International Bulletin of Drug Research., 5(8): 23-40, 2015
MOLECULAR DYNAMICS, DRUG LIKE FILTERS AND BOMAN
INDEX AS A TOOL IN RATIONAL DRUG DESIGN: A QUEST FOR
ANTITUBERCULAR TRIPEPTIDE LEADS
KANDASAMY NAGARAJAN1*, VINAY KUMAR2, SADAF JAMAL GILANI3,
ABHAY BHARDWAJ4, PARUL GROVER5, SALIL TIWARI6, UMAKANT BAJAJ7
___________________________________________________________________________
ABSTRACT
The main objective of our rational drug design is to predict the best novel tripeptide leads for
antitubercular effect from the designed set of templates with various descriptors such as
Boman index, lipinski rules and molecular dynamics simulations. We built 63 candidate
molecules from chemical templates and subjected to Boman index through online
antimicrobial peptide database. The best resulting 13 templates derived from this database
were passed them through empirical lipinski filters to assess drug like properties using
various molecular descriptors namely molecular mass, partition coefficient, Hydrogen bond
donors and Hydrogen bond acceptors respectively. Later, we conducted molecular dynamics
on the 13 templates with 20 nano second simulations of total atoms at an average temperature
298.15K in 100 steps with applying constraints to all bonds in pure water. Finally, the
conformational stability of the candidates were analyzed with measurement of total potential
energy, total Kinetic energy, Bond stretch energy, Bond angle blending energy, Torsion angle
energy, restraining torsional energy and Hydrogen bond energy. In addition, the protein–
ligand binding was tested with lennard jones energy and electrostatic energy for all the 13
templates. Among the results obtained with molecular dynamics for various non–bonded
interactions, we identified EHR, TYH, TRN, TFR, EWK, FRN and FKQ as potent tripeptide
leads with maximum conformational stability for antitubercular potency among the selected
63 template sets.
KEY WORDS
Tripeptides, Boman index, Lipinski rules, Molecular dynamics, Tuberculosis, Rational drug
design.
AUTHORS AFFILIATION
*Main author for correspondence: E-mail: [email protected]; Contact No.: +91
9997628670; Fax No.: 01232-27978 (India)
1.
Dr.Kandasamy Nagarajan, Professor & Head, Research Laboratory of Central
Instrumentation Frontiers, Department of Pharmaceutical Chemistry, KIET School of
Pharmacy, 13km. Stone, Ghaziabad Meerut Road, Ghaziabad-201206, India
2.
Dr.Vinay Kumar, Assoc. Professor & Head, ,Department of Pharmacology, KIET School
of Pharmacy, 13km. Stone, Ghaziabad Meerut Road, Ghaziabad-201206, India
3.
Dr. Sadaf Jamal Gilani, Assistant Professor, Department of Pharmaceutical Chemistry,
Glocal School of Pharmacy, New Bhagat Singh Colony, Bajoria Road, Saharanpur247001, India
4.
Mr. Abhay Bhardwaj, Assistant Professor, Department of Pharmaceutical Chemistry,
KIET School of Pharmacy, 13km. Stone, Ghaziabad Meerut Road, Ghaziabad-201206,
India.
23
International Bulletin of Drug Research., 5(8): 23-40, 2015
5.
Ms. Parul Grover, Assistant Professor , Department of Pharmaceutical Chemistry, KIET
School of Pharmacy, 13km. Stone, Ghaziabad Meerut Road, Ghaziabad-201206, India
6.
Mr. Salil Tiwari, Assistant Professor, Department of Pharmaceutical Chemistry, Saroj
Institute of technology and Management, Sitapur road, Lucknow, India.
7.
Dr. Umakant Bajaj, Professor & Principal, KIET School of Pharmacy, 13km. Stone,
Ghaziabad Meerut Road, Ghaziabad-201206, India.
INTRODUCTION
Lead molecular discovery and development is an expensive and lengthy process. For the
Pharmaceutical Industry, the number of years to bring a drug from discovery to market is
approximately 14 years, costing up to US$880 million per individual drug (Mathieu, 2001).
Given the vast size of organic chemical space (>1018 compounds) (Walters et al.,1988), drug
discovery cannot be reduced to a simple "synthesize and test" drudgery. There is an urgent
need particularly for life threatening diseases to identify and design lead like molecules from
the vast expense of what could be synthesized.
The development of new drugs with pharmacological efficacy and clinical utility is now a
major activity in the chemical and biological sciences. Successes in mid–century have led to
a level of confidence in our ability to make some predictions in the design of molecules with
desirable characteristics. Today, we are on the threshold of rational drug design based on the
occasional ability to recreate and model the molecular level scene of action of a ligand
molecule and an effector using computer graphic simulations. From such models, variations
in structure may be made with the objective of improving the drug–receptor encounters
leading to a better drug (Kier and Hall, 2002).
Molecular dynamics (MD) simulations can be used to model the motions of molecular
systems, including proteins, cell membranes and DNA, at an atomic level of detail. A
sufficiently long and accurate MD simulation could allow scientists and drug designers to
visualize for the first time many critically important biochemical phenomena that cannot
currently be observed in laboratory experiments, including the "folding" of proteins in to their
native three–dimensional structures, the structural changes that underlie protein function, and
the interactions between two proteins or between a protein and a candidate drug molecule
(Brooks and Case,1993). Such simulations could answer, some of the most important open
questions in the fields of Biology and Chemistry and have the potential to make substantial
contributions to the process of drug development.
Using MD simulations, the thermodynamic properties and time–dependent phenomena (i.e.
Kinetic) can be studied and this allows an understanding of various dynamic aspects of
biomolecular structure, recognition and function. The major strengths of MD simulations are
the possibility of its combination with statistical mechanics which connects the microscopic
simulation with macroscopic observables. Statistical mechanics can provide a rigorous
framework of mathematical expressions that can relate the distribution and motion of atoms
with macroscopic observables such as temperature, pressure, heat–capacity and free energies.
In this way, we are able to predict changes in the binding free energy of a particular drug
candidate or the mechanisms and energetic consequences of conformational changes in a
protein. Other aspects which can be studied by the aim of MD are macromolecular stability
(Tiana et al.,2004), the role of dynamics in enzyme activity (Wang etal.,2001), molecular
recognition and the properties of complexes (Brooijmans and Kuntz, 2003) and small
molecule transport (Roux, 2002).
24
International Bulletin of Drug Research., 5(8): 23-40, 2015
A novel class of Peptide deformylase (PDF) inhibitors such as N-alkyl urea hydroxamic
acids, were synthesized and evaluated for their antimycobacterial effects against the M.
tuberculosis PDF enzyme and found that some of the PDF-1 displayed antibacterial activity
against M. tuberculosis, including MDR strains with MIC90 values of < 1 µM (Copp,2003).
Thymidine monophosphate kinase from M. tuberculosis (TMPKmt) is an attractive target for
antitubercular chemotherapy. A pharmacophore hypothesis was developed based on the
substrate and known TMPKmt inhibitors and employed it to screen the Maybridge small
molecule database. The molecular docking was then performed in order to select the
compounds on the basis of their ability to form favourable interactions with the TMPKmt
active site, which results in selection of selection of 8 compounds and further evaluated for
antitubercular activity against M. tuberculosis H37 Rv in vitro. (Teo et al.,2006).
An approach was designed to access, utilize and manipulate peptide biosynthesis genes that
direct precursor construction and post assembly tailoring to create novel antitubercular
peptides such as analogues of viomycin, a peptide antibiotic produced by Streptomyces
vinaceus and a second line antitubercular agent that is often effective against MDR-TB
(Garcia,2012). Another approach based on targeting Mycobacterium tuberculosis
ribonucleotide reductase (RNR) was developed with the heptapeptide Ac-Glu-Asp-Asp-AspTrp-Asp-Phe-OH (Zabriskie,2006).
A tripeptide inhibitor has been identified using a structure based drug design approach based
on the high resolution crystal structure of mtDHFR complexed with known inhibitor
methotrexate (MTX) and cofactor NADPH is a selective inhibitor of mtDHFR (Nurbo,2007).
A docking experiment to M. tuberculosis dUTPase enzyme has been developed and resulted
in 10 new antitubercular drug candidates. (Kumar et al., 2010).
Incorporation of peptides in to Quinazolinone moiety has resulted in compounds with potent
antibacterial activities (Horvati et al., 2102). A series of shorter synthetic peptides analogues
of ovine bactenecin 5 (OaBac5), Phe-Arg-Pro-Xaa (Xaa= Phe, Met, Tyr and Trp) are
designed, synthesized and conjugated with the isonicotinic acid at the N-terminal of the tetra
peptides to study the change in the biological activity (Kaur, 2010).
In several organisms, the first barrier against microbial infections consists of antimicrobial
peptides (AMPs) which are molecules that act as components of the innate immune system.
Recent studies have demonstrated that AMPs can perform various functions in different
tissues or physiological conditions (Teixeira, 2013).
Naturally occurring proteins usually carry short peptide fragments that exhibit noticeable
biological activity against a wide variety of microorganisms such as bacteria, fungi and
protozoa. Traditional discovery of such antimicrobially active fragments (i.e. antimicrobial
peptides, AMPs) from protein repertoire is either random or led by chance (Yan, 2013).
Hence our main objective is to predict the best novel tripeptide from the designed set of 63
test analogues against the dreadful disease tuberculosis using certain molecular descriptors as
empirical filters, Boman index for protein binding potential and molecular dynamics for
conformational stability with various energy levels involving numerous atoms within the
designed molecule.
25
International Bulletin of Drug Research., 5(8): 23-40, 2015
MATERIALS AND METHODS
2.1 General
For all the selected test analogues, we have calculated the accessible topological descriptors
such as Lipinski rules as selective empirical drug filters for correlating their biopotency (Babu
et al., 2008). The above study was carried out at Supercomputing facility for Bioinformatics
and Computational Biology, IIT Delhi, Boman index of the test set analogues was calculated
directly from the online Antimicrobial Peptide Database (http://aps.unmc.edu/AP/main.php).
Molecular dynamics was performed with Schrodinger Suite 2011, Germany, supported by
Windows 7, 32 bit operating system with Intel (R) Core (TM) 2 Duo CPU E 4600 @ 2.40
GHZ and 1 GB RAM in HP 7540.
2.2 System Efficiency
We balance our design very differently from a general–purpose supercomputer architecture.
Relative to other high performance computing applications, Molecular dynamics uses much
communication and computation but surprisingly little memory. Consider an MD simulation
of 25,000 particles. If each particle requires 64 bytes of storage, then the entire architectural
state is just 1.6 MB. Divided among the 512 nodes of a typical machine, this is only 3.2 KB
per node, which would fit handily in to the L1 cache of any modern processor. We exploit
this property by using only 1GB RAM and small L1 caches on our ASIC, with all code and
data fitting on–chip in normal operation. Rather than spending silicon area on large caches
and aggressive memory hierarchies, we instead dedicate it to communication and
computation.
2.3 Template Library
Chemical templates are conceived as building blocks/structural frameworks for assembly and
generation of various tripeptide analogue leads. We have built a structure based template
library for the design of tri–peptide leads and created a set of 63 templates using 2D viewer in
Schrodinger Suite 2011. The selected test set are given subsequently in Table 1.
2.4 Protein Binding Potential for Preliminary Anti-tubercular Screening
All the 63 inbuilt templates have an input of 3 amino acid residues. The Boman index
estimates the potential for a protein to bind to other proteins (Ivanciuc,2002). In other words,
a high Boman index value indicates that the designed lead will be multifunctional or play a
variety of different roles within the cell due to its ability to interact with a wide range of
proteins. All the 63 tripeptide leads were subjected to Boman index according to the online
Antimicrobial Peptide Database (http://aps.unmc.edu/AP/main.php). Other design principles
taken into account included the total net charge of tripeptide leads, which are believed to be
important for establishing antimicrobial action (Boman,2003). Following their design, peptide
leads were selected for further empirical drug like lipinski filters by eliminating wrong
candidates through screening.
2.5 Molecular descriptors and drug like filters
A successfully lead discovery strategy must ensure bioavailability from the very start in
generating leads while eliminating wrong candidates from considerations. Lipinski rules of
26
International Bulletin of Drug Research., 5(8): 23-40, 2015
five (Yan and Adams,1998) helps in distinguishing between drug like and non–drug like
molecules. The Lipinski rules predicts the high probability of success or failure due to drug
likeness for molecules complying with 2 or more of the following rules:–
1. Molecular mass less than 500 Dalton.
2. High Lipophilicity (Log p less than 5).
3. Less than 5 Hydrogen bond donors.
4. Less than 10 Hydrogen bond acceptors.
5. Molar refractivity should be between 40–130.
We have introduced only 13 template leads to this stage which were already screened through
Boman index from the 63 templates and the structure inbuilt are given in Fig. 1. The selected
13 templates, viz., TFR, TWK, TFK, TYH, THR, TRN, FWK, FKQ, FHR, FRN, EWK, EHR
and TKQ was subjected to some empirical computational Lipinski filters based on drug like
properties/molecular descriptors.
The above mentioned molecular descriptors could act as computational filters based upon
their accepted limits to screen the candidate compounds. Introduction of filters facilitates
computational tractability by restricting the chemical space for potential candidates, saving
much time and cost in new lead discovery.
2.6 Molecular Dynamics
In rational drug design, one of the most important problems to be solved is the description of
the molecular aspects and the related energetic which are at the basis of the interactions
between proteins or more general macromolecular receptors, and molecules which could be
either endogenous natural ligands or a drug.
Given the structure of a biomolecular system i.e. the coordinates of the constituent atoms,
there are various computational methods able to investigate the dynamics of the molecular
system. All the dynamics methodologies employed are highly dependent upon the description
of a suitable potential energy function to describe the energy of the system with respect to the
molecular degrees of freedom.
In particular, the choice of an appropriate energy function for describing the intermolecular
and intramolecular interactions is critical for a successful MD simulation. In conventional
MD, the energy function is calculated by means of molecular mechanics methods, thus
considering the atomic motion only from a nuclear point of view according to the Born–
oppheneimer approximation (Lipinsky, 1997).
An MD computation simulates the motion of a collection of atoms (the chemical system)
over a period of time. The chemical system may consist of a peptide and its surrounding
environment (solvent, usually water), and may also include other types of molecules, such as
lipids, carbohydrates, nucleic acid or drug molecules. This MD computation breaks time in to
a series of discrete time steps, each representing a few femto seconds of simulated time. A
time step has two major phases. Force calculation computes the force on each particle due to
27
International Bulletin of Drug Research., 5(8): 23-40, 2015
other particles in the system. Integration uses the net force on each particle to update that
particle's position and velocity.
In recent years, molecular dynamics (MD) simulations have been increasingly used to
understand the complex conformational equilibria of polypeptides in solution and to predict
structural preferences (Born and Oppenheimer, 1927).
Thirteen systems were simulated with molecular dynamics. For each tripeptide, 20–ns
simulations for total atoms at an average temperature 298.15K in 100 steps with applying
constraints to all bonds in pure water was performed. All MD simulations were performed in
the constant surface area ensemble with the NAMD2 program (Garcia and
Sanbonmatsu,2001) and the CHARMM 27 force field (Philips et al.,2005).The binding free
energy between ligand and macromolecule complex was established to predict the
antitubercular potency in terms of both entropy and enthalpy and shall confirm the best
antitubercular activity with maximum binding energy using AMBER 94 program. A short–
range cutoff of 12A° was used for non–bonded interactions, and long–range electrostatic
interactions were treated with the Particle Mesh Ewald method (Mackerrell et al.,1998).
During the run, translational and rotational motions for the system are removed and the whole
simulations were integrated with verlet integrator.
The structural and dynamical properties of the resulting ensemble of structures were analyzed
accordingly with Schrodinger suite, 2011 in order to explore the conformational space of the
candidate molecule by measuring the total energy of system, total potential energy, total
kinetic energy, bond stretch energy, angle bending energy, torsion angle energy, restraining
energy for torsions, H–bond energy and various types of Lennard Jones energy and
electrostatic energy. Finally, the conformational stability of all the 13 tripeptide leads were
compared and a suitable template lead having maximum stability and antitubercular reactivity
was identified by this MD simulation techniques.
RESULTS & DISCUSSION
3.1 Boman Index
The results of Boman index and total net charge of subjected 63 tripeptide leads were
summarized in Table1. The peptide which has net negative charge has very little chance to be
an antimicrobial peptide. The peptide which has neutral charge may have antimicrobial
activity. But most active antimicrobial peptides have net charge between +3 to +8. If the
peptide sequence has very low ratio of hydrophobic amino acid, it has little chance to be an
antimicrobial peptide.
Table 1
Boman index and total net charge for template library analogues designed for
initial Antitubercular Screening.
Test Set
Total Net
Charge
Membrane Penetration
Power for Bioactivity
Boman Index
(K.cal/Mol)
Margin of
Interaction
with
Proteins
1.
TFK
+1
M
1.71
N
2.
TFR
+1
M
4.83
W
S.No.
28
International Bulletin of Drug Research., 5(8): 23-40, 2015
3.
TFH
+1
M
1.41
N
4.
TGA
0
L
–0.06
NS
5.
TAL
0
L
–1.38
NS
6.
TMF
0
L
–0.92
NS
7.
TLM
0
L
–1.56
NS
8.
TFW
0
L
–0.91
NS
9.
TWK
+1
M
1.92
N
10.
TKQ
+1
M
4.55
W
11.
TQE
–1
VL
4.97
W
12.
TES
–1
VL
4.26
W
13.
TSP
0
L
1.99
N
14.
TPV
0
L
–0.49
NS
15.
TVI
0
L
–2.13
NS
16.
TIC
0
L
–1.21
NS
17.
TCY
0
L
0.47
LS
18.
TYH
+1
M
2.45
N
19.
THR
+2
H
7.38
W
20.
TRN
+1
M
8.04
W
21.
TND
–1
VL
5.97
W
22.
TDT
–1
VL
4.62
W
23.
TTG
0
L
1.39
N
24.
FGA
0
L
–1.91
NS
25.
FAL
0
L
–3.23
NS
26.
FLM
0
L
–3.41
NS
27.
FMF
0
L
–2.77
NS
28.
FFW
0
L
–2.76
NS
29.
FWK
+1
M
0.07
LS
30.
FKQ
+1
M
2.7
N
31.
FQE
–1
VL
3.12
N
29
International Bulletin of Drug Research., 5(8): 23-40, 2015
32.
FES
–1
VL
2.4
N
33.
FSP
0
L
0.13
LS
34.
FPV
0
L
–2.34
NS
35.
FVI
0
L
–3.98
NS
36.
FIC
0
L
–3.06
NS
37.
FCY
0
L
–1.37
NS
38.
FYH
+1
M
0.6
LS
39.
FHR
+2
H
5.53
W
40.
FRN
+1
M
6.19
W
41.
FND
–1
VL
4.12
W
42.
FDT
–1
VL
2.77
N
43.
FTG
0
L
–0.45
NS
44.
EGA
–1
VL
1.35
LS
45.
EAL
–1
VL
0.02
LS
46.
ELM
–1
VL
–0.15
NS
47.
EMF
–1
VL
0.49
LS
48.
EFW
–1
VL
0.49
LS
49.
EWK
0
L
3.34
N
50.
EKQ
0
L
5.96
W
51.
EQE
–2
VL
6.38
W
52.
EES
–2
VL
5.67
W
53.
ESP
–1
VL
3.4
N
54.
EPV
–1
VL
0.92
LS
55.
EVI
–1
VL
–0.71
NS
56.
EIC
–1
VL
0.2
LS
57.
ECY
–1
VL
1.88
N
58.
EYH
0
L
3.86
N
59.
EHR
+1
M
8.79
W
60.
ERN
0
L
9.45
W
30
International Bulletin of Drug Research., 5(8): 23-40, 2015
61.
END
–2
VL
7.39
W
62.
EDT
–2
VL
6.03
W
63.
ETG
–1
VL
2.81
N
VL
 Very Low Penetration Power
4.0>
W
L
 Low Penetration Power
1.5–4  N
M
 Moderate Penetration Power
1<
 LS
H
 High Penetration Power
0<
 NS
W
 Wide margin of interaction
N
 Narrow margin of interaction
LS
 Less Satisfactory margin of interaction
NS
 No satisfactory interaction
The results from Table 1 clearly indicated that the templates subjected to drug design have
shown very low to high penetration power with their corresponding total net charge ranging
from –2 to +2 respectively. In addition, the Boman index values in K.cal/mole for the above
63 templates showed wide margin of interaction with proteins and narrow margin of
interaction with proteins. Some of the templates doesn't show satisfactory interaction with
proteins.
Among the 63 templates subjected to Boman index, only 13 template analogues are selected
with the rational drug design approach as far as the antitubercular potency is concerned and
these selected analogues along with their corresponding hydrophobic ratios are depicted as
final outcome with Table 2 with their structure inbuilt from Fig. 1.
Table 2:
Rational Drug Design outcome for Antitubercular Tripeptide leads with
their Hydrophobicity ratios.
Selected Lead Set
Hydrophobic
Ratio (%)
Total Net
Change
Boman Index
(K.cal/Mol)
1.
TFR
33
+1
4.83
2.
TWK
33
+1
1.92
3.
TFK
33
+1
1.71
4.
TYH
0
+1
2.45
5.
THR
0
+2
7.38
6.
TRN
0
+1
8.04
7.
FWK
66
+1
0.07
8.
FKQ
33
+1
2.7
9.
FHR
33
+2
5.53
10.
FRN
33
+1
6.19
11.
EWK
33
0
3.34
12.
EHR
0
+1
8.79
13.
TKQ
0
+1
4.55
S.No.
31
International Bulletin of Drug Research., 5(8): 23-40, 2015
Figure 1:
Inbuilt structure of selected tripeptide leads from their best predicted
total net charge and Boman index values.
As from the results of Table 2, we observed that certain tripeptide leads fulfill the designed
characteristic descriptors, viz., Hydrophobic ratio, total net charge and Boman index. Hence,
it is concluded that the designed tripeptide leads THR, TRN, FWK, FHR, EWK and TKQ are
found to have theoretically active antitubercular effect and has to be further tested in vitro to
predict their antitubercular effect after synthesizing the same corresponding leads.
3.2 Molecular descriptors and drug like filters.
The results of Lipinski rules for the 13 template leads are indicated in Table 3.
32
International Bulletin of Drug Research., 5(8): 23-40, 2015
Table 3: Molecular descriptors of test data set analogues by Lipinski rules.
Observed H-bonds
S.
No.
Test
Set
Number of Hbond donor
Number of Hbond acceptor
Drug like
fulfillment with
respect to
Molecular
Mass
Log P
H-bond H-bond
donor acceptor
Molar
Refractivity
Drug like fulfillment with
respect to
Mol.
Mass
Log
P
Molar
Refractivity
Final
Remark
1.
TFR
0
4
S
S
393
0
0
S
S
NS
PDL
2.
TWK
0
5
S
S
403
0
0
S
S
NS
PDL
3.
TFK
0
5
S
S
365
0
0
S
S
NS
PDL
4.
TYH
0
5
S
S
395
0
0
S
S
NS
PDL
5.
THR
0
4
S
S
385
0
0
S
S
NS
PDL
6.
TRN
0
5
S
S
69.117
–1.578
–69.117
S
S
S
DL
7.
FWK
0
3
S
S
4.47
0
0
S
S
NS
PDL
8.
FKQ
0
4
S
S
391
0
0
S
S
NS
PDL
9.
FHR
0
3
S
S
429
0
0
S
S
NS
PDL
10.
FRN
0
4
S
S
407
0
0
S
S
NS
PDL
11.
EWK
0
5
S
S
431
0
0
S
S
NS
PDL
12.
EHR
0
5
S
S
413
0
0
S
S
NS
PDL
13.
TKQ
0
5
S
S
347
–0.75
68.991
S
S
S
DL
S  Satisfied; NS  Not satisfied; DL  Drug like characteristics; PDL  Partial Drug like characteristics.
33
International Bulletin of Drug Research., 5(8): 23-40, 2015
As per Lipinski rules, we observed that 2 out of 13 test sets are found to be drug like
molecule, which clearly indicated that the drug design what we have performed is a rational
drug design. Among the test set of 13 analogues, TKQ and TRN were fulfilling all the
molecular descriptors, viz., H–bond donors, H–bond acceptors, Molecular mass and Log P
tested with empirical Lipinski filter. The remaining 11 test analogues namely TFR, TWK,
TFK, TYH, THR, FWK, FKQ, FHR, ERN, EWK and EHR exhibited partial drug like
characteristics as screened through empirical Lipinski drug filter, since they doesn't fulfill all
the molecular descriptors used in the study.
3.3. Molecular dynamics
All the 13 templates are subjected to molecular dynamics simulations and their corresponding
final energy reports are summarized in Table 4.
Computation of absolute binding free energies from atomic level descriptors of the systems is
a formidable task (Darden et al.,1998). In a phenomenon logical view, the net binding free
energy may be considered to be a sum of the free energy changes due to the following
contributions, (i.e.), the Vander Waals interactions between the protein and the inhibitor
indicating the influence of shape complementarities and packing effects and net electrostatics
which includes interactions between partial or full charges, hydrogen bonds and electrostatics
of desolvation upon binding and added salt effects.
Molecular dynamics simulations were configured on the unbound species (free protein and
free ligand), with explicit solvent under ambient temperature and pressure conditions. In a
nutshell, the electrostatic contribution to the interaction energy is computed via Coulomb's
law with a sigmoidal dielectric function. The vander waals interactions are modeled (Jayaram
et al.,2002) using a Lennard-Jones potential between the atoms of the protein and ligand.
The energy function described in Table 4 enables evaluation of the total non-bonded
interaction energy of a protein-ligand complex in aqueous environment from the cartesian
coordinates of all the atoms.
All the total atoms present in the tripeptide were subjected to molecular dynamics
simulations. The lower the potential energy of molecule, the more stability of the lead.
34
International Bulletin of Drug Research., 5(8): 23-40, 2015
Table 4: Final energy reports of Tripeptide template leads for conformational stability using MD simulations.
S.
No.
Test
Set
1.
TFR
2.
TWK
3.
TFK
4.
TYH
5.
THR
6.
TRN
7.
FWK
8.
FKQ
9.
FHR
10.
FRN
11.
EWK
12.
EHR
13.
TKQ
Total No. of
Atoms &
System
Temperature
(°K)
70
(314.408)
72
(311.743)
68
(318.977)
64
(290.389)
67
(305.370)
64
(290.670)
78
(305.172)
71
(308.613)
73
(263.436)
70
(311.929)
73
(315.580)
68
(336.811)
65
(309.179)
Total
Energy
of
System
(K.cal/
Mol)
7.663
–01
2.413
+01
2.811
+01
–1.571
+00
1.048
+01
–2.479
+01
2.885
+01
–1.417
+01
2.289
+01
–3.097
+01
–2.383
+01
–3.942
+01
–1.930
+00
Total
Potential
Energy
(K. cal/
mol)
Total
Kinetic
Energy
(K. cal/
mol)
Bond
Stretch
Energy
(K. cal/
mol)
Angle
Bending
Energy
(K. cal/
mol)
Torsion
Angle
Energy
(K. cal/
mol)
Restraining
Energy for
Torsion
–4.109
+01
–1.830
+01
–1.309
+01
–3.648
+01
–2.836
+01
–6.031
+01
–1.602
+01
–5.587
+01
–1.349
+01
–7.250
+01
–6.741
+01
–8.292
+01
–4.033
+01
4.186
+01
4.244
+01
4.120
+01
3.491
+01
3.884
+01
3.552
+01
4.488
+01
4.170
+01
3.638
+01
4.153
+01
4.359
+01
4.351
+01
3.84
+01
1.329
–11
1.301
–11
1.397
–11
1.260
–11
1.437
–11
1.376
–11
1.685
–11
1.272
–11
1.497
–11
1.422
–11
1.523
–11
1.821
–11
1.175
–11
1.845
+01
2.014
+01
1.439
+01
1.587
+01
2.23
+01
1.993
+01
2.225
+01
2.083
+01
4.471
+01
1.787
+01
2.98
+01
2.317
+01
1.996
+01
1.56
+01
2.397
+01
2.339
+01
2.785
+01
2.904
+01
2.779
+01
1.542
+01
1.163
+01
2.469
+01
1.713
+01
1.444
+01
1.707
+01
2.505
+01
0
+00
0
+00
0
+00
0
+00
0
+00
0
+00
0
+00
0
+00
0
+00
0
+00
0
+00
0
+00
0
+00
1, 4
Lennard
Jones
Energy
(K. cal/
mol)
2.654
+01
2.618
+01
2.452
+01
2.527
+01
1.984
+01
2.069
+01
2.935
+01
2.240
+01
2.635
+01
2.231
+01
2.086
+01
1.778
+01
1.970
+01
1, 4
Electro
Static
Energy
(K. cal/
mol)
–1.186
+02
9.778
+01
8.913
+01
6.391
+01
–1.319
+02
–1.561
+02
1.070
+02
7.941
+01
–1.353
+02
–1.553
+02
1.161
+02
–1.162
+02
7.354
+01
Lennard
Jones
Energy
(K. cal/
mol)
Electro
Static
Energy
(K. cal/
mol)
H–
bond
Energy
(K. cal/
mol)
–9.115
+00
–9.861
+00
–7.004
+00
–9.584
+00
–5.892
+00
–5.127
+00
–1.050
+01
–9.595
+00
–1.141
+01
–5.229
+00
–8.192
+00
–7.375
+00
–6.371
+00
2.598
+01
–1.720
+02
–1.575
+02
–1.598
+02
3.825
+01
3.247
+01
–1.796
+02
–1.805
+02
3.749
+01
3.076
+01
–2.336
+02
–1.732
+01
–1.722
+02
0
+00
0
+00
0
+00
0
+00
0
+00
0
+00
0
+00
0
+00
0
+00
0
+00
0
+00
0
+00
0
+00
Non–
bonded
Count
2058
2176
1923
1691
1872
1695
2590
2115
2256
2061
2242
1933
1743
35
International Bulletin of Drug Research., 5(8): 23-40, 2015
Table 5: Binding energy calculation by MD Simulation (Amber 94) using ligand and macromolecule comple.
(Dalton)
No. of
atoms
No. of
Bonds
No. of
Molecule
Intermolecular
Vander Waal’s
Energy(K.cal/mol)
Binding
Energy
(K.cal/mol)
Time
(Sec)
a=15.763,b=14.045, c=4
422.237
32
29
4
265.547
274.055
0.27
EWK
a=18.473,b=16.358, c=4
440.272
34
32
4
32.0292
49.898
0.55
3.
FHR
a=17.539,b=14.709 ,c=4
438.282
34
31
5
44.3253
44.138
0.39
4.
FKQ
a=16.787,b=16.516, c=4
402.243
31
29
3
107.845
107.527
0.34
5.
FRN
a-19.453,b=13.303, c=4
416.253
32
30
3
63.0252
91.051
0.27
6.
FWK
a=18.829,b=15.565, c=4
456.317
36
36
4
36.8418
36.484
0.38
7.
TFK
a=15.641,b=15.467, c=4
374.233
29
26
4
26.0304
29.958
0.31
8.
TFR
a=17.274,b=15.467, c=4
402.246
31
28
4
168.257
168.115
0.33
9.
THR
a= 14.634,b=15, c=4
394.227
30
26
5
86.787
86.674
0.84
10.
TKQ
a=14.156,b=17.176, c=4
358.188
27
23
4
37.926
37.802
0.23
11.
TRN
a=17.552,13.677, c=4
372.197
28
23
5
182.567
182.501
0.26
12.
TWK
a=16.455,b=16.348, c=4
412.262
32
29
5
100.98
100.809
0.34
13.
TYH
a=16.037,b=14.324, c=4
404.239
31
29
4
251.983
258.11
0.32
S.No.
Test
Compound
Crystal Cell
Mass
α=90; β=90; =90
1.
EHR
2.
36
International Bulletin of Drug Research., 5(8): 23-40, 2015
Among the 13 templates simulated, only 5 tripeptide leads have shown minimum potential
energy, viz., EHR (–8.292 K. cal/mol); FRN (–7.250 K. cal/mol); EWK (–6.741 K. cal/mol);
TRN (– 6.031 K. cal/mol); FKQ (– 5.587 K. cal/mol); and provded to have maximum
conformational stability.
Bond energies represent a state of potential chemical energy. Bond energy must be
introduced to break a bond. The more bond energy, the more stability has enhanced for the
lead. This has been confirmed with the Bond stretch energy obtained for following template
leads, viz., EHR (1.821 K. cal/mol); EWK (1.523 K. cal/mol); FRN (1.422 K. cal/mol); FHR
(1.497 K. cal/mol) and FWK (1.685 K. cal/mol) respectively.
Higher torsion angle energy leads to distorted local conformation. The low torsion angle
energy was observed FKQ (1.163 K. cal/mol); FWK (1.542 K. cal/mol); EHR (1.707 K.
cal/mol) and EWK (1.444 K. cal/mol) correspondingly. Without a Bond angle bending
energy, a stable configuration cannot be found. The maximum bond angle energy was
observed with FHR (4.471 K. cal/mol); EHR (2.317 K. cal/mol); EWK (2.298 K. cal/mol)
and FRN (1.787 K. cal/mol) respectively.
The loss in surface area from the minimum number of atoms (64) to maximum number of
atoms (78) among the 13 tripeptide leads indicating that the net loss in surface area is
favourable for binding. The net contribution of Vanderwaals component with electrostatic
interactions indicating higher negative values in EWK (– 10.528 K. cal/mol); FKQ (– 11.4 K.
cal/mol); TYH (–11.182 K. cal/mol) and TWK (– 11.581 K. cal/mol). This was further
observed with the non-bonded count for the above template leads as seen from Table 4. All
the above data’s have shown the potential energy prediction of protein-ligand design. The
Lennard-Jones parameters adopted for the system have improved the sensitivity of potential
energy calculation.
The maximum binding energy calculation by MD Simulation (Amber 94) using ligand and
macromolecule complex is between EHR (274.055 K.cal/mol) (Fig.2) followed by TYH
(258.11 K.cal/mol) (Fig.3),TRN (182.501 K.cal/mol) and TFR (168.115 K.cal/mol)
respectively (Table 5).
Figure 2: EHR: Ligand Interaction Diagram.
37
International Bulletin of Drug Research., 5(8): 23-40, 2015
Figure 3: TYH: Ligand Interaction Diagram.
The discovery of new pharamceuticals via computer modeling is one of the key challenges in
modern medicine. Computational methods are anticipated to play a pivotal role in exploiting
the structural and functional information to understand specific molecular recognition events
of the target peptide molecule with candidate hits leading ultimately to the design of
improved leads for the target. In this article, we sketch a realization of the various stages in
the pathway proposed with our own research to demonstrate the way in which an iterative
process of computer based rational drug design can aid in developing potent tripeptide leads
such as EHR, TYH, TRN, TFR, EWK, FRN and FKQ with maximum conformational
stability by MD simulations for potent antitubercular effect as final outcome from the
selected 63 inbuild templates.
ACKNOWLEDGEMENT
The authors are very much thankful to Dr. Sraben Mukherjee, Director, KIET Group of
Institutions and Dr. Umakant Bajaj, Principal, KIET School of Pharmacy for providing moral
support. We also remain thankful to Dr. R. Raghu, Senior Director, Schrodinger Ltd.,
Bangalore and Research Department, Jadavpur University for providing the technical datas in
timely manner.
38
International Bulletin of Drug Research., 5(8): 23-40, 2015
REFERENCES
1.
Babu Narendra S.N., Rangappa K.S. Ind. J. Chem. 47B (2):297-304, 2008.
2.
Boman H. J. Int. Med. 254:197-215, 2003.
3.
Born M., Oppenheimer J.R. Ann. Phys. 84: 457-484, 1927.
4.
Brooijmans N., Kuntz I.D. Ann. Rev. Biophys. Biomol. Struct. 32:355–373, 2003.
5.
Brooks C.L., Case D.A. Chem. Rev. 93: 2487–2502, 1993.
6.
Copp B.R. Nat. Prod. Rep. 20:535-557, 2003.
7.
Darden T., York D., Pedersen L. J. Chem. Phys. 98:10089-10092, 1993.
8.
Garcia A., Garcia V.B., Nicolas J.P.P., Rivera G. Eur. J. Med. Chem. 49:1-23, 2012.
9.
Garcia A.E., Sanbonmatsu K.Y. Proteins. Struct. Funct. Genet. 42:345-354, 2001.
10.
Ghalia M.B. Amr. A. Am. Acids. 26 (3): 283-289, 2004.
11.
Horvati K., Bacsa B., Szabo N., David S., Mezo G., Grolmusz V., Vertessy B.,
Hudecz F., Bosze S. Bioconjugate. Chem. 23 (5): 900-907, 2012.
12.
Ivanciuc O. Rev. Roum. Chim. 47: 675-686, 2002.
13.
Jayaram B., Mc Connell K.J., Dixit S.B., Das A., Beveridge D.L. J. Comput. Chem.
23: 1-14, 2002.
14.
Kaur P., Kaur A., Kaur R., Kaur K. J. Global Pharm. Tech. 2 (4): 35-39, 2010.
15.
Kier L.B., Hall L.H. Croatica Chem. Acta. 75(2): 371-382, 2002.
16.
Kumar A., Chaturvedi V., Bhatnagar S., Sinha S., Siddiqi M.I. J. Chem. Inf. Model.
49 (1): 35-42, 2009.
17.
Kumar M., Vijayakrishnan R., Subba Rao G. Mol. Divers. 14 (3): 595-604, 2010.
18.
Lipinsky C.A., Lombardo F., Dominy B.W., Feeney P.J. Adv. Drug Del. Rev. 23: 325, 1997.
19.
Mackerrell A.D., Bashford D., Bellott M., Dunbrack R.L., Evanseck J.D. J. Phys.
Chem. B.102: 3586-3616, 1998.
20.
Mathieu M.P. Parexel's Pharmaceutical R & D Statistical Sourcebook. Blackwell
Publishers. UK. 96, 2001.
21.
Nurbo J., Roos A.K., Muthas D., Wahlstrom E., Ericsson D.J., Lundstedt T., Unge T.,
Karlen A. J. Pept. Sci. 13 (12): 822-832, 2007.
39
International Bulletin of Drug Research., 5(8): 23-40, 2015
22.
Philips J.C., Braun R., Wang W., Gumbart J., Tajkhorshid E. J. Comput. Chem. 26:
1781-1802, 2005.
23.
Roux B. Acc. Chem. Res. 35: 366-375, 2002.
24.
Teixeira D.L., Silva N.O., Migliolo L., Fensterseifer C.M.I., Octavio L., Franco L.O.
Peptides. 42:144–148, 2013.
25.
Teo J.W. P., Thayalan P., Beer D., Yap A.S.L., Nanjundappa M., Ngew X.,
Duraiswamy J., Liung S., Dartois V., Schreiber M., Hasan S., Cynamon M., Ryder
N.S., Weidmann B., Bracken K., Dick T., Mukherjee K. Antimicrob. Agents
Chemother. 50 (11): 3665-3673, 2006.
26.
Tiana G., Simona F., De Mori G.M.S., Broglia R.A., Colombo G. Protein Sci. 13:
113–123, 2004.
27.
Walters P.W., Stahl M.T., Murcko M.A. Drug Discov. Today. 3:160-164, 1988.
28.
Wang A., Donini O., Reyes C.M., Kollman P.A. Ann. Rev. Biophys. Biomol. Struct.
30: 211–220, 2001.
29.
Yan L., Adams M. E. J. Biol. Chem. 273: 2059-2066, 1998.
30.
Yana L., Yanb Y., Liua H., Lv Q. BioSys. 113:1– 8, 2013.
31.
Zabriskie T.M. Biosynthesis of antitubercular nonribosomal peptides:Grant
5R01GM069320-04, 2006.
40