PRELIMINARY STUDY ON DRUG DEVELOPMENT AGAINST HORMONAL

DISORDERS IN FEMALE, IN-SILICO APPROACH

A DISSERTATION

Submitted to:
DEPARTMENT OF BIOTECHNOLOGY
SANN INTERNATIONAL COLLEGE
Chabahil, Kathmandu
Purbanchal University, Nepal

Submitted by:
Amit Kumar Yadav (017-3-2-05138-2017)
Anisha Regmi (017-3-2-05139-2017)
Bhawana Samal Magar (017-3-2-05142-2017)
Sandesh Prasad Koirala (017-3-2-05160-2017)

In partial fulfillment of the requirements for the degree of

BACHELOR OF SCIENCE IN BIOTECHNOLOGY
September 2022

1
 DECLARATION

We, hereby, declare that the report entitled “Preliminary study on drug development against
hormonal disorders in female, In silico approach” being submitted to the Department of
Biotechnology, SANN International College, Purbanchal University, in partial fulfillment of
the requirements for the award of the degree of Bachelor of Science in Biotechnology is our
original concept which was accomplished under the continuous guidance of our supervisor Dr.
Pramod Aryal, co-supervisors Prof. Dr. Rameshwor Adhikari, and Sameeran Subedi along
with internal supervisor Mr. Keshab Raj Budha at the Department of Biotechnology,
Tribhuvan University. Kathmandu, Nepal; Research Centre for Applied Science and Technology
(RECAST), Kathmandu, Nepal and Department of Biotechnology, SANN International College,
Kathmandu, Nepal. The works shown here has not been submitted either in part or full to any
other university or institution for the award of any degree or diploma. Also, in this report, proper
citation with due acknowledgement has been done for all the necessary data, procedures,
information or any type of work by others.

Amit Kumar Yadav

Anisha Regmi

Bhawana Samal Magar

Sandesh Prasad Koirala

2
 ACKNOWLEDGEMENT
We would like to express our heartily gratitude to our respected supervisor Dr. Pramod Aryal,
Visiting Senior Scientist, TU for his continuous guidance and support throughout the research.
He always helped us see a bigger picture, and improved the potential in all the team members.
Without his concern and care, we could not be able to complete the dissertation. We are highly
indebted towards him for this immense support and are very much thankful to him.

We would like to acknowledge Prof. Dr. Rameshwor Adhikari, Research Centre for Applied
Science and Technology (RECAST) for providing us with the facilities of internet and study
desks.

We would like to appreciate and express our special gratitude to our co-supervisor Mr. Samiran
Subedi and mentor Mr. Aashish Pokharel for their consistent mentoring and guidance throughout
our research.

Likewise, we would like to thank our internal supervisor Mr. Keshab Raj Budha for his
irresistible support and encouragement to us to keep moving during the research.

We would like to express our gratitude to Er. Dilip Bhattarai, Vice-principal and Co-ordinator,
SANN International College for the consistent motivation and support.

We are also grateful to our batch mates of 14th batch, B.Sc. Biotechnology for their crucial
support and co-operation and to our parents who always motivated us to pursue our goals. We
are hugely indebted for all the love and support they have showered on us. Also, we would like
to express a special gratitude to SANN International College family.

At the end, we would like to thank to all those who directly or indirectly supported us throughout
the thesis work.

3
 LIST OF ABBREVIATIONS
ADME Absorption, Disposition, Metabolism and Excretion
APC Adenomatous Polyposis Coli
BLAST Basic Local Alignment Search Tool
BMP4 Bone Morphogenetic Protein 4
CADD Computer Aided Drug Discovery
CaMKII Ca2+/calmodulin-dependent protein kinase II
cAMP Cyclic Adenosine Monophosphate
CK1 Casein Kinase 1
CREB cAMP response element-binding protein
DAAM1 DVL-associated activator of morphogenesis 1
DEP Dishevelled, Egl-10, Pleckstrin
DFT Density Functional Computations
DIX Dishevelled and Axin
DVL Dishevelled
ERK Extracellular signal-regulated kinase
FDA Food And Drug Administration
FGF-R1 Fibroblast Growth Factor Receptor–1
FSH Follicle Stimulating Hormone
FXI Factor XI
FZD Frizzled
GDP Guanosine Diphosphate
GnRH Gonadotropin releasing Hormone
GSK3 Glycogen Synthase Kinase 3
GUI Graphical User Interface
hCG Human Chorionic Gonadotropin
HOMO Highest Occupied Molecular Orbital

4
HTS High Throughput Screening
LBDD Ligand Based Drug Design
LH Luteinizing Hormone
LOMO Lowest Occupied Molecular Orbital
LRP 5/6 Low density lipoprotein receptor related protein 5/6
MAP Mitogen Activated Protein Kinase
MAT Methionine Adenosyltransferase
MEP Molecular Electrostatic Potential Maps
mTOR Mammalian Target of Rapamycin
NMR Nuclear Magnetic Resonance
NSAIDs Nonsteroidal Anti-Inflammatory Drugs
OSIRIS Open-Source Independent Review and Interpretation System
PAF Platelet Activating factor
PCP Planar Cell Polarity
PDB Protein Data Bank
PDBQT Protein Data Bank, Partial Charge (Q), And Atom Type (T)
PDZ Post synaptic density-95/Discs large/Zonula- occludens-1
PGE2 Prostaglandin E2
PKA Protein Kinase A
PKC Protein kinase C
PLC Phospholipase C
PP2A Protein Phosphatase 2A
RCSB Research Collaboratory For Structural Bioinformatics
Rhob Ras Homolog Family Member
RTV Retinovir
SAM S-adenosylmethionine
SAR Structure-Activity Relationship

5
SAS Stattistical Analysis System
SBDD Structure Based Drug Design
SF-1 Steroidogenic Factor-1
TCF Transcription Factor
TNF Tumor Necrosis Factor
TPSA Topological polar surface area
UCSF University of California, San Francisco’s department of pharmaceutical chemistry
UFF Universal Force Field
UORSY UkrOrgSntez. Ltd.
VS Virtual Screening
vWD von Willebrand Disease
vWF von Willebrand Factor

6
 ABSTRACT
Heavy bleeding is caused by hormonal imbalances caused by polycystic ovarian syndrome
[PCOS], obesity, insulin resistance, or thyroid disorders. Excessive growth of endometrium
occurs when there is a hormonal imbalance, and it is finally lost through heavy menstrual
bleeding. It is quite difficult to maintain one’s usual schedule during their period due to the
significant bleeding and cramps induced by this illness. It might lead to iron insufficiency if left
untreated. According to Johns Hopkins Medicine, the treatment for menorrhagia is dependent on
the reason of the abnormally high bleeding. Progesterone, birth control pills, nonsteroidal anti-
inflammatory drugs, iron supplements for anemia, or, in severe situations, surgical treatments
such as ablation, resection, or even a hysterectomy may be recommended by doctors. All these
treatments have their specific side-effects. Computer Aided Drug Design (CADD) could
ameliorate the development of novel disease-containment medication candidates. Use of
optimized protocol for protein identification and improved practice in selection of drug
compounds including their metabolic profiling helps in achieving overall good data coverage and
quality with reduced data noise in biological network of disease. In this paper, aromatase is our
prioritized protein that acts as a rate limiting enzyme in steroidogenesis. So, as to maintain its
expression and regulation for the control of the production of estrogen, we searched for the
compounds (natural products) from ZINC 15. The ligands were then filtered for ADME/TOX
properties and then further filtered with hMAT1A. In addition to it, the ligands underwent
filtration with CYP3A4, drug metabolizing enzymes along with other Phase I drug metabolizing
enzymes. Then the protein-ligand interaction of top hits with the aromatase were observed.
Using protein-protein interaction multiple target drug will be identified. Further DFT analysis
were performed to favor the molecular docking results.

7
Table of Contents
DECLARATION..................................................................................................................................2
1 CHAPTER ONE: INTRODUCTION...................................................................................10
1.1 Background....................................................................................................................10
1.2 Hypothesis......................................................................................................................10
1.3 Objectives.......................................................................................................................11
 To identify possible protein targets........................................................................................11
 To make ligand library...........................................................................................................11
 To perform structure based virtual screening.........................................................................11
 To study protein ligand interaction........................................................................................11
 To choose probable drug candidates......................................................................................11
 To perform DFT analysis.......................................................................................................11
2 CHAPTER – TWO LITERATURE REVIEW...............................................................12
2.1 Normal menstrual cycle....................................................................................................12
2.1.1 Menorrhagia....................................................................................................................13
2.1.3 Frequency of Menorrhagia............................................................................................14
2.1.4 Menorrhagia in Nepal:...................................................................................................15
2.2 von Willebrand Factor for Menorrhagia........................................................................15
2.2.1 vWD Cases.......................................................................................................................16
2.3 Role of aromatase in endometrial disease.......................................................................18
2.4 PP2A....................................................................................................................................20
2.4.1 Cholesterol biosynthesis.................................................................................................20
2.4.2 Identification and characterization of a novel Dvl-binding protein that suppresses
Wnt signalling pathway...........................................................................................................21
2.4.5 Nuclear Dishevelled: An enigmatic role in governing cell fate and Wnt signaling. .21
2.5 Drug Discovery...................................................................................................................22
2.6 Defining target protein (or preparation of drug target)................................................29
2.7 Binding site identification.................................................................................................29
2.8 Compound database selection and library preparation.................................................30
2.9 Filters..................................................................................................................................30
2.10 Docking and scoring........................................................................................................31
3 CHAPTER THREE: MATERIALS AND METHODOLOGY.......................................38

8
 3.1 Protein or target structure retrievation......................................................................38
3.2 Protein and ligand preparation:..................................................................................38
3.3 Preparation of ligand database:...................................................................................38
3.3.1 Ligand library:................................................................................................................39
3.4 ADME-TOX filter.........................................................................................................39
3.5 hMAT1A filter for hepatotoxicity:..............................................................................40
3.6 Initial virtual screening:...............................................................................................40
3.7 CYP3A4 filter for drug metabolism............................................................................40
3.7.1 Initial virtual screening:.........................................................................................................40
3.8 Binding site prediction..................................................................................................40
3.9 Final Structure based Virtual Screening....................................................................41
3.10 Preparation of final ligand dataset..........................................................................41
3.11 Analysis of docking results.......................................................................................41
4 CHAPTER FOUR: RESULTS............................................................................................44
4.1 Search for reference and target proteins....................................................................44
4.2 Protein Structure Preparation.....................................................................................44
4.3 Ligand Library Preparation........................................................................................46
4.4 Binding Site Prediction.................................................................................................49
4.5 Molecular Docking and Analysis.................................................................................51
4.6 Computational analysis (DFT).....................................................................................56
4.7 Frontier molecular orbital study and chemical reactivity descriptor analysis.......56

9
 1 CHAPTER ONE: INTRODUCTION

1.1 Background

Menorrhagia is one of the most prevalent gynecologic problems in modern gynecology, defined
as menstruation at regular cycle intervals but with excessive flow and duration. Menorrhagia is
defined as a total blood loss of more than 80 ml every cycle or a menstrual period lasting more
than 7 days. According to the World Health Organization, 18 million women aged 30-55 years
think their menstrual bleeding is excessive. According to reports, only 10% of these women have
blood loss significant enough to induce anemia or be clinically characterized as menorrhagia
(Shaw, 2018a). Menstrual blood loss is difficult to measure in practice. As a result, the diagnosis
is frequently made based on the patient’s medical history (Menorrhagia | Johns Hopkins
Medicine, n.d.). A typical menstrual cycle lasts 21-35 days, with bleeding lasting 7 days on
average and a flow of 25-80 ml.

Clinically, menorrhagia must be separated from other gynecologic disorders. Metrorrhagia,

which is infrequent and excessive flow, menometrorrhagia, which includes frequent and
excessive flow, polymenorrhea that is bleeding at intervals of less than 21 days, and
dysfunctional uterine bleeding are all examples of abnormal uterine bleeding without any
obvious structural or systemic abnormality) (Shaw, 2018a).

Heavy menstrual bleeding is why over 30% of all hysterectomies are performed in the United
States. Historically, the mainstay of treatment for menorrhagia has been definitive surgical
repair. Modern gynecology has shifted toward conservative treatment, owing to financial
concerns and the desire of many women to keep their uterus (Shaw, 2018a).

1.2 Hypothesis

1.2.1 Null hypothesis

Screened compounds showed no significant reaction against the prioritized proteins
addressing, hormonal disorders in female.

10
1.2.2 Alternate hypothesis
Screened compounds showed significant reaction against the prioritized proteins
addressing, hormonal disorders in female.

1.3 Objectives

1.3.1 General objective:

 To develop lead compounds, capable of regulating hormonal imbalance in females, at
threshold.

1.3.2 Specific objectives:

 To identify possible protein targets

 To make ligand library.

 To perform structure based virtual screening

 To study protein ligand interaction

 To choose probable drug candidates

 To perform DFT analysis

11
2 CHAPTER – TWO LITERATURE REVIEW

2.1 Normal menstrual cycle

A coordinated serial event that repeats every month at regular intervals in which the participation
of hypothalamus takes place along with the secretion of GnRH, secretion of follicle stimulating
hormone (FSH) and luteinizing hormone(LH) by the pituitary gland, ovary that responds to those
hormones by recruitment of dominant follicle and secretion of estradiol and inhibin A. The
proliferation of endometrium and production of cervix mucus is stimulated by estradiol. The
discharge of LH accountable for ovulation is triggered by peak of estradiol. This LH is also
responsible for posterior secretion of progesterone by the corpus luteum, that in turn,
involutionates 14 days later, if the stimulation of hCG (pregnancy)is not received. The duration
of normal menstrual cycle is 28 ± 7 days, that accepts the ±2 days fluctuation in the same
woman, as a normal pattern that is termed as a regular cycle. (1)

Source of figure: https://www.britannica.com/science/menstrual-cycle

12
2.1.1 Menorrhagia

Menorrhagia is a heavy menstrual period lasting more than 7 days or menstrual blood loss of
more than 80 ml throughout three successive menses. The average menstrual loss is 35 milliliters
every period, with the 90th percentile at 80 milliliters per period. If menstrual blood loss is
greater than 80 ml, it is termed excessive, and this loss is linked to the development of iron
deficiency anemia (Rees, 1987). It could be caused by systemic, local, or iatrogenic problems.
Endocrine disease (hypothyroidism, anovulation) and coagulation abnormalities (Haemophilia
carriage, Von Willebrand’s disease) are examples of systemic conditions. Endometrial
hyperplasia, endometriosis, pelvic inflammatory disease, benign leiomyomas, polyps, and
malignant (endometrial, cervical) tumors are all common disorders in the area. There is no
clinical etiology in more than half of the women who have menorrhagia. Abnormal endometrial
concentrations of prostaglandins and plasminogen activator have been linked to the excessive
menstrual loss (Rees, 1987).

2.1.2 Contrasting normal Menstruation with Menorrhagia

The reduction in estradiol and progesterone concentrations after luteolysis triggers regular
menstruation. Endometrial tissue breakdown and bleeding leads it and hemostasis, healing and
angiogenesis follow. Local factors play a key role in these processes (Pawitan, 2001). Due to an
immature hypothalamo-pituitary-ovarian axis, menstrual difficulties are prevalent during
adolescence and can linger for several years following menarch (Palep-Singh & Prentice, 2007).

Figure 1 Mechanism of normal menstruation (Pawitan, 2001)

13
Variables like: Endometrial bleeding associated factor (ebaf), migratory leucocytes,
macrophages & mast cells, lysosomal enzymes, prostaglandins & edothelins, growth factor
receptor, fibrinolysis and endometrial blood flow changes plays an important role in
menorrhagia.
Abnormal endometrial bleeding is correlated with increased ebaf gene expression. The copper
intrauterine device has been linked to an excessive leucocyte infiltration to cause menorrhagia.
The vasodilators: Platelet activating factor (PAF) and prostaglandin E may be released by
macrophages and also vasoactive substances released by mast cells might increase menstrual
blood loss. The heparin secreted by mast cells stimulates endometrial fibrinolysis. The vascular
fluid loss is increased through the gaps created between vascular endothelial cells due to
histamine. Increased menstrual bleeding might be caused by excessive tissue breakdown by
lysosomal enzymes. There is increase in prostaglandin receptor concentrations in women who
have menorrhagia. As a consequence, elevated vasodilatory prostaglandins may have an impact
on endometrial tissue. Interruptions in endothelin production pathway might cause menstrual
blood loss to be delayed or increased. Fibroblast Growth Factor Receptor–1(FGF-R1) is required
for endometrial maturation and normal endometrium regeneration after menstruation. The
hemostatic system can become unbalanced if the fibrinolytic process is overactive, resulting in
excessive blood loss (Pawitan, 2001).

2.1.3 Frequency of Menorrhagia

5% of women between age group 30-49 complain about menorrhagia each year in the UK. The
prostaglandin produced within the endometrium may be disordered to cause Idiopathic ovulatory
menorrhagia. In 10% of normal patients, Fibroids are found, whereas in 40% of severe patients
Fibroids are found. Even though 12% of gynecology problem is related to this disease
Menorrhagia, the specific cause is yet to be identified. Women affected with von-Willebrand’s
disease or impaired factor XI and carrying classic hemophilia gene are found to be menorrhagic
(Kadir et al., 1998). .
Kadir et al. carried out investigation of menorrhagia in 150 women having Menstrual Blood Loss
(MBL) of more than 80 ml. For all patients, the quantification of partial thromboplastin time,
factor VIII activity, von-Willebrand factor antigen & activity and factor XI were performed. In

14
17% of patients, inherited bleeding disorder was observed where, 13% patients had recurrence of
von-Willebrand’s disease and 4% patients had recurrence of FXI deficiency.

2.1.4 Menorrhagia in Nepal:

The observational study done in Nobel Medical College, Biratnagar, Nepal from June 2012 to
June 2015 found that, endometrial bleeding was associated with age, with the greatest prevalence
in peri-menopausal age group (Nepal et al., 2016).

The study conducted by Shrestha DN et al. over 427 females of reproductive age group (15-49
years) showed that 23.4% was menstrual abnormalities complaint, among which 21.7%
complaint was of menorrhagia (Shrestha et al., 2017).

2.2 von Willebrand Factor for Menorrhagia

The most prevalent congenital bleeding problem is von Willebrand disease (vWD). Menorrhagia
is the most prevalent symptom of bleeding in women, and it is disabling, resulting in iron
deficiency anemia, expensive healthcare costs, and a poor quality of life (Palep-Singh &
Prentice, 2007). von Willebrand disease (vWD) is caused by a lack of or deficiency in von
Willebrand factor (vWF), a glycoprotein that increases platelet adherence to vessel walls
following vascular damage. vWF is a carrier protein for factor VIII and plays an important role
in platelet adhesion and platelet plug formation during primary hemostasis. Mucosal bleeding
occurs when vWF is deficient in the oropharyngeal, gastrointestinal, and genitourinary tracts
(Ragni et al., 2016).

von Willebrand factor (vWF), which is usually provided when other treatments fail, has a
scarcity of evidence. The lack of effective menorrhagia therapy is still the most unmet medical
need among women with vWD (Ragni et al., 2016).

Twenty (24.1%) of the 83 questionnaires sent to HTC MDs yielded sufficient data for analysis.
816 (61.8 percent) of the 1321 women with vWD examined between 2011 and 2014 experienced
menorrhagia, for which combination oral contraceptives, tranexamic acid, and desmopressin
were the most prevalent first-line medications, although vWF was recorded as third-line therapy

15
in 13 women (1.6 percent). vWF safely reduced menorrhagia in 101 women at a dose of 33–100
IU kg1 on days 1–6 of the menstrual cycle, based on data from 88 women from six published
studies (Ragni et al., 2016).

2.2.1 vWD Cases

In a survey conducted by Ragni MV et al., 1321 women of age group 18 -45 with vWD seen
during 2011-2014, 816 (61.8%) had menorrhagia (Ragni et al., 2016). Approximately 10.7% of
women with menorrhagia will have vWD as the reason of their menorrhagia (El-Hemaidi et al.,
2007).

Many variables impact the amount of vWF in plasma, including age, blood type, hormones, and
smoking (Edlund et al., 1996). vWD affects people of all ethnicities. There would be about 100
clinically symptomatic patients per million people, according to estimates. According to this
estimate, there are roughly 620,000 persons worldwide who are substantially impacted by vWD.
The majority of these 496,000 people would live in underdeveloped nations (Srivastava &
Rodeghiero, 2005).

Only 6% of 1180 individuals with hereditary coagulation problem had vWD, while 90% had
hemophilia, according to a recent Chinese study. In Singapore, only 8% of patients had severe
disease; in Oman, 32% had severe disease; in Iran, 50% had severe disease; and in India, 52%
had severe disease (Srivastava & Rodeghiero, 2005).

2.2.2 Effects of Menorrhagia on Women’s Quality of Life

Due to high menstrual flow, daily activities are limited. Women can’t do your typical activities
during their period if they have menorrhagia, since they are losing so much blood and cramping.
Soaking through one or more sanitary pads or tampons for several hours at a time. To limit
menstrual flow, women need to utilize two types of sanitary protection. Having to change
sanitary protection in the middle of the night. Bleeding that lasts more than a week. Passing
blood clots that are more than a quarter of an inch in diameter. Anemia symptoms include
weariness, exhaustion, and shortness of breath. When comparing the women in the menorrhagia
group to the women in the control group, the researchers discovered that menorrhagia group
members were significantly affected in all subscales of the scale model SF-36 (physical

16
functioning, physical role functioning, pain, general health, vitality, social role functioning,
emotional role functioning, and mental health) (Gokyildiz et al., 2013).

2.2.3 Treatment of Menorrhagia

Medical or surgical treatment options are available. If there are no evident pelvic abnormalities
and the woman wants to keep her fertility, medical therapy is recommended. When dealing with
pelvic inflammatory disease or endometriotic masses, surgery may be required. Women who
want to keep their fertility should have their operations done as conservatively as possible. When
medicinal treatment has failed, surgical surgery may be required (Rees, 1987).

2.2.4 Medical treatment

Antifibrinolytic medicines and prostaglandin synthatase inhibitors, such as mefenamic acid, are
both given during menstruation to minimize menstrual blood loss. Progestogens, such as
norethisterone, are also used to lower menstrual bleeding during the luteal phase, however, the
impact has not been proven using blood loss measurements. Oral contraceptives are especially
helpful for women under the age of 35 who also need contraception. Clomiphene has been used
to trigger ovulation in cases of menorrhagia caused by anovulation, however, there is no
objective evidence that it reduces monthly blood loss. It’s probably a good idea to save
clomiphene for people who want to get pregnant. Using danazol on a regular basis lowers
menstrual blood loss and can cause amenorrhea. It is important to remember that menorrhagia
can lead to iron deficiency anemia, which needs treatment (Rees, 1987).

2.2.5 Surgical treatment

Removal of cervical or endometrial polyps, curettage, myomectomy, and finally hysterectomy is

among surgical options. In specialized centers, endometrial polyps and sub-mucous fibroids can
be removed under hysteroscopic management. Menstrual blood flow is reduced after dilatation
and curettage, although there appears to be no long-term influence on measured menstrual loss.
It’s better to think of it as a diagnostic test rather than a treatment. Because older women can
utilize prostaglandin inhibitors or anti-fibrinolytic medicines until menopause instantiates,

17
hysterectomy is more commonly administered to younger women with complete families (Rees,
1987).

2.3 Role of aromatase in endometrial disease

Aromatase is the master enzyme for the biosynthesis of estrogen, expressed in human skin,
ovary, brain, and adipose tissue. However, its activity is not distinguishable in normal
endometrium. In contrast, Aromatase is expressed abnormally in endometriosis and is stimulated
by PGE2 which results in the local production of estrogen which further induces PGE2
formation. It establishes a positive feedback cycle (10).

PGE2 - Prostaglandin E2

As a result of FSH binding to its G-protein-coupled receptor in the granulosa cell membrane,
intracellular cAMP levels increase, which facilitates the binding of two crucial transcription
factors, steroidogenic factor-1 (SF-1) and cAMP response element binding protein (CREB), to
the classically placed proximal promoter II of the aromatase gene. The pre-ovulatory follicle's
production of estrogen is subsequently triggered, increasing aromatase expression (10).

2.3.1 Regulation of Aromatase Expression in Estrogen- Responsive Breast and Uterine

Disease: From Bench to Treatment

Aromatase inhibitor therapy causes uterine leiomyomas to decrease and Estrogen may boost
proliferating leiomyoma cells directly or indirectly by improving progesterone action in human
leiomyomata (11).

An aromatase inhibitor was added to stop the androstenedione's stimulatory effects. According to
these findings, leiomyoma smooth muscle cells locally converts androstenedione into
biologically relevant amounts of estrone before converting it into estradiol (11).

18
Despite significantly low amounts of circulating estrogen in perimenopausal or postmenopausal
women, aromatase expression may play a significant role in the availability of estrogen to
leiomyoma tissue (11).

2.3.1.1 The evolving role of aromatase inhibitors in breast cancer

Anti-aromatase medications prevent the final stage of estrogen production in peripheral tissues
by inhibiting the cytochrome P-450 component of the aromatase enzyme complex. Type I
inhibitors are also known as aromatase inactivators because they irreversibly inactivate the
enzyme by blocking the substrate-binding site and have a steroidal structure comparable to
androgens (12).

2.3.1.2 Aromatase expression in health and disease

Aromatase belongs to the family 19 of the superfamily P450 which is responsible for converting
C19 androgenic steroids to corresponding estrogens, which is known as Aromatization as it
involves the conversion of the androgen delta 4-3-one A-ring to phenolic A-ring characteristics
of estrogens (13).

2.3.1.3 Regulation of aromatase expression: potential therapeutic insight into breast cancer
treatment.

Estrogenic steroid hormones play a vital role in many key physiological process including
growth, ossification, differentiation, neurologic and reproductive development hence, the
enzyme is expressed in diverse tissues; including skin; endometrium, bone, brain, ovary; ovary,
testis, placenta, mammary and dynamic and is dependent on the reproductive age. (Simpson)

Promoter II aromatase activity is regulated by cyclic AMP (cAMP) response element binding
protein (CREB) which binds DNA response elements like sequences CRE1 and CRE2 identified
on this promoter by mutational analyses in breast adipose tissues. This interaction has been
reportedly enhanced by the presence of cAMP; and PGE2.

19
However, in the ovary in addition to CREB; there is the requirement of an orphan member of the
nuclear receptor subfamily, the steroidogenic factor-1 (SF-1) for the basal transcriptional
regulation of aromatase promoter-II activity; which has been shown to be mediated by follicle
stimulating hormone (FSH) in the presence of the B-catenin. This implicates the involvement of
the Wnt pathway in the activation of steroidogenic signaling.

2.4 PP2A

PP2A is a crucial component of numerous biological processes. It controls the activity of the
enzymes involved in glycolysis, lipid metabolism, and catecholamine production and manage
cell metabolism (1). It controls practically all-important cell cycle checkpoints and pathways,
including Wnt, mTOR, and MAP Kinase by forming structurally unique families of holoenzymes
that are controlled geographically and temporally by particular regulators (2).

PP2A plays the part of a "gatekeeper," as its activation may control and restrain the expression
and activity of the oncogene. Inhibition of PP2A, on the other hand, causes an increase in
oncogene activity and sets off a chain of events that may promote the onset and progression of
the disease by enhancing the survival, proliferation, and self-renewal, impairing differentiation,
and raising genomic instability (3).

PP2A dysregulation can cause a decrease in cell apoptosis and an increase in cell proliferation
(4).

Other than the known interaction partners (Axin and CaMK IV), PP2Ac-specific interacting
proteins, such as tuberous sclerosis complex 2 (TSC2), R-Ras, Nm23H2, and RhoB, were
validated by pull-down tests in the presence of the Wnt3a as a ligand, a stimulus that can activate
the growth regulating signaling cascade of Wnt/beta-catenin and ERK (1).

The Wnt pathway controls both embryonic development and cell polarity as well as proliferation.
It triggers the transcription of vital cell differentiation promoters like cyclin Di and c-Myc. A
complex made up of axin, adenomatous polyposis coli, glycogen synthase kinase 3b (GSK3b),
and casein kinase 1 acts to break down the protein b-catenin in the absence of Wnt signaling.
Unique domains in APC and axin bind to CK1 and GSK3b and act as scaffolds to promote b-
catenin phosphorylation; these scaffolds are frequently mutated in malignancies (2).

An intracellular complex made up of the receptors disheveled, axin, CK1, and GSK3b is
produced when extracellular Wnt binds to the receptor frizzled and co-receptor LRP 5/6. This
prevents the phosphorylation and subsequent destruction of b-catenin (2)

20
2.4.1 Cholesterol biosynthesis

The cholesterol side-chain cleavage (P450scc) and aromatase (P450arom) enzymes are essential
regulators of progesterone and estradiol production in granulosa cells. FSH
increases progesterone production and P450 cc mRNA expression very slightly while maximum
synthesis occurs in the presence of LH.  Through differentiating cAMP production, FSH and LH
control the synthesis of estradiol and progesterone. When cAMP interacts with cAMP-dependent
protein kinase (PKA), its effects are transduced. Two cAMP molecules must bind to each
regulatory subunit in order to activate it. This causes a reversible conformational shift that frees
the active catalytic subunits. By producing various amounts of the second messenger cAMP,
FSH and LH have distinct impacts on the expression of the P450scc and P450arom genes (5).

It has been discovered that the cholesterol metabolite 27-hydroxycholesterol (27-OHC) functions
as a selective estrogen receptor modulator and fosters the development of estrogen receptor-
positive breast cancer. The endogenous cholesterol production pathway was upregulated as a
result of epigenetic alterations in cells that developed resistance to treatment with aromatase
inhibitors. Using statin medications to prevent cholesterol manufacturing could drastically
diminish estrogen receptor activation and cell invasion (6).

Malignant transformation and tumor development have been linked to dysregulation of

cholesterol metabolism via the mevalonate pathway. Breast cancer resistance to aromatase
inhibitors has been linked to cholesterol biosynthesis (7)

2.4.2 Identification and characterization of a novel Dvl-binding protein that suppresses

Wnt signalling pathway

DVL is a protein that is known for transmitting Wnt signal to canonical B-catenin as well as non-
Canonical PCP Pathways (8). It has three genes DVL 1-3 which have been isolated in mammals
but the differences in their regulation and relative function are yet to be fully understood (9).
DIX domain, PDZ domain, and DEP domain are highly conserved domains present in all DVL
family members (8).

To activate the Wnt/-catenin signaling, the DVL DEP domain and C terminal region (DEP-C)
must bind to the three-segmented discontinuous motifs in FZD, and the FZD C terminal tail (9).

21
2.4.5 Nuclear Dishevelled: An enigmatic role in governing cell fate and Wnt signaling

DVL is known to be a multifunctional protein and works as an intercessor of Wnt signaling.

DVL's role in the development and cytosolic function is clarified and well respected. However,
its nuclear role is still a mystery. The discovery of its translocation to the nucleus re-evaluates
how DVL is observed functionally. Likewise, further studies disclosed that multiple DVL
paralogs translocate to the nucleus, bind the novel genomic loci as well as regulate the
expression of genes that are not previously associated with Wnt signaling (9).

The phosphorylated tail of LRP5/6 interacts with DVL and Axin which leads to GSK3
inhibition. This rupture of signaling leads to the collection of unphosphorylated β-catenin which
then translocates to the nucleus and acts as a transcriptional co-activator alongside binding to the
LEF/TCF transcription factors at Wnt target genes (9).

DVL paralogs are also used to transmit non-canonical signals. DVL interacts with DVL-
associated activator of morphogenesis 1 (DAAM1), which communicates via Rho family small
GTPases, in response to the proper upstream stimuli. Additionally, DVL can interact with and
activate the exchange factor Tiam1, which changes the inactive GDP-bound version of Rac1 into
the active GTP-bound form (9).

The Wnt/Ca2+ pathway is involved in non-canonical signaling. The Wnt ligands interact with
heterotrimeric G-proteins through FZD, which is involved in this signaling. Similar to this, DVL
controls this signaling and arrange the activation of phospholipase C (PLC), which raises Ca2+
levels inside cells. Proteins that control cell adhesion and migration, including PKC (protein
kinase C), calcineurin, and CaMKII (Ca2+/calmodulin-dependent protein kinase II), are activated
in this way of signaling (9).

Nuclear DVL binds to several tissue-specific promoters of the CYP19A1 gene, which codes for
aromatase, an enzyme that catalyzes the conversion of androgens to estrogens. (9).

DVL1 and DVL3 perform a promoter-specific and cell type dependent role in activating or
inhibiting CYP19A1 transcription (9).

2.5 Drug Discovery

2.5.1 Discovery and resupply of pharmacologically active plant-derived natural products

Modern drugs are acquired from natural products and their derivatives as medicinal plants are
regarded as an important source of therapeutic agents. Natural products are obtained from living
organisms, they possess properties that are advanced for serving various biological functions.
(e.g., binding to specific target proteins or other biomolecules.
Sometimes therapeutically effective compounds in humans as well as documented
ethnopharmacological information about the traditional use is known to be the biggest
advantage, especially in the context of drug development.

22
Natural Products for Drug Discovery in the 21st Century: Innovations for Novel Drug
Discovery

Fig: Innovative technologies for natural product drug discovery. Application of these
technologies can potentially lead to novel drug candidates from natural products.

Computational drug design use and target prediction are now substantial and will carry on
shortly to guide drug development. However, only those previously studied proteins or targets
can be anticipated. To understand the molecular basis of natural product therapeutic qualities and
to predict potential derivatives that may improve activity, computer-based quantitative structure-
activity techniques can be used. The advantage of Computer-based drug design is that it guides
the optimization of lead compounds and decides whether to increase their affinity or
pharmacokinetic and pharmacodynamic properties.

23
2.5.2 Computer Aided Drug Design (CADD)

In combining with wet-lab techniques that elucidate the mechanism of drug resistance, one of the
areas that could support identification of novel antibiotics, the computer-aided drug design
(CADD) could be time and cost-efficient tools to search new antibiotic targets and to design vital
molecules for both known and new protein targets (Schneider & Fechner, 2005). To facilitate the
drug design process, CADD could additionally be used for interpretation at an atomic level the
structure-activity relationship (SAR) of particular molecule(s) for computational modeling in
predicting biological activity or toxicity of unknown or understudied new medical entities
(NMEs) on the basis of chemical structural similarities with other similar chemicals/drugs for
which such data exists. Thus, in drug discovery process, SAR is prominent in numerous aspects,
ranging from primary screening to lead optimization (Kandel., 2021) based on the interaction of
the respective pharmaceutical agents (ligands) screened. Mainly, CADD uses two approaches,
Structure Based Drug Design (SBDD) and Ligand Based Drug Design (LBDD) (Zappa et al.,
2010).

All the information from the CADD methods is then used to design compounds that are
subjected to chemical synthesis and biological assays. With the information from these
experiments used to further develop the SAR, yielding further improvements in the compounds
with respect to activity as well as absorption, disposition, metabolism and excretion (ADME)
considerations (Ekins, S., Boulanger, B., Swaan, P. W., & Hupcey, M. A., 2002). Computer
aided drug discovery also called in silico drug design and is very popular in recent years. Due to
change in life style of food habits and environmental issue cause outbreak of deadliest diseases
time to time. Pharmaceuticals and health fields require novel antibiotics all the time because

24
there is growing emergence of resistance against various viruses and bacteria. The idea of the
CADD was first accepted in the 20 th century by the scientific society since then the use of
combinatorial chemistry and merging with several bioinformatics tools has been the major part
of drug development process now (Talambedu et. al., 2017).

It is very complex, time consuming and costs billions for the development of drugs and the
chances of failure are always there from beginning to preclinical and clinical phases. It takes
roughly 10 to 16 years to introduce a new drug into the pharmaceutical industry, and there are
multiple steps that is needed to make the process successful which includes the identification of
the target, validation, lead identification, synthesis, candidate optimization, preclinical studies
and research, clinical trials (phase I, II, and III), evaluation, and getting FDA I and phase IV
studies approval (Paul et. al.,2010).

Computational approach has thus been part of interdisciplinary drug discovery approach and
considered more sophisticated. It is less laborious, cost-effective and less time consuming. Now
a days pharmaceutical world has been widely using computational tools for the design and
discovery of therapeutic products for different life threatening diseases (Talambedu et. al.,2017,
Sliwoski et. al.,2014). High throughput screening of molecules is one computational strategy that
is used to prioritize and identify small molecules for screening (Leelananda and Lindert,2016).
Further toxicity testing can be done in optimized conditional while saving cost and time,
minimizing experimental testing, optimization of overall resources including the large number of
animals used for clinical trials (Gillespie and McHugh, 1997). High throughput screening (HTS)
has been used for very long time and is a traditional approach of discovering novel therapeutics.
Virtual screening (VS) has emerged as a powerful complementary tool of CADD to HTS which
is less expensive and gives hit compound within less time. The number of ligands can be
significantly reduced making the screening procedure much convenient. New compounds or
ligands can be discovered based on biological structure using computational tools by virtual
screening methods (Shoichet et. al.,2004).

CADD provides a virtual platform in the journey of drug discovery process. The lead compound
is identified for testing, its effectiveness, side effects and bioavailability is predicted and is thus a
virtual shortcut in lengthy procedure of traditional drug discovery pathway. CADD has played a

25
major role in the discovery of many pharmaceutical drugs available that have obtained FDA
approval (Talele et. al., 2010; Clark 2006; Kitchen et.al., 2004). The carbonic anhydrase inhibitor
dorzolamide (1995) (Talele et. al., 2010;Clark 2006) angiotensin-converting enzyme (ACE)
inhibitor captopril (1981), a fibrinogen antagonist tirofiban (1998) and therapeutic agents for the
treatment of human immunodeficiency virus (HIV): saquinavir (1995), ritonavir (1996) and
indinavir (1996)are some successful drug discovered and approved novel therapeutics developed
by using the computational tools of CADD (Talele et. al., 2010

Fig: Schematic representation of drug discovery using Computer Aided

DrugDiscovery(Leelananda and Lindert,2016).

Computer Aided Drug Discovery (CADD) can be broadly studied into two categories; structure
based and ligand based drug discovery approaches.

26
2.5.3 Ligand Based drug discovery (LBDD)

LBDD is the approach of drug discovery which is used when the 3D structure of the target is not
available. The 3D structure of the receptor molecule or protein is not present and the ligands
binding to it act as pivotal in the drug discovery procedure. In ligand based drug designing path
the 3D quantitative structure activity relationships (3DQSAR) and pharmacophore modeling are
most widely used tools (Aparoy et. al., 2012). QSAR is the method which correlates molecular
structure with properties like invitro or in vivo biological activity. This method can also be
applied to toxicological data and it is called quantitative structure toxicological relationship
(QSTR). QSAR deals with the physiochemical properties of the compounds (Winkler, 2002).
The physical, chemical and biological properties of a molecule depend upon the geometric, steric
and electronic properties that are responsible for the structure of any molecule. QSAR is based
on this assumption (Eleni et. al., 2003) and it analyzes the set of similar compounds and proceeds
by correlating works by correlating the structural molecular properties of a particular molecule
with functions (i.e. physicochemical properties, biological activities, toxicity, etc.) (Merz
et.al.,2010)

2.5.4 Structure Based drug discovery (SBDD)

The three dimensional structure of the target protein related to the disease is known in structure
based approach of drug discovery. It is the most commonly used approach of CADD and has
been able to play important role in the discovery of some FDA approved drugs available in
market. SBDD methods study and analyze the information related to the structure of 3-
dimensional macromolecular target. This may include proteins or RNA, the key sites within the
macromolecule are identified and interactions that are important for their respective biological
functions are studied in details. This information is very important and plays key role in the
structure based drug discovery (Yu et. al.,2017).

Structure based drug discovery uses two methods in the procedure of drug discovery. They are
molecular docking and de novo designing of ligands. De novo designed ligands can be inhibitors,
antagonists, agonists. The X-ray or NMR structures of such molecules are used and these can be

27
retrieved from different structural databases such as PDB
(https://www.rcsb.org/pdb/home/home.do ) and others. Different online tools and software
available are utilized to generate the experimental structure through homology modeling if the
structure is not found in PDB. Different strategies are utilized to gain the required structure by
homology modeling. Some of them are very crucial and needs to be taken care of for better
results. General plan for generation of protein model includes selection of a suitable template;
align the target sequence to the template and finally building a model using the 3D co-ordinates
of each atom of the template (Sonnhammer et. al., 1998).

SBDD is comparatively effective and advantageous in the journey of drug discovery pathway.
Hundreds of thousands of ligands can be studied, described and virtually screened as potential
drug leads. The prior purchasing or synthesizing is not required which makes it relatively rapid
in comparison to in vitro screening and is cost effective also. The large scale drug discovery
programs also get benefited and SBDD is efficient for this as well. The computational strategies
and software identify optimal binding modes of small-molecule ligands in the structure of a
target and these binding modes are then scored for their non-covalent interactions ( Moitessier et.
al., 2008). The molecular basis of a disease can be understood by analyzing and utilizing the
knowledge of the three-dimensional (3D) structure of the biological target in the process.
Similarly using different computational methods and the information from 3D structure of the
protein target, the underlying molecular interactions involved in ligand-protein binding can be
investigated which helps in the interpretation of experimental results in atomic- level detail
(Lionta et al.,2014).

28
Fig: General plan in the drug discovery pathway through structural based drug discovery (Cheng
et.al.,2012)

The screening of compounds and finalization of specific compound during the course of drug
discovery by SBDD initially focuses on the determination of protein structure or target. With the
proper defining and understanding of target protein, numerous aspects are addressed including
establishment of database library and the application of filters along with structural issues. After
this docking and post processing is carried out which has post filters and other experimental
assays to maintain accuracy (Reddy et al.,2007;Cheng et al.,2012; Lavecchia et. al.,2013)

2.6 Defining target protein (or preparation of drug target)

A well-defined drug target structure with quality 3D structure is the good starting for SBDD. The
structure which is experimentally determined through X-ray crystallography or NMR techniques
and deposited in the PDB is considered ideal for docking. Target structures are determined with
greater efficiency by the application of structural genomics. Comparative models of target
proteins are generated and have been used successfully in many docking and drug discovery
process when experimentally determined structure are not available (Becker et. al., 2006; Budzik
et. al., 2010).

29
Usually a structure present in repository like PDB contains water molecules, cofactors,
activators, ligands, and metal ions as well as several protein subunits. However, the detailed
information regarding the information on bond orders, topologies, or formal
atomic charges are generally not given. Ionization and tautomeric states might not be assigned.
Because of low resolution of a particular protein area and steric clashes may exist. Thus, the
structure preparation for target protein can address these issues (Pitt etal., 2013; Cheng et. al.,
2012; Sastry et. al., 2013). Comparative modeling is used to predict target structure on the basis
of a template with a similar sequence. Homology modeling is a specific type of comparative
modeling which uses the template and target proteins with the same evolutionary origin. With
advancement in computer technologies many web server and computer programs have been
introduced which automates the comparative modeling procedure (Silowski et. al., 2014).
MODELLER is one of the computer programs that help in the generation of target model (Marti-
Renom et.al.,2000).

2.7 Binding site identification

The possible binding sites can be predicted from the interaction of natural or non-natural ligand
with 3Dcrystal structure of target protein or such similar protein structures (Silowski et. al.,
2014). The binding of ligand within specific area is associated with increase or decrease in its
activity within a protein molecule (Leelananda andLindert,2016). This interaction determines the
drug activity. Protein-ligand interaction isa prerequisite for drug activity (Laurie and Jackson,
2006; Henrich et. al., 2010). Active site water molecules also play important roles in ligand-
target binding (Thilagavathi and Mancera, 2010). Accounting for water molecules has been quite
a challenge in docking and different studies have focused on this issue recently (Abe et. al.,
2008). Computational methods have been using geometric algorithms, methods based on
energetic consideration and methods considering dynamics of protein structures (Silowski
et.al.,2014)

2.8 Compound database selection and library preparation

The construction of compound databases is also the important step in the SBDD process. More
diverse and structured database provides the efficiency in the screening process. Ligand libraries
are constructed by enriching ligands for drug likeness or certain desirable physiochemical
properties that are suitable for the target of interest (Silowskiet. al., 2014) Some of the commonly
used databases are ZINC (with 230 million compounds available in 3D format)
(http://zinc15.docking.org/ ),PubChem (with 30million compounds available in 3D format)
(http://pubchem.ncbi.nlm.nih.gov ).Databases are freely available or available via purchase or
synthesis, and these are often drug-like small molecules with desirable characteristics such as
stability and solubility in aqueous media and possess appropriate functional groups to interact

30
with biological targets along with other important characters like absence of toxic and
undesirable moieties. Sometimes a custom-made library is to be created by the users from the
vast databases available (Lionta et. al.,2014).

2.9 Filters

Physicochemical properties of the selected drug candidates, ligands and hits, affect their
absorption, distribution, metabolism, elimination, and toxicity (ADMET) and, consequently,
their drug-like properties which is very important for effectiveness of final product (W.P et. al.,
2011). The huge library needs to be screened for probable candidates thus to facilitate the
decision-making process, and to increase the speed and probability of rapidly finding and
developing high quality compounds, a variety of multiparametric guidelines, also named rules,
have been developed, including the Lipinski'srules ('rules of five', Ro5) (Lipinski et. al., 1997),
extended Ro5 (eRo5), and beyond Ro5(bRo5) (Doak et. al., 2016).

Similarly other rules like ligand-binding thermodynamic and kinetic profiles (Pan et al.,2014),
and, more recently, the use of LE indices, also known as LE metrics have been introduced for
this purpose (Abad-Zapatero and Metz, 2005). Similarly monitoring of molecular weight (MW)
and lipophilicity during optimization also simplify drug-likeness prediction during the multipara
metric optimization process as pharmacological compound has to penetrate through various
physiological barriers, such as blood-brain barrier (BBB) gastrointestinal barrier, and
microcirculatory barrier. Similarly determining the potency and safety drug profile of
compounds in humans is also analyzed (Gardiner, 2006).The in silico modeling of ADMET
properties predict the in-vivo deposition behavior of potential drug molecules in the human body
and gathers all the ongoing kinetic processes which helps to picture and gives understanding of
how the predicted drug compound acts and gives the tentative idea of response by the human
body towards the drug molecule (Mignani et. al., 2018). Most commonly followed in the drug
discovery purpose for screening of suitable compound is Lipinski's rule of five (Lipinski, 2004).
According to this rule the orally active drug has to follow given criteria for certain physio-
chemical properties (with no more than violation of one criterion mentioned below):

Molecular weight<500 Daltons

Calculated logP<5
Hydrogen bond donors<5
Hydrogen bond acceptors <10

An extension of the Ro5, bRo5, et al. and is also considered which is based on 475complex drugs
and clinical candidates against difficult targets that are outside the Ro5space (Doak et. al., 2016).
This rule has an appropriate balance between rigidity and flexibility to bind to the considered
targets, and can be defined as follows: MW >500, cLogP 7.5,HBD >5,HBA>10,PSA >200 Å2,

31
and NRB >20. The main targets involved in the bRo5 space include G-protein-coupled receptors
(GPCRs), proteases, hydrolases, transferases, isomerases, as well compounds such as enzyme
regulators (Mignani et al.,2018).

Online tools for library design such as CLEVER (Chemical Library Editing, Visualizing and
Enumerating Resource) helps in chemical library manipulation, combinatorial chemical library
enumeration using user-specified chemical components, chemical format conversion, as well as
chemical compound analysis and filtration with respect to druglikeness, lead-likeness, and
fragment-likeness based on the physicochemical properties computed from the derived
molecules. It aIso provides an integrated property-based graphing component that visually
depicts the diversity, coverage and distribution of selected compound collections (Lionta et. al.,
2014). Knowledge of the binding kinetics between a drug and its target is also taken as a
powerful parameter of drug-likeliness filter for drug discovery with good-quality compound
selection criteria during the hit and lead optimization phase. Slow offset, slow off-rate, slow
dissociation, insurmountable antagonism, ultimate physiological inhibition, tight binding, and
non-equilibrium blockade are the main binding kinetic profiles which are considered during drug
discovery for optimization steps (Keighley, 2011).

2.10 Docking and scoring

Molecular docking predicts the protein-ligand complex structure and is followed by scoring in
virtual screening procedure in SBDD in order to rank the compounds. Docking programs utilize
different methods of conformational search and explore the ligand conformational space; these
can be categorized as follows:
a) Systematic methods, which place ligands in the predicted binding site considering all degrees
of freedom,
b) Stochastic torsional searches about rotatable bonds, such as Monte Carlo and genetic
algorithms to determine low energy conformers,
c) Molecular Dynamics simulation methods and energy minimization for exploring the energy
landscape of a molecule (Guido et.al.,2008).

Large number of docking programs have been developed and used for docking purpose.
AutoDock (Morris et. al., 2008), Dock (Ewing et. al., 2001), Gold (Jones et. al.,
1997),FlexX(Rarey et. al., 1996), Surflex (Jain, 2003), and eHiTS (Zsoldos et. al., 2003) are
some of the programs used for docking.

Auto dock uses an interaction grid to account for receptor conformations and Monti Carlo
simulated annealing, evolutionary genetic and Lamarckian genetic algorithm methods to account
for ligand conformations (Morris et. al., 2009). It is also used with Auto Dock Tools (ADT)
(Cosconati et. al., 2010). Details will be discussed further.

32
GOLD (Genetic Optimization of ligand docking) uses genetic algorithms. It allows partial
protein flexibility and explores the full range of ligand conformational flexibility with partial
flexibility of the protein. It can work even in presence of water and ion molecules in the active
site of proteins (Jones et. al., 1997).
FlexX uses incremental construction for ligands (Rarey, 1996).
Docking programs utilize scoring functions in order to rank compounds and estimate the free
energy of binding of a ligand to a specific target and is based on a docked pose generated after
docking of protein with different ligands of a database. Several scoring functions have been
developed. They can be categorized as follows:
(a) Force field-based functions which estimate the binding free energy by adding the strength of
intermolecular van der Waals, electrostatic interactions and hydrogen bonding between all the
atoms of the protein and ligand involved in complex along with the Solvation and entropy.
(b) Empirical scoring functions that are based on counting the number of various types of
interactions between hydrophobic contacts, number of hydrogen bonds and number of rotatable
bonds immobilized in complex formation.
(c)Knowledge-based functions that use statistical observations of intermolecular contacts in
receptor-ligand complexes with known structural conformations (Lionta et a1.,2014).

2.10.1 Compound selection after docking (post processing)

In SBDD the expert chemist is required for compound selection and post processing after the
virtual screening to finalize the compounds for experimental trial and is considered the rate
limiting step. Simplified scoring functions and the inadequate sampling of the conformational
space for the ligand lead to unrealistic poses, intra-ligand steric clashes, twisted amides, E/Z
esters, imperfect hydrogen bonding network, and poses based on shape complementarity. This
result into an unreasonably high score so such compounds need to be discarded. Therefore,
visual inspection and careful analyzing of thousands of docking poses and interatomic contacts
between docked ligands and target is done for efficient and reliable outcomes (Athanasiadis et.
al.,2012;Waszkowycz,2008).

Softwares used in this study

I) Auto Dock
Auto dock is a suite of automated docking tools used to design to predict how small molecules
such as substrates or drug candidates, bind to a receptor of known 3Dstructure. This software is
very fast, provides high quality predictions of ligand conformations, and is able to show good
correlations between predicted inhibition constants and experimental ones. Auto Dock can also

33
be used in blind docking when the binding site is unknown (http://autodock.scripps.edu/ ).It was
developed in the 1990s at the Scripps Research Institute. It has been used widely in number of
researches, studies and for educational purpose (Morris et al., 2009). The graphical user interface
(GUI) of Auto Dock provides the opportunity for students to view proteins in three dimensions.
This allows the user to understand tertiary structures more effectively. Further students can view
surface representations of protein crystal structures to demonstrate that finite ligand binding
regions exist. This helps in understanding the mechanisms involved in protein-ligand interactions
and provides the unique opportunity for students to explore protein structures in three
dimensions (Helgren et al., 2017).

This docking tools tests and ranks the geometry of the drug compounds or the ligands under
study perhaps based on modifications of an existing lead compound, or to screen entire databases
of available molecules, searching for novel compound (Goodsell et al.,1996). Ligands are
regarded as rigid bodies or with torsional flexibility; scoring is made possible by shape
complementarity or with detailed energetic models; employing exhaustive searches or more
limited search techniques (Lybrand, 1995; Rosenfeld et al.,1995). It uses simulated annealing for
searching conformations, allowing several torsional degrees of freedom in a flexible ligand to be
searched. A grid-based technique is used for energy evaluation at each step of the simulation
(Goodsell and Olson, 1990).

Auto dock Tools (ADT) is a graphical user interface which helps to set up which bond is to be
treated as rotatable in the ligand and to analyze dockings. It has improved graphical front-end for
Auto dock and AutoGrid. It runs on Linux, Mac OSX, SGI IRIX and Microsoft Windows. With
this software we can view molecules in 3D, rotate and scale in real time, add all hydrogens or
just non-polar hydrogens, assign partial atomic charges to the ligand and the macromolecule
(Gasteiger or Kollman United Atom charges),merge non-polar hydrogens and their charges with
their parent carbon atom, Set up rotatable bonds in the ligand using a graphical version of
AutoTors, set up the AutoGrid Parameter File (GPF) using a visual representation of the grid
box, and slider-based widgets, set up the AutoDock Parameter File (DPF) using forms, Launch
AutoGrid and AutoDock, read in the results of an AutoDock job and graphically display them
and view is contoured AutoGrid affinity map (Morris, 2007). Autodock Tools-1.5.6 was used in
this study. It was developed by the Scripps Research Institute in python programming language.
The release of MGL tools 1.5.6 (ADT) was announced in 2012 February
2(http://mgltools.scripps.edu/News/mgltools-1-5-6-release-announcement ).

lI) Osiris datawarrior

Datawarrior is an open source data visualization and analysis program Data visualization and
analysis software with sufficient chemical intelligence. It runs on all major operating system and
was developed in Java programming language by Actelion Pharmaceuticals Ltd. In 1998 which

34
has been updated and added with novel features time to time at Actelion/ldorsia Pharmaceuticals
Ltd. DataWarrior combines dynamic graphical views and interactive row filtering with chemical
intelligence. Numerical and categorical data along with chemical information as shared scaffolds
and compound substitution patterns is visualized by scatter plots, box plots, and bar or pie charts
visualize numerical or category data along with. Chemical descriptors independently encode
various aspects of chemical structures, e.g. the chemical graph, chemical functionality and 3-
dimensional pharmacophore features. ADME/Tox properties are used as filtration measure to
generate library of preferred compounds with drug like properties. Different kinds of molecular
similarities are calculated and used in graphical views or for row filtering and other purposes.
Special features are inbuilt which support different stages of drug discovery from the screening
of compounds through structure activity analysis to the statistical interpretation of animal
experiments (Sander et al., 2015). In this study Osiris datawarrior version 5.2.1©2002-2020
Idorsia Pharmaceuticals Ltd. was used. It is highly integrated into the platform, connected to
databases and other tools.
(https://openmolecules.org/datawarrior/index.html).

iii)PyMol

PyMOL is a Python-enhanced molecular graphics tool which is open source visualization

software that can be used in drug discovery process for 3D visualization of proteins, small
molecules, density, surfaces, and trajectories. It also allows molecular editing, raytracing, and
movies (https://fossies.org/linux/privat/pymol-open-source 2.4.0.tar.gz/index_tp.html). This has
many features including the active sites determination within protein and protein sequence
visualization. It can be extended to python plugins because of the language it is written in.
PyMOL is one of the most widely used macromolecular visualization tools. This software can be
used not only in macromolecular visualization but also for macromolecule editing (Yuan et
al.,2017). PyMOL has been successfully used in the discovery of new drug candidates for
various targets such as in the discovery of a potent small molecule inhibitor for gankyrin
(Thakuret al.,2011) lead optimization for Cytochrome P450 enzymes (Danielson et al., 2011) and
VS of new drug candidates for the tumor suppressor protein P53 (Pereira et al., 2016).

PyMOL was created by Warren Lyford Delano and commercialized initially by DeLano
Scientific LLC. In 2010, after an agreement was done with Schrödinger Inc. Since then PyMOL
was developed, maintained, supported and sold and all the current subscription is regulated by
Schrödinger Inc (Yuan et al., 2017). The PyMol Molecular Graphic System Version 2.4.0
Copyright Schrödinger,LLC was used in this study.

lii) Discovery Studio Visualizer

35
The BIOVIA Discovery Studio Visualizer is a suite of software, free, feature-rich molecular
modeling application which allows viewing, sharing and analyzing protein and small molecule
data. It offers an interactive environment for viewing and editing molecular structures,
sequences, X-ray reflection data, scripts, and other data relevant to life science researchers.
Discovery studio visualizer supports a range of stereo graphical options (E.g., split screen,
hardware stereo). It is designed to enable hardware graphics acceleration which supports
advanced visualization options such as depth cueing, blur and shading capabilities. Some other
features include visualization capabilities such as:
Sequences, including Chain view support for multi-domain proteins (E.g,Antibodies)
2D and 3D Charting, Histograms, Heat maps and Data tables
Map interactions using a comprehensive set of favorable, unfavorable and unsatisfied non-bond
monitors
Interactivity between multiple graphical views on the same data (https://www.3ds.com/products-
services/biovia/products/molecular-modeling-simulation/biovia-discovery-studio/visualization/) .
It is developed and distributed by Dassault Systems BIOVIA. Discovery Studio
Visualiserv16.1.0.15350.2015 by Dassault Systemes Biovia Corp.was used in this study.

iv)PyRx

PyRx is open sources Virtual Screening software that can be used to screen libraries of
compounds against potential drug targets in Computational Drug Discovery approach. Virtual
Screening of compounds is possible from any platform and it helps users in every step
throughout this process. PyRx wizard features facilitate easy-to-use user interface and chemical
spreadsheet like functionality that makes it a valuable tool for Rational Drug Design in the drug
discovery process. It is widely common for structural based drug discovery process as it includes
chemical spreadsheet-like functionality and powerful visualization engine that are essential for
structure based virtual screening(https://sourceforge.net/projects/pyrx/) .

PyRx is written in python programming language. It is a GUI (Graphic User Interface) that uses
a large body of established open source software such as:

 AutoDock 4 and AutoDock Vina are used as docking software.

 AutoDockTools, used to generate input files.

 Python as a programming/scripting language.

 Python for cross-platform GUI.

 The Visualization ToolKit (VTK) by Kitware,Inc.

36
 Enthought Tool Suite, including Traits,for application building blocks.

 Opal Toolkit for running AutoDock remotely using web services.

 Open Babel for importing SDF files, removing salts and energy minimization.

 matplotlib for 2D plotting

(https://cac.queensu.ca/wiki/index.php/HowTo:pyrx)

AutoDock Vina is open source software used for docking, designed and implemented by Dr.
Olegg Trott in the Molecular Graphics Lab at The Scripps Research Institute. It is easy to use
and flexible. It only needs the structure of molecule to be docked and specification of search
space including the binding sites. Vina uses the PDBQT file format for both input and output as
used by AutoDock(http://vina.scripps.edu/).

Open babel is an established software within PyRx generates PDBQT file format. It uses
Lipinski rules of five for filtering molecules and converting the different format files and energy
minimization is performed with the universal force field (UFF) (Aissouq et al.,2021) using the
conjugate gradient algorithm. It can be done in interactively or in batch. It can read write and
convert it to about 110 file formats (http://openbabel.org/wiki/ Main Page)

vi)Ligplot+

LigPlot+ is a successor to the orignal LIGPLOT program which is used for automatic generation
of 2D ligand-protein interaction diagrams. It runs from an intuitive java interface which allows
on-screen editing of the plots. It is an interface to the LIGPLOT and DIMPLOT (a program for
plotting protein-protein or domain-domain interactions) programs which generates schematic
diagrams of protein-ligand and protein-protein interactions respectively. Users can flexibly select
the interface of interest can be selected and DIMPLOT generates a diagram showing the residue-
residue interactions across the interface. The residues in one of the interfaces can be optionally
displayed in sequence order for further interpretation
(https://www.ebi.ac.uk/thornton-srv/software/LigPlus/). LigPlot+ v.1.4.5 by Roman Laskowski,
2009 was used in our study.

vi)Gaussian

Gaussian is an electronic structure program. It is a computational chemistry software package. It

was initially released in 1970 by John People (Pople,2004) and his research group at Carnegie

37
Mellon University as Gaussian 70 7 It is widely used by chemists and scientists for analysis of
computational data. Gaussian 03 version 6.0 by Gaussian Inc. was used in this study.

vii)Chemcraft

Chemcraft is a graphical program for working with quantum chemistry computations. It is a

convenient tool for visualizing computed results and preparing new jobs for a calculation.
Chemcraft is mainly developed as a graphical user interface fór the GAMESS(US version and
the PCGamess/Firefly) and Gaussian program packages

3 CHAPTER THREE: MATERIALS AND METHODOLOGY

3.1 Protein or target structure retrievation

The completion of the human genome project increased the range of therapeutic targets in drug
discovery and drug design by enabling everyone to better understand the human genome and the
depth of genetic makeup within the human body. Nuclear magnetic resonance (NMR)
spectroscopy, high-throughput crystallography, protein purification, and other cutting-edge
techniques have been used to provide structural data on protein complexes and protein-ligand
interactions. This has benefited in the development of molecular docking and computer-aided
drug design (Bajorath, 2002).

38
When the target protein's structure is known, a docking technique based on the structure is used.
Additionally, it is possible to use existing knowledge and information about the target structure
by calculating the interaction energy for each molecule tested.
Aromatase is rate limiting step in hormonal pathway for production of estrogen. The crystal
structure of Aromatse protein (PDB ID: 3EQM) , catalyse the biosynthesis of all estrogens from
androgens. The crystallography structure with 5Å resolutions was obtained in PDB format.

3.2 Protein and ligand preparation:

Prior to docking simulations, the dockable protein structures and ligands require specific
preparation. The proteins and ligands must be prepared for molecular docking beforehand using
a variety of tools to ensure effective and accurate docking results. Hydrogen atoms are added,
non-polar bonds are combined, Gasteiger charges are added, and then the protein is prepared for
use in mgltools' (http://mgltools.scripps.edu/) docking studies. This format is called pdbqt.
Similar to this, ligand preparation was carried out in Openbabel GUI (O'Boyle et al., 2011) using
PyRx 0.9.8 setup by energy minimization using forcefield and converted to pdbqt file format,
which can then be used for docking.

3.3 Preparation of ligand database:

Structure-based virtual screening was carried out in the current work. A broad chemical database
with structural diversity may include diverse class of molecules that may be sufficient to develop
potential leads against the selected viral target, and structure-based virtual screening was
conducted with the primary goal of avoiding partiality among the class of molecules. To address
this urgency in the context of the global health crisis, it would be more efficient to use a library
of compounds that are already safe for use in humans and have low side effect profiles.

3.3.1 Ligand library:

Ligand library of natural products were selected for the present study. This ligand library was
downloaded from Zinc 15 ligand library(https://zinc15.docking.org/substances/home/) in .sdf
format.

3.4 ADME-TOX filter

The process of finding new drugs and developing them takes a long time and is very expensive.
Many promising lead compounds that were initially thought to be better fall short of making it to
the clinical stage of drug development due primarily to unfavorable pharmacokinetic
characteristics and toxicity risks. The OSIRIS program was used to predict the toxic profiles and
drug likeness of potential lead compounds obtained from docking studies (Nisha et. al., 2016).
Using the functional groups and chemical bonds found in the lead molecules' structures, OSIRIS

39
determines the lead molecules' molecular weight, cLogP, cLogS, Druglikeness, total polar
surface, toxicities including mutagenicity, tumorigenicity, reproductive effects, and irritant
effects, and rotatable bond counts (Sander et. al., 2015). The ligand library was subjected to Data
warrior software (Sander et al., 2015) where these compounds got filtered on the basis of certain
parameters:

Table 1: Parameters for filtration of ligands

Properties Lower Limit Upper Limit
Total Molecular Weight 200 500
Clog P -3 6
Clog S -4 -2
H acceptor 0 10
H donor 0 5
Polar Surface Area 0 120
Druglikeness 0 15
Rotable bond count 0 10
Mutagenic None
Tumorigenic None
Reproductive effectiveness None
Irritant None

Additionally, these ligands are transformed into .dwar files with the aid of datawarrior, and then
into sdf file formats with the aid of PyRx software (Dallakyan & Olson, 2015). The ligand
library was reduced to 35,534 ligands after the ligand library was filtered according to the
aforementioned parameters.

3.5 hMAT1A filter for hepatotoxicity:

The hMAT1A filter is typically used to remove ligands that interfere with cellular processes and
harm the liver by failing to replace the native ligand SAM of hMAT1A. Human S-
adenosylmethionine synthetase 1 (hMAT1A protein), crystal structure (ligand-free form), (RCSB
id:6SW5) was taken for filtration. After being modified in PymoL, the hMAT1A protein is
prepared by removing Chains C and D and only leaving Chains A and B in their place. Water

40
molecule and native ligands were eliminated. Prior to docking, the modified protein is converted
to the pdbqt format using Autodock tools.

3.6 Initial virtual screening:

The ligand library was used for initial docking with the hMAT1A protein where SAM (S-
Adenosylmethionine) was used as a reference inhibitor. These hMAT1A filtered ligands in
which ligands with binding energies lower than SAM inhibitor were taken and those with higher
energies were filtered out. Additionally, these filtered ligands were subjected to another
screening using the other protein CYP3A4 for lead compound identification to be developed as
probable drug candidate.

3.7 CYP3A4 filter for drug metabolism

The CYP3A4 was used as the another filter as it has the major role in the drug metabolism.

3.7.1 Initial virtual screening:

After the initial screening with hMAT1A, the filtered ligand library was used for initial docking
with the CYP3A4 protein. Ritonavir (RIT) was used as a reference inhibitor for the screening. In
addition to hMAT1A filtered ligands, these CYP3A4 filtered ligands were used for the final
molecular docking for lead target identification after the virtual screening, in which ligands with
binding energies lower than RIT inhibitor were taken and those with higher energies were
filtered out.

3.8 Binding site prediction

The ability of a molecule to interact with a particular pocket was found after the 3D structure of
the protein was obtained. The location of the binding site could then be found once the ligands
co-crystallized with the target protein (Cerqueira et al. 2010). PymoL software and literature
reviews were used to identify the target protein's active binding pocket.

3.8.1 PymoL:
BLAST was utilized to locate the template with the highest level of sequence coverage,
similarity, and length. PymoL was used to predict its active site. Now, the target protein's
sequence alignment with that specific template was performed. Using the template's PymoL, the
active binding residues were found. Comparing the active binding site residues, it was
discovered that template and target were perfectly aligned. As a result, that specific residue is
regarded as our target protein's active site.
3.8.2 Review:
Following a thorough review of the literature, the target protein's active binding sites were
provided. Our protein and the protein used in the relevant research or literature review had their

41
sequences aligned. Without protein, it was discovered that the active residues on the binding site
matched. The binding sites could thus be predicted.

3.9 Final Structure based Virtual Screening

The filtered ligands from CYP3A4 were used in molecular docking using Autodock Vina in
PyRx interface, version 0.9.8 (Dallakyan & Olson, 2015), along with the reference/native ligand
for our prioritized protein (CYP19A1). The docking parameters were set to have an
exhaustiveness of 32 and number of modes of 32. The trajectory for the ligand to move (SAS
figure) freely was decided based on active binding sites (reference) and the SAS results (figure
no.).

3.10 Preparation of final ligand dataset

After conducting docking studies, the top lead compounds were obtained. The 3D sdf structures
of these compounds were searched for and downloaded from the Pubchem database
(https://pubchem.ncbi.nlm.nih.gov/), and they were then renamed and combined using the Osiris
data warrior tool to create a final dataset of ligands.

3.11 Analysis of docking results

3.11.1 Binding interaction of ligands with protein target

Docking a lead compound to a target protein involves chemical interactions that contribute to the
lead molecule's stability in that association. Using the PymoL software version 4.6.0, the amino
acid residues involved in possible interactions with the protein structure were identified.

3.11.2 Bond length and bond types responsible for drug ligand interaction
The interaction of a chemical lead with its target protein is always determined by the various
types of chemical interactions that are possible. A ligand is bound to its target protein with
varying bond lengths and interactions, which were analyzed using Discovery Studio Visualizer.
The 2D diagram was used to analyze the ligand's 2D interaction with our protein.

3.11.3 Hydrophobic bond interactions

Different types of bonding interactions are critical in establishing a ligand's stable association to
the enzyme's binding pocket. One of the most important bond types is hydrophobic interaction,
which can be studied using the Ligplot software version. The ligand-protein interactions
involving hydrogen bonds and hydrophobic contacts were investigated using the Ligplot
software. PymoL software was used to create the PDB file of a protein interacting ligand. The

42
protein PBD file was opened with the Ligplus executable jar file, and the protein molecule was
chosen and run through it. The image then showed an interacting protein and ligand, as well as
hydrogen and hydrophobic bonds between them and various interacting residues and atoms.

3.11.4 Density Functional Computations (DFT) using Gaussian 03

DFT calculations were measured on GAUSSIAN 03 platform to interpret the atomic

arrangement of studied compound, and were optimized using B3LYP/ 631G basis set to establish
the geometry theoretically and study further. Gauss view was used to prepare and visualize
frontier molecular orbital studies and molecular electrostatic potential maps (MEP), and the
compounds under study were then analysed utilizing the results.

Initially, a Gauss input file was created. In PyMol, the docked ligand structure's pdb file was
created for this purpose by adding hydrogen. After that, it was opened in the Gauss view, where
the parameter for the calculate Gaussian calculator option was set. Subsequently, the job type
option's further parameters were selected as Optimization + frequency and the option for Raman
computation with the yes option. Similar to this, the DFT computation technique with the
B3LYP/631G basis set was selected under the method option. Another field in the title must be
filled out for the file name, and under link zero, the read-only and checkpoint files were left with
their default names and the memory limit was set to 1GB. After that, the file was saved with the
retain option selected. The file name must be saved with no spaces between the letters, and the
file type must be specified as a Gaussian input file.

Gaussian input file was then opened in Gauss 03 and the content of in % selection was copied
upto slash sign (/) and this was pasted in % rwf section. Below this under in route section option
genome+ connectivity and space before it was removed and finally in molecule specification
section all numbers and spaces below C and H was removed then under file option exit and run
was selected for Gauss run. For visualization and preparation of MEP and observation of
vibrations chk file obtained after Gauss run was used.

To determine the highest occupied molecular orbital (HOMO), lowest occupied molecular orbital
(LUMO), and HUMO-LUMO energy gap, frontier orbital studies were conducted. The Gaussian
output file was examined in a similar manner to observe the charge distribution. This provides us
with the areas of positive and negative electrostatic potential.

3.11.5 Ligand-based ADME prediction and Prediction of toxicological properties

Following the identification of top hits for each of the prioritized targets, the pharmacokinetic
properties of these hits are investigated further. To analyze these properties as well as the
toxicological properties of selected potential drugs, the online web tools SwissADME

43
(http://www.swissadme.ch/index.php) and pkCSM web tools
(http://biosig.unimelb.edu.au/pkcsm/prediction) are used.

44
4 CHAPTER FOUR: RESULTS

4.1 Search for reference and target proteins

For our study, hMAT1A and CYP3A4 were taken as the reference proteins. Along with it, the
Phase I drug metabolizing enzymes were also taken as references. Our target proteins were
determined by vast literature survey and correspondingly the use of software as maggielab. From
this, we got the proteins aromatase and fumarase as our target proteins.
4.2 Protein Structure Preparation

4.2.1 Retrieving from RCSB:

Structural information is crucial in the quest for novel diagnostic and therapeutic strategies as
structure comes before function in biology. For biological macromolecules, atomic-level three-
dimensional (3D) structural data are frequently essential for analyzing and comprehending the
specific mechanisms underlying hormonal disorders. They are also widely employed to find
targets that might be druggable and to speed up the discovery and creation of both small-
molecule and biologic medicines.
The spectrum of therapeutic targets in drug design and discovery has expanded since the human
genome project's completion; thanks to knowledge of the human genome and an insight of the
depth of genetic makeup within the human body. Nuclear magnetic resonance (NMR)
spectroscopy, high-throughput crystallography, protein purification, and other cutting-edge
techniques have been exploited to obtain structural data on protein complexes and protein-ligand
interactions. This has facilitated in the evolution of molecular docking and computer-aided drug
design. [Bajorath, 2002 (Guheshwori dd)]
Building on PDB data, the RCSB Protein Data Bank (RCSB PDB; http://rcsb.org) supports
research and learning in structural biology, computational biology, and other fields. The only
global repository for empirically determined, atomically detailed three-dimensional structures of
biological macromolecules is the Protein Data Bank (PDB). Multiple tools for structure inquiry,
browsing, analysis, and molecular visualization are available on the RCSB PDB website
(rcsb.org). Users can create sophisticated search combinations of parameters and criteria using
the Advanced Search interface, or can make simple searches using the top menu bar search
containing PDB ID, name, sequence, and ligand SMILES. [1]

45
4.2.2 Structure Validation:

When it comes to stereochemistry, biological molecules are frequently asymmetric, but proper
stereochemistry is crucial to their functionality. Stereochemical mistakes must be avoided even
when they do not result in such significant issues since they indicate aberrations of the simulated
system from the biological reality that is to be replicated. Structure validation, a crucial step in
having files ready for depositing in a coordinate database, is possible using a number of servers
and tools. In order to discover anomalies in geometry and structure, such as chirality, bond
angles, close contacts, or rotamer states of amino acid side chains, the SAVES server, which
provides an interface to tools like PROCHECK, WHAT CHECK, and other applications, might
be used as an example. [2]
The Structure analysis and verification server, SAVES v6.0 (https://saves.mbi.ucla.edu/),
examined the stability of the created model. The ERRAT score generated in SAVES, as shown
in Table, for each of the models generated by RCSB and prepared by iTASSER, was used to
confirm the accuracy of each modeled protein. The result showed that each model's non-bonded
interactions fell within a respectable typical range. Procheck is used to analyze the constructed
models' overall stereochemical quality as well as their overall and residue by residue geometry.
For each model, we looked at the Ramachandran plot as it's crucial to draw attention to
unrealistic conformations inside the model by analyzing bond geometry.
90% of the residues of a decent model should typically be found in the permitted areas of a
Ramachandran plot (Laskowski,1993).

Additionally, the ProSA web server (https://prosa.services.came.sbg.ac.at/prosa.php) was used to

generate the z scores for each model. The z score indicates the overall model quality. The
estimated "degree of nativeness" of the characteristics exhibited in a created model is reflected
by the QMEAN Z-score, which also reveals if our model parallel with experimental structures in
terms of quality. A model's Z-scores display any notable deviations from the predicted "natural"
behavior. Table 5 displays the z scores for each target protein, and the energy of ProSA with a
negative value indicates the model's dependability and quality.

46
4.3 Ligand Library Preparation

Zinc15 database, Natural Product – 2,00,000

4.3.1 Screening of Ligand Library against hMAT1A protein:

Figure X: 3D Structure of S-Adenosylmethionine Synthetase 1 (hMAT1A) prepared from

Discovery Studio 2016 (PDB: 6SW5) (ligand-free form)

47
S-adenosylmethionine (SAM) is formed through methionine catabolism which occurs mostly in
the liver in a reaction catalyzed by methionine adenosyltransferase (MAT). MAT is product of
two different genes, MAT1A and MAT2A, respectively. Decrease in SAM has shown adverse
effects in liver leading to many hepatic complexities including production of tumor necrosis
factor (TNF). Thus, SAM is critical for human. In human body, hepatic hMAT1A and the extra
hepatic MAT, hMAT2A biosynthesize SAM (Lu et al., 2002). It is necessary to protect the
interference with hepatic SAM biosynthesis pathway by preventing the inhibition of hMAT1A
while screening the ligand library through molecular docking. However, the SAM transporters
could also be used in some extreme conditions.
This is crucial because the drug targets discovered shouldn’t degrade human health and must be
safe for human consumption. Therefore, to ensure the hepatic safety of the screened compound,
the ligand libraries were screened against hMAT1A also. The 3D structure of hMAT1A (PDB:
6SW5) was retrieved form PDB. It was processed in Auto dock tools and converted into pdbqt
file format a readable file format for docking in PyRx inference later. The ligands with equal or
lower binding affininty to hMAT1A were screened and studied further.

Table X: Active binding sites obtained by PyMol in hMAT1A (PDB: 6SW5)

Amino acid Residue

ALA 55
GLU 70
GLN 112
SER 114
ILE 117
GLY 133
ASP 134
LYS 289
ASP 291

The following dimension were fixed in Å: Center (X, Y, Z) = (31.096, -0.571, 24.983),
Dimensions (X, Y, Z) = (27.625, 57.151, 31.622). The ligand molecules having binding energy
higher or similar than native ligand SAM were selected from docking of ligand library against
CYP3A4. Thus, ligand filtration was done based on the comparison of binding affinity of the
ligand library (Zinc database, NP) with human hepatic hMAT1A.

48
4.3.2 Reference Ligands

Table: Illustration of reference proteins with their corresponding native or reference ligands
Protein Reference/Native Ligands Pubchem Structure
ID (Source: PubChem)
hMAT1A S-Adenosyl-L-Methionine 34755
(Native ligand)

Cyp3a4 Ritonavir 392622

Table: Illustration of target proteins with their corresponding native or reference ligands

Protein Reference/Native Ligands Pubchem Structure

ID (Source: RCSB)
hCYP19a1 Androstenedione (asd) 6128

Fumarase 4-(2-HYDROXYETHYL)-1- 23831

PIPERAZINE
ETHANESULFONIC ACID
(EPE)

49
4.4 Binding Site Prediction
Total gasteiger charge added was -0.9965. Gasteiger (-Marsili) charge was added. For the given
molecular system when partial charges are not present in the protein. Gasteiger method assumes
an overall net neutral state for the respective molecular system. Polar hydrogens are hydrogen
atoms that are bonded to electronegative atoms like oxygen and nitrogen. Auto Dock tools
assumes that non-polar hydrogen is hydrogen bonded to carbon atoms. It creates cellular
environment by charge equilibration.
For protein Aromatase, active binding sites were determined by taking the amino acid residues
lying within 5 Å of protein and native ligand interaction sites as it was found to be consistent
with several literatures and results for webservers like 3DLigandSite.
For protein Fumarase, active binding sites were determined from the RCSB as some sites were
found missing while taking the amino acid residues lying within 5 Å of protein and native ligand
interaction sites.

Table: Active binding sites obtained by PyMol in Aromatase (PDB: 3EQM)

Amino acid Residue Amino acid Residue
ILE 133 THR 310
PHE 134 VAL 370
PHE 221 LEU 372
TRP 224 VAL 373
ILE 305 LEU 477
ALA 306 SER 478
ASP 309 HEM 600

Table: Active binding sites obtained by PyMol in Fumarase (PDB: 5UPP)

Amino acid Residue Amino acid Residue
THR 457 HIS 462
ALA 458 ILE 463
LEU 459 GLY 464
ASN 460 LYS 467
PRO 461

50
The binding site has to lie within the grid box in PyRx interface used for molecular docking. The
grid box size is defined and determined as mentioned in material and methodology section.
Calculation of binding energies was performed using Autodock vina PyRx. Initially, a grid box
was set to cover the active site of crystal structure of experimental protein within the grid box
and grid size was determined. This was kept constant for each docking process with all the
ligand libraries studied and used in virtual screening in this study. This helps to maintain similar
experimental rooms for each ligand library under study and the outcomes can be comparatively
studied as unbiased results.
For the protein Aromatase, the following dimensions were fixed in Å: Center (X, Y, Z) = (87.12,
53.099, 46.697), Dimensions (X, Y, Z) = (19.672, 19.739, 21.539) with an exhaustiveness of 32
and num mod 32.
For the protein Fumarase, the following dimensions were fixed in Å: Center (X, Y, Z) = (43.827,
-38.503, -16.032), Dimensions (X, Y, Z) = (17.712, 15.726, 16.203) with an exhaustiveness of
32 and num mod 32.
PyRx automatically advances to results page, where results of virtual screening computation can
be viewed.

51
4.5 Molecular Docking and Analysis

a) Selection of Top hits / Lead Compound

Compound Compound Name P.I. CYP19A CYP3A4 hMAT1A

ID (zinc 15,
np)
SAM -7.6
SAH -6.9
RTV -10.4
ASD -13.4
Compound 3-[(1S,9R)-6-oxo-7,11- 1.4 -10.1 -7.7 -7.3
31466 diazatricyclo [7.3.1.02,7] trideca-
2,4-dien-11-yl] propanoic acid

Compound (1R,3S,6S,7S,10S,11R)-7- 1.2 -10.6 -8.4 -6.9

31487 [(dimethylamino)methyl]-3,12-
dimethyl-2,9-dioxatetracyclo
[9.3.0.01,3.06,10] tetradec-12-en-
8-one

Compound 1-methyl-3-[(1R,2S,5R,6S,9R)- 0.6 -10.1 -8.4 -6.7

10084 6,8,9-trimethyl-3-oxabicyclo
[3.3.1] non-7-en-2-yl] pyridin-1-
ium

Compound (5R)-5-benzyl-5-ethyl-2-methyl- 1 -10.2 -8.5 -7.1

13617 3,4-dihydro-1H-pyrido[4,3-b]
indol-5-ium

52
4.3.3 Protein-Ligand Interaction

Hydrophobic Bond Length Ligand atom Protein atom

bond residues
4530 865
PHE 134A 3.83
4526 863
PHE 134A 3.64
4522 1713
PHE 221A 3.80
3.79 4517 1745
TRP 224A
3.52 4525 1743
TRP 224A
3.53 4533 3206
VAL 370A
VAL370A 3.78 4532 3205

3.84 4529 3238

MET 374A
3.56 4518 4271
LEU 477A

Table: Protein ligand interaction of compound 13617 with aromatase

53
 Hydrophobic bond Bond Length Ligand atom Protein
residues atom

ILE 133A 3.49 11 871

ILE 133A 3.62 12 872

TRP 224A 3.26 4 1764

ALA 306A 3.67 12 2596

ASP 309A 3.78 10 2615

Table: Protein
3.62 9 2625
ligand THR310A
interaction of compound 10084 with aromatase

Hydrophobic bond Bond Length Ligand atom Protein atom

residues
ALA 306A 3.75 7 2597

54
 THR 310A 3.80 4 2626
VAL 370A 3.97 12 3225

Table: Protein ligand interaction of compound 31466 with aromatase

Hydrophobic bond Bond Length Ligand atom Protein atom

residues
ILE 133A 3.98 4514 852
PHE 134 3.78 4527 865
PHE 221A 3.22 4525 1713
ASP 309A 3.56 4524 2596
THR 310A 3.66 4519 2606
VAL 370A 3.67 4520 3206
VAL 370A 3.58 4517 3205

Table: Protein ligand interaction of compound 31487 with aromatase

4.4 ADMET profiling of Selected Ligands

The prior passing or testing of ligands for certain essential parameters of drug molecules enables
to make the ligand library that is not only dockable but also has higher chances of clearing
further steps in drug discovery, including clinical trials and addressing toxic issues as well. The
toxic profiles and druglikeness and pharmacokinetics properties using a list of parameters
including Lipinski’s rule of five (Sander et. al., 2015) as mentioned in material and methodology
section.
Drug molecules need to be absorbed properly for their optimal functionality. Drug molecules are
absorbed via passive diffusion and connective volume flow through water filled intercellular
spaces. These are drug particles with lower molecular weight small molecules.
The cLogP value of a compound is the logarithm of its partition coefficient between n-octanol
and water log (coctanol/cwater), is a measure of the compound’s hydrophilicity. Thus, lower
hydrophilicity and higher LogP values decreases the absorbance and permeation of drug
molecule. It has been shown that compounds must have cLogP greater than 5.0 to have a
reasonable probability of being well absorbed.

55
In addition, cLogS correspond to 10-based logarithm of the solubility of a molecule measure in
mol/L and correlates to water solubility of a drug. The aqueous solubility of a compound
significantly affects its absorption and distribution characteristics. It plays important role in drug
uptake and elimination. A drug molecule needs to be absorbed optimally to exhibit
pharmachological activity. Poorly soluble drugs get eliminated before entering circulation and
thus do not exhibit pharmachological activity.
OSIRIS calculated the other essential features of drugs like: druglikeness and toxicities like
mutagenicity, tumorigencity, reproductive effects and irritant effects. And the compounds with
no such effects were filtered from the inbuilt parameters of OSIRIS.
Hydrogen bonding affects the membrane transport and distribution of drug within the biological
systems. The number of hydrogen bonds affects the absorption as well. It is unfavorable for a
compound to have more hydrogen bonds. These bonds need to be broken for the drug molecule
to pass the across the aqueous environment within the cells.
Topological polar surface area (TPSA) makes us of functional group contributions based on a
large database of structures. It is considered a convenient measure pf the polar surface area that
avoids the need to calculate ligand 3d structure or to decide which is the relevant biological
conformation/s (Prasanna and Doerksen, 2009). Low polar surface area is found to be important
predictors of good oral bioavailability.
Druglikeness is qualitative concept that is used in drug designing to determine how “drug-like” a
substance is in terms of bioavailability. It is estimated form the molecular structure before the
substance is synthesized and tested. It is calculated by using the sum of all drug scores in
OSIRIS. Parameter is set as 0 to positive value for druglikeness. A positive value states that our
molecule contains predominantly fragments which are frequently present in commercial drugs.
Ligand molecule with rotable bonds between 0-10 is taken as a selection parameter. Increase in
number of rotable bonds decrease permeation. This is important criteria that needs to be
addressed in drug screening for the oral bioavailability of drugs.
The compounds were also screened for mutagenicity, tumorigencity, irritant properties and effect
in reproductiveness by OSIRIS. These parameters were set at null value to make the ligand
library including the ligands non-mutagenic, non-tumorigenic, non-irritant and no any adverse
effect in reproductiveness. Thus, eliminating the ligand molecules which fail to meet the basic
criteria set for compound screening in this study.
Thus, ligand libraries were prepared which included the compounds which were filtered through
ADME/Tox screening. Finally, ligand library is converted into pdbqt file format after. Energy
minimization was performed with the universal force field (UFF) using the conjugate gradient
algorithm in Openbabel GUI available in PyRx interface. This is the usable file format for
molecular docking in PyRx.

56
4.6 Computational analysis (DFT)

DFT calculations were performed on Gaussian 03 platform to interpret the atomic arrangement
of compounds under study and optimized using B3LYP/G31G (6D, 7F) basis set for
establishment of theoretical geometry. The optimized structure parameter of studied compound
had singlet spin, zero charge and parameters were calculated by RB3LYP method with
OptimizTION+ Frequency job type. The calculated parameters of : 3-[(1S,9R)-6-oxo-7,11-
diazatricyclo[7.3.1.02,7]trideca-2,4-dien-11-yl]propanoic acid, (1R,3S,6S,7S,10S,11R)-7-
[(dimethylamino)methyl]-3,12-dimethyl-2,9-dioxatetracyclo[9.3.0.01,3.06,10]tetradec-12-en-8-
one, 1-methyl-3-[(1R,2S,5R,6S,9R)-6,8,9-trimethyl-3-oxabicyclo[3.3.1]non-7-en-2-yl]pyridin-1-
ium and (5R)-5-benzyl-5-ethyl-2-methyl-3,4-dihydro-1H-pyrido[4,3-b]indol-5-ium are tabulated
in table X:

Table X: Calculated parameters of the compounds studied

Compound Total Energy (in a.u.) Dipole moment (in RMS Gradient
Debye) norm (in a.u.)
31466 -879.26219990 5.1750 0.00000319
31487 -942.95777996 5.4341 0.00000323
10084 -787.85158978 2.8642 0.00000464
13617 -924.52245328 2.3803 0.00001273

The study of dipole moment of the compounds clearly indicates that all compounds possess
stronger dipole-dipole interactions. Dipole moment shows the molecular charge distribution and
is given as a vector in three dimensions and is used as descriptor to depict the charge movement
across the molecule depending upon the centers of positive and negative charges. Dipole
moments are strictly determined for neutral molecules. Compound 31487 was found to have
highest dipole moment with 5.2507 thus highest dipole-dipole moment with 5.4341; thus, highest
dipole-dipole interactions. A molecule is more stable if there are more attractive forces and less
repulsive forces. Attractive forces lower the potential energy of the molecule and repulsive
forces increase the potential energy of molecule. Therefore, molecules with lower energy are
more stable. All compounds exhibited negative value of energy indicating the stability of
complexes. Root Mean Square (RMS) Gradient is calculated as a root mean squared gradient,
sqrt (mean F_i dot. F_i), where F_i is the force on atom i and mean () takes an average over all
atoms. RMS Gradient norm is the RMS of the individual variables.

4.7 Frontier molecular orbital study and chemical reactivity descriptor analysis

Frontier molecular orbitals (FMO) are most important orbitals and play a vital role in the charge
transfer interactions with the binding site of target protein. FMO analysis helps in understanding
the reactions and active site conjugations. These are highest occupied molecular orbital (HOMO)

57
and lowest unoccupied molecular orbital (LUMO). The highest occupied molecular orbital
(HOMO) and the lowest unoccupied molecular orbital (LUMO) are the main orbital’s that are
responsible for the chemical stability of the complex. The difference in bond energy distinctly
explains the chemical reactivity and chemical stability of the active molecules (Govindarajan et
al., 2012). HOMO energy indicates the ability or character of donating electrons and LUMO
energy is responsible for accepting the electron. These regions are located over a protein-ligand
complex at different locations indicating the charge transfer in LUMO regions. Energy gap
between HOMO and LUMO energy is called energy gap values. The calculated energy gap
between 3-[(1S,9R)-6-oxo-7,11-diazatricyclo[7.3.1.02,7]trideca-2,4-dien-11-yl]propanoic acid,
(1R,3S,6S,7S,10S,11R)-7-[(dimethylamino)methyl]-3,12-dimethyl-2,9-
dioxatetracyclo[9.3.0.01,3.06,10]tetradec-12-en-8-one, 1-methyl-3-[(1R,2S,5R,6S,9R)-6,8,9-
trimethyl-3-oxabicyclo[3.3.1]non-7-en-2-yl]pyridin-1-ium and (5R)-5-benzyl-5-ethyl-2-methyl-
3,4-dihydro-1H-pyrido[4,3-b]indol-5-ium are given in table: X. Lower the gap energy higher the
chances of transition of electron and thus compound is reactive. Wide energy gaps negatively
affect the electron to move from the HOMO to the LUMO subsequently leading to a weak
affinity of the ligand for target protein.

Table X: Calculated energy differences (LUMO-HOMO) of the compounds:

Compound HOMO (eV) LUMO (eV) Energy Gap (eV)
31466 -0.20212 -0.02984 0.17228
31487 -0.19777 -0.00233 0.19544

Fig: Compound 31466

Fig: Compound 31487

58
References:
Abad-Zapatero, C., & Metz, J. T. (2005). (Mignani et al., 2018). Drug Discovery Today, 10(7),
464–469. https://www.sciencedirect.com/science/article/pii/S1359644605033866?via
%3Dihub
Abel, R., Young, T., Farid, R., Berne, B. J., & Friesner, R. A. (2008). Role of the active-site
solvent in the thermodynamics of factor Xa ligand binding. Journal of the American
Chemical Society, 130(9), 2817–2831. https://doi.org/10.1021/ja0771033
Andricopulo, A., Guido, R., & Oliva, G. (2008). Virtual Screening and Its Integration with
Modern Drug Design Technologies. Current Medicinal Chemistry, 15(1), 37–46.
https://doi.org/10.2174/092986708783330683
Aparoy, P., Kumar Reddy, K., & Reddanna, P. (2012). Structure and Ligand Based Drug Design
Strategies in the Development of Novel 5- LOX Inhibitors. Current Medicinal Chemistry,
19(22), 3763–3778. https://doi.org/10.2174/092986712801661112
Athanasiadis, E., Cournia, Z., & Spyrou, G. (2012). ChemBioServer: A web-based pipeline for
filtering, clustering and visualization of chemical compounds used in drug discovery.
Bioinformatics, 28(22), 3002–3003. https://doi.org/10.1093/bioinformatics/bts551
Becker, O. M., Dhanoa, D. S., Marantz, Y., Chen, D., Shacham, S., Cheruku, S., Heifetz, A.,
Mohanty, P., Fichman, M., Sharadendu, A., Nudelman, R., Kauffman, M., & Noiman, S.
(2006). An integrated in silico 3D model-driven discovery of a novel, potent, and selective
amidosulfonamide 5-HT1A agonist (PRX-00023) for the treatment of anxiety and
depression. Journal of Medicinal Chemistry, 49(11), 3116–3135.
https://doi.org/10.1021/jm0508641
Boligala, G. P., Yang, M. V., van Wunnik, J. C., & Pruitt, K. (2022). Nuclear Dishevelled: An
enigmatic role in governing cell fate and Wnt signaling. Biochimica et Biophysica Acta
(BBA) - Molecular Cell Research, 1869(10), 119305.
https://doi.org/10.1016/j.bbamcr.2022.119305
Budzik, B., Garzya, V., Shi, D., Walker, G., Woolley-Roberts, M., Pardoe, J., Lucas, A., Tehan,
B., Rivero, R. A., Langmead, C. J., Watson, J., Wu, Z., Forbes, I. T., & Jin, J. (2010). Novel
N-substituted benzimidazolones as potent, selective, CNS-penetrant, and orally active M1
mAChR agonists. ACS Medicinal Chemistry Letters, 1(6), 244–248.
https://doi.org/10.1021/ml100105x
Cheng, T., Li, Q., Zhou, Z., Wang, Y., & Bryant, S. H. (2012). Structure-based virtual screening
for drug discovery: A problem-centric review. AAPS Journal, 14(1), 133–141.
https://doi.org/10.1208/s12248-012-9322-0
Doak, B. C., Zheng, J., Dobritzsch, D., & Kihlberg, J. (2016). How Beyond Rule of 5 Drugs and
Clinical Candidates Bind to Their Targets. Journal of Medicinal Chemistry, 59(6), 2312–
2327. https://doi.org/10.1021/acs.jmedchem.5b01286
Edlund, M., Blombäck, M., Von Schoultz, B., & Andersson, O. (1996). On the value of
menorrhagia as a predictor for coagulation disorders. American Journal of Hematology,

59
 53(4), 234–238. https://doi.org/10.1002/(SICI)1096-8652(199612)53:4<234::AID-
AJH4>3.0.CO;2-Z
El-Hemaidi, I., Gharaibeh, A., & Shehata, H. (2007). Menorrhagia and bleeding disorders.
Current Opinion in Obstetrics and Gynecology, 19(6), 513–520.
https://doi.org/10.1097/GCO.0b013e3282f1ddbe
Gillespie, S. H., & McHugh, T. D. (1997). The biological cost of antimicrobial resistance.
Trends in Microbiology, 5(9), 337–339. https://doi.org/10.1016/S0966-842X(97)01101-3
Gokyildiz, S., Aslan, E., Beji, N. K., & Mecdi, M. (2013). The Effects of Menorrhagia on
Women’s Quality of Life: A Case-Control Study. ISRN Obstetrics and Gynecology,
2013(2010), 1–7. https://doi.org/10.1155/2013/918179
Goodsell, D. S., Morris, G. M., & Olson, A. J. (1996). Automated docking of flexible ligands:
Applications of AutoDock. Journal of Molecular Recognition, 9(1), 1–5.
https://doi.org/10.1002/(SICI)1099-1352(199601)9:1<1::AID-JMR241>3.0.CO;2-6
Goodsell, D. S., & Olson, A. J. (1990). Automated docking of substrates to proteins by simulated
annealing. Proteins: Structure, Function, and Bioinformatics, 8(3), 195–202.
https://doi.org/10.1002/prot.340080302
Helgren, T. R., & Hagen, T. J. (2017). Demonstration of AutoDock as an Educational Tool for
Drug Discovery. Journal of Chemical Education, 94(3), 345–349.
https://doi.org/10.1021/acs.jchemed.6b00555
Jones, G., Willett, P., Glen, R. C., Leach, A. R., & Taylor, R. (1997). Development and
validation of a genetic algorithm for flexible docking. Journal of Molecular Biology,
267(3), 727–748. https://doi.org/10.1006/jmbi.1996.0897
Kadir, R. A., Economides, D. L., Sabin, C. A., Owens, D., & Lee, C. A. (1998). Frequency of
inherited bleeding disorders in women with menorrhagia. Lancet, 351(9101), 485–489.
https://doi.org/10.1016/S0140-6736(97)08248-2
Keighley, W. (2011). The need for high throughput kinetics early in the drug discovery process.
Drug Discovery World, 12(3), 39–45.
Kimbung, S., Lettiero, B., Feldt, M., Bosch, A., & Borgquist, S. (2016). High expression of
cholesterol biosynthesis genes is associated with resistance to statin treatment and inferior
survival in breast cancer. In Oncotarget (Vol. 7, Issue 37, pp. 59640–59651).
https://doi.org/10.18632/oncotarget.10746
Kitchen, D. B., Decornez, H., Furr, J. R., & Bajorath, J. (2004). Docking and scoring in virtual
screening for drug discovery: Methods and applications. Nature Reviews Drug Discovery,
3(11), 935–949. https://doi.org/10.1038/nrd1549
Leelananda, S. P., & Lindert, S. (2016). Computational methods in drug discovery. Beilstein
Journal of Organic Chemistry, 12, 2694–2718. https://doi.org/10.3762/bjoc.12.267
Lionta, E., Spyrou, G., Vassilatis, D., & Cournia, Z. (2014). Structure-Based Virtual Screening
for Drug Discovery: Principles, Applications and Recent Advances. Current Topics in

60
 Medicinal Chemistry, 14(16), 1923–1938.
https://doi.org/10.2174/1568026614666140929124445
Lipinski, C. A., Lombardo, F., Dominy, B. W., & Feeney, P. J. (2012). Experimental and
computational approaches to estimate solubility and permeability in drug discovery and
development settings. Advanced Drug Delivery Reviews, 64(SUPPL.), 4–17.
https://doi.org/10.1016/j.addr.2012.09.019
Menorrhagia | Johns Hopkins Medicine. (n.d.). Johns Hopkins Medicine. Retrieved February 28,
2022, from https://www.hopkinsmedicine.org/health/conditions-and-diseases/menorrhagia
Moitessier, N., Englebienne, P., Lee, D., Lawandi, J., & Corbeil, C. R. (2008). Towards the
development of universal, fast and highly accurate docking/scoring methods: A long way to
go. British Journal of Pharmacology, 153(SUPPL. 1), 7–26.
https://doi.org/10.1038/sj.bjp.0707515
Nepal, N., Chaudhary, P., & Mainali, N. (2016). Histopathological analysis of endometrial
biopsies in dysfunctional uterine bleeding. Journal of Pathology of Nepal, 6(11), 910–913.
https://doi.org/10.3126/jpn.v6i11.15670
Oshita, A., Kishida, S., Kobayashi, H., Michiue, T., Asahara, T., Asashima, M., & Kikuchi, A.
(2003). Identification and characterization of a novel Dvl-binding protein that suppresses
Wnt signalling pathway. Genes to Cells, 8(12), 1005–1017. https://doi.org/10.1111/j.1365-
2443.2003.00692.x
Palep-Singh, M., & Prentice, A. (2007). Epidemiology of abnormal uterine bleeding. Best
Practice and Research in Clinical Obstetrics and Gynaecology, 21(6), 887–890.
https://doi.org/10.1016/j.bpobgyn.2007.03.012
Pan, A. C., Borhani, D. W., Dror, R. O., & Shaw, D. E. (2013). Molecular determinants of drug-
receptor binding kinetics. Drug Discovery Today, 18(13–14), 667–673.
https://doi.org/10.1016/j.drudis.2013.02.007
Paul, S. M., Mytelka, D. S., Dunwiddie, C. T., Persinger, C. C., Munos, B. H., Lindborg, S. R.,
& Schacht, A. L. (2010). How to improve RD productivity: The pharmaceutical industry’s
grand challenge. Nature Reviews Drug Discovery, 9(3), 203–214.
https://doi.org/10.1038/nrd3078
Pawitan, J. A. (2001). Mechanism of normal menstruation and abnormality associated with
menorrhagia. Medical Journal of Indonesia, 10(2), 121.
https://doi.org/10.13181/mji.v10i2.20
Pereira, D., Lima, R. T., Palmeira, A., Seca, H., Soares, J., Gomes, S., Raimundo, L., Maciel, C.,
Pinto, M., Sousa, E., Helena Vasconcelos, M., Saraiva, L., & Cidade, H. (2019). Design and
synthesis of new inhibitors of p53–MDM2 interaction with a chalcone scaffold. Arabian
Journal of Chemistry, 12(8), 4150–4161. https://doi.org/10.1016/j.arabjc.2016.04.015
Perrotti, D., & Neviani, P. (2008). Protein phosphatase 2A (PP2A), a drugable tumor suppressor
in Ph1(+) leukemias. Cancer and Metastasis Reviews, 27(2), 159–168.
https://doi.org/10.1007/s10555-008-9119-x

61
Pople, S. J. A. (2004). Publisher’s Note. Journal of Computational Chemistry, 25(14), i–i.
https://doi.org/10.1002/jcc.20136
Ragni, M. V., Machin, N., Malec, L. M., James, A. H., Kessler, C. M., Konkle, B. A., Kouides,
P. A., Neff, A. T., Philipp, C. S., & Brambilla, D. J. (2016). Von Willebrand factor for
menorrhagia: A survey and literature review. Haemophilia, 22(3), 397–402.
https://doi.org/10.1111/hae.12898
Rarey, M., Kramer, B., Lengauer, T., & Klebe, G. (1996). A fast flexible docking method using
an incremental construction algorithm. Journal of Molecular Biology, 261(3), 470–489.
https://doi.org/10.1006/jmbi.1996.0477
Reddy, A. S., Pati, S. P., Kumar, P. P., Pradeep, H. N., & Sastry, G. N. (2007). Virtual screening
in drug discovery -- a computational perspective. Current Protein & Peptide Science, 8(4),
329–351. http://www.ncbi.nlm.nih.gov/pubmed/17696867
Rees, M. (1987). Menorrhagia. British Medical Journal (Clinical Research Ed.), 294(6574),
759–762. https://doi.org/10.1136/bmj.294.6574.759
Rosen, R., Vajda, S., & Delisi, C. (1995). DESIGN. 677–700.
Seshacharyulu, P., Pandey, P., Datta, K., & Batra, S. K. (2013). Phosphatase: PP2A structural
importance, regulation and its aberrant expression in cancer. Cancer Letters, 335(1), 9–18.
https://doi.org/10.1016/j.canlet.2013.02.036
Shaw, J. A. (2018a). Menorrhagia: Practice Essentials, Background, Pathophysiology.
Medscape. https://emedicine.medscape.com/article/255540-overview#a5
Shaw, J. A. (2018b). Menorrhagia Treatment & Management: Medical Care, Surgical Care.
Medscape. https://emedicine.medscape.com/article/255540-treatment
Shoichet, B. K. (2004). Successes from virtual screening. 432(December).
Shrestha, D. B., Thapa, S., Singh, R., Khatiwada, R., Pandey, S., Srivastava, A., Aryal, B. B.,
Upreti, T., & Gurung, A. (2017). Presenting complaints in Gynecology Outpatient
Department (OPD) and Prevalence and Assessment of Dysmenorrhea. Medical Journal of
Shree Birendra Hospital, 16(1), 47–51. https://doi.org/10.3126/mjsbh.v16i1.16027
Sliwoski, G. R., Meiler, J., & Lowe, E. W. (2014). Computational Methods in Drug Discovery
Prediction of protein structure and ensembles from limited experimental data View project
Antibody modeling, Antibody design and Antigen-Antibody interactions View project.
Computational Methods in Drug Discovery, 66(1), 334–395.
Srivastava, A., & Rodeghiero, F. (2005). Epidemiology of von Willebrand disease in developing
countries. Seminars in Thrombosis and Hemostasis, 31(5), 569–576.
https://doi.org/10.1055/s-2005-922229
Stadler, S. C., Hacker, U., & Burkhardt, R. (2016). Cholesterol metabolism and breast cancer.
Current Opinion in Lipidology, 27(2), 200–201.
https://doi.org/10.1097/MOL.0000000000000285
Talele, T., Khedkar, S., & Rigby, A. (2010). Successful Applications of Computer Aided Drug

62
 Discovery: Moving Drugs from Concept to the Clinic. Current Topics in Medicinal
Chemistry, 10(1), 127–141. https://doi.org/10.2174/156802610790232251
Thakur, P. K., & Hassan, M. I. (2011). Discovering a potent small molecule inhibitor for
gankyrin using de novo drug design approach. International Journal of Computational
Biology and Drug Design, 4(4), 373–386. https://doi.org/10.1504/IJCBDD.2011.044404
Usha, T., Shanmugarajan, D., Goyal, A. K., Kumar, C. S., & Middha, S. K. (2018). Recent
Updates on Computer-aided Drug Discovery: Time for a Paradigm Shift. Current Topics in
Medicinal Chemistry, 17(30), 3296–3307.
https://doi.org/10.2174/1568026618666180101163651
Walters, W. P., Green, J., Weiss, J. R., & Murcko, M. A. (2011). What do medicinal chemists
actually make? A 50-year retrospective. Journal of Medicinal Chemistry, 54(19), 6405–
6416. https://doi.org/10.1021/jm200504p
Waszkowycz, B. (2008). Towards improving compound selection in structure-based virtual
screening. Drug Discovery Today, 13(5–6), 219–226.
https://doi.org/10.1016/j.drudis.2007.12.002
Winkler, D. A. (2002). The role of quantitative structure--activity relationships (QSAR) in
biomolecular discovery. Briefings in Bioinformatics, 3(1), 73–86.
https://doi.org/10.1093/bib/3.1.73
Wlodarchak, N., & Xing, Y. (2016). PP2A as a master regulator of the cell cycle. Critical
Reviews in Biochemistry and Molecular Biology, 51(3), 162–184.
https://doi.org/10.3109/10409238.2016.1143913
Yong, E. L., Hillier, S. G., Turner, M., Baird, D. T., Ng, S. C., Bongso, A., & Ratnam, S. S.
(1994). Differential regulation of cholesterol side-chain cleavage (P450scc) and aromatase
(P450arom) enzyme mRNA expression by gonadotrophins and cyclic AMP in human
granulosa cells. In Journal of Molecular Endocrinology (Vol. 12, Issue 2, pp. 239–249).
https://doi.org/10.1677/jme.0.0120239
Yu, W., & Mackerell, A. D. (2017). Computer-aided drug design methods. Methods in
Molecular Biology, 1520, 85–106. https://doi.org/10.1007/978-1-4939-6634-9_5
Yuan, S., Chan, H. C. S., & Hu, Z. (2017). Using PyMOL as a platform for computational drug
design. Wiley Interdisciplinary Reviews: Computational Molecular Science, 7(2).
https://doi.org/10.1002/wcms.1298
Zappa, S., Li, K., & Bauer, C. E. (2010). The tetrapyrrole biosynthetic pathway and its regulation
in Rhodobacter capsulatus. Advances in Experimental Medicine and Biology, 675, 229–250.
https://doi.org/10.1007/978-1-4419-1528-3_13
Zhang, H., Liu, J., Zhao, X., Liang, B., Wen, Z., & Li, M. (2014). Loss of PP2A and PTEN
immunoexpression coexists with survivin overexpression in adenomyosis. Reproductive
Biology, 14(3), 200–205. https://doi.org/10.1016/j.repbio.2014.04.004

63
64