The object of this module is to visualize and identify a molecule that will bind with high affinity and specificity to a biological target of known (or predictable) three-dimensional (3D) structure of protein.

Need For 3D Visualization of Proteins
Drug-target interaction (DTI) is the basis of drug discovery and design. It is time consuming and costly to determine DTI experimentally. Hence, it is necessary to develop computational methods for the prediction of potential DTI. X-ray crystallography and NMR Spectroscopy can reveal 3D structure of drugs and bound compounds. Visualization of these “complexes” of proteins and potential drugs can help scientists understand the mechanism of action of the drug and to improve the design of a drug.
Critical in drug design -- yields insight into how the protein might interact with ligands at active sites

Predicting Drug target Interactions
A number of programs convert atomic coordinates of 3-d structures into views of the molecule. These programs allow the user to manipulate the molecule by rotation, zooming, etc. They are available both as a standalone application (like RasMol) and a browser applet.
Most popular Programs for viewing 3-D structures:
Protein explorer:
http://www.umass.edu/microbio/chime/pe_beta/pe/protexpl/frntdoor.htm?q=microbio/chime/pe/protexpl/frntdoor.htm
Rasmol:
http://www.umass.edu/microbio/rasmol/
Chime:
http://www.umass.edu/microbio/chime/
Jmol:
http://jmol.sourceforge.net/
Swiss 3D viewer:
http://spdbv.vital-it.ch/

Protein Ligand Interactions
“Ligand” comes from “ligare” meaning a “band” or “tie.” It is currently used to mean a chemical molecule (of any size) that binds or interacts with another molecule“target” or “receptor” through noncovalent forces—that is, the interaction does not involve formation of chemical bonds.
The vast majority of the currently available drugs act via non-covalent interactions with the target protein.The nature of the interaction between the ligand and the target depends upon the chemical/physical between each other and the solvent environment. The chemical and physical forces arise mainly from the interaction of electrons and is studied using quantum mechanics (QM).The goal is to develop a rapid method to predict the bound conformation and binding affinity of the small molecule.

Major Parts of a Protein Ligand Interaction

There is now a large body of experimental data available on 3-D structures of protein-ligand complexes and binding affinities in protein data bank(PDB). These data clearly indicate that there are several features found basically in all complexes of tightly binding ligands:
  • There is a high level of steric complementarity between the protein and the ligand. This observation is also described as the lock-and-key paradigm.
  • There is usually high complementarity between surface properties of protein and the ligand. Lipophilic parts of the ligands are most frequently found to be in contact with lipophilic parts of the protein. Polar groups are usually paired with suitable polar protein groups to form hydrogen bonding or ionic interactions.Also the H bond geometry is strongly preserved.
  • The ligand usually binds in an energetically favorable conformation given in the figure below.
Screen Shot 2013-12-06 at 3.44.53 PM.png

The binding of a ligand to a specific protein is determined by the structural and energetic recognition of a ligand and a protein.The most important non bonded interactions for protein ligand complexes are :
  • Hydrogen Bonds
hbond.png
  • Ionic Interactions
ionic.png
  • Hydrophobic Interactions
hydrophobic.png
  • Cation - pie interactions
cataionic.jpg
The selective binding of a low-molecular-weight ligand to a specific protein is determined by the structural and energetic recognition of a ligand and a protein.The binding affinity can be determined from the experimentally measured binding constant Ki.

Screen Shot 2013-12-06 at 3.34.33 PM.png
where R is the gas constant and T is the absolute temperature in Kelvin. Of course, this is of great interest in drug design because prediction of Ki is a direct prediction of ligand affinity.The pharmacological literature frequently reports the dissociation constant, Kd , which is simply the reciprocal of the equilibrium constant, or the IC50 , the concentration of ligand that achieves a 50% change in the normal activity.The binding equation is calculated and given below.
Screen Shot 2013-12-06 at 9.01.53 PM.png

Methods in computer aided molecule design:
Computational ligand design can be divided into two different strategies: ligand-based (analog-based) or target-based(protein structure based) design.
Ligand-based design depends on a set of known ligands(compounds) and is particularly valuable if no structural information of the receptor or protein target is available. Hence, it is generally applicable to all classes of drugs. Target-based design usually starts with the structure of a receptor site, such as the active site in a protein.Since this module is mostly focused on structures based perspective we will discuss this in detail.

Docking
The aim of molecular docking is to evaluate the feasible binding geometries of a putative ligand with a target whose 3D structure is known. The binding geometries, often called binding modes or poses include, in principle, both the positioning of the ligand relative to the receptor (ligand configuration) and the conformational state(s) of the ligand and the receptor. There are three basic tasks any docking procedure:
  • characterization of the binding site;
  • positioning of the ligand into the binding site (orienting); and
  • evaluating the strength of interaction for a specific ligand-receptor complex (“scoring”).

Target Preparation for docking
In structure based virtual screening identification of compounds for automated docking does a given receptor-binding pocket. The 3D structure of targets is typically obtained by X-ray crystallography; NMR and homology models can also be used. In some cases, the target is not a protein but can be RNA or DNA. Creating a homology model is generally carried out in several steps:
  • Finding a known template structure (or structures) that has a high sequence similarity/identity with the target sequence (experimental target structures are deposited at the Protein Data Bank);
  • Aligning the target sequence and the template sequence taking into account structural data;
  • Building the model (the core and loop segments if possible),
  • Refinement of the model structure and assessing the quality of the resulting 3D model.


Table below list the Receptor 3D Structures databases, Homology Modeling tools, 2D/3D Structure Prediction of the Receptor and Macromolecular Interaction Databases.
Name
URLs
Type
The Entrez Structure Database
http://www.ncbi.nlm.nih.gov/Structure/MMDB/mmdb.shtml
Macromolecule database
PDBCat can be used to manipulate and process PDB files using commonly available tools such as Perl, awk, etc.
http://www.ks.uiuc.edu/Development/MDTools/pdbcat/
Tools to manipulate PDB
files
Atlas of protein side-chain interactions within known protein structures and interactions with DNA
http://www.biochem.ucl.ac.uk/bsm/sidechains/
Macromolecule interaction
database
BIND: The Biomolecular Interaction Network Database
http://www.bind.ca/Action
Macromolecule interaction
database
Database of interacting proteins: DIP
http://dip.doe-mbi.ucla.edu/hold/main.html
Macromolecule interaction
database
MINT database stores data on functional interactions between proteins
http://mint.bio.uniroma2.it/mint/Welcome.do
Macromolecule interaction
database
Many structural bioinformatics tools
http://www.expasy.org/
Tools to analyze macromolecules
The UCLA-DOE Structure Evaluation server is a tool designed to help in the refinement of crystallographic structures and models
http://nihserver.mbi.ucla.edu/Verify_3D/
Validate protein structure
ERRAT is a protein structure verification algorithm
http://nihserver.mbi.ucla.edu/ERRATv2/
Validate protein structure
Modeller: software package for homology or comparative
modeling of protein 3D structures
http://www.salilab.org/modeller/modeller.html
Homology modeling
PredictProtein : webserver for homology modeling and protein function prediction
http://www.predictprotein.org/newwebsite/
Homology modeling &
related tools
Robetta: web server for protein structure prediction and
analysis
http://robetta.bakerlab.org/
Protein structure prediction
(PS)2 : automated homology modeling server using a consensus strategy between psi-blast, impala and T-coffee with a final 3D structure modeled with Modeller
http://ps2.life.nctu.edu.tw/
Homology modeling &
related tools
FATCAT: web server for flexible structure comparison and structure similarity searching
http://fatcat.burnham.org/
Protein structure analysis, Structural similarity search
SVMProt: protein functional family prediction
http://jing.cz3.nus.edu.sg/cgi-bin/svmprot.cgi
Protein Structural analysis
Protein structure analysis
http://kinemage.biochem.duke.edu/~jsr/html/anatax.2a.html
Protein structure analysis
Monster web application for inferring potentially stabilizing non-bonding interactions in macromolecular structures
http://monster.northwestern.edu/monster.jsp
Protein Structural analysis, mutations

The target 3D structures are prepared by addition of hydrogen atoms and the prediction of the correct protonation state for the titratable residues (pKa prediction electrostatic computations), adding or removing tightly-bound water molecules, counterions, metal ions, cofactors, sugar molecules, removing subunits not involved in ligand binding or far from the binding site, introducing corrections for the tautomeric states of histidine residues and re-orientations of hydroxyl groups. Next, it is important to identify binding pockets and “hot spots” on selected targets. Some of the methods are geometry based or energy based. Structural variations of the protein binding site help to study the site directed mutagenesis of receptor using docking.
The primary aim of docking is to predict the ligand conformation and binding affinity (scoring). Docking protocols can be described as a combination of two components a search strategy and a simple scoring/fitting function. The search algorithm generates the ligand conformations, which include experimentally determined binding mode. The scoring functions commonly used are said to be: force field based, empirical and knowledge-based. The use of ligand efficiency and consensus docking scores is also practiced. Finding novel hits is always difficult and in most of the cases it is important to combine several methods such and structure and ligand based. Multi-step procedures are often used to facilitate the in silico screening process; they can combine different packages, use molecular dynamics after docking.

Table below shows some the docking and scoring methods.

Name
URLs
Type
AutoDock , small molecule docking
http://www.scripps.edu/mb/olson/doc/autodock
Small molecule docking
BDT (is an easy-to-use front-end application for automation
of massive docking tasks and complex docking strategies with
AutoDock
http://www.quimica.urv.cat/~pujadas/BDT/
Graphics interface for
small molecule docking
eHits: Small molecule docking
http://www.simbiosys.ca/ehits/index.html
Small molecule docking
FRED ,small molecule docking. See the online demos
to try many OpenEye applications
http://www.eyesopen.com
Small molecule docking
DOCK,Small molecule docking
http://dock.compbio.ucsf.edu/
Small molecule docking
ViewDock to analyze Dock data and tools for post-processing DOCK results
http://www.cgl.ucsf.edu/chimera/docs/UsersGuide/index.html
Small molecule docking
post-docking processing
Surflex
http://www.biopharmics.com
Small molecule docking
Plants
http://www.tcd.uni-konstanz.de/research/plants.php
Small molecule docking
PSI-DOCK
ftp://ftp2.ipc.pku.edu.cn/pub/software
Small molecule docking
GFscore :A General Non-Linear Consensus Scoring
Function for High-Throughput Docking
http://gfscore.cnrs-mrs.fr/index.htm
Scoring
SCORE is an empirical method developed for estimating the binding affinity of protein-ligand complex with known three-dimensional structure
ftp://ftp2.ipc.pku.edu.cn/
Scoring
FOLD-X can compute binding energy
http://fold-x.embl-heidelberg.de:1100/cgi-bin/main.cgi
Scoring
Xscore : tool to predict binding energy between ligand
and receptor
http://sw16.im.med.umich.edu/software/xtool/
Scoring
CLIBE
http://bidd.nus.edu.sg/group/CLiBE/CLiBE.asp
Help for scoring functions

Thermodynamics of docking
Before complex formation, the ligand and the active site of the receptor both make interactions with the solute molecules. The polar and charged groups on the surface of the unbound receptor and the unbound ligand form hydrogen bonds with the water molecules in the solvent, and all groups interact through Van der Waals interactions. On complex formation, the receptor binding site and the ligand become at least partially desolvated and the hydrogen bonds with the solvent are replaced with hydrogen bonds between the receptor and the ligand. It has been shown that burying a hydrogen bond donor or acceptor group (either neutral or charged) in the protein or complex interior without formation of a hydrogen bond can be detrimental to stability (Bogan and Thorn, 1998; Hendsch et al., 1996). Polar and non-polar groups of the receptor and the ligand form Van der Waals interactions on complex formation, and charged groups interact strongly through Coulomb interactions. The release of the ordered water molecules around the ligand and in the receptor active site on complex formation and the resulting increase in entropy of these water molecules favors binding and is what underlies the hydrophobic effect. Entropy losses that occur on complex formation are partially due to the reduction in translational and rotational degrees of freedom of the ligand. The translational and rotational degrees of freedom of the complex are also slightly different from those of the receptor in isolation.

Manual Docking
Molecular modeling can be used to dock or fit a molecule to a target into its binding or active site. The docking program searches for the idea fit of the molecule around the binding site to try to get the best fit such that the groups in the ligand and protein are within the preferred bonding distances of each other. Also while fitting both ligand and the protein remain in the same conformation throughout the docking process and this method is known as rigid fit. Once a ligand has been successfully docked then optimization like energy minimization is carried out. Different molecular conformations can be docked into the protein target and the interaction energies helps to identify which conformation fit the best.

Automated docking
There are several docking programs available that can automatically dock ligands into a binding site with the minimum of input from the operator. The application of automated docking programs is to carry out large-scale virtual screening of several millions of molecules with the aim of identifying new lead compounds. For each molecule studied in docking automated programs generated several docking poses for a particular ligand. All of the binding modes are scored in order to identify the most likely binding mode in terms of how well it fits the space available and how many intermolecular interactions it can form with the binding site. The calculation of the docked scores should be accurate and rapid in order to process large number of molecules. This is a difficult problem since increasing the speed of which the algorithm is working involves some assumptions and short cut methods that reduce the accuracy of the calculation. The simplest method for automated docking is to treat the ligand and the protein target as rigid bodies .The next of complexity comes when the ligand is considered rigid and the protein target in flexible. The most complex situation arises is where both the ligand and the protein target is considered flexible and this situation takes maximum amount of time.

Rigid body docking
In rigid body docking a molecular surface is defined, one way to do that is by defining the binding site by its van der waals radius. Defining by van der Waals radius results in calculating extensive surface area ,in place of that only the solvent accessible surface area is defined.Now a probe sphere of radius 1.4-1.5 A is rolled over the surface of the binding site containing the convex surface and the concave surface (represents the surface that how far the probe atom can access the space between the atoms of the binding site). In concave surface the probe atoms is in contact with 2-3 atoms. This surface is also known as Connolly surface. The Dock program is one of the earliest docking programs, which uses Connolly surface. The ligand atoms and the sphere pseudo atoms of the binding site are overlaid and a matching operation takes place called the distance matching. In distance matching, the distances between the ligand atoms and the spheres (pseudoatoms) in the binding cavity are measured. These distances are then used to identify which ligand atoms and receptor spheres can be matched. Graphs are prepared for both the cases and list of pairs of ligand atoms and spheres are selected.The minimum number of pairings required for an acceptable docking is four. Later it is extended to see if more spheres match. After an acceptable binding mode is found optimization is carried out which fine-tunes the position of the ligand in the binding site. This minimizes unfavorable interactions between the ligand and the binding site. The binding energy of the ligand is measured and a score is assigned to the ligand-binding mode.

Another method for rigid docking is by matching the hydrogen bonding groups. The method is similar to the distance matching technique where the hydrogen binding groups at binding site are matched with the hydrogen bonding groups of the ligand. There are two important factors to consider here first; hydrogen-bonding group on the ligand must be the correct distance from the hydrogen-bonding group in the binding site. Second, the two groups concerned must have correct orientation with respect to each other. The positions of atoms in the binding site are identified by creating a sphere around each of the hydrogen-bonding group. The surface of the sphere represents the optimum distance at which a complementary group on the ligand should be placed in order to form a good hydrogen bond. A series of uniformly spaced points is placed over the surface of the sphere to define the surface. These points are known as interaction points onto which the complementary binding groups of ligand will be positioned.

Flexible docking
The drawback of rigid body docking is that they fail to give satisfactory answer for flexible ligands, which forms varieties of flexible conformations. One of the ways to handle this problem is to dock different conformations. Another method to dock the ligand is to fragment the ligand and use the rigid anchor fragment for docking. Then the ligand is grown sequentially and added in layers and energy is minimized of the construct.

Below shows the working of Autodock VINA (docking software) . The tools you require to do docking and viewing the results are
1. Autodock Vina /Autodock4.2
2. MGL Tools
3. Pymol for educational purpose