Structure based design, stability study and synthesis of the dinitrophenylhydrazone derivative of the oxidation product of lanosterol as a potential P. falciparum transketolase inhibitor and in-vivo antimalarial study

Article abstract of DOI:10.1007/s40203-021-00097-8

The growing resistance to the current antimalarial drugs in the absence of a vaccine can be effectively tackled by identifying new metabolic pathways that are essential to the survival of the malaria parasite and developing new drugs against them. Triterpenes and steroids are the most abundant group of natural products with a great variety of biological activities. However, lanosterol is not known to possess any significant biological activity. In this study the binding and interactions of a dinitrophenyl hydrazine (DNP) derivative of lanosterol, LAN (a derivative that incorporates a substantially polar moiety into the steroid) with P. falciparum?transketolase was studied by molecular docking and MD simulation with the view to exploit the DNP derivative as a lead in antimalarial chemotherapy?development considering that the P. falciparum?transketolase (PfTk) is a novel target in antimalarial chemotherapy. The enzyme catalyses the production of ribose sugars needed for nucleic acid synthesis; it lacks a three-dimensional (3D) structure necessary for docking because it is difficult to obtain a crystalline form. A homology model of PfTk was constructed using Saccharomyces cerevisiae transketolase (protein data bank ID of 1TRK) as the template. The compound was observed to have Free Energy of Binding higher than that of the cofactor of the protein (Thiamine Pyrophosphate, TPP) and a synthetic analog (SUBTPP) used as reference compounds after MD Simulation. The compound was synthesized in a two-step, one-pot reaction, utilizing a non-acidic and mild oxidant to oxidize the lanosterol in order to avoid the rearrangement that accompanies the oxidation of sterols using acidic oxidants. The LAN was characterized using IR spectroscopy and NMR experiments and tested in-vivo for its antimalarial chemo suppression using a murine model with Chloroquine as a standard. The LAN at a concentration of 25?mg/kg was found to have a comparable activity with Chloroquine at 10?mg/kg and no mortality was observed among the test animals 24?days post drug administration showing that the compound indeed has potential as an antimalarial agent and a likely inhibitor of PfTk considering that there is a strong agreement between the in-silico results and biological study.

Full text of DOI:10.1007/s40203-021-00097-8

In Silico Pharmacology
(2021) 9:38
https://doi.org/10.1007/s40203-021-00097-8
ORIGINAL RESEARCH
Structure based design, stability study and synthesis
of the dinitrophenylhydrazone derivative of the oxidation product
of lanosterol as a potential P. falciparum transketolase inhibitor
and in‑vivo antimalarial study
Olatomide A. Fadare1 · Nusrat O. Omisore2 · Oluwaseun B. Adegbite1 · Oladoja A. Awofsayo3 ·
Frank A. Ogundolie4 · Julius K. Adesanwo1 · Craig A. Obafemi1
Received: 30 April 2021 / Accepted: 31 May 2021
© The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2021
Abstract
The growing resistance to the current antimalarial drugs in the absence of a vaccine can be efectively tackled by identifying
new metabolic pathways that are essential to the survival of the malaria parasite and developing new drugs against them.
Triterpenes and steroids are the most abundant group of natural products with a great variety of biological activities. However,
lanosterol is not known to possess any signifcant biological activity. In this study the binding and interactions of a dinitro-
phenyl hydrazine (DNP) derivative of lanosterol, LAN (a derivative that incorporates a substantially polar moiety into the
steroid) with P. falciparum transketolase was studied by molecular docking and MD simulation with the view to exploit the
DNP derivative as a lead in antimalarial chemotherapy development considering that the P. falciparum transketolase (PfTk)
is a novel target in antimalarial chemotherapy. The enzyme catalyses the production of ribose sugars needed for nucleic
acid synthesis; it lacks a three-dimensional (3D) structure necessary for docking because it is difcult to obtain a crystalline
form. A homology model of PfTk was constructed using Saccharomyces cerevisiae transketolase (protein data bank ID of
1TRK) as the template. The compound was observed to have Free Energy of Binding higher than that of the cofactor of the
protein (Thiamine Pyrophosphate, TPP) and a synthetic analog (SUBTPP) used as reference compounds after MD Simula-
tion. The compound was synthesized in a two-step, one-pot reaction, utilizing a non-acidic and mild oxidant to oxidize the
lanosterol in order to avoid the rearrangement that accompanies the oxidation of sterols using acidic oxidants. The LAN was
characterized using IR spectroscopy and NMR experiments and tested in-vivo for its antimalarial chemo suppression using
a murine model with Chloroquine as a standard. The LAN at a concentration of 25 mg/kg was found to have a comparable
activity with Chloroquine at 10 mg/kg and no mortality was observed among the test animals 24 days post drug administra-
tion showing that the compound indeed has potential as an antimalarial agent and a likely inhibitor of PfTk considering that
there is a strong agreement between the in-silico results and biological study.
Keywords Transketolase · Molecular dynamics · Homology modelling · Synthesis · Chemo suppression
Introduction
* Olatomide A. Fadare
tofadare@oauife.edu.ng
Malaria is a mosquito-borne infectious disease caused by a
1
parasitic protozoan belonging to the plasmodium type. This
parasite gets transmitted via a bite of an infected female
Anopheles mosquito. Malaria is one of the diseases that
pose a signifcant threat to human lives. Malaria is promi-
nent in most parts of Africa, Asia, and South America due
to the weather condition of these areas. According to the
2019 World Malaria Report (WHO 2010), Nigeria had
the highest number of global malaria cases (25% of global
Department of Chemistry, Obafemi Awolowo University,
Ile-Ife, Nigeria
2
3
4
Department of Pharmacology, Faculty of Pharmacy, Obafemi
Awolowo University, Ile-Ife, Nigeria
Department of Pharmaceutical and Medicinal Chemistry,
University of Uyo, Uyo, Nigeria
Enzymology and Enzyme Biotechnology Unit, Department
of Biochemistry, Federal University of Technology, Akure,
Nigeria
Vol.:(0123456789)
1 3

38
Page 2 of 16
In Silico Pharmacology
(2021) 9:38
malaria cases) in 2018 and accounted for the highest num-
ber of deaths (24% of global malaria deaths). Malaria is a
major public health problem in Nigeria where it accounts
for more cases and deaths than any other country in the
world. Malaria is a risk for 97% of Nigeria’s population.
The remaining 3% of the population live in the malaria free
highlands. There are an estimated (average) 100 million
malaria cases with over 300,000 deaths per year in Nigeria.
This compares with 215,000 deaths (average) per year in
Nigeria from HIV/AIDS. It is worthy to note that malaria
contributes to an estimated 11% of maternal mortality as
well in Nigeria (Nigeria Malaria Fact sheet—us embassy in
Nigeria). By virtue of the fact that Malaria is a leading cause
of death and disease in many developing countries especially
in Nigeria, where young children and pregnant women are
the groups most afected, and the propensity of the malaria
parasite to develop resistance to known antimalarial drugs,
especially, recently, that there have been reported cases of
resistance to the most dependable and commonly used anti-
malarial drug, Artemisinin and derivatives (WHO 2017;
Amato et al. 2018), it has become expedient that new anti-
malaria drugs are developed fast to counter the efect of
artemisinin resistance, hence, this study. A more efective
way to resolve the antimalarial resistance issue is to identify
new metabolic pathways in the parasite that are essential for
its survival and developing new drugs against them (Olliaro
and Yuthavong 1999; Ridley 2002; Rosenthal 2001). The
Pentose phosphate pathway (PPP) is an essential pathway
in the parasite, this pathway is responsible for the replica-
tion of the parasite, PPP produces ribose sugar used in the
synthesis of nucleotide and nucleic acid. Pentose phosphate
pathway is an enzyme driven pathway and one of the critical
enzyme in this pathway is transketolase, it is a novel tar-
get in the anti-malarial drug discovery efort even though it
lacks an experimentally determined 3D structure. However,
homology modelling has helped in bridging the gap between
sequence and structure space, enabling scientists to construct
protein models that are hard to crystallize or for which NMR
spectroscopy structure determination is not tractable. It helps
to construct a three-dimensional (3D) model of an unknown
“target” protein from its amino acid sequence and an experi-
mental 3D structure of a related homologous protein which
is referred to as the “template”.
triterpenoid “tetracyclic triterpenoid” isolated from Momor-
dica balsamina L. (Cucurbitaceae) having the lanostane
skeleton (steroids) exhibit in-vitro anti-plasmodial activity
against blood stages of parasites, hence this study of exploit-
ing (lanosterol also a steroid and endogenous) as an anti-
malarial agent. It is thought that lanosterol, being lipohilic
and endogenous would be harmless to the human body and
if derivatized in such a way that it posses polar functional
groups, it might have some medicinal properties. This pre-
sent research envisaged to use computational techniques to
investigate the interactions of the proposed compound with
the P. falciparum transketolase enzyme and synthesize dini-
trophenylhydrazone derivative of oxidized lanosterol as a
potential inhibitor of the enzyme by occupying the cavity
that the substrates would normally occupy and binding with
critical polar residues within the active site of the enzyme
such that the metabolic function of the enzyme is diminished
and ultimately leading to the inhibition of the plasmodium
causing malaria knowing that the P. falciparum transketo-
lase enzyme has least homology with the human transketo-
lase enzyme.
Materials and methods
A model of Plasmodium falciparum transketolase was con-
structed by homology modelling using the software pack-
age, Modeller v9.1. The homology model was optimized by
molecular dynamic simulation using the GROMACS v5.0
software package and the optimized model was visualized
and analysed using Visual Molecular Dynamics (VMD)
v1.9. The docking study was carried out using Autodock
vina v12.0. The structures of the compounds were drawn in
Chemdraw v12.0 and energy minimization of the structures
done by Chem 3D, visualization of the protein and struc-
tures was done in Pymol v1.7.4.5. The molecular dynam-
ics simulation of the modelled protein with the proposed
lanosterol derivative (LAN) was done in a box of water using
the GROMACS software and grommos 54A7 forcefeld to
establish the trajectory after 50 ns from which informa-
tion on protein–ligand complex stability was determined,
root mean square deviation (rmsd) and fuctuation (rmsf),
radius of gyration (ROG) and solvent accessible surface area
(SASA), after which the g_MMPBSA module was used to
estimate the free energy of binding. The MD simulation (as
well as free energy estimate) was replicated for two reference
compounds, thiamine pyrophosphate (TPP) and a synthetic
analog (SUBTPP) that are cofactors of the transketolase
enzyme, for comparison (Fig. 1). The lanosterol deriva-
tive was synthesized and characterized using spectroscopic
techniques (NMR experiments and UV–Vis spectrocopy).
The synthesized compound was then tested for antimalaria
chemosuppression using a murine model.
Triterpenoids and steroids are the most abundant group
of natural products with a great variety of biological activi-
ties. Their diverse use in the pharmaceutical industry have
distinguished them as economically important substances
that generates annual sales of 12.4 billion USD (Dzubak
et al. 2006; Nazaruk and Borym-Kluczyk 2015). Triterpe-
noids have been shown to possess various activities like
anti-parasitic, anti-fungal, anti-bacterial, anti-infammatory,
anti-proliferative, anti-diabetics etc. Nogueira and Lopes
(2011) and Nasomjai et al. (2014) reported that a class of
1 3

In Silico Pharmacology
(2021) 9:38
Page 3 of 16
38
Fig. 1 The structure of the
compounds studied as ligands in
the molecular docking and MD
simulation
constructing the Ramachandran plot for the model using the
Zlab server for Ramachandran plot generation (University
of Massachusetts Medical School). The RMSD of the mod-
elled protein structure was also estimated as a measure of
the validity of the model.
In‑silico study of the interaction of study compound
with transketolase enzyme
This was done prior to wet lab synthesis using available
software, this intends to gain insight into the possible inter-
action of the DNP derivative of the oxidized lanosterol with
the transketolase enzyme to see if it will have any potential
as an inhibitor (in-silico) before going ahead to synthesize
and carry out any biological test.
Protein structure optimization
The P. falciparum transketolase homology model was sub-
jected to molecular dynamic simulation using the software
GROMACS (Abraham et al. 2015). A topology was frst
generated: topology contains all the information necessary
to defne the molecule within a simulation, the force feld,
OPLS AA was used and a box (rhombic-dodecahedron)
flled with water molecules was used for the simulation,
ions were added to the system to balance the charges of the
solvated protein, the homology model (3D structure) was
relaxed by energy minimization in order to ensure that the
system has no steric clashes or inappropriate geometry, then
the solvent and ions around the protein were equilibrated to a
desired temperature, pressure and density. After the system
was well equilibrated, the production molecular dynamics
was run for data collection and the simulated protein was
analysed for root mean square deviation (rmsd), stability
by assessment of radius of gyration and solvent accessible
surface area.
Model construction (homology modelling)
The amino acid sequence of Plasmodium falciparum
transketolase (target protein) was obtained in FASTA for-
mat from National Centre for Biotechnology Information
(NCBI), a similarity search was performed in the Protein
Data Bank (PDB) using protein-basic local alignment search
tool (BLASTp) in order to get a homologous protein which
will serve as the template protein. The percentage homol-
ogy with the target and the resolution of the 3D structure as
well as the statistical signifcance (E-value) of the alignment
between the target and the template protein were taken into
consideration when selecting a template that was used in the
construction of the target protein. The amino acid sequence
of the target protein (P. falciparum transketolase) was
aligned with the amino acid sequence of the template protein
(Saccharomyces cerevisiae transketolase) using the python
script align2d.py of the software, MODELLER, in order to
obtain optimal alignment. P. falciparum transketolase model
was constructed using the python script Build_profle.py of
the same MODELLER software (Šali and Blundell 1993).
This software was used to construct fve similar models of
which the one with the lowest Discreet Optimized Poten-
tial Energy (DOPE) and highest GA341 score was chosen
as the best model. The model built was evaluated by the
python script, evaluate_model.py; it was also validated by
Molecular docking
The docking studies were carried out using Autodock tools
v1.5.4 (Morris et al. 2009) and AutodockVina v1.1.2. (Trott
and Olson 2010). The chemical structures were drawn using
Chemdraw v12.0 and energy minimization was done using
Chem 3D v12.0. The lanosterol derivative (LAN), thia-
mine pyrophosphate (TPP) and 6-methyl thiamine pyroph-
osphate (SUBTTP) were subjected to molecular docking
1 3

38
Page 4 of 16
In Silico Pharmacology
(2021) 9:38
using Autodock tools. TPP (a cofactor of transketolase) and
6-methyl TPP (SUBTPP) served as reference compounds.
The protein was converted from.pdb to.pdbqt format, after
which its binding site was analyzed and assigned a grid box
(search space area) using AutoDOCK Tool. The ligands
were prepared for docking using Chem3D and AutoDock
Tools. After the preparation of the ligands and protein, each
ligand was docked into the protein using AutoDock Vina to
estimate the binding afnity and best binding pose.
calculate the free energy of binding between LAN, TPP and
SUBTPP and the modelled transketolase. The free energy of
binding calculation is based on the following theory:
(
)
Gbinding = Gcomplex− Gprotein + Gligand
where Gcomplex is the total free energy of the protein–inhibi-
tor complex and Gprotein and Gligand are total free energies
of the separated form of protein and inhibitor in solvent,
respectively (Egan et al. 2000). The average binding energy
calculations were done by a python script provided in g_
mmpbsa stand-alone program.
MD simulation and free energy estimation
Molecular dynamics study of the protein–ligand complexes
were done using GROMACS 2018.3 (Abraham et al. 2015)
software which was installed in ubuntu 18.04 LTS. Docked
structures of the protein–ligand complex were used in the
simulation study. The protein was processed and the topol-
ogy fle was prepared by using pdb2gmx and GROMO-
S54a7atb.f (imported from the automated topology builder
website) force feld while the ligand topology fle was pre-
pared by using the Automated Topology Builder (ATB)
version 3.0. The solvent addition was done in a cubic box
by using a box distance 1.0 nm from closest atom in the
protein. The addition of Cl− ions was used to neutralize the
system. Energy minimization was done by using steepest
descent algorithm taking 50,000 steps and 5 kJ/mol maxi-
mum force and Verlet cut-of scheme taking Particle Mesh
Ewald (PME) columbic interactions. Position restraints were
applied in the equilibration step. After that, NVT equilibra-
tion was done in 300 K and 100 ps of steps and NPT equili-
bration taking Parrinello-Rahman (pressure coupling), 1 bar
reference pressure and 100 ps of steps. LINCS algorithm was
applied to constraint all the bonds length. For long-range
electrostatics, Particle-mesh Ewald (PME) algorithm was
used. And fnally the production MD of the protein–ligand
complex was run for 50 ns.
Synthesis
All the chemicals and the reagents used were of synthetic
and analytical grades. The synthesis of the dinitrophenyl-
hydrazone of oxidized lanosterol was carried out by frst
oxidizing lanosterol with freshly prepared quinoliniumchlo-
rochromate (QCC), stirring at room temperature for 18 h,
after which the oxidized product was reacted with 2,4-dini-
trophenylhydrazine using an acid catalyst (5 drops of conc
H2SO4) by stirring at room temperature for 24 h. After the
synthesis, the product was purifed using column chromatog-
raphy and then recrystallized using appropriate solvent. The
dinitrophenylhydrazone derivative of oxidized lanosterol
(Compound A) was characterised by analysing the Reten-
tion factor (Rf), Melting point, Ultraviolet–Visible spectrum,
Infra-red Spectrum, 1H and 13C NMR Spectra. The melting
point of the DNP derivative of the oxidized lanosterol was
determined in an open capillary tube, the compound was
also confrmed for its purity on TLC plate using suitable sol-
vent system. IR Spectrum of the synthesized compound was
recorded using KBR pellets in the range of 400–4000 cm−1
on Agilent Cary 630 FT-IR spectrometer. 1H NMR and 13
C
of the synthesized compound were recorded in CDCl3 by
Bruker 400 MHz FT-NMR Spectrometer.
After successful completion of molecular dynamic simu-
lation for 50 ns, Root mean square deviation (RMSD) of
backbone residues, number of Hydrogen bonds, Root mean
square fuctuations (RMSF), Radius of gyration (RG), Sol-
vent accessible surface area (SASA) (Marsh and Teichmann
2011) were calculated. Free energy of binding calculations
were done following the g_mmpbsa protocol by Kumari
and Kumar (2014), developed for GROMOS96 43a1 force
feld. Molecular Mechanics Poisson-Boltzmann surface area
(MM-PBSA) is a widely used method to calculate the free
binding energy and to predict the stability of the complex.
Modifed Molecular Mechanics–Poisson Boltzmann Surface
Area (MM-PBSA) is open-source software, and employed
to calculate the free energy of binding between two defned
groups. Recently, MM-PBSA algorithm has been utilized as
a scoring function in computational drug design (Pant et al.
2020). In this study, MM-PBSA method was employed to
Procedure for the formation
of 2,4‑dinitrophenylhydrazone derivative
of oxidized lanosterol (LAN)
This synthesis involves a two-step, one-pot reaction (inter-
mediate ketone was not isolated).
Step 1. Oxidation of lanosterol
One molar equivalent of Lanosterol (10 g; 23.67 mmols)
was added to the freshly prepared quinolinium chlo-
rochromate (QCC) (6.22 g; 23.67 mmols) in dichlo-
romethame (DCM) and stirred at room temperature for
18 h. The reaction mixture was worked up by extrac-
tion with dichloromethane and concentrated in-vacuo
to get the crude product which was dried and weighed
1 3

In Silico Pharmacology
(2021) 9:38
Page 5 of 16
38
(70% yield), melting point of 116–120 °C. IR (KBr):
1709 cm−1 (C=O), 2926 cm−1 (sp3-C–H stretch), 1464
and 1377 cm−1 (C–H bend of hydrocarbon).
were administered for four (4) consecutive days. On the
ffth day, thin blood smear was prepared and the blood flms
were stained with Giemsa solution and then microscopically
examined with 100 × magnifcation (oil immersion). The
percentage chemo suppression of parasitaemia was calcu-
lated for each dose level by comparing the % parasitaemia
of the controls with those of treated mice (Knight and Peters
1980). The percentage parasitaemia of the experimental ani-
mals after the four days of drug administration was estimated
as count of parasitized red blood cells divided by the total
red blood cells counted within 20 microscopic felds (having
an average of 2000 total red blood cells counted).
Step 2. Formation of the DNP derivative of oxidized
lanosterol
One molar equivalent of oxidized lanosterol was added
together with 2,4-dinitrophenylhydrazine in ethanol
using an acid (1 mL of conc. H2SO4) as a catalyst. The
mixture was stirred at room temperature for 24 h. Being
monitored with thin layer chromatography (TLC), the
reaction mixture was fltered, dried and packed on a col-
umn to purify. The product of the reaction was concen-
trated, dried and weighed, melting point of 198-200 °C.
IR (KBr): 3372 cm−1 (N–H), 3103 cm−1 (Sp2-C–H)
2951 cm−1 (Sp3-C–H), 1618 cm−1 (C=N), 1518 and
1334 cm−1 (NO2), 1591 cm−1 (C=C Ar), 1471 cm−1
The percentage chemo suppression (Chart 2) was calcu-
lated as a relative value taking into account the % parasitae-
mia of the negative control. The formula used is as described
below
1
(C=C–H bend), 1132 cm−1 (C–N). The H NMR data
Parasitized red blood cells
is presented as follows; 1H, d, δ11.20 (J=4.3 Hz); 1H,s,
δ11.10; 1H,d, δ9.13 (J=2.6 Hz); 1H,dd, δ8.30 (J=8.9,
5.6, 2.6 Hz); 1H,dd, δ7.95 (J = 8.4, 5.9, 2.4 Hz); 1H,s,
δ5.30; 1H,t, δ5.10; 2H,m, δ 2.46–2.85; 1H,t δ 2.32
(J = 6.8 Hz); 2H, s, δ 2.18; 4H,m, δ 1.98–2.13; 2H, m,
δ1.79–1.98; 7H,m, 1.52–1.76; 3H,m, 1.36–1.51;4H, m,
1.28–1.35; 4H,d δ 1.24 (J=6.5 Hz); 4H, m, δ1.10–1.20;
2H,dd, δ 1.03 (J=26.8, 6.9 Hz); 7H, pd δ 0.88 (J=7.9,
7.0, 3.9 Hz), 2H,d, δ0.70 (J=9.5 Hz). And the 13C NMR
data is also presented as follows (δ values, ppm); 35.55,
27.81, 38.85, 50.35, 18.22, 26.46, 134.35, 134.35, 36.97,
20.97, 30.94, 44.43, 49.76, 30.82, 28.18, 50.35, 15.71,
19.11, 36.24, 18.61, 36.33, 24.89, 125.22, 130.91, 25.72,
17.62, 27.93, 15.40, 24.43.
% Parasitaemia =
× 100
Total red blood cells
(X − Y)
% Chemosuppression =
× 100
X
X=% Parasitaemia of the negative control Y=% Parasitae-
mia of the test compound.
Result and discussion
Homology modelling
The target protein, P. falciparum transketolase, has 672
amino acid residues in its sequence. Homology modelling
was used to construct a 3D structure for the target protein.
Saccharomyces cerevisea transketolase having 49% homol-
ogy with the target protein was used as the template. The
alignment of the amino acid sequence of the target and the
template done by Clustal Omega is shown in Fig. 2. Five
diferent models of this target protein were successfully con-
structed by the MODELLER software, the model with the
lowest DOPE score of − 80,725.75 was selected. The model
constructed was evaluated by plotting the energy profle,
Fig. 3, the energy profle also shows the alignment between
Biological study
Lethal dose determination
The acute toxicity of Compound A was determined using
Lorke’s method (Lorke 1983). Twelve mice were divided
into four groups (A, B, C and D). The four groups were
administered orally with graded concentrations (10, 100,
1000, 5000 mg kg−1 respectively) of the compound. These
animals were watched for 24 h.
Antimalarial chemo suppressive test
Table 1 Binding afnities of the ligands
Swiss white mice were divided into groups of fve and inoc-
ulated with 1.0 *107 P. berghei parasites intraperitoneal. Dif-
ferent doses of Compound A were dissolved in 1:1 mixture
of DMSO and H2O. Three of the groups were given oral
doses of 25, 50 and 100 mg/kg of the compound respec-
tively. Two other groups were administered with 10 mg/
kg chloroquine (reference drug) and 0.2 ml of vehicle of
preparation (negative control) respectively. The test agents
S/N
Compounds docked
Binding
afnity
(kcal/mol)
1
DNP derivative of oxidized
lanosterol (compound A)
− 8.0
2
3
TPP
− 7.0
− 7.0
Synthetic Analog of TPP
1 3

38
Page 6 of 16
In Silico Pharmacology
(2021) 9:38
Fig. 2 A screenshot showing alignment between the amino acid sequence of P. falciparum transketolase and Saccharomyces cerevisea transketo-
lase. The target is represented as CAG25349.1 and template as 1TRK-A (PDB code)
the template protein and the target model in a graphical form,
the energy profle in Fig. 3 shows close alignment for most
regions of the proteins. A Ramanchandran plot that shows
the stereo chemical assessment of the model constructed is
shown in Fig. 4. The Ramanchandran plot, Fig. 4, shows that
95.1% and 4.2% of the amino acid residues are in favoured
and allowed regions respectively, only 0.7% was found to
be in the outlier region, this also confrmed that the model
1 3

In Silico Pharmacology
(2021) 9:38
Page 7 of 16
38
Fig. 3 Energy profle of the
alignment between the target
protein model and the template
is a very good one. The superimposition of the model and
the template gives a RMSD value of 1.03 Å which further
buttress the fact that the model is a good model since an
RMSD less than 7 is said to signify a good model. The pro-
tein model was optimized by molecular dynamic simula-
tion in a box of water with OPLS force feld. The model,
after being subjected to molecular dynamic simulation was
also assessed using a Ramanchandran plot, Fig. 5, and the
number of amino acid residues in the favoured and allowed
regions are 87.6% and 10.6% respectively. Only 1.8% are
found in the outlier region. The superimposition of the
homology model and the 3D structure obtained after opti-
mization with molecular dynamic simulation gives a RMSD
value of 2.65 Å.
Fig. 4 Ramachandran plot of
the homology model of the
target protein (Green crosses
- highly preferred (favoured);
Brown triangles - preferred
(allowed) and Red circles -
questionable observations
(outlier region))
1 3

38
Page 8 of 16
In Silico Pharmacology
(2021) 9:38
Docking studies
Molecular docking provided insight into the interactions
and the binding modes of the protein and compound A.
The active site of the target protein model was determined
by visual inspection and superimposition of the templates
3D structure on that of the protein model. The proteins
and the compound were prepared as described earlier. The
3D structure of compound A was docked into the target
model with the gridbox (dimension of the grid box) centred
around the active site. The cofactor (TPP) and its synthetic
analog (6-methyl TPP) were also docked into the protein
to serve as reference. Compound A has the highest bind-
ing afnity of − 8.0 kcal/mol relative to that of the cofac-
tor (TPP, − 7.0 kcal/mol) and synthetic analog (6-methyl
TPP, − 7.0 kcal/mol). The higher the binding afnity, the
stronger the interaction between the protein and the ligand.
The resultant binding afnities from all the interactions for
the docked compounds are shown in Table 1. This implies
that the enzyme has higher afnity for compound A relative
to that of cofactor, which suggest that compound A has the
capacity to displace the cofactor and eventually disrupting
the activity of the enzyme, thereby efectively inhibiting the
enzyme, and preventing the enzyme from carrying out its
catalytic function and subsequently preventing the parasite
from carrying out its metabolic activity which will eventu-
ally lead to the death of the parasite. This observation was
corroborated from a follow-up in-vivo study using animal
model.
MD simulation and free energy estimation
The conformational changes that took place during the simu-
lation were estimated by determining the RMSD, RMSF,
ROG, SASA and H-bonds. The extent of conformational
change was used as a measure of the stability of the system
and extent of interaction between the protein and the ligands.
The rmsd plot of the protein backbone ftted to the ligands
(is shown in Fig. 3) shows that the simulations reached equi-
librium at around 25 ns for TPP and SUBTPP and at 35 ns
for LAN. The RMSF plots for the protein (C_alpha) ftted to
the ligands shows that the ligands fuctuated around the same
set of residues within the protein except for the residues
between 250 and 350 that showed marked diferences in the
fuctuations for the three ligands as observed in the plots for
RMSF (Fig. 6). The radius of gyration (ROG) plots (Fig. 6)
which indicates the compactness of the protein–ligand com-
plex decreased steadily showing that the protein–ligand
complex for TPP and the lanosterol derivative (LAN) was
getting more compact as the simulation progressed but the
complex of SUBTPP with the protein did not achieve the
kind of compactness that TPP and lanosterol had over the
course of the simulation.
1 3

In Silico Pharmacology
(2021) 9:38
Page 9 of 16
38
Fig. 5 The Ramachandran plot
of the simulated protein in a
box of water, for optimiza-
tion (Green crosses - highly
preferred (favoured); Brown
triangles - preferred (allowed)
and Red circles - questionable
observations (outlier region))
The gradual decrease in the solvent accessible area of the
complexes as the simulation progressed show that the bind-
ing of the ligands with the protein caused the protein to con-
strict in such a way that the surface area available for interac-
tion of water molecules reduced which is typical for proteins
having strong interactions with ligands. The plots for hydro-
gen bonding show that TPP and SUBTPP have plenty of
hydrogen bond interactions with the protein throughout the
course of the simulation which was expected because of the
high number of polar functional groups (considering the
heterocyclic core of the TPP and SUBTPP is also charged)
which facilitated strong electrostatic interactions Fig. 7. The
plot of resdiues contributing to free energy (Fig. 8) shows
some strong repulsive interactions between some residues
and ligands (LYS191 and ASP195 for TPP; ASP189 and
LYS191 for SUBTPP)—these repulsive interactions must
have been ofset by some other strong attractive interactions
(though smaller in magnitude but more) which culminated
in an overall strong electrostatic interaction unlike the lanos-
terol derivative which did not have much of the electrostatic
interactions, observed for TPP and SUBTPP, due to being
largely hydrophobic. The Van der Waals forces of attraction
between the lanosterol derivative and protein appears to be
more pronounced than its electrostatic interactions, however,
TPP and SUBTPP also have signifcant VDW interactions
(Fig. 7c). The Overall free energy of binding of the three
ligands shows that the lanosterol derivative has the highest
free energy of binding relative to TPP and SUBTPP which
had markedly reduced Free energy of binding despite their
high electrostatic interaction contributions. The overall free
energy must have been ofset by the high cost of desolva-
tion of the ligands. Their charged/polar nature must have
caused them to associate strongly with the water molecules
such that the penalty for stripping them of water molecules
at the binding site must have been very high as observed
in the breakdown shown in Fig. 9. It is a well-known fact
that the traditional, enthalpy-dominated way of regarding
ligand–protein interactions essentially ignores solvation
efects, which has now been establihsed to strongly afect
the thermodynamic profle of a binding event (Bronowska
2011). Several detailed binding free energy studies have
suggested that diferences in solvation may play a criti-
cal role in the observed diferences in the free energy of
binding between relatively similar compounds (Jiao et al.
1 3

38
Page 10 of 16
In Silico Pharmacology
(2021) 9:38
Fig. 6 The plots of RMSD, RMSF, ROG and SASA
2008; Reddy and Erion 2001). Two molecules might have
similar interactions with a protein, similar strain energies,
etc., but have diferent solvation properties in water, lead-
ing to solvation-driven diferences in binding free energies
(Bronowska 2011). Recently it became clear that studying
the hydration state of a protein binding pocket in the apo
(unbound) state should be a routine procedure in rational
drug design, as the role of solvation in tuning binding afn-
ity is critical (Bronowska 2011). Solvation costs appear to be
a valid reason why some ligands, despite ftting into a bind-
ing site, fail during experimental tests as inhibitors. Young
and co-workers showed that an optimised inhibitor of factor
Xa turns virtually inactive when the isopropyl group inter-
acting in the S4 pocket of factor Xa is replaced by hydrogen.
Substitution of this group by hydrogen, apart from trimming
down the number of favourable hydrophobic interactions,
leads to unfavourable solvation of the binding pocket (Young
et al. 2007).
Desolvation of the ligand itself may sometimes control
the binding free energy. Several detailed binding free energy
studies have suggested that diferences in solvation may
play an important role in diferences in binding free energy
between relatively similar compounds (Jiao et al. 2008;
Reddy and Erion 2001; Mobley and Dill 2009). Two mol-
ecules might have similar interactions with a protein, similar
strain energies, etc., but have diferent solvation properties
in water, leading to solvation-driven diferences in binding
free energies (Mobley and Dill 2009).
For highly hydrophilic ligands, the desolvation costs may
be very high and make unfavourable contributions to the
binding (Daranas et al. 2004; MacRaild et al. 2007; Syme
et al. 2010). Essentially, a charged ligand may make favora-
ble electrostatic interactions in a polar binding site, but it
may also cost a huge amount of energy to remove it from
water (Brenk et al. 2006; Gilson and Zhou 2007; Shoichet
et al. 1999). In other cases, a small modifcation to a ligand
1 3

In Silico Pharmacology
(2021) 9:38
Page 11 of 16
38
Fig. 7 a TPP showing residues within close proximity in the binding
pocket with the diferent types of interactions. Particularly the unfa-
vourable positive-positive interaction between Lysine 191 and a posi-
tively charged phosphorus (enhanced electrophilicity) in the phos-
phate group. b SUBTPP showing residues within close proximity in
the binding pocket showing more electrostatic interactions around the
phosphate groups. c The lanosterol derivative within the binding site
of the transketolase, showing more of Van der Waals interactions. d
The protein rendered as an hydrophobic surface, showing the lanos-
terol derivative (CPK rendering) with the hydrocarbon part hovering
over the entrance of the active site with the polar groups occupying
positions that are tending towards hyrophilic on the surface—noting
that the TPP and SUBTPP have preference for deeper in the bind-
ing site. The images are obtained from a midpoint of a cluster from
geometric clustering performed using the gmx cluster module in the
GROMACS program (2018.3) with the trajectory from the 50 ns sim-
ulation of the ligand with the protein
can potentially lead to afnity gains due to a change in the
desolvation cost (Kangas and Tidor 2001). Therefore, in this
study, it would have been expected that the strong electro-
static interactions between the reference compounds, TPP
and SUBTPP within the deep binding pocket would amount
to high Free Energy of Bininding but it turns out that the
cost of desolvation played a signifcant role in the reduction
of the binding energy. Furthermore, the SUBTPP being a
synthetic derivative of TPP with a methyl substituent at the
6-position hard a markedly reduced reduction in the free
energy of binding relative to the TPP (TPP is − 33 1.529
while SUBTPP is − 6.152 1.529 kJ/mol) which confrms
the observations and conclusions made by Jiao et al. (2008)
as well as Reddy and Erion (2001), for polar/charged ligands
within a deep binding pocket flled with water. The overall
binding energy for the lanosterol derivative being 5 times
1 3

38
Page 12 of 16
In Silico Pharmacology
(2021) 9:38
the sterol such that, starting with one sterol will yield an
expected ketone and other ketones borne from rearrange-
ment of double bond positions. It is expected that using
QCC which is not acidic will minimize the incidence of rear-
rangement or prevent it completely (Dhar and Singh 1977).
The observation from the oxidation of lanosterol with QCC
in this study shows that the incidence of rearrangement is
greatly minimized, since the TLC of the oxidized product
shows one major ketone and another present in trace amount.
The oxidized lanosterol was reacted with 2,4-dinitrophenyl-
hydrazine to produce the lanosterol derivative (LAN). The
lanosterol derivative was purifed by packing on a column
for chromatography (eluted with a mixture of n-hexane and
ethyl acetate at diferent ratios) which was also recrystal-
Fig. 8 Chart of amino residues that made signifcant contributions to
the Free Energy of Binding
1
lized. The LAN was characterised by UV, FTIR, H and
13C NMR.
The infrared spectra of the lanosterol derivative (LAN),
intermediate ketone and starting sterol were compared. The
absorption profles confrm that the expected reactions from
sterol to ketone to DNP derivative occurred (Table 2). A
sharp band at 3372 cm−1 characteristic of N–H stretching
vibration (which must be due to the hydrazinyl moiety) was
observed in the IR spectrum of LAN, a band at 1618 cm−1
(sharp with medium intensity) was also observed which is
attributed to the C=N bond stretching vibration. Two char-
acteristics absorption for –NO2 vibration at 1518 cm−1 and
1334 cm−1 were also observed as well as bands at 1591 cm−1
and 1471 cm−1 for C=C of benzene which show that the
DNP moiety has been incorporated into the target molecule.
The 1H NMR spectrum shows three signals in the aromatic
region which is expected for the DNP moiety: 1H, d at
δ9.13 ppm; 1H, dd at δ8.30 ppm (J=8.9 Hz) and 1H, dd at
7.95 ppm (8.4 Hz). The number of protons in the hydrocar-
bon region is characteristic of a steroid (lanosterol in par-
ticular), this represents the steroidal part of the molecule.
The region further downfeld, beyond the aromatic region is
expected to have one signal corresponding to the hydrazinyl
proton but shows two signals, close to each other, and their
integrations are fractions of a whole which suggest a form
of tautomerism occurring in solution such that the com-
pound switches between an imine form and eneamine form
(Fig. 10). This suspected tautomerism is confrmed by the
second observed signal in the alkene region of the spectrum,
where only one signal is expected but two are observed. The
13C NMR spectrum also shows the same trend (suggesting
a form of tautomerism). The hydrocarbon region matches
with that of lanosterol from literature (Emmons et al. 1989)
(Table 3).
more than that of TPP (a cofactor) shows that the lanosterol
derivative would be a potent inhibitor of the P. falciparum
transketolase and a possible antimalaria agent whit high ef-
cacy if it is bioavailable at the site of action.
Synthesis
Fig. 9 Breakdown of components of Free energy that make up the
total Free energy of binding of the ligands
The synthetic strategy of the 2,4-dinitrophenylhydrazone
derivative of oxidized lanosterol is depicted in Fig. 7. Lanos-
terol was oxidized by freshly prepared QCC which is a mild
oxidizing agent to minimize the incidence of rearrangement
or prevent it completely (Dhar and Singh 1977). Attempts
to oxidize sterols in literature, using acidic oxidants such
as Jones reagent or chromic acid usually led to a mixture
of products due to the possibility of rearrangements of
double bond position that the acidic medium afords. This
rearrangement will normally accompany the oxidation of
1 3

In Silico Pharmacology
(2021) 9:38
Page 13 of 16
38
Table 3 Comparison of 13
C
S/N
Carbons
Dinitrophenylhy-
drazone derivative
of lanosterol (pure,
experimental)
Lanosterol (Emmons
et al. 1989)
spectroscopic data of lanosterol
from literature with that of
the synthesized compound
(2, 4-dinitrophenylhydrazone
derivative of oxidized
1
2
3
4
5
6
7
8
C1
C2
C3 (C=N)
C4
C5
C6
C7
C8
C9
C10
C11
C12
C13
C14
C15
C16
C17
C18
C19
C20
C21
C22
C23
C24
C25
C26
C27
C28
C29
40.15
28.90
167.98
45.45
51.86
19.19
28.57
146.18
146.10
45.01
24.46
40.07
48.36
50.48
37.58
36.81
50.93
18.35
23.40
42.80
21.66
42.94
25.51
124.15
131.55
26.34
19.09
31.46
16.43
24.85
117.03
129.45
129.69
133.54
164.98
136.13
137.99
155.77
145.71
140.29
35.55
27.81
Absent
38.85
50.35
18.22
26.46
134.35
134.35
36.97
20.97
30.94
44.43
49.76
30.82
28.18
50.35
15.71
19.11
36.24
18.61
36.33
24.89
125.22
130.91
25.72
17.62
27.93
15.40
24.43
Absent
Absent
Absent
Absent
Absent
Absent
Absent
Absent
Absent
Absent
lanosterol)
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
C30
DNP Carbon
DNP Carbon
DNP Carbon
DNP Carbon
DNP Carbon
DNP Carbon
DNP Carbon
DNP Carbon
DNP Carbon
C=C (transient)
guidelines). This gives the beneft for dosing in the use of
the substance since any LD50 beyond 5000 mg/kg is of no
practical interest.
In‑vivo study
Lethal dose determination
The oral safety of the test compound (LAN) was ascertained
by the outcome of the LD50 assay, none of the animals died
even at the stipulated highest dose of 5000 mg/kg (OECD
1 3

38
Page 14 of 16
In Silico Pharmacology
(2021) 9:38
Fig. 10 Imine-enamine tautom-
erism of Compound A to show
the occurrence of the second
N–H group and the olefn
hydrogen
Antimalarial chemo suppressive test
Figure 11 shows the percentage parasitemia of the com-
pound at diferent doses, the lowest percentage parasitemia
for LAN, 1.33% is observed at 25 mg/kg, which is compa-
rable with that of the positive control chloroquine which
is 1.28%. As it can be seen from Fig. 12, the survival rate
observed in 25 mg/kg which has the optimal activity com-
pared favourably with that of the reference drug (chloro-
quine) too. This shows the ability of the compound to sustain
the host (murine model) after the dosing is stopped. The %
chemo suppression, Fig. 12, of the test compound (LAN) at
25 mg/kg (53.13%) is comparable with that of the positive
control (CQ) having % chemo suppression of 54.93%, while
at 50 mg/kg, it is 50.41% and at 100 mg/kg, it is 35.67%.
This shows that the compound compares favourably with the
Fig. 11 Graphical representation of % Parasitemia
Fig. 12 Graphical Representa-
tion of % Chemosuppression
with Survival Index
1 3

In Silico Pharmacology
(2021) 9:38
Page 15 of 16
38
SoftwareX 1–2:19–25. https://doi.org/10.1016/j.softx.2015.06.
001
reference drug, chloroquine, at both 25 mg/kg and 50 mg/kg.
The % Chemosuppression may be declining as the concen-
tration of the dosage increases due to a form of toxicity (at
high doses) which must have led to a reduction of survival
index (eventual death of the test animals).
Amato R, Pearson RD, Almagro-Garcia J, Amaratunga C, Lim P, Suon
S, Sreng S, Drury E, Stalker J, Miotto O, Fairhurst RM, Kwiat-
kowski DP (2018) Origins of the current outbreak of multidrug-
resistant malaria in southeast Asia: a retrospective genetic study.
Lancet Infect Dis 18:337
Brenk R, Vetter SW, Boyce SE, Goodin DB, Shoichet BK (2006) Prob-
ing molecular docking at a charged model binding site. J Mol Biol
357:1449–1470
Conclusion
Bronowska AK (2011) Thermodynamics of ligand-protein interactions:
implications for molecular design. Thermodyn Interact Stud Sol-
ids Liquids Gases. https://doi.org/10.5772/19447
In this study, the 3D model of P. falciparum transketolase
was constructed and made available for docking which gave
insight into its interaction with the synthesized compound.
The MD simulation also gave insight into the stability of
the complex with the study ligands and revealed the advan-
tage of the lanosterol derivative (LAN) binding Free Energy
in terms of solvation energy contribution to the Total Free
Energy of Binding. The oxidation of lanosterol with a mild
oxidizing agent, QCC, also reduces the extent of rearrange-
ment of pi-bond rearrangement (shift) and degradation,
this was confrmed by the TLC and the spectroscopic data
obtained for the DNP derivative (LAN). The in-vivo experi-
ments corroborated the result from in-silico study, it showed
that the dinitrophenylhydrazone derivative of oxidized
lanosterol possesses antimalarial activity comparable to
that of chloroquin at the lowest dose administered for LAN,
25 mg/kg and all the animals tested survived throughout the
course of the experiment (at the 25 mg/kg dose for LAN)
which shows that the compound is active as well as less toxic
at low dosage, which is one of the goals set in drug design.
2,4-dinitrophenylhydrazone derivative of oxidized lanosterol
can be exploited as lead in the antimalarial chemotherapy,
the compound can be modifed to get a better activity by
replacing the DNP group with other groups and creating a
variety of derivatives. The compound may also be subjected
to an in-vitro enzyme inhibition assay for transketolase in a
further study—which would require the cloning and over-
expression of the transketolase to obtain sufcient amount
for in-vitro study.
Daranas AH, Shimizu H, Homans SW (2004) Thermodynamics of
binding of Dgalactose and deoxy derivatives thereof to the L-ara-
binose-binding protein. J Am Chem Soc 126(38):11870–11876
Dhar DN, Singh AK (1977) Pyridinium Chlorochromate oxida-
tion of some steroidal system. Zeitschrift Fur Naturforschung
32B:1476–1477
Dzubak P, Haajduch M, Vydra D, Hustova A, Kvasnica M, Biedermann
D, Markova L, Urban M, Sarek J (2006) Pharmacological activi-
ties of natural triterpenoids and their therapeutic implications. Nat
Prod Rep 23(3):394–411
Egan WJ, Merz KM, Baldwin JJ (2000) Prediction of drug absorption
using multivariate statistics. J Med Chem 43:3867–3877. https://
doi.org/10.1021/jm000292e
Emmons GT, Wilson WK, Shroepfer GJ (1989) 1H and 13C NMR
assignments for lanostan-3B-ol Derivatives: revised assignments
for lanosterol. Magn Reson Chem 27:1012–1024
Gilson MK, Zhou H-X (2007) Calculation of protein-ligand binding
afnities. Annu Rev Biophys Biomol Struct 36:21–42
Jiao D, Golubkov PA, Darden TA, Ren P (2008) Calculation of protein-
ligand binding free energy using a polarizable potential. Proc Nat
Acad Sci USA 105:6290–6295
Kangas E, Tidor B (2001) Electrostatic complementarity at ligand
binding sites: application to chorismate mutase. J Phys Chem B
105:880–888
Knight DJ, Peters W (1980) The antimalarial activity of N-benzy-
loxydihydrotriazines. I. The activity of clociguanil (BRL 50216)
against rodent malaria, and studies on its mode of action. Ann
Trop Med Parasitol 74:393–404
Kumari R, Kumar R (2014) Open Source Drug Discovery Consortium,
Lynn A. g_mmpbsa–a GROMACS tool for high-throughput MM-
PBSA calculations. J Chem Inf Model 54(7):1951–1962. https://
doi.org/10.1021/ci500020m
Lorke D (1983) A new approach to tropical acute toxicity testing. Arch
Toxicol 53:275–285
MacRaild CA, Daranas AH, Bronowska A, Homans SW (2007) Global
changes in local protein dynamics reduce the entropic cost of car-
bohydrate binding in the arabinose-binding protein. J Mol Biol
368(3):822–832
Supplementary Information The online version contains supplemen-
tary material available at https://doi.org/10.1007/s40203-021-00097-8.
Marsh J, Teichmann S (2011) Relative solvent accessible surface area
predicts protein conformational changes upon binding. Structure
19:859–867. https://doi.org/10.1016/j.str.2011.03.010
Mobley DL, Dill KA (2009) Binding of Small-Molecule Ligands to
Proteins: “What You See” Is Not Always “What You Get.” Struc-
ture 17(4):489–498. https://doi.org/10.1016/j.str.2009.02.010
Morris GM, Huey R, Lindstrom W, Sanner MF, Belew RK, Goodsell
DS, Olson AJ (2009) AutoDock4 and AutoDockTools4: auto-
mated docking with selective receptor fexibility. J Comput Chem
30(16):2785–2791. https://doi.org/10.1002/jcc.21256
Nasomjai P, Arpha K, Sodngam S, Brandt SD (2014) Potential anti-
malarial derivatives from astraodorol. Arch Pharmacal Res
37(12):1538–1545
Acknowledgements The authors would like to acknowledge the Berk-
man’s Lab, Washington State University, USA for assisiting with all
the NMR experiments in this study.
Funding The entire project was self-funded.
References
Abraham MJ, Murtola T, Schulz R, Páll S, Smith JC, Hess B, Lindahl
E (2015) GROMACS: High performance molecular simulations
through multi-level parallelism from laptops to supercomputers.
1 3

38
Page 16 of 16
In Silico Pharmacology
(2021) 9:38
Nazaruk J, Borym-Kluczyk M (2015) The role of triterpenes in the
management of diabetes mellitus and its complications. Phyto-
chem Rev 14(4):675–690
Shoichet BK, Leach AR, Kuntz ID (1999) Ligand solvation in molecu-
lar docking. Proteins 34:4–16
Syme NR, Dennis C, Bronowska A, Paesen GC, Homans SW (2010)
Comparison of entropic contributions to binding in a “hydro-
philic” versus “hydrophobic” ligand−protein interaction. J Am
Chem Soc 132(25):8682–8689. https://doi.org/10.1021/ja101362u
Trott O, Olson AJ (2010) AutoDock Vina: improving the speed and
accuracy of docking with a new scoring function, efcient optimi-
zation and multithreading. J Comput Chem 31:455–461
WHO Status report on artemisinin and ACT resistance (2017) WHO/
HTM/GMP/2017.9
Nogueira CR, Lopes LMX (2011) Antiplasmodial natural products.
Molecules 16(3):2146–2190
Olliaro PL, Yuthavong Y (1999) An overview of chemotherapeutic tar-
gets for antimalarial drug discovery. Pharmacol Ther 81:91–110
Pant R, Joshi A, Maiti P, Nand M, Pande V, Chandra S (2020) Iden-
tifcation of potential mycolyltransferase Ag85C inhibitors of
Mycobacterium tuberculosis H37Rv via virtual High Throughput
Screening and binding free energy studies. J Mol Graph Model.
https://doi.org/10.1016/j.jmgm.2020.107584
World Health Organization (2010) World Malaria Report. http://www.
who.int/malaria/publications/worldmalariareport/en/
Reddy MR, Erion MD (2001) Calculation of relative binding free
energy diferences for fructose 1,6-bisphosphatase inhibitors using
the thermodynamic cycle perturbation approach. J Am Chem Soc
123:6246–6252
Young T, Abel R, Kim B, Berne BJ, Friesner RA (2007) Motifs for
molecular recognition exploiting hydrophobic enclosure in pro-
tein–ligand binding. Proc Natl Acad Sci 104(3):808–813. https://
doi.org/10.1073/pnas.0610202104
Ridley RG (2002) Medical need, scientifc opportunity and the drive
for antimalarial drugs. Nature 415:686–693
Rosenthal PJ (2001) Antimalarial chemotherapy: mechanisms of
action, resistance, and new directions in drug discovery. Humana
Press. Google Scholar, Totowa
Publisher’s Note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional afliations.
Sali A, Blundell TL (1993) Comparative protein modelling by satisfac-
tion of spatial restraints. J Mol Biol 234:779–815
1 3