Rational drug design

Question 1

Challenges of the drug discovery pipeline

Answer 1

Increasing costs as move towards large scale clinical trials (and even in toxicity tests)

Low success rate - lots of compounds have off target effects

late stage attrition

Question 2

The drug discovery pipeline

Answer 2

1. identify a target - basic reserch, understanding signalling pathways and receptors/proteins
2. Identify hits - compounds that have affinity for the target of interest, start to improve affinity until you get to high affinity drug binders
3. Prelinical studies - toxicity, pharmacological profiles done on animals
4. Drug development - clinical trials I-IV (Humans), review for regulatory approval

Question 3

Goal of drug descovery pipeline

Answer 3

Identify successful candidates early

minimise late stage attrition

Question 4

Methods for drug design

Answer 4

Bichemical assays
- thermal stability assays (screening)
- binding assays e.g. raiolabeled
- quantitative structure activity relationship (QSAR) - find common pharmacophores
Computational
- molecular docking
- molecular dynamics
allows screening for array of compounds - lower cost and effort
Structure based
- X-raycrystallography
- NMR
high cost associated, but rewards are great as high resolution info

Question 5

Give examples of experimental drug screening methods and what they help discover

Answer 5

Diverse e.g. isothermal calorimetry, mass spec, SPR

Characterise binding e.g.
finding Kd (to get affinity)
understand mode of binding
screening using QSAR

Idealling methods are high throughput and have low utilisation of the target (e.g. only micrograms of protein) e.g. SPR allows this

Question 6

whats the disadvantage of the approaches to computing drug binding

Answer 6

Demands we have structure or homology

Question 7

explain molecular dynamics (MD)

Answer 7

Uses newtons second law of motion and applying it to all the particles of a system

with a protein we know where all the atoms are and their weight, and the forces inbetween the different atoms

sum all these things up in a force field

take protein, take drug, do an MD simulation

Question 8

Advantages of MD for drug design

Answer 8

inherently encodes molecular flexibility

solvent included - we know where all our water molecules are

Atomistic description - can see atomic level interactions

Question 9

Challenges of MD simulations

Answer 9

Calculations at ns (nanoseconds) - only ran over a short period of time - protein binding is normally microseconds or milliseconds - so we have to have tricks to overcome this cant just simply do it

slow to sample a wide range of potential binding sites

newtonian physics knows nothing about electrons - so any interactions reliant on movment of electrons in the system, it doesnt explicitly deal with it

difficult to obtain delta G

Question 10

explain docking

Answer 10

have target, take crystal structure, and compound of interest and see if we can get it to fit into it

Question 11

How do you do a docking calculation

Answer 11

find the binding site:
- determine all possible conformations
- calculate the energy of the resulting complex

Question 12

Explain the different ways of determining the binding sites in docking

Answer 12

local - pre-existing knowledge of binding site, sononly relative orientation and conformation are varied, problem is not going to find allosteric binding sites

systematic search - systematically place the ligand at sites on the protein in all possible conformations and measure interaction energy

Random - random conformation of the drug are placed randomly about the protein and the interaction scored

Stimulated - MD driven drug conformations, followed by simulated annealing based docking

Question 13

after finding out possible binding sites, how do we identify which molecules binded optimally? in docking

Answer 13

Use a scoring functions

Question 14

Types of scoring functions

Answer 14

Forcefeild:
More chemically accurate, based on the forcefield used in MD, Dock, autodoc

Empirical/knowledge based:
based what we know experimentally e.g. if it forms a strong H bond - give it a higher score, Chemscore, drugscore ect.

Question 15

Types of molecular docking

Answer 15

Rigid body:
- assumes drug and target geometry is fixed
- Ad: Computationally very efficient, only 6 degrees of freedom
-DisAd: If sidechains need to move to accomedate ligand, or vice versa, this won’t work

Alternative
Flexible:
1. make ligands flexible
2. make target flexible (specify residues ie in exposed binding site, so that the whole protein isn’t all moving like in MD)
DisAd- 100/1000s degrees of freedom, makes it computationally expensive

Question 16

Explain how molecular docking and MD can be combined

Answer 16

Use MD - Identify different conformational states of target protein, allows identification of different representative drug binding sites

then do a cluster analysis, take representative structure from each cluster, and then. do a docking analysis for each of the conformational states

Question 17

When do computational methods work well?

Answer 17

• when binding pockets are well defined
• when theres limited flexibility in target and ligand

Question 18

What do computational methods struggle with?

Answer 18

Cryptic binding sites

Question 19

What can information from computational methods lead to?

Answer 19

Experimental screens
Lead optimisation

Question 20

NMR based methods for drug design

Answer 20

Ligand based:
- observing the ligand to see if it is binding to the target molecule

Target based:
- looking at target to see if the ligand changes what the target looks like

Question 21

Advantages of ligand based NMR methods in drug design

Answer 21

• very small quantities of target protein needed
• don’t need to label molecules with NMR sensitive isotopes
• Fast
• no size restriction

Question 22

Examples of ligand based NMR methods

Answer 22

Saturation-Transfer Difference NMR (STD-NMR)

Transferred NOE (TrNOE)

Question 23

Features of target based NMR methods for drug design

Answer 23

• need larger quantities of target (mg)
• Typically limited to smaller targets
• Provide with info on drug binding site (identification of interacting residues)

Question 24

Features of Saturation-Transfer Difference NMR (STD-NMR)

Answer 24

Ligand based NMR
- very easy
- acquire a proton spectrum of ligand in solution of target protein
- Can then saturate one of the signals from the protein, and suppress signals from another protein, if we subtract these, get a spectrum of the drug thats binding
- Using a mixture of ligands and using subtraction, can see which ligand is giving signals and is therefore bound

can also start to see which parts of the molecule are binding

Question 25

Explain protein based NMR screening in more detail

Answer 25

NMR resonances sensitive to local environment - we can see where binding sites are based on which peaks of the spectra move upon ligand binding

Take this one step further using titration, this allow for identification of Kds of the ligands being tested

Question 26

What is crystallography useful for in drug design

Answer 26

Lead optimisation and identifying initial compounds

Question 27

Ways of finding drug/target structure via crystallography

Answer 27

Co-crystallization:
- relatively slow
- trying to get drug and target to crystallise at the same time
- difficulty - have to establish the crystallisation conditions for each target with its drug

Soaking:
- Grow crystals in a well defined environment
- take these and put them into a solution containing compound of interest
-with a high enough conc of ligand, it will diffuse into the crystal, and hopefully into the binding site
- difficulty - need a very soluble drug to drive diffusion into crystal
- good for screening as can generate a stock of crystals to put with compound of interest

Question 28

What is fragment based drug design?

Answer 28

Uses structural techniques to try and identify what molecules should be designed in the first place

take fragments or small molecules, screen using NMR and crystallography, find binding with lower affinity, then build up the fragments into more complete molecules (essentially grow new drug)

e.g. used to target SARS-CoV-2 - target spike proteins and RNA polymerases