Transcript slides - CRIStAL

Structural biology and drug design
as simple as this …
de Ruyck Jérôme
30/11/2015
Lille - France
Introduction
https://en.wikibooks.org/wiki/Structural_Biochemistry
Introduction
Pre-clinical studies
Chemistry
Pharmacology
2-4 years
Scalingup
2 months
- 1 year
Clinical trials
Phase 1
Phase 2
3-5 years
Phase 3
NDA
Phase 4
2-3 years
Introduction
Pre-clinical studies
Scientific challenge
Chemistry
Pharmacology
Clinical trials
Phase 1
Scalingup
Phase 2
Phase 3
Phase 4
150 M $
Money
2-4 years
2 months
- 1 year
3-5 years
• Challenging
• Time consuming
• Expensive
NDA
2-3 years
• Efficiency increased
Multidisciplinary
approaches
• Time saved
• Cost effectiveness
Introduction
In-vitro drug-design
In-silico drug-design
Medicinal chemistry
High-throughput screening
Structural biology
Bioinformatics
Virtual screening
Molecular modelling
Introduction
In-vitro drug design
In-silico drug design
Medicinal chemistry
High-throughput screening
Structural biology
Bioinformatics
Virtual screening
Molecular modelling
Introduction
In-vitro drug design
In-silico drug design
Medicinal chemistry
High-throughput screening
Structural biology
Bioinformatics
Virtual screening
Molecular modelling
Molecular biology /
Bioinformatics
Target
Structural biology /
Molecular modelling
HTS
Hits
Medicinal chemistry
Leads
Drugs
Pharmacology /
PK-PD predictions
Introduction
In-silico drug design
Ligand-based
drug design
Target
Structure-based
drug design
Hits
Leads
Drugs
Outline
Virtual screening
Ligand-based drug design
Pharmacophore
Structure-based drug design
Ligand
Docking
Optimization
Protein
Lead design
In situ design
Fragment
Pharmacophore
Evaluation
Nomenclature
Different kind of inter- and intramolecular interactions
Polar interactions
Dipoles-Dipoles
Hydrophobic interactions
VdW – H-Bond
Aliphatic
Electrostatic
Aromatic
Salt bridges
Nomenclature
Aromatic interactions
Quadropular interactions
π – π interactions
π – cation interactions
Nomenclature
Quantum Mechanics (QM)
Computational Chemistry
Theoretical Chemistry
Molecular Mechanics (MM)
Molecular Modelling
Molecular Dynamics
Deriving information on
molecular systems
without really synthesizing them !
Quantum mechanics
• Nuclei and electrons separated
• Time consuming
• Applied to small molecules
• Not suitable for proteins
Method
Accuracy
Max atoms
Semiempirical
Low
2000
HF & DFT
Medium
500
Perturbation
methods
High
50
Coupled-cluster
Very High
20
Molecular mechanics
𝐸 𝑡𝑜𝑡 =
𝑏𝑜𝑛𝑑𝑠 𝐾𝑟 𝑟 − 𝑟0
2
+
Streching
𝑎𝑛𝑔𝑙𝑒𝑠 𝐾Θ Θ − Θ0 +
Bending
• Spheres and springs model
• Very quick
• Can be applied to small molecules
• Suitable for proteins
• Accuracy depends on a force field
𝑉𝑛
𝑑𝑖ℎ𝑒𝑑𝑟𝑎𝑙𝑠 2
1 + 𝑐𝑜𝑠 𝑛τ − 𝛾
Torsion
+
𝐴𝑖𝑗
𝑖<𝑗
12
𝑟𝑖𝑗
−
𝐵𝑖𝑗
6
𝑟𝑖𝑗
+
Non-bonded
𝑞𝑖 𝑞𝑗
𝜀𝑟𝑖𝑗
Force field
𝐸 𝑡𝑜𝑡 =
𝑏𝑜𝑛𝑑𝑠 𝐾𝑟 𝑟 − 𝑟0
Streching
2
+
𝑎𝑛𝑔𝑙𝑒𝑠 𝐾Θ Θ − Θ0 +
𝑑𝑖ℎ𝑒𝑑𝑟𝑎𝑙𝑠 2
Bending
• Different force fields for different systems
•
•
•
•
Proteins
Sugars
Metals
…
• Different parameterization
• Empirical
• Semi-empirical (including QM)
𝑉𝑛
1 + 𝑐𝑜𝑠 𝑛τ − 𝛾
Torsion
+
𝐴𝑖𝑗
𝑖<𝑗
12
𝑟𝑖𝑗
−
𝐵𝑖𝑗
6
𝑟𝑖𝑗
+
Non-bonded
𝑞𝑖 𝑞𝑗
𝜀𝑟𝑖𝑗
Direct vs Indirect approaches
Virtual screening
Pharmacophore
Ligand
IC50 / Ki
2D Structures
Docking
Ligand-based drug design
(Indirect approach)
Protein
3D structures
Lead design
In situ design
Fragment
2D Structures
Pharmacophore
Structure-based drug design
(Direct approach)
Direct vs Indirect approaches
Indirect approach
Direct approach
Don’t know receptor
Know Ligands
Know receptor
Don’t know ligand
Direct vs Indirect approaches
Indirect approach
Direct approach
Don’t know receptor
Know Ligands
Know receptor
Don’t know ligand
?
Statistical methods
Structural Biology
Pharmacophore
3D-QSAR
Protein - ligand interactions
Docking
Ligand-based drug design
Virtual screening
Ligand-based drug design
Pharmacophore
Ligand
IC50 / Ki
2D Structures
Optimization
Lead design
Fragment
2D Structures
Pharmacophore
Selection
Evaluation
Pharmacophore
A pharmacophore is a geometrical description of molecular features
which are necessary for molecular recognition of a ligand by a
biological macromolecule.
Typical pharmacophore features include hydrophobic centroids, aromatic
rings, hydrogen bond acceptors or donors, cations, and anions.
Pharmacophore
Inhibition data
generation
Structural superimposition
Example (1)
“A four-point pharmacophore of COX-2 selective inhibitors was derived from a training set of 16 compounds, using the Catalyst program. It
consists of a H bond acceptor, two hydrophobic groups and an aromatic ring, in accordance with SAR data of the compounds and with
topology of the COX-2 active site. This hypothesis, combined with exclusion volume spheres representing important residues of the COX-2
binding site, was used to virtually screen the Maybridge database. Eight compounds were selected for an in vitro enzymatic assay. Five
of them show COX-2 inhibition close to that of nimesulide and rofecoxib, two reference COX-2 selective inhibitors. As a result, structurebased pharmacophore generation was able to identify original lead compounds, inhibiting the COX-2 isoform.”
Example (2)
“Apolar trisubstituted derivatives of harmine show high antiproliferative activity on diverse cancer cell lines. However, these molecules
present a poor solubility making these compounds poorly bioavailable. Here, new compounds were synthesized in order to improve
solubility while retaining antiproliferative activity. First, polar substituents have shown a higher solubility but a loss of antiproliferative activity.
Second, a Comparative Molecular Field Analysis (CoMFA) model was developed, guiding the design and synthesis of eight new
compounds. Characterization has underlined the in vitro antiproliferative character of these compounds on five cancerous cell lines,
combining with a high solubility at physiological pH, making these molecules druggable. Moreover, targeting glioma treatment, human
intestinal absorption and blood brain penetration have been calculated, showing high absorption and penetration properties.”
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
908
717
526
408
343
290
236
11211
10817
10263
9757
9220
8697
67843
60473
53286
46615
40410
76020
84763
93645
101667
NMR
8176
7609
6956
34256
X-Ray
196
155
136
5991
28733
24295
19898
16306
13776
11389
5130
4254
110
87
75
3525
2994
46
2553
22
2139
13
Structural biology improvement
PDB STATISTICS
CryoEM
2015
Structure-based drug design
Virtual screening
Structure-based drug design
Ligand
IC50 / Ki
2D Structures
Docking
Optimization
Protein
3D structures
Lead design
In situ design
Fragment
2D Structures
Selection
Evaluation
Structure-based drug design
Virtual screening
Structure-based drug design
Ligand
IC50 / Ki
2D Structures
Docking
Optimization
Protein
3D structures
Lead design
In situ design
Fragment
2D Structures
Selection
Evaluation
Molecular docking
Protein-ligand docking is a computational method that mimics the
binding of a ligand to a protein to form a complex.
It predicts the pose of the molecule in the binding site and calculates a
score representing the strength of the binding.
Protein
Ligand
Docking
Binding site
Complex
How does it work ?
Protein-ligand docking software works in two different steps
Search algorithm
Scoring function
Generates a large number
of poses of a molecule in
the active site
Calculates a score or
binding affinity for a
particular pose
Genetic
Lamarckian
Simulated annealing
Forcefield-based
Empirical
Knowledge-based potentials
Performs automated docking with
full acyclic ligand flexibility, partial
cyclic ligand flexibility and partial
protein flexibility in and around
active site.
Scoring: includes H-bonding term,
pairwise dispersion potential
(hydrophobic interactions),
molecular and mechanics term for
internal energy.
Example (1)
“Crystal structures of Thermus thermophilus and Bacillus subtilis type 2 IPP isomerases were combined to generate an almost complete
model of the FMN-bound structure of the enzyme. In contrast to previous studies, positions of flexible loops were obtained and carefully
analyzed by molecular dynamics. Docking simulations find a unique putative binding site for the IPP substrate.”
Example (2)
“A total of 1,990 compounds from the National Cancer Institute (NCI) diversity set with nonredundant structures have been tested to inhibit
cancer cell lines with unknown mechanism. Cancer inhibition through EGFR-TK is one of the mechanisms of these compounds. In this work,
we performed receptor-based virtual screening against the NCI diversity database. Using two different docking algorithms, AutoDock and
Gold, combined with subsequent post-docking analyses, we found eight candidate compounds with high scoring functions that all bind to
the ATP-competitive site of the kinase. None of these compounds belongs to the main group of the currently known EGFR-TK inhibitors.
Binding mode analyses revealed that the way these compounds complexed with EGFR-TK differs from quinazoline inhibitor binding and the
interaction mainly involves hydrophobic interactions. Our results suggest that these molecules could be developed as novel lead
compounds in anti-cancer drug design.”
Structure-based drug design
Virtual screening
Structure-based drug design
Ligand
IC50 / Ki
2D Structures
Docking
Optimization
Protein
3D structures
Lead design
In situ design
Fragment
2D Structures
Selection
Evaluation
Fragment-based drug design
Fragment-based drug design is the screening of libraries of fragments
with low chemical complexity.
The fragments usually bind the protein target with low affinity (high mM). The
fragments selected for follow-up are then optimized by addition of chemical moieties
or linked together with the aim of obtaining a highly potent drug or inhibitor.
Examples (1)
“The search for new drugs is plagued by high attrition rates at all stages in research and development. Chemists have an opportunity to
tackle this problem because attrition can be traced back, in part, to the quality of the chemical leads. Fragment-based drug discovery (FBDD)
is a new approach, increasingly used in the pharmaceutical industry, for reducing attrition and providing leads for previously intractable
biological targets. FBDD identifies low-molecular-weight ligands (~150 Da) that bind to biologically important macromolecules. The threedimensional experimental binding mode of these fragments is determined using X-ray crystallography or NMR spectroscopy, and is used
to facilitate their optimization into potent molecules with drug like properties. Compared with high-throughput-screening, the fragment
approach requires fewer compounds to be screened, and, despite the lower initial potency of the screening hits, offers more efficient and
fruitful optimization campaigns.”
Examples (2)
“X-ray crystallography is an established technique for ligand
screening in fragment-based drug-design projects, but the
required manual handling steps – soaking crystals with ligand and
the subsequent harvesting – are tedious and limit the throughput
of the process. Here, an alternative approach is reported:
crystallization plates are pre-coated with potential binders prior to
protein crystallization and X-ray diffraction is performed directly
‘in situ’ (or in-plate). Its performance is demonstrated on distinct
and relevant therapeutic targets currently being studied for ligand
screening by X-ray crystallography using either a bending-magnet
beamline or a rotating-anode generator. The possibility of using
DMSO stock solutions of the ligands to be coated opens up a
route to screening most chemical libraries.”
The future is now …
Quantum Mechanics (QM)
Computational Chemistry
Theoretical Chemistry
Molecular Mechanics (MM)
Molecular Modeling
Molecular Dynamics
Deriving information on
molecular systems
without really synthesizing them !
Hybrid QM/MM
Simulations
Computational Biology
The ONIOM method
𝐸 𝑂𝑁𝐼𝑂𝑀, 𝑅𝑒𝑎𝑙 = 𝐸 𝑙𝑜𝑤, 𝑟𝑒𝑎𝑙 − 𝐸 𝑙𝑜𝑤, 𝑚𝑜𝑑𝑒𝑙 + 𝐸(ℎ𝑖𝑔ℎ, 𝑚𝑜𝑑𝑒𝑙)
S. Dapprich, et al. 1999 THEOCHEM. 461-462: 1
Inside the mechanism
“Here, we report an integrated quantum
mechanics/molecular mechanics (QM/MM) study
of the bioorganometallic reaction pathway of the
reduction
of
(E)-4-hydroxy-3-methylbut-2-enyl
pyrophosphate (HMBPP) into the so called
universal
terpenoid
precursors
isopentenyl
pyrophosphate
(IPP)
and
dimethylallyl
pyrophosphate (DMAPP), promoted by the IspH
enzyme. Dehydroxylation of HMBPP is triggered by
a proton transfer from Glu126 to the OH group of
HMBPP. The reaction pathway is completed by
competitive proton transfer from the terminal
phosphate group to the C2 or C4 atom of HMBPP.”
Mechanism-based drug design
“Development of novel influenza neuraminidase inhibitors is critical for preparedness against influenza outbreaks. Knowledge of the
neuraminidase enzymatic mechanism and transition state analogue, 2-deoxy-2,3-didehydro-N-acetylneuraminic acid, contributed to the
development of the first generation anti-neuraminidase drugs, zanamivir and oseltamivir. However, lack of evidence regarding influenza
neuraminidase key catalytic residues has limited strategies for novel neuraminidase inhibitor design. Here, we confirm that influenza
neuraminidase conserved Tyr406 is the key catalytic residue that may function as a nucleophile; thus, mechanism-based covalent
inhibition of influenza neuraminidase was conceived. Crystallographic studies reveal that 2a,3ax-difluoro-N-acetylneuraminic acid forms a
covalent bond with influenza neuraminidase Tyr406 and the compound was found to possess potent anti-influenza activity against both
influenza A and B viruses. Our results address many unanswered questions about the influenza neuraminidase catalytic mechanism and
demonstrate that covalent inhibition of influenza neuraminidase is a promising and novel strategy for the development of next-generation
influenza drugs.”
Acknowledgment
•
Computational Molecular Systems Biology team
Dr. M. Lensink
Dr. R. Blossey
Dr. J. Bouckaert
Dr. T. Dumych
Dr. E.-M. Krammer
Ir. G. Brysbaert
•
Fundings