Discovery and Design of Drugs
Choosing a drug target
• Drug targets – agonist, antagonist for receptor or enzyme inhibitor
• The more selective and specific, the better the drug as there’s a less chance of interaction with different
targets (also means less side effects)
• Ideally, enzyme inhibition needs to be specific to a single enzyme
Identifying a bioassay
• The bioassay should be simple, quick and relevant
• In-vitro tests – isolated cells, tissues. This is preferred as it’s cheap and easy to carry out
• In-vivo tests – in animals
• High throughput screening – involves robotics and miniaturisation of in-vitro tests. Many compounds
are automatically tested against different targets. Positive hits have activity in 30μM-1nM but there can
be false positive hits too.
Finding a lead compound
• Natural: sources include plants (salicylic acid, nicotine, galantamine), microorganisms (penicillin,
cephalosporins, aminoglycosides), marine sources (antivirals, anticancer drugs), animals (Botox).
Typically isolated as fermentation broths, plant extracts or animal fluids
• Synthetics: typically, small molecules made from a variety of sources
• Biologics: natural products, recombinant products
• Methods = starting from a natural ligand, combinatorial and parallel synthesis, computer aided design of
leads, serendipity, computerised searching, fragment-based lead discovery
Properties of lead compounds
• Lipinski’s rule of 5 – molecular weight <500amu, <5 HBD groups, <10 HBA groups, calculated logP of <5
(measure of hydrophobicity)
• Typically need to increase the size and hydrophobicity, have more aromatic rings and/or HBAs and lastly
consider ligand binding/efficiency of the lead
Optimising target interactions
• There are various aims for drug design to obtain the final drug – good selectivity, easily synthesised,
chemically stable, non-toxic, pharmacodynamically stable, minimal side effects
• Pharmacodynamic properties optimise the interaction between the drug and target to obtain
• Pharmacokinetic properties to increase the ability to reach the target
• PD and PK properties should have equal priority in influencing drug design strategy
Structure-activity relationships (SARs)
• Aim: to identify which parts of the molecule are essential for activity
• If crystals can be grown of the lead in the binding site, then the x-ray crystallography structure can be
solved
, Drug Discovery Advances
1. Receptor Structure Determination
• X-ray crystallography can determine the 3D structure:
o High-resolution detail (atomic resolution)
o Bond distances, angles, stereochemistry and configuration
o Proteins or protein-drug complexes (if co-crystallised)
• Many receptors are membrane proteins. This means it’s difficult to identify 3D structures as there
are isolation problems since these are found within the membranes
Applications of crystal structure in drug design
1. Priori drug design (aka de novo) – crystal structure of the target is used to create the initial
structurally novel lead compound with the aid of molecular graphics
2. Posteriori analysis – combines existing structure-activity knowledge with the x-ray structure and
proposes design improvements
Example: HIV protease (HIV-1 PR) Inhibitors
• X-ray studies of recombinant expressed HIV-1 PR suggest a substrate binding site (priori).
• The crystal structure of a complex of an inhibitor bound to HIV-1 PR was determined as was a
close contact map detailing hydrogen-bonding interactions.
• Improvements in future inhibitor design (posteriori).
• Currently, there are several HIV protease inhibitors in the market – ritonavir, saquinavir
2. Multi-dimensional NMR
Disadvantage of x-ray = the structure is not under the biological condition in solution
✓ NMR is conducted in solution and provides fast, structural information
3. Bioinformatics
• Information technology and innovative software is applied to genomics and molecular biology as the
molecules and processes are studied in silico (with computers)
Molecular Modelling
• Software offers techniques beyond traditional graphs.
• The properties of molecules can be visualised:
o Mapped onto the surface of the molecule using colours
o Hydrophobicity can be mapped onto the surface
o Non-physical properties i.e. how highly conserved the residue is or similarity to another
protein can be visualised
Computational Chemistry
• Required for accurate modelling or molecular behaviour.
• Properties should be calculated and assigned and should accurately estimate experiment
• Primary computations = energy (thermodynamics) and kinetics
• Methods for calculations = quantum mechanics or molecular mechanics
o Quantum mechanics – derived from the quantum theory, these calculations are based on
the extended Schrodinger equation that relates the motion of electrons and nucleus to the
potential energy. However, this is time-consuming and resource intensive. Therefore,
approximations are used: semi-empirical methods.
o Molecular mechanics – based on simple empirical approximations, here the focus is on the
motion of the molecules. The energy of the molecule is calculated from the bond angle and
other interatomic interaction energies. The process is fast but less accurate compared to
quantum mechanics.
Choosing a drug target
• Drug targets – agonist, antagonist for receptor or enzyme inhibitor
• The more selective and specific, the better the drug as there’s a less chance of interaction with different
targets (also means less side effects)
• Ideally, enzyme inhibition needs to be specific to a single enzyme
Identifying a bioassay
• The bioassay should be simple, quick and relevant
• In-vitro tests – isolated cells, tissues. This is preferred as it’s cheap and easy to carry out
• In-vivo tests – in animals
• High throughput screening – involves robotics and miniaturisation of in-vitro tests. Many compounds
are automatically tested against different targets. Positive hits have activity in 30μM-1nM but there can
be false positive hits too.
Finding a lead compound
• Natural: sources include plants (salicylic acid, nicotine, galantamine), microorganisms (penicillin,
cephalosporins, aminoglycosides), marine sources (antivirals, anticancer drugs), animals (Botox).
Typically isolated as fermentation broths, plant extracts or animal fluids
• Synthetics: typically, small molecules made from a variety of sources
• Biologics: natural products, recombinant products
• Methods = starting from a natural ligand, combinatorial and parallel synthesis, computer aided design of
leads, serendipity, computerised searching, fragment-based lead discovery
Properties of lead compounds
• Lipinski’s rule of 5 – molecular weight <500amu, <5 HBD groups, <10 HBA groups, calculated logP of <5
(measure of hydrophobicity)
• Typically need to increase the size and hydrophobicity, have more aromatic rings and/or HBAs and lastly
consider ligand binding/efficiency of the lead
Optimising target interactions
• There are various aims for drug design to obtain the final drug – good selectivity, easily synthesised,
chemically stable, non-toxic, pharmacodynamically stable, minimal side effects
• Pharmacodynamic properties optimise the interaction between the drug and target to obtain
• Pharmacokinetic properties to increase the ability to reach the target
• PD and PK properties should have equal priority in influencing drug design strategy
Structure-activity relationships (SARs)
• Aim: to identify which parts of the molecule are essential for activity
• If crystals can be grown of the lead in the binding site, then the x-ray crystallography structure can be
solved
, Drug Discovery Advances
1. Receptor Structure Determination
• X-ray crystallography can determine the 3D structure:
o High-resolution detail (atomic resolution)
o Bond distances, angles, stereochemistry and configuration
o Proteins or protein-drug complexes (if co-crystallised)
• Many receptors are membrane proteins. This means it’s difficult to identify 3D structures as there
are isolation problems since these are found within the membranes
Applications of crystal structure in drug design
1. Priori drug design (aka de novo) – crystal structure of the target is used to create the initial
structurally novel lead compound with the aid of molecular graphics
2. Posteriori analysis – combines existing structure-activity knowledge with the x-ray structure and
proposes design improvements
Example: HIV protease (HIV-1 PR) Inhibitors
• X-ray studies of recombinant expressed HIV-1 PR suggest a substrate binding site (priori).
• The crystal structure of a complex of an inhibitor bound to HIV-1 PR was determined as was a
close contact map detailing hydrogen-bonding interactions.
• Improvements in future inhibitor design (posteriori).
• Currently, there are several HIV protease inhibitors in the market – ritonavir, saquinavir
2. Multi-dimensional NMR
Disadvantage of x-ray = the structure is not under the biological condition in solution
✓ NMR is conducted in solution and provides fast, structural information
3. Bioinformatics
• Information technology and innovative software is applied to genomics and molecular biology as the
molecules and processes are studied in silico (with computers)
Molecular Modelling
• Software offers techniques beyond traditional graphs.
• The properties of molecules can be visualised:
o Mapped onto the surface of the molecule using colours
o Hydrophobicity can be mapped onto the surface
o Non-physical properties i.e. how highly conserved the residue is or similarity to another
protein can be visualised
Computational Chemistry
• Required for accurate modelling or molecular behaviour.
• Properties should be calculated and assigned and should accurately estimate experiment
• Primary computations = energy (thermodynamics) and kinetics
• Methods for calculations = quantum mechanics or molecular mechanics
o Quantum mechanics – derived from the quantum theory, these calculations are based on
the extended Schrodinger equation that relates the motion of electrons and nucleus to the
potential energy. However, this is time-consuming and resource intensive. Therefore,
approximations are used: semi-empirical methods.
o Molecular mechanics – based on simple empirical approximations, here the focus is on the
motion of the molecules. The energy of the molecule is calculated from the bond angle and
other interatomic interaction energies. The process is fast but less accurate compared to
quantum mechanics.
