Structure Based Drug Design

Drugs and Proteins

Drugs either inhibit or activate a particular protein.

A high affinity (low Kd and negative ΔGbind) means a proportionally low dose is needed to bind to the target protein

Specificity is required, otherwise drugs will bind to other proteins and cause side-effects.

Drug Development Pipeline

  1. drug discovery
  2. preclinical trials
  3. clinical trials
  4. FDA review
  5. manufacturing

Drug Discovery

A target is a protein through to be related to the disease.

A hit is a small molecule that binds to the protein, altering function.

A lead is a hit that shows promise as a drug. It has drug like properties and alters the protein function.

pre-clinical testing is done on cell culture and animals.


Target identification

Determine which proteins are related to the disease or the pathogen.

Determine if target will bind to small molecules. Protein-protein interacting proteins are less likely to bind to small drug molecules.

Determine if existing drugs bind to similar proteins

Risk of side-effects, are there similar proteins with essential functions

Check for competitor drug companies.


GWAS can determine if variants are related to the disease or susceptibility.

Pathogen genomes have also been sequenced. sequence and structural analysis may reveal proteins that are unique to these pathogens and there could be multiple (hopefully new) protein targets.


Healthy and diseased tissue will have different quantities of proteins. Some may be important to the disease.

Large amounts of data will be gained, use stats to figure out which are important.

Target validation

Genetic experiments, mice or human cell culture.


Use experiments to find out which drugs alter activity in the target. Automated robots can help achieve high throughput testing, each time using a different drug on the target in vitro.

Pharma will have millions of compounds available for testing.

Use chemoinformatics to select compounds from databases based on their descriptors.

This can help reduce number of compounds that must be tested.

Screened using docking tests, or pharmocophore search.

Automated screening needs to prepare the compounds. Add hydrogens, model in 3D, tautomer and protonation states.

Filter based on relevant properties.

Docking is a molecular simulation that searches for a target protein + compound low energy state. This is computationally expensive and error prone.

3D pharmacophore search can be done based on geometrical info of the binding site. The search will look for complementary compounds. (e.g. hydrogen bond donors vs acceptors, charged molecules). It is much faster but has less detail and doesn't account for flexibility.

Leads (from Hits)

Hits must be filtered again.

Check for affinity and use experiments to confirm activity on the target protein.

Assess hit properties (solubility, binding specificity, stability, affinity, activity)

Patent check

hits based on similiar hits.

QSAR modelling of hits

more chemistry again based on QSAR model.

Check ADME - Tox

  1. Absorption - drug and be absorbed into the bloodstream
  2. Distribution - drug can travel to where it is required in the body.
  3. Metabolism - can be broken down efficiently
  4. Excretion - can be removed from the body
  5. Toxic - not toxic.

These functions are carried out by proteins.

Often clinical trials will fail because of these properties.

Lead Optimization

Explore chemistry

Improve drug-like props

Solve binding structures of lead to target protein

Explore structures

Affinity and Specificity

Both are sensitive to structural details and may need to be manually modelled.