Structural Biology and Structure-based Drug Design
On 16 October 2007 we will hold an eCheminfo Community of Practice conference session at Bryn Mawr College, Philadelphia to discuss latest advances in structural biology related to drug discovery. The session will be chaired by Max Cummings (Tibotec Pharmaceuticals) and includes a knowledgeable panel of speakers and discussion leaders: Charles Lesburg (Schering-Plough), Heather Carlson (University of Michigan), Gerard Kleywegt (University of Uppsala), Paul Labute (CCG), Alan Cheng (Amgen), and Ajay Jain (UCSF). A description of the session with presentation abstracts follows:
Structural Biology
http://echeminfo.com/COMTY_conferencesprog07structbio
Structural biology efforts
bring the 3D structures of proteins to light, and thereby greatly enable many
drug discovery efforts. In the affinity optimization phase of a medicinal
chemistry effort the timely determination of the 3D structure of a relevant
protein-ligand complex can have a huge positive impact. At the same time it is
important to note that a wide range of structural information can be useful in
guiding drug design efforts - from a single lowly 2D ligand structure to a set
of high resolution 3D structures of protein-ligand complexes.
Talks in this session cover different topics in the areas of structural biology
and/or structure-based drug design. Ligand structures and their known
activities can help to predict side effects, as well as possible new uses for
known drugs. How reliable is a given 3D protein structure, and how can users
evaluate this for themselves? Positioning of hydrogen atoms for biomolecular
calculations is a longstanding issue, and a new approach to solving this problem
is presented. Aspects of the recently popular concept of druggability are
explored by two of the speakers. Protein flexibility is discussed in the
context of protein-ligand structures related to the regulation of drug
metabolism. Attendees will be exposed to various ways in which structural
information is used in drug design, and should gain an appreciation for a few
currently emerging challenges in the fields of structural biology and
structure-based drug design.
(Please follow continuation to read abstracts of sessions talks.)
Abstracts
Charles Lesburg,
Schering-Plough
The pregnane X receptor (PXR) is responsible for upregulating cytochrome P450
expression in response to a wide variety of xenobiotics. This induction of
metabolism negatively impacts drug exposure and forms the basis for detrimental
drug-drug interactions. We have engineered a constitutively-active PXR ligand
binding domain and demonstrated that it is capable of binding known PXR
ligands. Moreover, we used this construct in an inverse structure-assisted drug
design effort to understand the structural basis of ligand binding to PXR. During
this analysis, we were presented with unexpected binding modes which underscore
the plasticity of the PXR ligand binding pocket.
Fundamental
differences between high- and low-affinity complexes of enzymes and non-enzymes
Heather A.
Carlson, University of Michigan
Stark differences between enzymes and non-enzymes were found through mining the
protein-ligand database, Binding MOAD (Mother of All Databases, pronounced
"mode" as a pun on the ligand's mode of binding to its protein
target). Ligand efficiencies were found to be much higher in non-enzymes. Every
atom, every square Ångstrom of contact gives more free energy in non-enzymatic
binding sites. This implies they may be more "druggable" targets than
enzymes because drug-like affinities can be obtained with smaller molecules
which are more easily absorbed and have less functional groups for potential
toxicity concerns. Additional data also suggest that divergent approaches may
be needed to improve the affinity of ligands for the two classes of proteins. High-affinity
ligands are much larger than low-affinity ligands for enzyme complexes. The
addition of complementary functional groups is likely to improve the affinity
of an enzyme inhibitor through more contact with the pocket, but this process
may not be as fruitful for ligands of non-enzymes. High- and low-affinity
inhibitors are the same size in non-enzymes. The inherent differences between
enzymes and non-enzymes have significant ramifications for scoring functions
and structure-based drug design.
Protein
crystallography: not as simple as ABC then?
Gerard J
Kleywegt, Uppsala University, Uppsala, Sweden
The recent debacle with the ABC transporter structures has shown (not for the
first time, unfortunately - and probably not for the last time either) that the
mere fact that a pretty crystal structure is reported in the pages of a glossy
journal does not guarantee that the structure is even close to being correct. Some
of the causes of serious errors in biomacromolecular crystal structures will be
discussed. In addition, a few examples of such errors will be given, and simple
ways in which non-experts can assess the overall reliability of a protein
crystal structure will be discussed. However, even when the overall structure
is reliable, this is not necessarily true for each and every detail. Particularly
relevant in this respect are the protein residues that interact with substrates
or inhibitors and the interacting molecules themselves. A few examples will be
used to demonstrate that the structures of ligands etc. are often less reliable
than those of their proteinaceous hosts.
Obviously, for users of protein structures these observations have important
implications. First and foremost, crystal structures should be treated with
healthy scepticism rather than reverence. The best way to assess the
reliability of critical aspects of a structure (ligands, active-site residues,
metal-binding sites, interface residues, etc.) is to inspect the experimental
electron density maps. Our efforts to provide such maps, and information derived
from them, for all crystal structures for which structure factors have been
deposited with the wwPDB will be discussed. They have resulted in EDS - the
Uppsala Electron Density Server (http://eds.bmc.uu.se/).
Assignment
of Protonation States and Geometries to Macromolecular Structures using Unary
Quadratic Optimization
Paul Labute, CCG
Many computational methodologies relating to macromolecular structures depend
significantly on the assigment of protonation states and proton coordinates. The
complexity of hydrogen bond networks, tautomers, ionization states and metal
ligands often make automated protonation assignment quite difficult. A method
based upon Unary Quadratic Optiimization is presented to address this problem. The
algorithm and theormodynamic theory is described along with some validation
results on high resolution crystal structures.
Structure-based
prediction of small-molecule druggability
Alan C. Cheng,
Amgen
Over half of drug discovery efforts fail in lead identification or in
optimization of leads for drug-like properties. We set out to reduce the
failure rate by identifying the more druggable targets early on, and discovered
that a model based on basic biophysical principles does this very well. The
basic idea is to quantitatively estimate the maximal affinity achievable at a
given binding site by a small molecule with certain ‘drug-like’ properties. These
estimates turn out to correlate reasonably well with drug discovery outcomes
and are useful in the prioritization of small-molecule drug targets. Computer-aided
approaches that have potential in tackling difficult druggability targets will
also be presented.
Drug
Target Modeling: Ligands Tell Us More Than We Think
Ajay N. Jain,
Ph.D., Professor, Cancer Research Institute and Depts. of Biopharmaceutical
Sciences and Laboratory Medicine; Director, Informatics Core, UCSF Cancer
Center
There are just over 1000 small molecule therapeutics approved for human use in
the United States. Systematic annotation of their primary targets reveals that
over 700 of these modulate approximately 85 biological targets. The results of
multiple analyses, based exclusively on ligand-focused modeling, will be
discussed. Drug/drug similarities and target/target similarities were computed
on the basis of three-dimensional ligand structures. Drug pairs sharing a
target had significantly higher similarity than drug pairs sharing no target. Also,
target pairs with no overlap in annotated drug specificity shared lower
similarity than target pairs with increasing overlap. Clustering analysis
suggested that side effects and drug-drug interactions might be revealed by
modeling many targets. Ligand-based models of diverse targets were constructed
and tested in virtual screening protocols. Excellent enrichment was possible
against backgrounds of screening molecules. More interesting, however, was that
by crossing all drugs against all targets, it becomes possible to identify a
number of known side effects, drug specificities, and drug-drug interactions
that have a rational basis in molecular structure.
Barry Hardy
eCheminfo Community of Practice
Structural Biology eCheminfo cheminformatics chemoinformatics bioinformatics Medicinal Chemistry Computational Chemistry Virtual Screening Docking Molecular Modelling Molecular Modeling pharmaceutical pharma meeting workshop training Oxford Critical Path toxicology Bursary Life Sciences Pharma Drug Discovery Research and Development Drug Development Healthcare Innovation Knowledge Management events
Comments