hexahydro derivatives.
This family of
compounds has numerous pharmacological effects within a notably
broad spectrum of diseases, and thus are regarded as privileged
First isolated in 1855, its pharmacological activities have been
widely studied in the past two decades. α‐Mangostin has 9H‐xanthene
backbone substituted by a hydroxyl group at positions C‐1, C‐3, and
C‐6, a methoxy group at position C‐7, an oxo group at position C‐9, and
[6]
structures. Their pharmacological activities are associated with the
tricyclic scaffold but vary depending on the nature and/or position of
Arch Pharm. 2020;e2000156.
wileyonlinelibrary.com/journal/ardp © 2020 Deutsche Pharmazeutische Gesellschaft
|
1 of 9
https://doi.org/10.1002/ardp.202000156

2
of 9
KHAW ET AL.
|
prenyl group at positions C‐2 and C‐8 (PubChem CID: 5281650).
α‐Mangostin has been widely reported for its capabilities against several
its derivatives were predicted using Accelrys Discovery Studio
software.
[10]
diseases.
inflammatory,
hibitory activity,
Among some of its notable effects include anti‐
[11,12]
[13,14]
[15]
antioxidant,
antidiabetic,
neuraminidase in-
[17–19]
[16]
and neuroprotective properties.
Therefore,
2 | RESULTS AND DISCUSSION
2.1 | Chemistry
further development of α‐mangostin as a therapeutic lead compound is
of great interest. However, α‐mangostin has also been reported to be
cytotoxic against a wide range of cell lines such as human breast cancer
(
MCF‐7 and MDA‐MB‐231), human lung cancer (NCI‐H460 and NCI‐
Semisynthesis of natural compounds is of great interest to a medicinal
chemist for one or more of the following reasons: improve bioactivities
or simplify the structure while retaining the bioactivity, reduce toxi-
city, and enhance ADME (absorption, distribution, metabolism, and
excretion) properties. A large number of clinically used drugs were
semisynthesized natural product analogs. For example, vorapaxar, an
[20,21]
H187), glioblastoma (SF‐268),
human colon cancer (DLD‐1 and
[
22,23]
HT‐29),
human melanoma (SK‐MEL‐28), squamous cell carcinoma
[
24]
[25]
(A‐431),
and human leukemia (HL60).
In our continuous search for cholinesterase inhibitors from natural
resources, we discovered that α‐mangostin is a potent dual cholines-
terase inhibitor (IC50: acetylcholinesterase [AChE], 2.14 µM; butyr-
antithrombotic agent was semisynthesized from himbacine (a tricyclic
[27]
[26]
ylcholinesterase [BChE], 5.41 µM).
Preliminary structure–activity
lactone from Galbulimima baccata).
α‐Mangostin is a naturally
relationship suggests the importance of the C‐8 prenyl and C‐7 hy-
droxyl group for potent AChE and BChE inhibitory activities of the
prenylated xanthones from G. mangostana. It is reasoned that the
cholinesterase inhibitory activities of α‐mangostin were lower than
that of γ‐mangostin due to methoxylation of the hydroxyl group at C‐7.
Nevertheless, molecular docking revealed that the hydroxyl group of
α‐mangostin at C‐3 and C‐6 formed hydrogen bonding within the ac-
occurring xanthone derived from G. mangostana. It possesses favorable
pharmacological properties; however, the toxicity feature of this
compound may limit its usefulness as a therapeutic drug. Several
groups have attempted to semisynthesize α‐mangostin derivatives to
[28,29]
improve its potency and drug‐likeness.
In the present study, the
α‐mangostin, obtained from the chloroform fraction of G. mangostana
fruit pericarp, had undergone simple structural modification of
the hydroxyl group at positions C‐3 and C‐6. The semisynthesized
derivatives were obtained via O‐alkylation reaction with correspond-
ing alkyl halides (methyl‐butyl) in acetone under reflux (Scheme 1).
The structural characterization of the compounds was deter-
mined by nuclear magnetic resonance (NMR) spectroscopy and mass
spectrometry.
[26]
tive site of the AChE and BChE enzymes.
However, its cytotoxicity
is of great concern and could limit its potential for further develop-
ment as a cholinesterase inhibitor.
We hypothesized that the pharmacological effects and cyto-
toxicity of α‐mangostin might be attributed at least, in part, by the
hydroxyl group at C‐3 and C‐6 positions. Therefore, in the present
study, the impact of simple structural modifications of α‐mangostin at
C‐3 and C‐6 on its pharmacological effect and toxicity was in-
vestigated. We aimed to semisynthesize α‐mangostin derivatives
2.1.1 | Crystal structure of A1
(A1–A4) via alkylation and evaluated their cholinesterase inhibition
versus cytotoxicity. Molecular docking study was carried out to
better understand the differences in interactions between enzymes
and derivatives. The physicochemical properties of α‐mangostin and
Compound A1 crystallizes in the triclinic space group P−1. The mo-
lecular view of A1 is shown in Figure 1. The 9H‐xanthene ring system
is essentially planar, the root mean square deviations being 0.046 Å
SCHEME 1 General reaction scheme of the synthesis of α‐mangostin derivatives

KHAW ET AL.
3 of 9
|
FIGURE 1 The molecular structure of A1
showing 50% probability displacement ellipsoids
for non‐H atoms and the atom‐numbering
scheme. An intramolecular hydrogen bond is
shown as a dashed line
and its least‐square plane makes dihedral angles of 63.20(6)° and
A4, the IC50 values against BChE are greatly reduced, hence their
79.60(6)° with the two planar 2‐methylbut‐2‐ene (root mean square
selectivity also reduces.
deviations = 0.023 and 0.006 Å) moieties, respectively. The two
benzene rings make a dihedral angle of 4.73(6)°. The crystal structure
is stabilized by an intramolecular O–H···O hydrogen bond, which
forms an S(6) ring motif (Figure 1). The crystal structure was found to
exhibit no intermolecular hydrogen bonding but instead possesses a
weak C–H···π and a π–π stacking interactions between the benzene
and 4H‐pyran rings (centroid to centroid distance = 3.5794(7) Å).
2.3 | Molecular docking
In silico molecular docking was performed to investigate the inter-
actions between the derivatives A1–A4 and residues at the active
site of the enzymes (Figure 2). AChE has an active site that is not on
the protein's surface but located at the bottom of a 20‐Å‐deep gorge,
[30]
lined largely by aromatic residues.
The results showed that A1–A4
2
.2 | Cholinesterase inhibitory activities
could accommodate inside the gorge of AChE. A1 formed π–π
The cholinesterase inhibitory activities of α‐mangostin and its deri-
vatives were determined by the Ellman method and the results are
summarized in Table 1. In general, the derivatives A1–A4 had com-
parable AChE selective inhibitory activities with IC50 values ranging
from 4.15 to 6.73 µM that are 1.7‐ to 2.7‐fold less potent compared
with α‐mangostin. The AChE inhibitory activities of A1–A4 were
improved from methyl to butyl substitutions and A4 was the most
potent inhibitor among the four. However, the inhibitory activities of
A1–A4 against BChE were greatly reduced with the IC50 values
ranging from 26.99 to 129.13 µM, that is, 4.6‐ to 22.0‐fold less potent
compared with α‐mangostin. It is interesting to note that the AChE
selectivity of derivatives was increased up to 21.7‐fold by one‐step
alkylation, the selectivity reduced consistently with an increase in the
aliphatic side chain length. For example, the AChE selectivity for
derivative A4 is 6.5 (Table 1), derived from the formula where the
IC50 against BChE is the nominator and IC50 against AChE is the
denominator. As the aliphatic side chain length increases from A1 to
TABLE 1 Cholinesterase inhibitory activities of the dialkylated
derivatives of α‐mangostin
AChE
inhibition,
IC50 (µM)
BChE inhibition,
IC50 (µM)
AChE
a
Compound
selectivity
A1
A2
A3
A4
6.73 ± 0.07
5.04 ± 0.14
4.89 ± 0.10
4.15 ± 0.18
129.13 ± 0.07
109.48 ± 0.99
86.93 ± 1.57
26.99 ± 0.65
5.87 ± 0.09
19.2
21.7
17.8
6.5
α‐Mangostin 2.48 ± 0.11
2.4
Galantamine 0.27 ± 0.07
5.55 ± 0.24
20.4
Note: Data are presented as mean ± SD (n = 3).
Abbreviations: AChE, acetylcholinesterase; BChE, butyrylcholinesterase;
SD, standard deviation.
a
Selectivity against AChE: IC50BChE/IC50AChE.

4
of 9
KHAW ET AL.
|
FIGURE 2 Dialkylated derivatives of α‐
mangostin docked into the binding site of
acetylcholinesterase
interaction with Phe330 at the choline‐binding site. In contrast,
A4 was unable to penetrate deep into the gorge, probably attributed
to the bulkiness of the compound and narrow active site gorge of
AChE. The molecular docking result showed that A4 accommodated
close to the choline‐binding site, forming σ–π interaction with
Phe330 (3.29 Å), which anchored this compound to the TcAChE ac-
tive site gorge (Table 2). Compound A4 accommodates in the same
orientation as α‐mangostin at the middle of the AChE gorge but with
TABLE 2 Binding interactions data for A1 and A4 docked into active site gorge of TcAChE
Ligand
A1
Binding energy (kcal) Enzyme interacting site Residue
Type of interaction Distance (Å) Ligand‐interacting moiety (ring)
−11.28
−12.47
−12.69
Choline‐binding site
Choline‐binding site
Choline‐binding site
Choline‐binding site
Choline‐binding site
Peripheral anionic site
Oxyanionic hole
Phe330
Phe330
Trp84
Hydrophobic
Hydrophobic
Hydrophobic
Hydrophobic
Hydrophobic
Hydrophobic
Hydrogen
4.64
3.29
4.75
3.51
5.47
2.84
2.01
C
A4
C
a
α‐Mangostin
B
Trp84
A
Trp84
C
Asp72
Gly117
C
A (OH at C‐6)
a
[26]
Data taken from Khaw et al.

KHAW ET AL.
5 of 9
|
lesser interactions due to the hydroxyl groups at C‐3 and C‐6 posi-
tions being alkylated. Furthermore, the interaction of A1 and A4 with
choline‐binding site was found to be different compared with
derivatives of α‐mangostin showed much lower cytotoxicity, being
6.4‐ to 9.0‐fold and 3.8‐ to 5.5‐fold less toxic than their parent
compound on neuroblastoma and liver cancer cells, respectively
(Table 3). Likewise, the selectivity index of the derivatives was 1.9‐ to
4.4‐fold higher, in particular, A2 and A4 showed improved selectivity
index compared with α‐mangostin, since both derivatives had a larger
ratio of LC50(on respective cells)/IC50(AChE), indicating lower leth-
ality. In other words, derivatives A2 and A4 have a bigger therapeutic
index, thus they may offer a larger range of safe and effective drug
dosing to work with compared with α‐mangostin. This finding in-
dicates that the hydroxyl group at positions 3 and 6 may, in part, be
responsible for the cytotoxicity of α‐mangostin against the tested
cells. Derivatization of the hydroxyl group at those positions lowered
the cytotoxicity of the parent compound. Our findings are consistent
[
26]
α‐mangostin.
Choline‐binding site is located at the middle narrow
neck of the gorge, lined by Phe330 and Tyr121 at the upper site and
Trp84 at the further down site of the gorge, approximately 15 Å
[
30]
deep.
Due to the bulkiness of the derivatives, especially A4, they
could interact with Phe330 but not Trp84, which is seen with
α‐mangostin, because α‐mangostin could penetrate deeper into the
gorge. On the other hand, derivatives A2 and A3 did not demonstrate
any interactions at AChE active site. Despite being able to dock into
the active site there was no hydrogen bonding within a distance of
3.5 Å or hydrophobic interaction within a distance of 7.0 Å between
the hydroxyl group or phenolic rings of A2 and A3 with the protein
residues at the active site. The molecular docking showed an inverse
relationship between AChE inhibitory activities (IC50) of the deriva-
tives with their free energy of binding as shown in Figure 3, in which
A1–A4 demonstrated good free energy of binding.
[
29]
with a study by Fei et al.,
which showed that the phenol groups at
C‐3 and C‐6 positions are critical for the anticancer activities of
α‐mangostin and C‐4 modification is capable to improve both antic-
ancer activity and drug‐like properties. The findings of the present
study prompt a renewed interest in the search for the new choli-
nesterase inhibitors from α‐mangostin with a higher safety margin.
2
.4 | Cytotoxicity
Few studies reported on the cytotoxicity of α‐mangostin in various
2.5 | Theoretical prediction of ADME properties
[
25]
cell lines. Matsumoto et al.
stated that α‐mangostin was cytotoxic
against human leukemia cells with LC50 of 10 µM, while it was highly
cytotoxic on breast cancer (LC50 of 0.92 µg/ml) and epidermoid
carcinoma of the mouth (LC50 of 2.08 µg/ml) cell lines. In addition,
The ADME properties of α‐mangostin and dialkylated derivatives
were predicted in silico using the Accelrys Discovery Studio and the
data are summarized in Table 4. Compounds A1–A4 showed com-
parable ADME properties to α‐mangostin with good absorption and
plasma protein binding, but lower solubility. In addition, all deriva-
tives followed Lipinski's rule of five (molecular weight, logP, number
of hydrogen donors, and acceptors) except for A4 with higher logP
value and molecular weight over 500. One of the major hurdles for a
medicinal chemist is the penetration of potential compounds through
the complex blood–brain barrier. Most of the central nervous system
active drugs prompt to penetrate into the brain through passive
[
23]
Nakagawa et al.
reported decreased viability of DLD‐1 cells upon
[23]
treatment with 20 µM of α‐mangostin.
α‐Mangostin together with
its derivatives, A1–A4, were tested for their toxicity on human
neuroblastoma (SH‐SY5Y, SK‐N‐SH) and human liver cancer (HepG2)
cells. Table 3 summarizes the LC50 values and the selectivity index of
the tested compounds on the three types of cells. Similar to the
findings from previous studies, α‐mangostin was found to be cyto-
toxic toward SH‐SY5Y, SK‐N‐SH, and HepG2 cells with LC50 of 11.20,
[
31,32]
12.16, and 12.71 µM, respectively. Interestingly, all four dialkylated
diffusion.
Interestingly, compound A1 is predicted to have high
FIGURE 3 Relationship between free
energies of binding (FEB) and IC50 of isolated
compounds. AChE, acetylcholinesterase

6
of 9
KHAW ET AL.
|
TABLE 3 Cytotoxicity of the dialkylated
derivatives of α‐mangostin
SH‐SY5Y
SK‐N‐SH
HepG2
a
a
a
Compound
LC50 (µM)
SI
LC50 (µM)
SI
LC50
SI
A1
71.92 ± 3.44
100.94 ± 1.18
75.98 ± 1.53
74.00 ± 10.79
11.20 ± 0.09
10.7
20.0
15.5
17.8
4.5
103.99 ± 1.80
107.85 ± 0.36
79.36 ± 0.34
88.18 ± 0.67
12.16 ± 0.16
15.4
21.4
16.2
21.2
4.9
67.79 ± 0.97
48.46 ± 1.17
49.52 ± 0.38
69.38 ± 0.84
12.71 ± 0.26
10.1
9.6
A2
A3
10.1
16.7
5.1
A4
α‐Mangostin
Note: Data are presented as mean ± SD (n = 3).
Abbreviations: AChE, acetylcholinesterase; SI, selectivity index.
a
SI is defined as LC50(on respective cells)/IC50(AChE).
b
c
d
e
TABLE 4 Theoretical prediction of
ADME properties of α‐mangostin and its
dialkylated derivatives
Compound
LogP Molecular weight BBB level
Absorption level
LogS
PPB
A1
A2
A3
A4
4.97
5.67
6.71
7.63
439.2
467.2
495.3
523.3
410.1
287.3
1
4
4
4
4
3
0
1
2
3
1
0
−5.647
−5.992
−6.597
−7.036
−4.913
−2.668
1
2
2
2
1
0
a
α‐Mangostin 4.52
Galantamine 1.06
Abbreviations: ADME, absorption, distribution, metabolism, and excretion; BBB, blood–brain barrier;
PPB, plasma protein binding.
a
The data were determined with Accelrys Discovery Studio.
b
0
, 1, 2, 3, and 4 indicate very high, high, medium, low, and undefined penetration, respectively.
c
0
, 1, 2, and 3 indicate good, moderate, low, and very low absorption, respectively.
d
LogS [log(molar solubility)] < −8.0, extremely low; −8.0 < logS < −6.0, very low; −6.0 < logS < −4.0, low;
−
4.0 < logS < −2.0, good; −2.0 < logS < 0.0, optimal; logS > 0.0, too soluble.
e
0
, 1, and 2 indicate <90% binding, ≥90% binding, and ≥95% binding with plasma protein,
respectively.
blood–brain barrier penetration that is better than the standard
drug, galantamine. However, the other derivatives and α‐mangostin
were predicted to have an undefined blood–brain barrier penetra-
tion. For the human intestinal absorption of the compounds, A1 is
predicted to have good absorption similar to galantamine, while A2 is
predicted to have moderate absorption similar to α‐mangostin. The
derivatives A3 and A4 are predicted to have low to very low ab-
sorption. LogS is the log of molar solubility, which is directly related
to the solubility of a compound and defined as the number of moles
of the solute that can be dissolved per liter of a solution before
saturation. Galantamine is categorized in the good solubility cate-
gory; α‐mangostin, A1 and A2 are categorized in the low solubility
category whereas A3 and A4 are categorized in the very low solu-
bility category. Nevertheless, the low solubility and absorption
properties of the derivatives are not a major concern and could be
overcome by appropriate formulation to enhance their solubility and
absorption. The derivatives A2–A4 are predicted to have >95%
protein binding, while A1 and α‐mangostin are predicted to have
polypharmacy to avoid possible drug–drug interactions via dis-
placement from protein binding.
3 | CONCLUSION
The present study reported on the impact of simple structural
modification of α‐mangostin on its cholinesterase inhibitory ac-
tivities and toxicity toward neuroblastoma and liver cancer cells
for the first time. All semisynthesized derivatives remained as
potent inhibitors of AChE with IC50 lower than 10 µM, but
moderate to weak BChE inhibition with IC50 values up to 130 μM
compared with α‐mangostin. The nature of protein–ligand inter-
action with AChE is mainly hydrophobic. In contrast, all the
derivatives were less cytotoxic and had a much higher selec-
tivity index than the parent compound. Thus, modification of
α‐mangostin at position C‐3 and C‐6 greatly impact its BChE in-
hibitory activities and cytotoxicity but not AChE inhibitory ac-
tivities. The dialkylated derivatives are viable candidates for
further structural modification and refinement in the search of
new AChE inhibitors with higher safety margins.
>90% protein binding. From the pharmacological point of view,
compounds having lesser protein binding are considered favorable,
especially in a disease state with lower plasma proteins, or in cases of

KHAW ET AL.
7 of 9
|
4
4
| EXPERIMENTAL
s), 3.37 (2H, d, J = 5.8 Hz), 1.94 (2H, m), 1.85 (3H, s), 1.80 (3H, s), 1. 67
13
(
3H, s), 1.12 (3H, t, J = 5.9 Hz), 1.09, (3H, t, J = 5.9 Hz). C NMR
(CDCl , 125 MHz, δ ppm): 182.0, 162.8, 159.8, 157.4, 155.2, 155.1,
144.0, 137.0, 131.6, 131.4, 123.3, 122.5, 111.8, 111.4, 103.8, 98.7,
.1 | Isolation of α‐mangostin and semisynthesis of
the alkyl derivatives
3
89.2, 70.3, 69.9, 60.8, 26.1, 25.9, 25.8, 22.5, 22.3, 21.2, 18.1, 17.8,
+
The fruit pericarp of G. mangostana was collected from Penang,
Malaysia, and a voucher specimen (No. 11247) of the plant was de-
posited at the School of Biological Sciences, Universiti Sains Malaysia.
α‐Mangostin was isolated according to the procedure described
10.6. Mass 496.3 [M + H] .
4.1.4 | Synthesis of A4
[26]
previously.
The semisynthesis of α‐mangostin derivatives was con-
ducted by O‐alkylation reaction in which the hydroxyl group of the
α‐mangostin was replaced by the alkyl chain from the alkyl halide
The title compound A4 was obtained as yellowish powder (yield
1
43.2%) with melting point of 88–90°C. H NMR (CDCl
3
, 500 MHz, δ
(
methyl iodide, ethyl iodide, propyl iodide, and 1‐chlorobutane). The
ppm): 6.70 (1H, s), 6.28 (1H, s), 4.13 (2H, d, J = 7.1 Hz), 5.26 (1H, m),
1.80 (3H, s), 1.86 (3H, s), 3.36 (2H, d, J = 7.1 Hz), 1. 67 (3H, s), 3.80
(3H, s), 13.50 (1H, s), 4.03 (2H, t, J = 6.35 Hz), 4.08 (2H, t, J = 6.35 Hz),
flow of the reaction is as follows: α‐Mangostin was dissolved in acetone
and refluxed at about 60–65°C with constant stirring overnight;
potassium carbonate was used as a catalyst. The reaction was mon-
itored by developing the product and reactant on thin layer chroma-
tography with the solvent system of hexane and ethyl‐acetate (6:4).
The InChI codes of the investigated compounds, together with
some biological activity data, are provided as Supporting Informa-
tion Data.
1
3
1.91 (2H, m), 1.58 (2H, m), 1.02 (3H, t, J = 7.35 Hz). C NMR in
CDCl : 182.0, 162.8, 159.8, 157.4, 155.3, 155.1, 144.0, 137.0, 131.6,
3
131.3, 123.3, 122.5, 111.8, 111.4, 103.8, 98.67, 89.2, 68.5, 68.1, 30.9,
31.1, 26.1, 25.9, 25.8, 21.4, 17.8, 18.1, 19.3, 19.0, 13.8. Mass
+
524.3 [M + H] .
4
.1.5 | Crystallographic data of A1
4
.1.1 | Synthesis of A1
Crystal was placed in the cold stream of an Oxford Cryosystems
[
33]
The title compound was obtained as pale yellow solid (yield 80.0%)
Cobra open‐flow nitrogen cryostat
operating at 100.0 K. Crystal-
1
with melting point of 89–95°C. H NMR (CDCl
3
, 500 MHz, δ ppm):
lographic data were collected using a Bruker SMART APEX II CCD
[
34]
13.47 (1H, s), 6.73 (1H, s), 6.32 (1H, s), 5.23 (1H, m), 4.13 (2H, d,
diffractometer. SADABS and SAINT software
were used for ab-
J = 7 Hz), 3.35 (2H, d, J = 7 Hz), 3.79 (3H, s), 3.89 (3H, s), 1.84 (3H, s),
sorption correction and data reduction, respectively. The structure
1
3
2
1
.79 (3H, s), 1.67 (3H, s), 1.57 (3H, s). C NMR (CDCl3, 125 MHz, δ
was refined by full‐matrix least squares on F and solved by direct
ppm): 182.0, 163.4, 158.0, 155.3, 155.1, 159.8, 144.0, 137.3, 131.7,
methods using the SHELXTL software package. O‐bound H atom was
in a difference Fourier map and refined freely (O–H = 0.92(2) Å).
The remaining hydrogen atoms were positioned geometrically
1
2
31.6, 123.2, 122.3, 104.0, 111.5, 112.1, 98.2, 88.6, 60.9, 55.7, 55.9,
+
6.1, 25.9, 25.8, 21.3, 18.1, 17.7. Mass 440.2 [M + H] .
(C–H = 0.95–0.99 Å) and were refined using a riding model, with
Uiso ~ (H) = 1.2 Ueq(C). A rotating‐group model was applied for the
4
.1.2 | Synthesis of A2
methyl groups. Crystallographic data for A1 has been deposited at
the Cambridge Crystallographic Data Center (No. CCDC 868385).
The title compound was obtained as pale yellow solid (yield 75.0%) with
1
melting point of 100–102°C. H NMR (CDCl
3
, 500 MHz, δ ppm): 13.49
(1H, s), 6.68 (1H, s), 6.26 (1H, s), 5.27 (1H, m), 1.85 (3H, s), 1.80 (3H, s),
4.2 | Biological assays
3.36 (2H, d, J = 7.4 Hz), 1.68 (3H, s), 3.80 (3H, s), 4.17 (2H, m), 1.48 (3H,
13
t, J = 6.9 Hz), 1.51 (3H, t, J = 6.5 Hz). C NMR (CDCl3, 125 MHz, δ ppm):
4.2.1 | Cholinesterase inhibition assay
1
1
2
82.0, 162.7, 159.8, 157.2, 155.2, 155.1, 144.0, 137.1, 131.6, 131.3,
23.3, 122.4, 111.9, 111.4, 103.8, 98.6, 89.2, 64.4, 64.0, 60.7, 26.1, 25.9,
Cholinesterase inhibitory activities of the synthesized derivatives
were evaluated following Ellman's microplate assay. Briefly, for AChE
assay, 140 μl of 0.1 M sodium phosphate buffer (pH 8) was first
added to each well of a 96‐well microplate followed by 20 μl of the
test sample (in 10% methanol) and 20 μl of 0.09 unit/ml AChE. After
+
5.8, 21.4, 18.1, 17.8, 14.7, 14.5. Mass 468.2 [M + H] .
4
.1.3 | Synthesis of A3
15 min of pre‐incubation at 25°C, 10 μl of 10 mM 5,5′‐dithiobis(2‐
The title compound A3 was obtained as yellowish powder (yield
nitrobenzoic acid) was added to each well followed by 10 μl of 14 mM
acetylthiocholine iodide. The absorbance of the colored end product
was measured at 412 nm at designated intervals for 30 min after the
initiation of enzymatic reaction by Tecan Infinite 200 ProMicroplate
1
3
4.1%) with melting point of 87–89°C. H NMR (CDCl
3
, 500 MHz, δ
ppm): 13.49 (1H, s), 6.70 (1H, s), 6.28 (1H, s), 5.26 (1H, m), 4.13 (2H,
d, J = 5.8 Hz), 4.03 (2H, t, J = 6.6 Hz), 3.98 (2H, t, J = 6.6 Hz), 3.80 (3H,

8
of 9
KHAW ET AL.
|
spectrometer (Switzerland). For the BChE inhibitory assay, the same
procedure was followed except that the enzyme and substrate used
were BChE from the equine serum and S‐butyrylthiocholine chloride,
respectively. Galantamine was used as a reference standard. Each
sample test was conducted in triplicates. The absorbance of the test
sample was corrected by subtracting the absorbance of its respective
blank. A set of five concentrations was used to estimate the
ORCID
Kooi‐Yeong Khaw
http://orcid.org/0000-0003-2360-6235
REFERENCES
[1] L. Gales, A. M. Damas, Curr. Med. Chem. 2005, 12, 2499.
[
[
2] K. S. Masters, S. Brase, Chem. Rev. 2012, 112, 3717.
3] M. M. Pinto, M. E. Sousa, M. S. Nascimento, Curr. Med. Chem. 2005,
2, 2517.
1
50% inhibitory concentration (IC50) of the compounds.
[4] M. E. Sousa, M. M. Pinto, Curr. Med. Chem. 2005, 12, 2447.
[5] L. M. Vieira, A. Kijjoa, Curr. Med. Chem. 2005, 12, 2413.
[6] B. Lesch, S. Bräse, Angew. Chem., Int. Ed. 2004, 43, 115.
[7] D.‐J. Jiang, Z. Dai, Y.‐J. Li, Cardiovasc. Drug Rev. 2004, 22, 91.
[8] V. Peres, T. J. Nagem, Phytochemistry 1997, 44, 191.
[9] J. S. Negi, V. K. Bisht, P. Singh, M. S. M. Rawat, G. P. Joshi, J. Appl.
Chem. 2013, 2013, 9.
4
.2.2 | Cytotoxicity assay
SH‐SY5Y, SK‐N‐SH, and HepG2 cells were seeded at 10,000 cells/
well in a 96‐well plate, in Dulbecco's modified Eagle's medium sup-
plemented with 50% Ham's F‐12 nutrient mixture, 10% fetal bovine
serum, penicillin (100 U/ml), and streptomycin (100 μg/ml), and were
grown at 37°C in a humidified chamber containing 95% air and
[10] M. Y. Ibrahim, N. M. Hashim, A. A. Mariod, S. Mohan, M. A.
Abdulla, S. I. Abdelwahab, I. A. Arbab, Arab. J. Chem. 2016,
9
, 317.
[
[
11] L. G. Chen, L. L. Yang, C. C. Wang, Food Chem. Toxicol. 2008, 46, 688.
12] S. Tewtrakul, C. Wattanapiromsakul, W. Mahabusarakam,
J. Ethnopharmacol. 2009, 121, 379.
2
5% CO . The cells were allowed to grow for 24 hr. The cell viability
[
13] H. A. Jung, B. N. Su, W. J. Keller, R. G. Mehta, A. D. Kinghorn, J. Agric.
Food Chem. 2006, 54, 2077.
was determined using MTT [3‐(4,5‐dimethylthiazol‐2‐yl)‐2,5‐diphenyl
tetrazolium bromide] colorimetric assay method. The cells were
incubated with different concentrations of compounds ranging
from 0.39 to 200 µg/ml for 48 hr. Thereafter, 10 μl of MTT solution
[
14] B. Márquez‐Valadez, R. Lugo‐Huitrón, V. Valdivia‐Cerda, L. R.
Miranda‐Ramírez, V. Pérez‐De La Cruz, O. González‐Cuahutencos, I.
Rivero‐Cruz, R. Mata, A. Santamaría, J. Pedraza‐Chaverrí, Nutr.
Neurosci. 2009, 12, 35.
(1 mg/ml) was added to the wells and the plate was further incubated
[
[
15] G. B. Nelli, A. S. K, E. K. Kilari, Syst. Biol. Reprod. Med. 2013, 59,
for 4 hr at 37°C. The medium was then removed and 100 μl of iso-
propanol was added to each well and vigorously mixed to dissolve the
formazan crystals. The absorbance was measured at 570 nm.
3
19.
16] H. W. Ryu, M. J. Curtis‐Long, S. Jung, Y. M. Jin, J. K.
Cho, Y. B. Ryu, W. S. Lee, K. H. Park, Bioorg. Med. Chem. 2010, 18,
6
258.
[
17] J. Pedraza‐Chaverri, N. Cardenas‐Rodriguez, M. Orozco‐Ibarra, J. M.
Perez‐Rojas, Food Chem. Toxicol. 2008, 46, 3227.
4
.3 | Molecular docking
[
18] L. M. Reyes‐Fermin, S. Gonzalez‐Reyes, N. G. Tarco‐Alvarez, M.
Hernandez‐Nava, M. Orozco‐Ibarra, J. Pedraza‐Chaverri, Nutr. Neu-
rosci. 2012, 15, 34.
Molecular docking was performed by AutoDock 3.0.5 along with
[35]
[19] Y. Wang, Z. Xia, J. R. Xu, Y. X. Wang, L. N. Hou, Y. Qiu, H. Z. Chen,
AutoDockTools (ADT) using the Lamarckian genetic algorithm.
Neuropharmacology 2012, 62, 871.
Energy minimization of the compounds was performed using the Tripos
force field with a distance gradient algorithm and a convergence cri-
terion of 0.05 kcal/(mol A) MOE (Chemical Computing Group, Montreal,
Canada). The crystal structure of TcAChE in complex with galantamine
[
20] A. Kijjoa, M. J. Gonzalez, M. M. Pinto, M. S. Nascimento, N. Campos, I.
O. Mondranondra, A. M. S. Silva, G. Eaton, W. Herz, Planta Med. 2008,
74, 864.
[
21] S. Laphookhieo, J. K. Syers, R. Kiattansakul, K. Chantrapromma,
Chem. Pharm. Bull. 2006, 54, 745.
[36]
was obtained from Protein Data Bank (PDB code: 1W6R).
TcAChE
[
22] A. R. Han, J. A. Kim, D. D. Lantvit, L. B. Kardono, S. Riswan, H. Chai, E.
J. C. D. Blanco, N. R. Farnsworth, S. M. Swanson, A. D. Kinghorn,
J. Nat. Prod. 2009, 72, 2028.
protein was edited using ADT to remove all water molecules and hy-
drogen atoms were added. Nonpolar hydrogen and lone pairs were then
merged and each atom within the macromolecule was assigned a Gas-
teiger partial charge. A grid box of 41 × 53 × 41 points, with a spacing of
[
23] Y. Nakagawa, M. Iinuma, T. Naoe, Y. Nozawa, Y. Akao, Bioorg. Med.
Chem. 2007, 15, 5620.
[
24] J. J. Wang, Q. H. Shi, W. Zhang, B. J. Sanderson, Food Chem. Toxicol.
0.375 Å was positioned at the active site gorge. One hundred in-
2
012, 50, 3004.
dependent dockings were carried out per docking experiment.
[25] K. Matsumoto, Y. Akao, H. Yi, K. Ohguchi, T. Ito, T. Tanaka, E.
Kobayashi, M. Linuma, Y. Nozawa, Bioorg. Med. Chem. 2004, 12,
5
799.
ACKNOWLEDGMENTS
[
26] K. Y. Khaw, S. B. Choi, S. C. Tan, H. A. Wahab, K. L. Chan, V. Mur-
The authors would like to thank Universiti Sains Malaysia, Malaysia,
for providing financial support and research facilities. This project
was funded by the Research University Grant (RUI) from Ministry of
Higher Education Malaysia and Universiti Sains Malaysia (Grant No.
ugaiyah, Phytomedicine 2014, 21, 1303.
[27] Z. Guo, Acta Pharm. Sin. B 2017, 7, 119.
[
[
[
[
28] M. Biagiotti, V. M. Pappalardo, M. Rabuffetti, G. Speranza, Nat. Prod.
Res. 2015, 29, 750.
29] X. Fei, M. Jo, B. Lee, S.‐B. Han, K. Lee, J.‐K. Jung, S. Y. Seo, Y.‐S. Kwak,
Bioorg. Med. Chem. Lett. 2014, 24, 2062.
1001/PFARMASI/815093).
30] H. Dvir, I. Silman, M. Harel, T. L. Rosenberry, J. L. Sussman, Chem.‐
Biol. Interact. 2010, 187, 10.
CONFLICTS OF INTERESTS
31] Z. Rankovic, J. Med. Chem. 2015, 58, 2584.
The authors declare that there are no conflicts of interests.

KHAW ET AL.
9 of 9
|
[
32] Z. Rankovic, Blood‐Brain Barrier in Drug Discovery: Optimizing Brain
Exposure of CNS Drugs and Minimizing Brain Side Effects for Peripheral
Drugs, John Wiley & Sons, Hoboken, NJ 2015, p. 3871.
How to cite this article: Khaw K‐Y, Kumar P, Yusof SR,
Ramanathan S, Murugaiyah V. Probing simple structural
[
[
[
33] J. Cosier, A. M. Glazer, J. Appl. Cryst. 1986, 19, 105.
modification of α‐mangostin on its cholinesterase inhibition
and cytotoxicity. Arch Pharm. 2020;e2000156.
https://doi.org/10.1002/ardp.202000156
34] G. Sheldrick, Acta Crystallogr. Sect. C 2015, 71, 3.
35] G. M. Morris, D. S. Goodsell, R. S. Halliday, R. Huey, W. E. Hart, R. K.
Belew, A. J. Olson, J. Comput. Chem. 1998, 19, 1639.
[36] H. M. Greenblatt, C. Guillou, D. Guenard, A. Argaman, S. Botti, B.
Badet, C. Thal, I. Silman, J. L. Sussman, J. Am. Chem. Soc. 2004, 126,
15405.