Structure-based drug design (SBDD) is the design and optimisation of a hit or lead compound using structural information of target protein obtained from either X-ray crystallography, cryo-EM or NMR

Structure based drug design is the workhorse of CADD (computer aided drug design) With that we can:

• Predict druggability
• Identify ligand binding sites
• Virtual screen for novel chemical matter
• Optimize potency of leads
• Reduce off-target effects

Here we are going to use open source software to do all the works. Which can be done using commercial softwares i.e. Schrodinger suites (Maestro)

• Sructure Quality

• Conformantional State

• Experimental Design

• Rigid (both protein and ligand)
• Flexible (both protein and ligand)
• Constrained (protein is constrained)

• Sampling Algorithm - Samples the conformations of ligand in the binding site/s
• Scoring Algorithm - Measures the binding energy of all poses of the ligand (only accounts for non-bonded interactions(electrostatic interactions, Hydrogen Bonds, Salt bridges, van der Waals interactions))

• Shape complementary (PatchDock)
• Genetic Sampling Algorithm (AutoDock4)
• Force Field algorithm
• Empirical Scoring function
• Knowledge based scoring function
• Consensus-based scoring function

• Assume receptor is rigid
• Assumes ligand is flexible

Ensemble of receptor-ligand poses produced by docking, which can be ranked by score. More negative score is better.

Scoring Functions:

• do not correlate with IC50, EC50, Kd etc
• do not provide a rank-ordering of ligands
• are optimized to give good enrichment
  • separates "good" ideas from "bad"
  • Limit the number of ligands to be investigated further

Homology modeling, also known as comparative modeling of protein, refers to constructing an atomic-resolution model of the "target" protein from its amino acid sequence and an experimental three-dimensional structure of a related homologous protein (the "template") Homology modeling relies on the identification of one or more known protein structures likely to resemble the structure of the query sequence, and on the production of an alignment that maps residues in the query sequence to residues in the template sequence.

I-tasser output file have information about:

• Predicted secondary structure (H-Helix, S-Strand, C-Coil, 0-10 (close to 10 means better prediction) ),
• Predicted Solvent solvent accessibility (0-buried residue, 9-highly exposed residue)
• Predicted normalized B-factor -> provides the information about flec=xibility of the residues (0 stands for rigid residue)
• Provide the multiple templates (look into z-score)
• Predicted models for proteins (look into C-score)
• Prediction of Ligand binding sites based on templates used to develop the models

The structure we used has ligand with it.We can load .pdb into mgltools, remove water(edit -> delete water), remove unnecessary part(go to protein left side and to the bottom and select them edit -> Delete -> delete selected atoms), missing atoms can be checked(edit-misc-missing atom ->repair missing atom), add charges(edit-charge-kollman charges) and after that check charges and any residue mentioned in the list will be selected (equally distributed charges through out the protein), Save(grid -> macromolecule -> select the protein -> ok -> save the protein as .pdbqt file)

Ligand structure(Can be cocrystallized with the protein, download 3D structure from PubChem, Zinc15, the Cambridge structural database ) or drawing(chemdoodle and saved as .mol file and convert it to 3D structure with open-babel(installed) and iBabel(add the path of open-babel and save ligand as .mol2 file format)—then minimize and optimize the structure in Avogadro—convert the ligand.mol2 to .pdbqt through open-babel for docking)Ligand preparation

Blind docking(whole protein surface has been used to find the active sites) or specific docking (particular area would be specified—from literature or prediction of binding site) To find the ligand binding residues of the protein-article or coach server(take 10 hours) ->upload the fasta sequence and protein.pdb structure and provide academic email to find the binding sites residues—specific docking and otherwise blind docking

Go to left side receptor click + and will see all the residues where you can select the binding sites residuesafter selecting the binding sites residues select GRID and we change the value in 3 direction of the grid box

In the case of blind docking grid box will cover the whole protein

In the uplaoded jupyter notebook (Docking_with_autodock.ipynb) I have shown how to use AutoDock to dock single ligand to our target protein. in the following I will show how to use Autodock vina to dock a ligand library:

Receptor = receptor.pdbqt Ligand = ligand.pdbqt Out = out.pdbqt center_x = X center (from grid box information) center_y = Y center (from grid box information) center_z = Z center (from grid box information) size_x = number of points in x dimension size_y = number of points in y dimension size_z = number of points in z dimension exhaustiveness = 8 (high is good)---save this as conf in the desired folder)

"~/autodock_vina_1_1_2_mac_catalina_64bit/bin/vina"(path of vina executable) --config conf.txt(configuration file) --log log.txt(output file)

vina.exe –-config conf.txt –-log log.txt

Load ligand -> Analyze -> docking -> docking results from autovina -> ouput -> multiple Conformations Load protein -> Analyze -> macromolecule -> receptor.pdbqt file Pressing right directed arrow we can see different conformations

• Shape Complementary Docking
• Protein and ligand should be in PDB format
• Output contains Score, Area of ligand-protein interface, ACE (atomic content energy, more the negative better the results), rotational transformation