Transcript Combinatorial Chemistry and Library Design

Combinatorial Chemistry and
Library Design
C371
Chemical Informatics Lecture
Based largely on the C&EN story
published October 27, 2003, pp. 45 ff.
Combinatorial Chemistry
• Definition: the synthesis of chemical compounds
as ensembles (libraries) and the screening of
those libraries for compounds with desirable
properties
• Potentially speedy route to new drugs, catalysts,
and other compounds and materials
• Technique invented in the late 1980s and early
1990s to enable tasks to be applied to many
molecules simultaneously
Combichem Techniques
• Tools
– Solid-phase synthesis
– Resins
– Reagents (Monomers)
– Linkers
– Screening methods
Combichem Methods
• Use of solid supports for peptide synthesis
led to wider applications
• Products from one reaction are divided
and reacted with other reagents in
succession
– Split-mix scheme: library size increases
exponentially
DIVERSE AND FOCUSED
LIBRARIES
• Many early disappointments led to:
– Design of smaller, more focused libraries with much
information about the target
• May concentrate on a family of targets (e.g., proteases or
kinases)
– Use of more diverse libraries when little is known
about the target
• “Primary screening libraries
• Give broad coverage of chemistry space
– Selection of compounds with “drug-like”
physicochemical properties
Problems with Early
Combichem Libraries
• Many compounds had undesirable
properties:
– Size
– Solubility
– Inappropriate functional groups
Criticism of the Technique
• Early libraries often based on a single skeleton
(basic structure)
• Limited number of skeletons accessible
• Individual library members were structurally
similar
• Compounds tended to be achiral or racemic
• Initial emphasis on creating mixtures of very
large numbers of compounds now out of favor
LIBRARY ENUMERATION
• Process by which the molecular graphs of
the product molecules are generated
automatically from lists of reagents (using
connection tables or SMILES strings)
– Fragment marking – Central core template
and one or more R groups
– Reaction transform approach – Transform is a
computer-readable representation of the
reaction mechanism: atom mapping
Advantages/Disadvantages
• Fragment marking generally a very fast
enumeration once core template and R
group fragments are defined.
– May be difficult to generate the core and to
generate fragments automatically
Combichem Techniques (cont’d)
• Markush-based approaches to
enumeration
– Ideally suited when a common core can be
identified
– Certain subsets of the product structures may
have features in common
COMBINATORIAL LIBRARY
DESIGN STRATEGIES
• Two Main Strategies:
– Monomer-based selection:
• Subsets of monomers selected without
consideration of the products
– Product-based selection:
• Properties of the resulting product molecules
influence the selection of the monomers
• Much more computationally demanding than
monomer-based selection, but can be more
effective when wanting to optimize the properties
of a library as a whole
APPROACHES TO PRODUCTBASED LIBRARY DESIGN
• Identify lists of potential reagents, filter them as
needed, and enumerate the virtual library
• Subject virtual library to virtual screening to
evaluate and score each structure
• Select reagents from results of virtual screening
plus additional criteria (degree of structural
diversity required, degree of similarity or
dissimilarity to existing collections)
– Usually done with optimization techniques (e.g.,
genetic algorithms or simulated annealing)
Alternatives to Product-Based
Library Design
• Molecule-based methods
– Appropriate for targeted or focused libraries
– Relatively fast, especially when combined
with optimization based on 2D properties
MULTIOBJECTIVE LIBRARY
DESIGN
• Optimizes multiple properties
simultaneously
• Balances diversity and focus
• Could search for drug-like properties
• Multiobjective Genetic Algorithm (MOGA)
PRACTICAL EXAMPLES OF
LIBRARY DESIGN
• See examples in the text for
– Structure-Based Library Design
– Library Design in Lead Optimization
TRENDS
• Design of smaller, more focused libraries
with as much information about the
therapeutic target as possible
– May use docking methods if target structure is
known
– Use pharmacophoric methods, 2D or
physicochemical properties if some actives
are known
• Focus on compounds with “drug-like”
physicochemical properties
New Combichem Techniques
• Current emphasis on arrays of fewer, wellcharacterized compounds
• Movement toward complex naturalproduct-like compounds
Recent Advances
•
•
•
•
Natural-product-like libraries
Dynamic combinatorial chemistry
Combinatorial optimization of catalysts
Multi-component reactions
New Approaches
• Use biologically relevant building blocks
• Use branching networks of reactions
• Produce libraries of natural-product-like
compounds
• Make all possible combinations of both
core skeletal structures and peripheral
groups
New Approaches
• Dynamic Combichem (DCC)
• Used to ID molecules that bind with high
affinity to macromolecular receptors OR
• Synthetic receptors that bind tightly to
small molecules
• Uses equilibrium forces to amplify
compounds that bind well to targets
New Approaches
• Combi Catalysis
– To discover and optimize catalysts
• Novel Methods for Combinatorial
Synthesis
– New linkages for solid-phase synthesis
– New multi-component reactions
New Combichem Techniques
•
•
•
•
Make compounds in parallel
Test them in parallel
Obtain new properties rapidly
Discrete compounds are produced by
parallel synthesis or by mixing synthesis
with directed sorting
Benefits to the Pharmaceutical
Industry
• Provides a stimulus for robot-controlled
and immobilization strategies that allow
high-throughput and multiple parallel
approaches to drug discovery
Benefits to Materials Science
• Combinatorial approaches now being
applied to solid-state and materials
applications
• Also to search for new catalysts
NIH Roadmap
• http://nihroadmap.nih.gov/
• Roadmap for Medical Research in the 21st
Century
• Includes: Molecular Libraries and Imaging
– NIH will assemble a huge combinatorial
library as a source of new drug candidates
– PubChem Database
• http://pubchem.ncbi.nlm.nih.gov/
CombiChem Web Sites
• CombiChem Lab
http://www.combichemlab.com
• Combinatorial Chemistry and High
Throughput Screening (Wendy Warr)
http://www.warr.com/ombichem.html