MOLECULAR DOCKING

Mokhammad Fajar Pradipta

Austrian Indonesian Centre (AIC) for
Computational Chemistry
Chemistry Department
Gadjah Mada University
 Introduction
Drug discovery take years to decade for
discovering a new drug and very costly
Effort to cut down the research timeline
and cost by reducing wet-lab experiment
use computer modeling
 Drug discovery

Chemical + biological system desired response?

 TRADITIONAL DRUG DESIGN

Lead generation:
Natural ligand / Screening

Biological Testing

Drug Design Cycle

If promising
Synthesis of New Compounds

Pre-Clinical Studies
 Finding lead compound
A lead compound is a small molecule that serves as the
starting point for an optimization involving many small
molecules that are closely related in structure to the lead
compound
Many organizations maintain databases of chemical
compounds
Some of these are publically accessible others are
proprietary
Databases contain an extremely large number of compounds
(ACS data bases contains 10 million compounds)
3D databases have information about chemical and
geometrical features
Hydrogen bond donors
Hydrogen bond acceptors
Positive Charge Centers
Aromatic ring centers
Hydrophobic centers
 Finding lead compound
There are two approaches to this
problem
A computer program AutoDock (or
similar version Affinity (accelrys)) can
be used to search a database by
generating fit between molecule
and the receptor
Alternatively one can search 3D
pharmacophore
 Structure based drug design
Drug design and development
Structure based drug design exploits the
3D structure of the target or a
pharmacophore
Find a molecule which would be expected to
interact with the receptor. (Searching a data
base)
Design entirely a new molecule from
SCRATCH (de novo drug/ligand design)
In this context bioinformatics and
chemoinformatics play a crucial role
 Structure-based Drug Design (SBDD)

Natural ligand / Screening

Molecular Biology & Protein Chemistry

3D Structure Determination of Target

and Target-Ligand Complex
Modelling

Drug Design Cycle

Structure Analysis
Biological Testing
and Compound Design
If promising

Synthesis of New Compounds

Pre-Clinical
Studies
 Structure based drug design
SBDD:
drug targets (usually proteins)
binding of ligands to the target
(docking)

rational drug design

(benefits = saved time and $$)

Schematics for structure based drug design
Select and Purify the
target protein
Obtain known
inhibitor

X-Ray structural
determination of native
protein

X-Ray structural
determination of
inhibitor complex

Model inhibitor Synthesis, Evaluate

with preclinical, clinical,
computational invitro, invivo, cells,
tools animals, & humans
Determine IC50

Drug
Structure Based Drug Design have the potential to shave off years
and millions of dollars
 Working at the intersection
Structural Biology
Biochemistry
Medicinal Chemistry
Toxicology
Pharmacology
Biophysical Chemistry
Natural Products Chemistry
Chemical Ecology
Information Technology
 Molecular docking-definition
It is a process by which two molecules
are put together in 3 Dimension

Best ways to put two molecules

together

Using molecular modeling and

computational chemistry tools
 Molecular docking
Docking used for finding binding modes
of protein with ligands/inhibitors
In molecular docking, we attempt to
predict the structure of the
intermolecular complex formed between
two or more molecules
Docking algorithms are able to generate
a large number of possible structures
We use force field based strategy to
carry out docking
Oxygen transport molecule (101M)
with surface and myoglobin ligand
Influenza virus b/beijing/1/87 neuraminidase
complexed with zanamivir
Influenza virus b/beijing/1/87 neuraminidase
complexed with zanamivir
 Plasma alpha antithrombin-iii and
pentasaccharide protein with heparin ligand
 Steps of molecular docking
Three steps
(1)Definition of the structure of the
target molecule

(2) Location of the binding site

(3) Determination of the binding

mode
Best ways to put two molecules
together
Need to quantify or rank solutions

Scoring function or force field

Experimental structure may be

amongst one of several predicted
solutions

-Need a Search method

 Questions
Search
What is it?
When/why and which search?
Scoring
What is it?
Dimensionality
Why is this important?
 Spectrum of search
Local
Molecular Mechanics
Short - Medium
Monte Carlo Simulated Annealing
Brownian Dynamics
Molecular Dynamics
Global
Docking
 Details of search
Level-of-Detail
Atom types
Terms of force field
Bond stretching
Bond-angle bending
Torsional potentials
Polarizability terms
Implicit solvation
 Kinds of search
Systematic
Exhaustive
Deterministic
Dependent on granularity of
sampling
Feasible only for low-dimensional
problems
DOF, 6D search
 Kinds of search
Stochastic
Random
Outcome varies
Repeat to improve chances of success
Feasible for higher-dimensional
problems
AutoDock, <~40D search
 Stochastic search methods
Simulated Annealing (SA)
Evolutionary Algorithms (EA)
Genetic Algorithm (GA)
Others
Tabu Search (TS)
Hybrid Global-Local Search
Lamarckian GA (LGA)
 Simulated annealing
One copy of the ligand (Population = 1)
Starts from a random or specific
postion/orientation/conformation
(=state)
Constant temperature annealing cycle
(Accepted & Rejected Moves)
Temperature reduced before next cycle
Stops at maximum cycles
 Search parameters
Simulated Annealing
Initial temperature (K)
Temperature reduction factor (K-
1cycle)

Termination criteria:
accepted moves
rejected moves
cycles
 Genetic function algorithm
Start with a random population (50-200)
Perform Crossover (Sex, two parents ->
2 children) and Mutation (Cosmic rays,
one individual gives 1 mutant child)
Compute fitness of each individual
Proportional Selection & Elitism
New Generation begins if total energy
evals or maximum generations reached
 Search parameters
Population size
Crossover rate
Mutation rate
Local search
energy evals
Termination criteria
energy evals
generations
 Dimensionality of molecular
docking
Degrees of Freedom (DOF)
Position or Translation
(x,y,z) = 3
Orientation or Quaternion
(qx, qy, qz, qw) = 4
Rotatable Bonds or Torsions
(tor1, tor2, torn) = n
Total DOF, or Dimensionality,
D=3+4+n
 Docking score
DGbinding = DGvdW + DGelec + DGhbond +
DGdesolv + DGtors
DGvdW
12-6 Lennard-Jones potential
DGelec
Coulombic with Solmajer-dielectric
DGhbond
12-10 Potential with Goodford Directionality
DGdesolv
Stouten Pairwise Atomic Solvation Parameters
DGtors
Number of rotatable bonds
 Molecular mechanics: theory
Considering the simple
harmonic approximation, the
potential energy of molecules
is given by
V= VBond+ VAngle + VTorsion + Vvdw
+ Velec+ Vop
VBond = 1/2Kr (rij-r0)2
Where Kr is the stretching
force constant
VAngle =1/2K (ijk-0)2
Where K is the bending force
constant
VTorsion =V/2 (1+ Cos n(+0))
Where V is the barrier to
rotation, is torsional angle
 Molecular mechanics: Theory
Lennard-Jones type of 6-12 potential is
used to describe non-bonded and weak
interaction
Vvdw= (Aij/rij12-Bij/rij6)
Simple Columbic potential is used to
describe electrostatic interaction
Velec=(qiqj/rij)
Out of plane bending/deformation is
described by the following expression
Vop= 0.5 Kop 2
 The forcefield
The purpose of a forcefield is to describe the
potential energy surface of entire classes of
molecules with reasonable accuracy
In a sense, the forcefield extrapolates from the
empirical data of the small set of models used
to parameterize it, a larger set of related
models
Some forcefields aim for high accuracy for a
limited set of elements, thus enabling good
predictions of many molecular properties
Others aim for the broadest possible coverage
of the periodic table, with necessarily lower
accuracy
 Components of a forcefield
The forcefield contains all the necessary elements
for calculations of energy and force:
A list of forcefield types
A list of partial charges
Forcefield-typing rules
Functional forms for the components of the
energy expression
Parameters for the function terms
For some forcefields, rules for generating
parameters that have not been explicitly defined
For some forcefields, a way of assigning
functional forms and parameters
The energy expression
 Valence interactions
The energy of valence interactions is generally
accounted for by diagonal terms:
bond stretching (bond)
valence angle bending (angle)
dihedral angle torsion (torsion)
inversion, also called out-of-plane interactions
(oop) terms, which are part of nearly all
forcefields for covalent systems
A Urey-Bradley (UB) term may be used to
account for interactions between atom pairs
involved in 1-3 configurations (i.e., atoms bound
to a common atom)
Evalence=Ebond + Eangle + Etorsion + Eoop + EUB
 Non-bond interactions

The energy of interactions between

non-bonded atoms is accounted for
by
van der Waals (vdW)
electrostatic (Coulomb)
hydrogen bond (hbond) terms in
some older forcefields
Enon-bond=EvdW + ECoulomb + Ehbond
 Molecular dynamics (MD)
simulations
A deterministic method
based on the solution of
Newtons equation of
motion
F i = mi ai
for the ith particle; the
acceleration at each step
is calculated from the
negative gradient of the
overall potential, using
Fi = - grad Vi - = - Vi

Vi = Sk(energies of
interactions between i
and all other residues k
located within a cutoff
distance of Rc from i)
 Classical molecular dynamics
Constituent molecules obey
classical laws of motion
In MD simulation, we have to
solve Newton's equation of
motion
Force calculation is the time
consuming part of the
simulation
MD simulation can be
performed in various ensembles
NVT, NPT and NVE are the
ensembles widely used in the
MD simulations
Both quantum and classical
potentials can be used to
perform MD simulation
Calculation of interaction energy
MM total energy can be used to get
interaction energy of the ligands with
biomolecules
In order to compute the interaction
energy, calculations have to be
performed for the biomolecule,
ligands and the biomolecule-ligand
adduct using the same force field
Eint= Ecomplex - {Ebiomolecule+Eligand}
 Integration of equation of motion
and time step
A key parameter in the integration algorithm is
the integration time step
The time step is related to molecular vibration
The main limitation imposed by the highest-
frequency motion
The vibrational period must be split into at
least 8-10 segments for models to satisfy the
Verlet algorithm that the velocities and
accelerations are constant over time step used
In most organic models, the highest vibrational
frequency is that of C-H stretching, whose
period is of the order of 10-14 s (10fs). Therefore
integration step should be 0.5-1 fs
 Stages and duration in MD
simulation
Dynamics simulations are usually carried out in two
stages, equilibration and data collection
The purpose of the equilibration is to prepare the
system so that it comes to the most probable
configuration consistent with the target temperature
and pressure
For large system, the equilibration takes long time
because of the vast conformational space it has to
search
The best way to judge whether a model has
equilibrated is to plot various thermodynamic
quantities such as energy, temperature, pressure
versus time
When equilibrated, the system fluctuate around their
average
 Durations of some real
molecular events
Event Approximate duration

Bond stretching 1-20 fs

Elastic domain modes 100 fs to several ps

Water reorientation 4 ps

Inter-domain bending 10 ps-100 ns

Globular protein tumbling 1-10 ns

Aromatic ring flipping 100 s to several seconds

Allosteric shifts 2 s to several seconds

Local denaturation 1 ms to several seconds

 Free energy simulations
Ability to predict binding energy
Free energy perturbation and
thermodynamic integration
Computational demand and issues related
to sampling prevent this technique in
probing structure based drug design
Free Energy equation
 De nova design of inhibitor for HIV-I
protease

An impressive example of the

application of SBDD is was the
design of the HIV-I protease
inhibitor
 De nova design
It is a member of the aspartyl protease
family with the two active sites
Structure has tetra coordinated water
molecules tat accepted two hydrogen
bond from the backbone amide
hydrogens of isoleucine in the flaps
Two hydrogen bonds to the carbonyl
oxygens of the inhibitor
 Application of structure based drug
design: HIV protease inhibitors
The starting point is the series of
X-ray structures of the enzyme
and enzyme-inhibitor complex
The enzyme is made up of two
equal halves
HIV protease is a symmetrical
molecule with two equal halves
and an active site near its center
like butterfly
For most such symmetrical
molecules, both halves have a
"business area," or active site,
that carries out the enzyme's job
But HIV protease has only one
such active site in the center of
the molecule where the two
halves meet
 Structure based drug design: HIV
protease inhibitors
The single active site was plugged with a
small molecule so that it is possible shut
down the whole enzyme and theoretically
stop the virus' spread in the body
Several Inhibitors have been designed based
on
Peptidic inhibitor
Peptidomemitic compounds
Non-peptide inhibitors
Further work has demonstrated the success
of this approach
 Some examples
Ritonavir (trade name Norvir) is one
of a class of anti-HIV drugs called
protease inhibitors
Saquinavir
Indinavir is another example of very
potent peptidomimetic compound
discovered using the elements of 3D
structure and Structure Activity
Relationship (SAR)
 De nova design
The first step was a 3D database
search of a subset of the Cambridge
Structural Database
The pharmacophore for this search
comprised of two hydrophobic groups
and a hydrogen bond donor or
acceptor
The hydrophobic groups were intented
to bind to the catalytic asp residues
 De nova design
The search yielded the hit which contained
desired element of the pharmacophore but it
also had oxygen that could replace the bound
water molecules
The benzene ring in the original compound was
changed to a cyclohexanone, which was able to
position substituents in a more fitting manner
The DuPont Merck group had explored a series
of peptide based diols that were potent
inhibitors but with poor oral bioavailability
 De nova design
They have retained the diol
functionality and expanded the six me
member ring to a seven membered diol
The ketone was changed to cyclic urea
to enhance the hydrogen bonding to the
flaps and to help synthesis
The compound chosen further studies
including clinical trails was p-
hydroxymethylbenzyl derivative
 P1 8.5-12 P1

3.5-6.5 3.5-6.5

H-bond donor or acceptor

3D hit
3D pharmacophore
Symmetric diol docked into
HIV active site Initial
idea for
inhibitor

Final Molecule selected Stereochemistry required Expand ring to give diol

for clinical Trials for optimal binding and incorporate urea
Host-Guest Interactions with
Collagen: As molecules
Dominated by Geometrical
factors and Solvent
Accessible Volumes
Energy minimized structure of 24-
mer collagen triple helix
 Complex Formation of poly phenols
at various collagen sites

Aspargine of
T.Helix and
gallic acid

Aspartic acid of
T.Helix and
catechin
Lysine of T.Helix
and
epigallocateching
allate
 Binding energies different complexes
between polyphenols and triple helix
Binding Energy (Kcal/mol)
Binding Sites in Epigallocatechi
triple helix Gallic acid Pentagalloyl
Catechin (Cat) ngallate
(Gal) glucose (PGG)
(EGCG)
9th residue Ser
of C-chain (2) 16.5 22.5 35.2 56.6

6th residue Hyp

of A-chain (1) 14.5 20.8 34.5 48.4

12th residue Lys

of B-chain (1) 19.2 23.8 37.9 41.1

21st residue Asp

of A-chain (1) 18.4 20.0 38.2 59.8

17th residue Asn

of C-chain (2) 14.1 23.7 34.3 52.8
 Interfacial interacting volume Vs Binding
energy of the collagen-poly phenol complex

Interacting Interfacial Volume (3)

 Effective solvent inaccessible contact volume
Vs Binding energy of the collagen-poly phenol
complex
Inset: effective solvent inaccessible contact surface area Vs Binding energy of the complex
 Plot of inverse of interacting interfacial volume
(1/Int.Vol.) Vs inverse of binding energy(1/B.E) of the
complexes
 Others have done the
work. Some have used
the work. I have spoken
only on behalf of their
behalf.