Download Document

G. Narahari Sastry
Molecular Modelling Group
Organic Chemical Sciences
Indian Institute of Chemical Technology
Hyderabad – 500 007
[email protected]; [email protected]
http://203.199.182.73/gnsmmg
National Seminar on BioInformatics - Pondicherry
Drug Discovery & Development
It starts with disease identification
Isolate protein
involved in
disease (2-5 years)
Find a drug effective
against disease protein
(2-5 years)
Scale-up
Human clinical trials
(2-10 years)
Preclinical testing
(1-3 years)
Formulation
FDA approval
(2-3 years)
Discovery and Development of Drugs
Discover mechanism of action of disease
Identify target protein
Screen known compounds against target or
Chemically develop promising leads
Find 1-2 potential drugs
Toxicity, pharmacology
Clinical Trials
Genomic Approach to Drug Discovery
Target Discovery
Genome data
Existing Chemical and
biochemical knowledge
Functional & comparative Genomics
Target gene annotation
Literature
Target
Prioritization
Biochemical & Cell
Experimental Based Assays
Validation
A
B
C
GO terms
1. Molecular Function
2. Biological process
3. Cellular component
Translated gene products
Functionally validated
target
A
B
C
Comparative Proteomics
Role of targets
in disease
Drug Development
Sequence-structure
analysis
Small
molecule lead
HTS+/- in silico SBDD
Screening and improvement
Therapeutic Application
Screening and Optimization Cycle with in-silico
components
Database
clustering
Target
Selected
Similarity
analysis
QSAR
pharmaco Structure
phore
based
design
Assay
HTS Chemistry Target
developed
begins
structure
obtained
Candidate
taken
forward
Docking
Indirect
Drug Design
Protein Modeling
Protein Analysis
Nucleotide Sequence
Analysis
BI
Virtual Screening
106 small-molecule compounds
vHTS: MM + scoring
functions
N x 102 leads
Filters: ADMET / QSAR
M x 101 leads
Filters: synthesis /
manufacturing / IP / patent
/ biological assays
1 - 5 leads
Integration of Chemoinformatics and Bioinformatics
Genomic
Biology
Large Molecule
Targets
Bioinformatics
Assays
High
Throughput
Screening
In silico
Small
Molecules
Computational
chemistry
Cheminformatics
Much About Structure
• Structure
Function
• Structure
Mechanism
• Structure
Origins/Evolution
• Structure
Anything!!!
Quantum Mechanics
“The underlying physical laws necessary for the
mathematical theory of…the whole of chemistry are thus
completely known, and the difficulty is only that the exact
application of these laws leads to equations much too
complicated to be soluble.”
-P. A. M. Dirac
• Exact solutions are available only for Hydrogen atom.
• Modeling any realistic system needs approximations
(mathematically not solvable)
• Plenty of approximations were put forward to tackle
mathematic complexity
Chemistry is an experimental science
But alternative routes are attractive at times!!!
Experimental
Computational
X-Ray
NMR
Structure, Stability
and Reactivity
Thermochemistry
…
…
Semiempirical
Ab Initio
DFT
Molecular Dynamics
Simulations
Monte Carlo
…
Results
Factual Data!!!
Understanding, Patterning and Predicting
Qualitative theory, Concepts, Rules, Correlations
Basis for Doing Science and Doing it Better
The Jargon of nomenclature
•
•
•
•
•
•
•
•
Molecular Modeling
Computational Chemistry
Theoretical Chemistry
Simulations
Quantum Chemistry
Computational Biology
Molecular Dynamics
Mathematical Chemistry
Central Paradigm: Deriving information on molecular systems
without really synthesizing them.
Computational Chemistry
Quantum Mechanics (QM)
Hybrid QM / MM
Molecular Mechanics (MM)
Semi-empirical (SE)
The current scenario in chemistry
• Computation has become an effective
alternative to explore the structural,
energetic, mechanistic and other properties
of small molecules (say less than 8-10
atoms).
SOMETIMES THE COMPUTATIONAL
ACCURACY SUPERCEDES THE
EXPERIMNTAL ACCURACY
Every Computational Experiment at Any
Level of Theory Yields an Answer…
Usually Answers for Many Questions
Judging the Reliability is the Crucial Task
Just Like Experiments Fail, Computations
Fail
The paradigm shift …
However, the challenges are of
different kind in modeling
chemistry and biology!!
It is not only the size but the
philosophy!!!..!!!
Biological Structure
Sequence
3D
structure
MESDAMESETMESSRSMYN
AMEISWALTERYALLKINCAL
LMEWALLYIPREFERDREVIL
MYSELFIMACENTERDIRATV
ANDYINTENNESSEEILIKENM
RANDDYNAMICSRPADNAPRI
MASERADCALCYCLINNDRKI
NASEMRPCALTRACTINKAR
KICIPCDPKIQDENVSDETAVS
WILLWINITALL
Structural Scales
polymerase
SSBs
Complexes
helicase
primase
Organism
Assemblies
Cell
Structures
System Dynamics
Cell
Bottlenecks in developing
Structure – Function Relationships



Structures determined by NMR, computation,
or X-ray crystallography are static snapshots
of highly dynamic molecular systems
Biological process (recognition, interaction,
chemistry) require molecular motions and time
dependent.
To comprehend and facilitate thinking about
the dynamic structure of molecules is crucial.
Relevant timescales
Bond
vibration
10-15
femto
MD
step
Isomeris- Water
ation dynamics
10-12
pico
Helix Fastest
forms folders
10-9
nano
long
MD run
Conformati
onal
transitions
10-6
typical
folders
10-3
100
micro
milli
where we
need to be
Enzyme
catalysis
slow
folders
seconds
where we’d
love to be
Protein folding
Ligand
binding
How does the drug differ from an
inhibitor?
 Selectivity
 Less toxicity
 Bioavailability
 Reach the target
 Ease of synthesis
 Low price
 Slow (or) no development of resistance
 Stability upon storage as tablet or solution
 Pharmacokinetic parameters
 No allergies
Bioavailability (ADMET)
• ADMET
• Adsorption
• Distribution
• Metabolism
• Excretion
• Toxicity
• Model and Predict:
• Biotransformations & metabolites
• Catalytic reactions
• Drug-receptor interactions
• GI physiology
• Transepithelial transport
• Epithelial permeability
• Solubility
• Toxicity
Which Strategy?
• Do you have a validated target?
• Do you have active ligands?
• Do you have both?
Computer Aided Drug Design
Science
Support
Drug Design
Structure based
Ligand based
Target (structure)
based drug design
Ligand (analog)
based drug design
Receptor
Receptor
structure is
known
Mechanism is known
Ligands and their
biological activities are
known/ unknown
structure is not
known
Mechanism is known/
unknown
Ligands and their
biological activities are
known
Various Steps Involved
•
•
•
•
•
•
Get the structure of the receptor
Identify the active site
Build a library of possible ligands
Docking & Scoring
Understand receptor-ligand interactions
Design new ligands
Structure Based Ligand Design
H N
O
Docking
Linking
Building
H N
O
O
H
?
?
?
H N
H N
?
O
H
O
O
H
O
O
O
?
H N
S
O
?
O
N
H N
O
O
H
H
O
O
H N
S
O
H
O
CADD Success Stories
• FKBP Ligand
• docking and scoring
• P. Burkhard et al., J. Mol. Biol. 287, 853-858, 1999
+
• K ion channel blocker
• fragment-based evolutionary design
• G. Schneider et al., J. Computer-Aided Mol. Design 14, 487-494, 2000
2+
• Ca antagonist / T-channel blocker
• pharmacophore similarity search
• G. Schneider et al., Angew. Chem. Int. Ed. Engl. 39, 4130-4133, 2000
• Glyceraldehyde-phosphate DH inhibitors
• combinatorial docking
• J.C. Bressi et al., J. Med. Chem. 44, 2080-2093, 2001
• Thrombin inhibitor
• docking, de-novo design
• H.J. Bohm et al., J. Computer-Aided Mol. Design 13, 51-56, 1999
• HIV-1 RNA TAR inhibitor
• docking, database search
• A.V. Filikov et al., J. Computer-Aided Mol. Design 14, 593-610, 2000
• Aldose reductase inhibitors
• 3-D database searching
• Y. Iwata et al., J. Med. Chem. 44, 1718-1728, 2001
• DNA gyrase inhibitor
• structure-based virtual screening
• H.J. Boehm et al., J. Med. Chem. 43, 2664-2674, 2000
• Let us look at some of recent interests
Broad Objectives: Aiding the experimentalists in
Drug/Molecule/Reaction design
We strongly believe that while chemistry and biology are experimental sciences
THEORY-EXPERIMENT INTERPLAY IS INDISPENSABLE
• Theoretical/computational approaches to bring insights
which might trigger interest of the prospective
experimental groups
(Usually with no collaboration with experimentalists)
• Rationalizing the experimental finding with
computations and participate in the designing of
experiments
(In collaboration with experimentalists or groups of
experimentalists)
Non-availability of the receptor structure
is a bottleneck…
In our pursuit to engage with experimentalists for
lead discovery or optimization, our efforts become
restricted in the absence of an experimental
structure of the receptor protein/enzyme.
When we analyze, it occurred to us that most of
these ‘important target receptors’ whose structures
are not available belong to the class of ‘membrane
proteins’.
MEMBRANE PROTEINS – What are they
• Membrane
proteins are those that exist in cell
membranes.
• They can serve as structural supports, as both passive
and active channels for ions and chemicals, or serve more
specialized functions such as light reception.
• Membrane proteins form about 25% of all protein
sequences.
(They constitute close to 70% of drug targets)
• Only 2% of PDB structures belong to membrane
proteins!
Sastry et al, Computational Biology and Chemistry, 2006, in press
Membrane proteins form about 25% of all protein sequences.
Only 2% of PDB structures belong to this class!
Membrane Proteins: Classification…
• Receptors for extracellular ligands
Ex :- G-Protein coupled receptors
Tyrosine kinase receptors
• Transport proteins
Ex :- Molecular translocators
Ion channels
• Membrane-bound enzymes
Ex :- Lipid synthases
Cytochrome P-450 enzymes
• Proteins associated with cytoskeletal network
Ex :- Cytoskeletal attachments
• Proteins associated with energy production
Ex :- Photosynthetic complexes
Respiratory chain complexes
Challenges in computer simulations
of membrane proteins.
• Heavy molecular weight and size.
• Their association with lipid bilayer.
• Technical limitations related to the accuracy of the empirical
potential function.
• Difficulties with accurately incorporating important variables
such as pH, transmembrane potential.
• Starting configuration of a simulation may also bias the
results in undesirable ways.
• Comparative
protein
modelling
approaches
are
very
essential
Sastry et al, Computational Biology and Chemistry, 2006, in press
HUMAN AROMATASE: A PERIPHERAL MP
PLAY A MAJOR ROLE
IN STEROID AND
INHIBITOR BINDING.
ACIDIC RESIDUES
HEME
HOMOLOGY MODEL
•Membrane bound microsomal cytochrome P450 enzyme.
•Converts androgens to estrogens by aromatisation of A-ring of steroids.
•Estrogens and their carcinogenic metabolites are responsible for progression of breast cancer
WHAT IS THE ROLE OF THESE ACIDIC RESIDUES IN THE AROMATIZATION MECHANISM?
Sastry et al, J. Com. Aided Mol. Design, 2006, in press
Our Attempts of Modeling Aromatase
• A protein model is constructed (based on
CYP 2C5 (pdb code: 1NR6, sequence
identity is found to be 28%)
• The role of acidic residues in controlling
the function(substrate binding with
androstenedione, testosterone and norandrogens) is studied.
• Studies help in designing putative
inhibitors to control the aromatase
activity.
Sastry et al, J. Com. Aided Mol. Design, 2006, in press
PROPOSED AROMATIZATION MECHANISM
A-ring of ANDROGENS
O
ANDROGEN
A
O
MOLECULAR DYNAMICS SIMULATIONS
Before complexation to steroidal substrates
No H-bond interaction
Environment suitable for carboxylate formation
High conformational flexibility
ACTIVE SITE ACIDIC RESIDUES
MOLECULAR DOCKING
After complexation to steroidal substrates
H-Bond formation
Repulsive interaction
predicted.
CLAMPED !
Flexibility decreases. Environment
suitable for carboxylate formation.
A MOLECULE WHICH ARRESTS THESE PROPERTIES IS PROPOSED TO BE AN INHIBITOR
Inhibition of aromatase activity by 4-hydroxy androstenedione
(formestane)
Critical H-bond between inhibitor and T310
hampering its’ role in the mechanism.
ONE COULD DESIGN A MOLECULE
BY ADDING OR DELETING
A GROUP FROM ANDROGEN
SKELETON TO ARREST THE
PROPERTIES OBSERVED FOLLOWING
COMPLEXATION.
O
ANDROSTENEDIONE
(Substrate)
A
O
O
FORMESTANE
(Inhibitor)
A
O
OH
ACTIVE SITE
Human 5-lipoxygenase (5-LO)-Peripheral MP
Catalytic domain
Non-heme iron
MODEL
Ca(2+) binding
Mg(2+) binding
Tryptophan residues
β-barrel domain
•5-LO catalyses the rate limiting steps in leukotriene synthesis.
•Calcium binds reversibly to 5-LO, triggering its translocation from
the cytoplasm to the nuclear membrane.
Sastry et al, Biophys. Biochem. Res. Comm, 2004, 320, 461-467
-barrel domain
•Two calcium binding sites are identified ; ligating residues: F14, A15, G16, D18,
D19, L76 and D79.
Ca(2+) location
Important residues
which affect activity
are marked.
Gastric Proton Pump H(+)K(+)-ATPase – Integral MP
ANTI-ULCER TARGET
ATP binds here
Cytoplasmic
Phosphorylation.
E1
E2
Inhibitor binding sites.
Cation binding sites
Transmembrane
Lumenal
•Expose ion binding sites sequentially to each side of the membrane.
Sastry et al, Biophys. Biochem. Res. Comm, 2004, 319, 312-320; Biophys. Biochem. Res.
Comm, 2005, 336, 961-966
Inhibitor Binding in TM region
Inhibitor binding sites
CYS323
Covalent linkage
Omeprazole
CYS815
However, the large SBA in E2 precludes the covalent binding of
Cys815 to omeprazole. This suggested another intermediate
conformation with slightly more exposed Cys815. The existence of
stable intermediate structures has been proved in 2004.
Cation binding in E1 conformation
T825
Q941
D826
E822
H3O+
N794
E345
H3O+
V343
A341
V340
E797
Proposed hydronium binding.
Cα – carbons of arenes
in the pump.
Regular disposition
aids hydronium
transport.
Amino acid ligands (D,E,N,Q) that bind to metal ions in proteins
# of Binding structures for metals
PDB (June 2004)
Ca2+ : 2020;
Cu(II) : 298 Ni(II) : 118
Na+ : 678;
Mn (II) : 454 Co(II) :101
K+
Fe (II) : 100 Fe(III) :269
:
258;
Typical non-covalent binding to cations (from
Mg2+ : 1167;
PDB). The distances between the ligating
atoms and ion vary for different cations.
Asp
Glu
Asn
Zn (II) : 1545
Gln
In general, the acidic amino acid or their amides (ASP, GLU, ASN, GLN)
are present in the ligating sphere of the cations (Ca, Na, K, Mg, etc.) .
Additional ligating amino acid residues: Ala, Val, Thr, Leu, Phe etc.
An investment in knowledge
pays the best interest.
Benjamin Franklin
CAUTION….
•Don't be a naive user!?!
•When computers are
applied to biology, it is
vital to understand the
difference between
mathematical & biological
significance
•computers don’t do
biology, they do sums
quickly
macromolecular structure
methods
protocols
Structure determinations
methods
Traditional Approach
Rational Approach
It’s like a game of LUDO
Done
99
98
97
96
95
94
93
92
91
81
82
83
84
85
86
87
88
89
90
80
79
78
77
76
75
74
73
72
71
61
62
63
64
65
66
67
68
69
70
60
59
58
57
56
55
54
53
52
51
41
42
43
44
45
46
47
48
49
50
40
39
38
37
36
35
34
33
32
31
21
22
23
24
25
26
27
28
29
30
20
19
18
17
16
15
14
13
12
11
1
2
3
4
5
6
7
8
9
10
Drug Discovery
“This isn’t rocket science.
This is much harder.”
-- Alan Holmer
-- President, PhRMA
GNS, Dr. G. Madhavi Sastry, Dr. Y. Soujanya, Srinivas Reddy, Punnagai,
Gayatri, Srivani, Sateesh, Nagaraju, Dolly, Srinivasa Rao, Prasad, Mukesh,
Murty, Usha Rani, Srinivas, Janardhan, Bharat, Upendra.
Past Ph.D. students: Dr. U. Deva Priyakumar, Mr. T.C. Dinadayalane