Академический Документы
Профессиональный Документы
Культура Документы
156
CHAPTER -5
MOLECULAR DOCKING STUDIES
158
5.2.5. Other physical factors - Conformational changes in the protein and the
ligand are often necessary for successful docking.
the experimental binding modes from all other modes explored through the
searching algorithm (Kitchen et al. 2004).
For example:
Empirical scoring function of any docking program
Fitness = vdW + Hbond + Elec
Binding Energy
G bind = Gvdw + Ghbond + Gelect + Gconform + G tor + G sol
5.3. Types of Docking -The following are type of docking used often.
5.3.1. Lock and Key or Rigid Docking In rigid docking, both the internal
geometry of the receptor and ligand is kept fixed during docking.
5.3.2. Induced fit or Flexible Docking - In this model, both the ligand and side
chain of the protein is kept flexible and the energy for different conformations of
the ligand fitting into the protein is calculated. For induced fit docking, the main
chain is also moved to incorporate the conformational changes of the protein
upon ligand binding. Though it is time consuming and omputationally
expensive, yet this method can evaluate many different possible conformations
which make it more exhaustive and possibly simulate real life phenomenon and
hence trustworthy.
5.3.3. Applications
A binding interaction between a small molecule
ligand and
160
scoring functions, GoldScore, ChemScore, ASP and user defined Score which
allows users to modify an existing function or implement their own scoring
function, with respect to using the GoldScore, ChemScore or ASP scoring
functions one may give a successful prediction where the other fails, but their
overall success are about the same.
5.4.2. Schrodinger Glide
Glide docks flexible ligands into a rigid/flexible receptor structure by
rapid sampling of the conformational, orientational, and positional degrees of
freedom of the ligand. There are three modes of running Glide which differ in
how ligand degrees of freedom are sampled and in the scoring function
employed. All three modes generate an exhaustive set of conformers for a ligand
and employ a series of hierarchical filters to enable rapid evaluation of ligand
degrees of freedom. The SP GlideScore scoring function is used to rank
compounds docked by SP or HTVS Glide. XP Glide begins with SP Glide
docking and then refines the predicted docking modes using an anchor-and-grow
algorithm to more thoroughly sample ligand degrees of freedom. The XP
GlideScore scoring function includes special recognition terms to identify and
reward structural motifs important to binding.
5.4.3. AutoDock Vina
AutoDock Vina is a comparatively new open-source program for drug
discovery, molecular docking and virtual screening, offering multi-core
capability, high performance, enhanced accuracy and ease of use. AutoDock
Vina has been designed and implemented by Dr. Oleg Trott (2010) in the
Molecular Graphics Lab at The Scripps Research Institute. AutoDock Vina
automatically calculates the grid maps and clusters the results in a way
transparent to the user.
162
5.4.3.1. Features
Accuracy
AutoDock Vina significantly improves the average accuracy of the
binding mode predictions compared to AutoDock4. Additionally, AutoDock
Vina has been tested against a virtual screening benchmark called the Directory
of Useful Decoys by the Watowich group, and was found to be "a strong
competitor against the other programs and at the top of the pack in many cases".
It should be noted that all six of the other docking programs, to which it was
compared, are distributed commercially.
For its input and output, Vina uses the PDBQT molecular structure file
format used by AutoDock. PDBQT files can be generated (interactively or in
batch mode) and viewed using MGL Tools. Other files, such as the AutoDock
and AutoGrid parameter files (GPF, DPF) and grid map files are not needed.
c) Metal ionization states are corrected to ensure proper formal charge and
force field treatment
d) Bond orders are enumerated to HET groups
e) Co-crystallized water molecules are removed at the user's discretion.
f) Residues with missing atoms or multiple occupancies are highlighted.
g) Quickly and easily determine the most likely ligand protonation state as
well as the energy penalties associated with alternate protonation states
h) Optimal protonation states for histidine residues are determined.
i) Potentially transposed heavy atoms in arginine, glutamine, and histidine
side chains are corrected.
j) The protein's hydrogen bond network is optimized by means of a
systematic, cluster-based approach, which greatly decreases preparation
times.
k) A restrained minimization is performed that allows hydrogen atoms to be
freely minimized, while allowing for sufficient heavy-atom movement to
relax strained bonds, angles, and clashes.
The ligands are prepared using Ligprep module with the following functions:
Chemically correct models: LigPrep generates accurate, energy
minimized 3D molecular structures. LigPrep also applies sophisticated rules to
correct Lewis structures and to eliminate mistakes in ligands in order to reduce
downstream computational errors.
Maximum diversity: LigPrep optionally expands tautomeric and
ionization states, ring conformations, and stereoisomers to produce broad
chemical and structural diversity from a single input structure.
The prepared protein is loaded into maestro environment and the active
site is defined. Grid centre is defined for the active site and box sizes are set. The
next step is to generate glide grid. After successful generation of the grids,
164
prepared ligands are loaded into maestro. Ligands are kept flexible, while the
protein is rigid and docking started with extra precision mode (XP mode). The
docking calculation generated few poses for each ligand. The selection of the
best pose was done on the interaction energy between the ligand and the protein
as well as on the interactions the ligand shows with experimentally proved
important residues.
The docking results for all the inhibitory compounds under study are
reported in Table.5.1. The compounds bind in the pocket defined by Asn 178,
Met 254, Lys 250, cytosine 111 and adenine 11 from the DNA. All the
compounds are observed to exhibit hydrogen bonds with the DNA molecule
(Figure.5.1.). The best compound forms hydrogen bond with amino group of
both adenine 11 and cytosine 111 (Figure 5.2.). Only one of the naphthoquinone
monomer moieties participate in the hydrogen bond formation while the other
one does not. The N-dimethylamino compound, apart from forming the
hydrogen bonds binds electrostatically with the DNA molecule. The molecular
electrostatic potential map of the binding site in the receptor has been generated
to prove this. The partial positive charge of the protonated nitrogen in the ligand
is completely surrounded by the predominantly negatively charged surface in the
binding site (Figure.5.3.)
Ligand
IC50
simultaneous
Glide score
Kcal/mole
51
-6.26
37.5
-9.00
74
-4.28
70
-8.69
No inhibition
N.D.
Lawsone Dimer
No inhibition
N.D.
165
166
167
docking calculation generated ten poses. The selection of the best pose was done
on the interaction energy between the ligand and the protein as well as on the
interactions the ligand shows with experimentally proved important residues.
Blue doted lines are showing hydrogen bond interaction with COX 2
enzyme. Apart from this, the cationic side chain of Arg 106 forms -cation
(guanidinium moiety of Arg 106) interaction with thiazole ring of the ligand.
Moreover, Ser 339 also forms a - interaction with one of the naphthoquinone
rings. These interactions are marked with yellow lines.
Figure.5.4. Interaction of best compound with COX-2. Here yellow and blue
doted lines showing hydrogen bond interaction with enzyme.
Yellow solid lines show pie-pie and pie-cation interactions.
168
topoisomerase-I
is
having
four
major
domains.
1)
The NH2 terminal domain is comprised between Met-1 and lys-197, and seems
dispensable for in-vitro activity. Residues Glu-198 to Ile-651 form the highly
conserved core domain followed by a short un-conserved linker (asp 652-glu
696).This linker has been found to be highly positive charged and may bind
directly to DNA. C-terminal domain, comprised between Gln-697 and Phe-795,
is highly conserved and contain the active site Tyr- 723 (Staker.et al.2005). The
catalytic residues of human DNA topoisomerase-I is Asn-722, Lys-532, Asp533, Arg-364, Asn-352 and Tyr-723.
169
It has been proposed that diospyrin and its derivatives form a direct
interaction
with
enzyme
and
interferes
with
camptothecin-dependent
topoisomerase-I mediated DNA cleavage and thus inhibit the kinase activity of
topoisomerase-I (Tazi et al. 2005). For our study we have used only
topoisomerase-I enzyme after removal of DNA from PDB complex (PDB:
1SC7) (Staker.et al.2005).We have used Gold software 5.1 (Jones. et al. 1995) to
dock all the compounds into the active site of DNA-Topo-1(PDB: 1SC7)
(Staker.et al.2005).Gold is a well known genetic algorithm program for docking
flexible ligands into protein binding sites (Lauria. A. and Ippolito, M. 2007).
The binding site was defined to include all residues within 10 of the
ligand in original complex of human DNA-Topo-1. Preparation of protein for
docking included extraction of DNA using Discovery Studio (Accelary,San
diego,CA) and removal of ligand, water molecules and addition of hydrogens
were performed with GOLD. Preparation of ligands for docking included energy
minimization using MMFFs in Vlife program (Thomas A. Halgren. 1999).
Addition of hydrogens and the protonation of charged group were set by GOLD
(Jones. et al. 1995) as default. The default calculation mode which provides the
best docked results was selected for calculations. Chemscore was used as the
scoring function. Results were saved in mol.2 file. The final choice of the
models was based on interactions with key residues and correlation with the
biological activity. Pymol (The PyMOL Molecular Graphics System), V 1.5.0.4
Schrdinger. LLC) was used for the purpose of visualization. Diospyrin is a
binaphthoquinone so for our convenience we have divided diospyrin into 4 rings,
first and second naphthoquinone moiety known as 1/2 and 3/4 rings respectively.
The highly active compounds (D1, D14, D2, D7) were showing hydrogen bond
interactions with essential residues like Arg-364, Arg-488, Tyr-723, Asp-533,
and Asn-722 and additionally Van-der Waals interactions with Asp-533, Asn480, Asn-722, Glu-356, Tyr-426, Asn-352, Asn-430 were also observed. D14 and
diospyrin were the most active compounds where in D14 (acetyl amino
170
derivative) quinone (C=O) of ring 3/4 was showing hydrogen bond interactions
with Arg 488 and Tyr-723 (Figure-5.5A). The -NH group of -NHCOCH3 in
position 3 of ring 4 was also forming hydrogen bond interaction with Asp-533,
Van- der Waals interaction also found in ring 3/4 with Asp-533. No hydrogen
bond interaction in ring 1/2 was observed but Van-der Waals interactions were
found between quinone and Asn-722. In the case of diospyrin, hydroxyl group
of ring 1/2 was showing hydrogen bond interaction with Asn-722 and the of
quinone and hydroxyl group
171
Fig-A
Fig.B
Figure.5.5 [A] The binding interaction of the most active compound [D14]
[B] [3a] (Valine methyl diospyrins dimethyl ether) right side
against human DNA Topoisomerase-I of 1SC7
172
Residues are shown in green color. Docking poses were visualized with
PYMOL molecular graphics software. Interaction of amino acid residues with
compound D14 and 3a with highest score stimulation are shown. Hydrogen bonding
is shown through blue dotted lines.D14 and 3a are shown yellow in color.
5.8. Molecular docking studies of best compound 2 with the H5N1
neuraminidase active site
5.8.1. Docking methodology
Docking studies were performed using GOLD 5.1 (Jones.et al. 1995)
software. The crystal structures of H5N1 neuraminidase (PDB ID: 2HTY
(Russell et al. 2006) and (PDB ID: 2HU4 (Russell et al. 2006) where loop-150
were in open and closed conformation, respectively, were used in the study. At
first all the water molecules, metals and ligands are removed from both the PDB
protein structures and was loaded in the Hermes module of GOLD. Subsequently
hydrogen atoms were also added. The histidine protonation states are also
determined and fixed in the protein structure. Binding site is determined using
the previous knowledge of the original ligand interaction site. Goldscore was
taken as the scoring function to rank the compounds to be investigated.
In docking stimulations each ligand was kept flexible but the amino
acid residues of the proteins were held rigid. Preparation of protein and ligands
(removal of water molecule, extraction of original ligands from the protein
active site, addition of hydrogen and protonation state of charge group) were
done with GOLD as per default settings. For the simulation runs default
parameter values were taken. The selection of atoms in the active site within 6
of original ligand was chosen as default. The minimum genetic algorithm run of
10,000 selected as default. The number of generated poses was set to 10 and top
ranked solutions were kept, with the early termination option turned on. The
Chemscore was selected for scoring function. The results were saved in mol.2
file.
173
5.8.2. Loop 150 dynamics and its implications in contemporary antineuraminidase research
As already reported in literature, The N1 and N2 neuraminidases of
viruses currently circulating in humans belong to two phylogenetically distinct
groups. Group-1 contains N1, N4, N5 and N8 subtypes, group-2, on the other
hand, contains N2, N3, N6, N7 and N9.In (2006) Russell et al. reported the
crystal structures of N1, N4 and N8 group-1 neuraminidases and when
comparison of active sites with N9 neuraminidase (group-2) were done,
specifically, on the 150-loop (residues 147152) the following differences in
conformation were observed: 1] The C position of from Val 149 of group-1 is
about 7 distant from the equivalent isoleucine residue in group-2 and
hydrophobic side chain at position 149 is pointed away from the active site in
group-1 but towards it in group-2 (Landon et al. 2008).
There is a difference of 1.5 in the side-chain position of the
conserved aspartic acid residue at position 151 between group-1 and group-2
neuraminidases (Cheng et al.2008). In group-2 structures Glu 119 forms a
hydrogen bond with Arg 156 but in group-1 it adopts a conformation such that
its carboxylate points in approximately the opposite direction. Due to this
difference in conformational aspects, a cavity observed to be forming, more
known as 150-cavity adjacent to the active site in group-1 but not in group-2
neuraminidases (Figure-5.6). Evidences were also found where the loop-150
changes its conformation upon inhibitor binding, a striking feature, which can
form basis of a different school of thought in future in context of drug design
against neuraminidases.
This differential inhibitor-binding concept was later supported by
Zhang and co-workers who utilized molecular docking, molecular dynamics
simulations, and MM/PBSA free energy calculation to confirm this. Inspired
from this discovery, Amaro et al. (2007) and conducted molecular dynamics
174
simulations with explicit solvent system , taking apo form as well as oseltamivir
bound into the active site, proposing that, the loop 150 has the capability to open
wider than that was shown in crystal structures. This motion of loop 150 is
simultaneously with loop 430 (comprising of residues Arg 430-Thr 439), which
makes the active site even wider; however, the loop movements tend to form a
closed conformation when oseltamivir is bound. Continuing to work further, the
same group, Cheng et al.(2008) proposed the presence of novel hot spots within
flexible binding regions (150 and 430 loop) of the N1 neuraminidase extensive
MD simulations, conformational clustering, and CS-Map and if utilized, novel
inhibitors can be discovered with enhanced oral bioavailability and less
susceptible to structural mutations.
In the same year, Amaro et al. (2007), identified 27 drug like
compounds, the best three being NSC 109836, NSC 211332 and NSC 45583.
The work utilized ensembles from MD simulations on crystal structures and the
proposed location of hot spots from the previous work (Russell et al 2006;
Landon et al. 2008) and the flexible regions from loop 150 and loop 430. Jo and
co-workers, similarly, utilized the 150-loop region of the H5N1 subtype to
design novel oseltamivir derivatives with proper shape and atomic charge to fit
inside the 150 cavity.
Attachment of chemical groups at the C3 position of oseltamivir
successfully improved the binding affinity with neuraminidase subtype N1.
Wang and Zhang in 2010 also proposed that ligand with a small basic group,
such as amino (as in oseltamivir), favor the closed conformation of H5N1 NA
otherwise, for those inhibitors possessing a large, positively charged group, such
as guandinium, binding to the open conformation of H5N1 NA is favored.
175
Figure 5.6. 150-cavity adjacent to the active site in group-1 but not in
group-2 neuraminidases
Until then, all group 1 neuraminidases have been reported as having an
open conformation and all group 2 neuraminidases have been reported as having
a closed conformation of the loop 150. Perhaps, one of the most surprising
discoveries of this year was the finding that group specific 150 cavity is absent
in H9N1 crystal structure (Li et al.2010). This finding implies that
neuraminidase inhibitors targeted to the 150-cavity will probably be less
effective against 09H1N1 variants. Recently, the single most unsolved structure
of group-1 neuraminidase, ie, N5 was also solved (Wang et al. 2011). The results
demonstrate that N5 possesses the common characteristics of the reported typical
group 1 NAs, including the presence of loop 150, which is in open conformation
but the loop closes when the protein is bound with zanamivir.
However, upon closer comparison of the uncomplexed N5 active site
with those of all other known structure group 1 NAs, it was observed that the N5
150-cavity is extended relative to those of all other group 1 structures
(Figure-5.7.). Although crystallography studies proved that 09N1 does not have
a 150 cavity, but contrary to this experimental evidence, long-term molecular
176
177
178
179
180