DSpace at VNU: An investigation of antidiabetic activities of bioactive compounds in Euphorbia hirta Linn using molecular docking and pharmacophore

MEDICINAL
CHEMISTRY
RESEARCH

Med Chem Res (2014) 23:2033–2045
DOI 10.1007/s00044-013-0794-y

ORIGINAL RESEARCH

An investigation of antidiabetic activities of bioactive compounds
in Euphorbia hirta Linn using molecular docking
and pharmacophore
Quy Trinh • Ly Le

Received: 20 May 2013 / Accepted: 12 September 2013 / Published online: 2 October 2013
Ó Springer Science+Business Media New York 2013

Abstract Herbal remedies have been considered as
potential medication for diabetes type 2 treatment. Bitter
melons, onions, or Goryeong Ginsengs are popular herbals
and their functions in diabetes patients have been well
documented. Recently, the Euphorbia hirta has been
shown to have strong effects on diabetes in mice, however,
there has been no research clearly indicating what the
active compound is. The main purpose of the current study
was therefore to evaluate whether a relationship exists
between various bioactive compounds in E. hirta Linn and
targeted protein relating diabetes type 2 in human. In view
of this, extraction from E. hirta Linn was tested if they
contained the bioactive compounds. This process involved
the docking of 3D structures of those substances (ligand)

into targeted proteins: 11-b hydroxysteroid dehydrogenase
type 1, glutamine: fructose-6-phosphate amidotransferase,
protein phosphatase, and mono-ADP-ribosyltransferase
sirtuin-6. Then, LigandScout was applied to evaluate the
bond formed between ligand and the binding pocket in the
protein. These test identified in eight substances with high
binding affinity (\-8.0 kcal/mol) to all four interested
proteins of this article. The substances are quercetrin, rutin,
myricitrin, cyanidin 3,5-O-diglucoside, pelargonium 3,5diglucose in ‘‘flavonoid family’’ and a-amyrine, b-amyrine,
Electronic supplementary material The online version of this
article (doi:10.1007/s00044-013-0794-y) contains supplementary
material, which is available to authorized users.
Q. Trinh Á L. Le
School of Biotechnology, International University–Vietnam
National University, Ho Chi Minh City, Vietnam
L. Le (&)
Life Science Laboratory, Institute of Computational Science
and Technology, Ho Chi Minh City, Vietnam
e-mail:

taraxerol in ‘‘terpenes group.’’ The result can be explained
by the 2D picture which showed hydrophobic interaction,
hydrogen bond acceptor, and hydrogen bond donor forming between carbonyl oxygen molecules of ligand with free
residues in the protein. These pictures of the bonding
provide evidence that E. hirta Linn may prove to be an
effective treatment for diabetes type 2.
Keywords Diabetes type 2 Á Euphorbia hirta Linn Á
Molecular docking Á Pharmacophore analysis

Introduction

Diabetes, one of the metabolic diseases that have high
blood sugar as a pathognomonic symptom, is spreading
like an epidemic. Worldwide, the number of patients
climbed steeply from 171 million in 2000 to 366 million in
2030 (Wild et al., 2004) and *90 % are of type 2 (International Diabetes Federation, 2006). A person with this
type of diabetes suffers a combination of insulin resistance
and a weakness in insulin production. Insulin resistance is
considered as stage one in diabetes type 2. In this phase, the
glucose, energy molecule of the cell cannot cross the cell
membrane due to blocking of the insulin receptor at the cell
surface. This result is a high glucose concentration in the
blood stream. To solve the problem, the pancreatic beta
cells produce extra insulin to maintain glucose in the normal range. However, this process is only effective in the
short term as burnout beta cell occurs. The failure for beta
cell to produce the extra insulin is the second stage of
diabetes type 2.
Determination of the best treatment for diabetes type 2
is complicated because this is a progressive disease. Currently, insulin combined with other drugs is the preferred

123


2034

treatment method. Recently, natural herbal medicines are
preferable options. A study by Modak and coworkers have
provided a list of several medicinal plants used for diabetes
treatment (Modak et al., 2007). Several of them, such as
Caesalpinia bonducella (L) Roxb (Chandramohan et al.,
2008), Allium cepa (Onion) (G. B. Kavishankar et al.,

2011), Vitis vinifera and Euonymus alatus (Chan et al.,
2012), share a high concentration of Flavonoid compounds
including Quercetin, Kaemferol, Cyanidin, and Pelargonium. Our study focuses on Euphorbia hirta, one member
of Euphorbiaceae family which has high concentration of
these bioactive compounds in their extraction. E. hirta has
been reported to be effective in reducing diabetes in mice
in vitro studies (Anup et al., 2012; Sunil and Rashmi,
2010). When, the ethanol extracted compound from the
leaves, stems, and flowers of E. hirta was applied to mice
which had induced diabetes by a single intraperitoneal
injection of streptozotocin (150 mg/kg), the result revealed
that compounds displayed antihyperglycemic activity in
the diabetic mice. To further understand this result, the
current study focuses on identifying the bioactivity of the
antidiabetes components of the ethanol extracts of E. hirta
by using them as ligand molecules for four targeted proteins to determine which compound is an effective binder.
E. hirta contains three families of biomolecular compounds such as tannin, flavonoid, and terpenes (Mohammad et al., 2010; Sandeep and Chandrakant, 2011). Tannin
and flavonoid are strong antioxidants (Pietta, 2000; Rield
and Hagerman, 2001). Quercitrin, one compound in Flavonoid group, was good illustration. In the thiobarbituric
acid (TBA) experiment quercitrin showed strong antioxidant activity, giving 92.5 % inhibition and the IC50 was
calculated to 23.40 lM (Basma et al., 2011). Products of
oxidation have been shown to play an essential role in the
pathogenesis of diabetes type 1 and 2 (Maritim et al.,
2003). In addition, the combination of high level of free
radicals and inactivation of antioxidant defense have been
shown to cause damage in cellular organelles and to the
production of insulin (Maritim et al., 2003). Therefore,
antioxidants such as tannin and flavonoid are considered to
have potential as therapeutic drugs for diabetes treatment.
Both flavonoid and terpenes from medicinal plants have

already been shown to have strong effects on diabetes
(Mankil et al., 2006). In light of this evidence, the current
study will screen a range of bioactive compounds from all
the three families to determine if and how they interact
with proteins important to human diabetes type 2.
Several proteins which were involved in glucose
metabolism consequently related to diabetes type 2. From
our intensive review on targeted proteins for antidiabetic
drug development (Trang and Ly, 2012), these important
target proteins including 11-b hydroxysteroid dehydrogenase type 1 (11b-HSD1), Glutamine fructose-6-phosphate

123

Med Chem Res (2014) 23:2033–2045

amidotransferase (GFPT or GFAT), protein phosphatase
(PPM1B), and Mono-ADP-ribosyltransferase sirtuin-6
(SIRT6) were selected as receptors in this study.

Material and methodology
Molecular docking
Receptor
11-b HSD1, GFAT, PPM1B, and SIRT6 are the proteins
relating to diabetes type 2 in humans (Hasan et al., 2002;
Trang and Ly, 2012; Vogel, 2002; Shi, 2009; Nerlich
et al., 1998). The 3D structures of these molecules taken
from Protein Data Bank are as follows: 11b-HSD1 (PDB
code 1XU7), GFAT (PDB code 2ZJ4), PPM1B (PDB
code 2P8E), and SIRT6 (PDB code 3K35). All these
structures were tested again at the binding site to verify the

capacity of the model in reproducing experimental observations with new ligand. In view of this, 11b-HSD1
(PDB code 1XU7) was tested again with molecule:
NADPH dihydro-nicotinamide-adenine-dinucleotide phosphate; GFAT (PDB code 2ZJ4) was tested with 2-deoxy-2amino glucitol-6-phosphate; SIRT6 (PDB code 3K35) with
adenosine-5-diphosphoribose; and PPM1B (PDB code
2P8E) with cysteine sulfonic acid. They served as control
docking models illustrated in supplementary Table 4. This
work was done by Autodock vina in molecular docking
experiment and VMD in visualization (Humphrey et al.,
1996).
Bioactive compounds in E. hirta
Most of the 3D structures of drug molecules in E. hirta
were downloaded from PubChem Compound section of
National Center for Biotechnology Information (NCBI).
For molecules with unknown structure, the 3D models
were built based on 2D picture by GaussView 5.0, optimized by Gaussian with Hatree-Fock method, and the
basis-set 6-31G* to increase reliability of structure. The 2D
structures of 27 ligands are illustrated in Table 1.
Docking simulations
The docking process was done using Autodock Vina (Oleg
and Arthur, 2009).
Autodocktool, one section in Molecular Graphic Laboratory, was applied to build a complete pdbqt file name of
ligands and receptors. Receptor preparation was carried out
by four major sub-steps: (i) Adding polar hydrogen, (ii)
Removing water molecule, (iii) Computation of Gasteiger
charges, and (iv) Location of Grid box (supplementary


Quercitrin

Rutin

Pelargonium 3,5-diglucose

Quercetin

Cycloartenol

Cyanidin 3,5-O-diglucoside

Table 1 2D structures of 27 drug candidates suggested from NCBI

Kaemferon

Leucocyanidin

Quercitol

Camphol

Myricitrin

Rhamnose

Med Chem Res (2014) 23:2033–2045
2035

123


3.4-di-O-galloyquinic acid

Stigmasterol

Ingenol triacetate

Neuchlogenic acid

Beta sitosterol

Resiniferonol

Table 1 continued

123
12-Deoxy-phorbol-13-dodecanoate-20-acetate

a-Amyrine

b-Amyrine

Campesterol

12-deoxy-phorbol-13-phenylacetate-20-acetate

Benzyl gallate

2036
Med Chem Res (2014) 23:2033–2045

Med Chem Res (2014) 23:2033–2045

2037
Table 2 Position of the Grid box center in four protein molecules
Protein molecule

PDB code

˚)
X, Y, Z coordination (A

Friedelin

X

Y

Z

11b-HSD1

1XU7

18.125

-27.72

-0.34

GFAT

2ZJ4

8.27

4.54

-7.67

PPM1B

2P8E

-11.72

-18.53

9.86

SIRT6

3K35

14.5

-18.02

17.04

Fig. 7). The site of Grid Box is illustrated in Table 2. For

setting the ligands, the 3D structure in pdb file-type was
loaded into Autodocktool to detect the root and convert it
to pdbqt.
Before switching on the Autodock Vina, one configure
file was built to encode information for starting this program. The content of configure file was determined as
position of receptor file, ligand file, data of Grid-box’s
three coordinates (Table 2), the size of Gridbox which was
set up in 30 9 30 9 30 points, number of modes which
were ten, and the energy range which was set up at 9 kcal/
mol.

Taraxerone

Pharmacophore modeling
This part of process was carried out using the pharmacophore tool included in LigandScout. The program showed
us the 2D and 3D structure with the position and interaction
of ligand in the binding pocket of the receptor. From these
2D pictures, some types of bond were identified by color
and symbol. Four features namely hydrogen bond acceptor
(HBA), hydrogen bond donor (HBD), negative ionizable
area, hydrophobic interaction were labeled as red arrow,
green arrow, red star, and orange bubble (supporting
information), respectively.

Result and discussion

Taraxerol

Table 1 continued

Free energy binding of bioactive compound to targeted
protein related to diabetes type 2
In order to investigate the binding capacity of bioactive
compounds in E. hirta Linn on proteins related to diabetes
type 2 in humans, we docked the compounds to the proteins. Results showed that the absolute value of binding
energy ranged from 7.0 to 12.8 kcal/mol (Fig. 2). The
group of terpenes including a-amyrine, b-amyrine, friedelin, taraxerol, taraxerone, and cycloartenol showed the best
results. All receptor for terpenes group had particularly

123


2038

Med Chem Res (2014) 23:2033–2045

high binding affinities with the highest at 11b-HSD1 (PDB
code 1XU7) which being 100 % larger than 11 (kcal/mol).
The next highest positions were SIRT6, GFAT, and
PPM1B (Fig. 1). For the terpenes group, the line for 11bHSD1 stayed at the upper level when compared to the other
three receptors. For the ligands tested, terpenes were
therefore considered to be the best drug candidate for
diabetes type 2 and the three compounds that had [8 kcal/
mol in terms of absolute value in binding affinity were
chosen for pharmacophore modeling. They were a-amyrine, b-amyrine, and taraxerol. The high binding efficiency
is thought to be due to the multiple methyl groups in the
structure as these functional groups have a strong ability to
construct hydrophobic bonds with the free residue of the
receptor.
The flavonoid family had the largest number of ligands

and some of these also had high binding affinity to all four
receptors. Five of these quercitrin, rutin, myricitrin, cyanidin 3,5-O-diglucoside, pelargonium 3,5-diglucose were
selected for pharmacophore modeling step. Unsurprisingly,
the five molecules had multiple aromatic phenol rings in
their structure which is characteristic of polyphenol family.
This structure contains a high number of hydroxyl groups
which serve to facilitate ligands in forming hydrogen bonds
with free residue of receptor. In addition, to containing a
high number of ligands with high binding capacities, the
flavonoid family also contained three compounds (quercitol, rhamnose, and camphol) which had the lowest binding
affinity. The absolute value for these three ligands is shown
sequentially in Table 1. They all share a simple structure

with only one ring and few hydroxyl groups outside which
may explain their low binding affinity. Thus, these molecules appear to have a low capacity to form a complex with
the four target proteins.
The tannin family also had molecules which bound well
to the receptors, but there was no representative molecule
for pharmacophore docking. However, they displayed
strong interaction with 11b-HSD1, GFAT1, SIRT6, and low
interaction with PPM1B. Neuchlogenic acid and 3,4 dio
galloy-quinic acid are illustrated in supplementary Table 3.
From the results of this section, we determined that eight
compounds showed strong binding capacity (|binding
energy| [8.0 kcal/mol) to all four 11b-HSD1, SIRT6,
GFAT, and PPM1B receptors. Three of them belong to
terpenes group (a-amyrine, b-amyrine, and taraxerol), the
other five are members of flavonoid family (quercitrin,
rutin, myricitrin, cyanidin 3,5-O-diglucoside, and pelargonium 3,5-diglucose). Five of them have structure of polyphenol family which had previously considered as potential
drug candidate for diabetes type 2 patients (Kati et al.,

2010). Besides that, overall viewing Fig. 1, the line of 11bHSD1 stayed in highest level in most of the case. It means
that there is stronger interaction of ligand on this protein,
compared to other three receptors. Figure 2 shows 24 of the
27 tested (89 %) were higher than 8 kcal/mol and the
friedelin molecule in the terpenes group had better binding
capacity than the controls. Thus the results provide strong
evidences that 11b-HSD1 is a suitable receptor for diabetes
type 2 patients being treated with bioactive compounds
derived from E. hirta.

Fig. 1 Absolute value of binding energy of 27 ligands to 4 receptors.
The short name of these ligands was written as QTin Quercetin, QTrin
quercitrin, QTol quercitol, RhNose Rhamnose. RTn Rutin, LDin
Leucocyanidin, MTrin Myricitrin, CyGlu cyanidin-3,5-diglucose,
KRon kaemferon, PeGlu pelargonium-3,5-diglucose, CPhol camphol,
Ngenic Neuchlogenic acid, GQnic 3,4 dio galloy-quinic acid, BGlate

Benzyl gallate, BSrol Betasitosterol, CSrol Campesterol, SSrol
Stigmasterol, DodeAte 12 deoxyphor-13 dodecanoate-20 acetate,
phenylAte 12 deoxyphor-13 phenylacetate-20 acetate, InTate Ingenol
triacetate, RNol Resiniferonol, ARine a-amyrine, BRine b amyrine,
Flin Friedelin, TRol Taraxerol, TRone Taraxerone, CyNol
Cycloartenol

123


Med Chem Res (2014) 23:2033–2045

2039

Fig. 2 Absolute value of
binding energy between E.
hirta’s ligand and 11b-HSD1
protein

Pharmacophore modeling
11b-HSD1
High binding affinity of the ligand to the receptor (Fig. 2)
was explained clearly by interaction analysis in Fig. 3.
According to the molecular framework, there is a tenable
pharmacophore identified between flavonoid family and
non-flavonoid family (terpenes group). Structure of flavonoid contained high number of hydroxyl group which can
form strong hydrogen bonds with receptors. Five molecules
(cyanidin 3,5-O-diglucose, myricitrin, pelargonium 3,5diglucose, quercitrin, and rutin) were frequently within
hydrogen contact with residues Tyr 183, Thr 124, and Ala
172. From this observation, three residues seemed to play a
critical role in catalytic activity of 11b-HSD1 (PDB code
1XU7). This conclusion is strongly supported by studies on
crystal structures and biochemical of 11b-HSD1 (Malin
et al., 2006; David et al., 2005). In Fig. 3d–h, the Tyr 183
subunit has an important function in the bonding to the
hydroxyl hydrogen of all five ligands whereas Thr 124
could form close vicinity to the ligand surface, and from
there, the hydrogen bond could be set up between them.
The same kind of interaction also happened in case of Ala
172 but this residue was also within hydrophobic contact
with hydrophobes part on ligand (Fig. 3d, f). Moreover,
cyanidin 3,5-O-diglucose, pelargonium 3,5-diglucose, and
rutin could link to the receptor with a high number of

hydrogen bonds compared to myricitrin and quercitrin.
This action can be explained by the affinity of each steroidal hydroxyl group for the receptor. For example, this
functional group in cyanidin 3,5-O-diglucose could donate
two or three hydrogen bonds with different residues such as
Ser 169, Ser 170, Tyr 183, and Leu 215.
In case of terpenes group which has many hydrophobic
components (CH3 group, benzene ring). Thus, terpenes can
form many hydrophobic interactions with other hydrophobic residues in receptors’ active site. a-amyrine, bamyrine, and taraxerol seemed to be rich on hydrophobic

contact at position of the methyl group which is non-polar.
The compounds cyanidin 3,5-O-diglucose, pelargonium
3,5-diglucose, and quercitrin were also in contact with this
receptor because of the presence of the benzene ring.
Previous studies using crystal structure analysis have
reported, Ser 261 and Arg 269 are reported as largely
hydrophobic residues in previous study involving crystal
structure analysis (Malin et al., 2006) but in the figures
from our study, these hydrophobic interactions were not
present. Ile 46, Ile 121, Leu 217, Leu 126, Thr 220, Thr
222… were frequently observed in ligand–receptor interactions between, so they can be a critical part in binding
pocket.
GFAT
There were similarities in the binding mode of 11bHSD1 and the steroidal hydroxyl group of cyanidin 3,5O-diglucose, myricitrin, pelargonium 3,5-diglucose,
quercitrin, and rutin. All established a hydrogen bond
with GFAT1 (PDB code 2ZJ4) at position of Ser 420,
Ser 376, Gln 421, Thr 375, and Ser 422 in the binding
pocket. This result was validated in previous studies
(Kuo-Chen 2004; Vedantham et al., 2007; Yuichiro
et al., 2009). In particular, pelargonium 3,5-diglucose
was seen to have a similar binding mode to the Glc6P

which is a strong inhibitor of GFAT1 (Vedantham et al.,
2007). Besides that, Fig. 4a–c, f, g displayed Thr 425
which was close to not only methyl groups but also to
the hydroxyl groups of a-amyrine, b-amyrine, quercitrin,
rutin, and taraxerol.
In addition, all of these ligands had hydrophobic interactions with receptors at positions of residue Leu 673, Val
677, Leu 556, and Thr 425. The mechanism of these
interactions, however, differed among the ligands. a-amyrine, b-amyrine, myricitrin, and taraxerol developed
hydrophobic bonds with the hydrophobic receptor from
methyl group. Meanwhile, the link between the benzene
ring and interested part of receptor was decisive tendency

123


2040

Fig. 3 Binding modes of selective compounds with 11b-HSD1. a a amyrine,
b b-amyrine, c taraxerol, d myricitrin, e pelargonium 3,5-diglucose, f quercitrin, g rutin, and h cyanidin 3,5-O-diglucose. a–c belong to terpenes family

123

Med Chem Res (2014) 23:2033–2045

and the rest are members of Flavonoid family. Hydrogen Bond Acceptor
(HBA) was shown asgreen vectors, Hydrogen Bond Donor (HBD) was
drawn as red vectors. Hydrophobic (H) was illustrated as yellow spheres


Med Chem Res (2014) 23:2033–2045

Fig. 4 Binding modes of selective compounds with GFAT. a a amyrine,
b b-amyrine, c taraxerol, d myricitrin, e pelargonium 3,5-diglucose, f quercitrin, g rutin, and h cyanidin 3,5-O-diglucose. a–c belong to terpenes family

2041

and the rest are members of Flavonoid family. Hydrogen Bond Acceptor
(HBA) was shown as green vectors, Hydrogen Bond Donor (HBD) was
drawn as red vectors. Hydrophobic (H) was illustrated as yellow spheres

123


2042

Fig. 5 Binding modes of selective compounds with PPM1B. a aamyrine, b b-amyrine, c taraxerol, d myricitrin, e pelargonium 3,5diglucose, f quercitrin, g rutin, and h cyanidin 3,5-O-diglucose. a–c belong
to terpenes family and the rest are members of Flavonoid family. Hydrogen

123

Med Chem Res (2014) 23:2033–2045

Bond Acceptor (HBA) was shown as green vectors, Hydrogen Bond
Donor (HBD) was drawn as red vectors. Hydrophobic (H) was illustrated
as yellow spheres


Med Chem Res (2014) 23:2033–2045

Fig. 6 Binding modes of selective compounds with SIRT6. a a

amyrine, b b amyrine, c taraxerol, d myricitrin, e pelargonium 3,5diglucose, f quercitrin, g rutin, and h cyanidin 3,5-O-diglucose. a–
c belong to terpenes family and the rest are members of Flavonoid

2043

family. Hydrogen Bond Acceptor (HBA) was shown as green vectors,
Hydrogen Bond Donor (HBD) was drawn as red vectors. Hydrophobic (H) was illustrated as yellow spheres

123


2044

in cyanidin 3,5-O-diglucose, pelargonium 3,5-diglucose,
quercitrin, and rutin.
PPM1B and SIRT6
PPM1B (PDB code 2P8E) had low binding affinity to
ligand when compared to 11b-HSD1, GFAT, and SIRT6
but not compared to rutin. This could be explained first by
a low number of bonds between the ligand and the receptor. a-amyrine, b-amyrine, cyanidin 3,5-O-diglucose, and
pelargonium 3,5-diglucose are good illustrations. The
binding energy of a-amyrine to 11b-HSD1, GFAT, SIRt6,
and PPM1B was -11.5, -9.6, -10.4, -8.6 (kcal/mol),
respectively, and the numbers of bonds for their ligand
interaction with the receptor were 23, 12, 11, and 8,
respectively. Moreover, the number of hydrophobic and
hydrogen bonds was also significantly reduced in the
arrangement from 11b-HSD1 to PPM1B. For rutin, the
total number of bonds in PPM was lower than GFAT but
higher than 11b-HSD1 and SIRT6. However, binding

affinity did not follow this pattern and to understand this
finding required a molecular dynamic (MD) and hydrogen
bond analysis step to show. The duration time of the
interaction between ligand and receptor is high frequency
of residues Ala 197, Leu 196, Asp 286, Asp 60, and Asn
287 seemed to play an important role in binding at mode of
PPM1B (Fig. 5). This result differed to Shi (2009) result
which showed Asp 119, Asp 231, Asp 34, Asp 18, Arg 13,
and Gly 35 as the key residue in binding site. This difference can be explained due to different in chain we tested
on.
By describing the crystal structure of SIRT6 (PDB code
3K35), Fig. 6 revealed the different positions of each
ligand in the binding pocket of 8. Figure 6e, g, h supported
this finding. Although there is similarity in the structure of
the molecules, three compounds bound to different residues
with different mechanisms. The benzene ring in cyanidin
3,5-O-diglucose and rutin contacted Trp 255 and Ala 56
through hydrophobic interaction, but in pelargonium 3,5diglucose, the Trp 186 had this function. The hydroxyl
group of the benzene ring in Fig. 6h was the HBD to Thr
55, in contrast with HBA of Tyr 255 in Fig. 6g. From this,
SIRT6 is seen to have a high number of residue which
could form interactions with the functional group of the
ligand. However, most of ligand could link with Trp 186
and Leu 184 which was previously found by Patricia et al.
(2011) in their study of the structure and biochemical
function of SIRT6.
In SIRT6, the total number of bonds did not used to
explain the differences in binding affinity among the three
other receptors in most of situation. For example, there
were 11 bonds between rutin and SIRT6, this number was

lower than 16 bonds in GFAT and 13 bonds in PPM1B but

123

Med Chem Res (2014) 23:2033–2045

rutin had a stronger binding affinity to SIRT6 with -10
(kcal/mol) in binding affinity which was lower than -8.6
(kcal/mol) in GFAT and -8.6 (kcal/mol) in PPM1B. This
result for rutin can, however, be explained by MD and
hydrogen bond analysis in PPM1B. These analysis will
figure out stable hydrogen bond and hydrophobic interaction between ligands and receptors.

Conclusion
Docking simulation of 27 drug candidates extracted from
E. hirta showed that the flavonoid and terpenes families
including cyanidin 3,5-O-diglucose, myricitrin, pelargonium 3,5-diglucose, quercitrin, rutin, a-amyrine, b-amyrine, and taraxerol have high binding affinity to all four
interested receptors which are strongly relevant to diabetes
type 2 in humans. These binding results were shown by
LigandScout to consist of a high number of hydrogen bond
and hydrophobic interactions. However, with the differences in pharmacophore features, the flavonoid family
shows more advantages in binding to these receptors than
terpenes due to relatively strong hydrogen bonds. The
binding pocket of each receptor: Tyr 183, Thr 124, Ala
172, Ile 46, Ile 121, Leu 217, Leu 126, Thr 220, Thr 222 in
11b-HSD1, Ser 420, Ser 376, Gln 421, Thr 375, Ser 422,
Leu 673, Val 677, Leu 556, Thr 425 in GFAT1, Ala 197,
Leu 196, Asp 286, Asp 60, Asn 287 in PPM1B and Trp 186
and Leu 184 in SIRT6 is in agreement with the previous
research. Moreover, five molecules from the flavonoid

family have the polyphenol structure indirectly confirming
the strong capacity of the polyphenol family as a treatment
for diabetes type 2 (Kati et al., 2010). Also the binding
affinity of three of the terpenes compounds also suggest
that this family is also a good prospect for the treatment of
type 2 diabetes. Finally, the comparison of the binding
affinity among the four receptors indicates that 11b-HSD1
is the best receptor for accepting of these bioactive compounds derived from E. hirta.
This study has partially demonstrated the effect of E.
hirta on some proteins relating to diabetes type 2. By
calculating the binding energy and pharmacophore modeling, we have obtained the list of 8 promising compounds
in E. hirta. However, further research, using the MD to
determine more accurate binding affinities and the stability
of ligand–proteins’ interactions, is highly suggested. In
addition, experiment study to determine the concentration
of these compounds in E. hirta extraction and their antidiabetic activity should be done for drug formulation.
Acknowledgments This research was funded by the Ho Chi Minh
International University-Vietnam National University. The computing resources and support by the Institute for Computer Science and


Med Chem Res (2014) 23:2033–2045
Technology (ICST) at the Ho Chi Minh City were gracefully
acknowledged.The authors wish to thank Prof. Ho Thanh Phong for
his encouragements and Dr Gay Marsden for proof reading our
manuscript.

References
Anup KM, Smriti T, Zabeer A, Ram K, Sahu (2012) Antidiabetic and
antihyperlipidemic effect of Euphorbia hirta in streptozotocin
induced diabetic rats. Der Pharmacia Lett 4(2):703–707

Basma AA, Zakaria Z, Latha LY, Sasidharan s (2011) Antioxidant
activity and phytochemical screening of the methanol extracts of
Euphorbia hirta L. Asian Pac J Trop Med 4(5):386–390
Chan CH, Ngoh GC, Yusoff R (2012) A brief review on anti-diabetic
plants: global distribution, active ingredients, extraction techniques and acting mechanisms. Phcog 6:22–28
Chandramohan G, Ignacimuthu S, Pugalendi KV (2008) A novel
compound from Casearia esculenta (Roxb.) root and its effect on
carbohydrate metabolism in streptozotocin-daibetic rats. Eur J
Pharmacol 590:437–443
David JH, Yiqin W, Robert JS, Mark H, Andy J, Gyorgy PS, Kathleen
A (2005) Conformational flexibility in crystal structure of human
11b- hydroxysteroid dehydrogenase type 1 provide insights into
glucocorticoid interconversion and enzyme regulation. J Biol
Chem 280(6):4639–4648
Hasan S, El-Andaloussi N, Hardeland U, Hassa PO, Burki C, Imhof
R, Schar R, Hottiger MO (2002) Acetylation regulates the DNA
end-trimming activity of DNA polymerase beta. Mol Cell
10(5):1213–1222
Humphrey W, Dalke A, Schulten K (1996) VMD: visual molecular
dynamics. J Mol Graph 14:33–38
International Diabetes Federation (IDF) (2006) Diabetes Atlas, 3rd
edn. International Diabetes Federation, Brussels
Kati H, Riitta T, Isabel BP, Jenna P, Marjukka K, Hannu M, Kaisa P
(2010) Impact of dietary polyphenols on carbohydrate metabolism. Int J Mol Sci 11:1365–1402
Kavishankar GB, Lakshmidevi N, Mahadeva Murthy S, Prakash HS,
Niranjanh SR (2011) Diabetes and medicinal plants—a review.
Int J Pharm Biomed Sci 2(3):65–80
Kuo-Chen C (2004) Molecular therapeutic target for type-2 diabetes.
J Proteome Res 3:1284–1288
Malin H, Naeem S, Bjorn E, Doreen M, Stefan S, Margareta F, Tjeerd

B, Jerk V, Lars A, Udo O (2006) Active site variability of type 1
11b-hydroxysteroid dehydrogenase revealed by selective inhibitors and cross-species comparisons. Mol Cell Endocrinol
248:26–33
Mankil J, Moonsoo P, Hyun CL, Yoon-Ho K, Eun SK, Sang KK
(2006) Antidiabetic agents from medicinal plants. Curr Med
Chem 13:1203–1218

2045
Maritim AC, Sanders RA, Watkins JB (2003) A review: diabetes,
oxidative stress, and antioxidants. J Biochem Mol Toxicol
17(1):24–38
Modak M, Dixit P, Londhe J, Ghaskadbi S, Paul ADT (2007) Indian
herbs and herbal drugs used for the treatment of diabetes. J Clin
Biochem Nutr 40(3):163–173
Mohammad ABR, Zakaria Z, Sreenivasan S, Lachimanan YL, Santhanam A (2010) Assessment of Euphorbia hirta L. Leaf, flower, stem
and root extracts for their antibacterial and antifungal activity and
brine shrimp lethality. J Mol 15:6008–6018
Nerlich AG, Sauer U, Kolm-Litty V, Wagner E, Koch M, Schleicher
ED (1998) Expression of glutamine:fructose-6-phosphate amidotransferase in human tissues: evidence for high variability and
distinct regulation in diabetes. Diabetes 47:170–178
Patricia WP, Jessica LF, Mark KD, Aiping D, Aled ME, John MD
(2011) Structure and biochemical function of SIRT6. J Biol
Chem 286(16):14575–14587
Pietta PG (2000) Flavonoids as antioxidants. J Nat Prod 63(7):
1035–1042
Rield KM, Hagerman AE (2001) Tannin protein complexes as radical
scavengers and radical sinks. J Agric Food Chem 49:4917–4923
Sandeep BP, Chandrakant SM (2011) phytochemical investigation
and antitumour activity of Euphorbia hirta Linn. Eur J Exp Biol
1(1):51–56

Shi Y (2009) Serine/threonine phosphatases: mechanism through
structure. Cell 139:1016
Sunil K, Rashmi DK (2010) Evaluation of antidiabetic activity of
Euphorbia hirta Linn in streptozotocin induced diabetes mice.
Indian J Nat Prod Resour 1(2):200–203
Trang N, Ly L (2012) Targeted proteins for diabetes drug design. Adv
Nat Sci 3:013001
Vedantham S, Narasimhan S, Rangasamy S, Syed F, Viswanathan M,
Muthuswamy B (2007) Glutamine fructose-6-phosphate amidotransferase (GFAT) gene expression and activity in patients with
type 2 diabetes: inter-relationships with hyperglycaemia and
oxidative stress. Clin Biochem 40:952–957
Vogel HG (2002) Drug discovery and evaluation. Pharmalogical
assays, 2nd edn. Springer, Berlin, pp 1030–1036
Wild S, Roglic G, Green A, Sicree R, King H (2004) Global
prevalence of diabetes: estimates for the year 2000 and
projections for 2030. Diabetes Care 27:1047–1053
Oleg T, Arthur JO (2009) Software news and update AutoDock Vina:
improving the speed and accuracy of docking with a new scoring
function, eficient optimization, and multithreading. Wiley InterScience, New York. doi:10.1002/jcc.21334
Yuichiro N, Masahiko B, Hiroshi S, Kenji W, Fumitaka G, Hideaki T,
Kazumi K, Makoto K (2009) Structural analysis of human
glutamine: fructose-6-phosphate amidotransferase, a key regulator in type 2 diabetes. FEBS Lett 583:163–167

123