The importance of drug 3D-shapes
Drug molecules ameliorate diseases by binding to target proteins to modulate their biological function. The 3D-shape information of the binding interaction often available during a drug discovery programme is illustrated in the figure below, using the drug lisinopril (Prinivil®) and its target angiotensin-converting enzyme (ACE). For particular classes of proteins, X-ray crystallography is routinely available to support drug discovery projects by providing experimental 3D-information on both the free and bound states of the protein.
Free drug ligand
Free target Protein
Bound drug-target Co-complex
While the importance of the 3D-shape of both free and bound protein targets is well recognised and routinely used in structure-based drug design (SBDD), the critical role of the 3D-shape of the free ligand is generally overlooked.
For a drug ligand to bind strongly to its target and have a strong effect, it needs to be able to adopt the right shape to fit into the target binding site. If a molecule has to dramatically change shape to bind to its target, it is likely to bind poorly and therefore be unsuitable as a drug. In contrast, if a molecule is already the right binding shape (the ‘bioactive conformation’), it is more likely to bind strongly to the target and be a good drug. Knowledge of the 3D-shape of the free ligand is therefore extremely valuable in guiding the design of the best drug molecules.
In conventional drug discovery, the shape of the free ligand is not known, so the majority of early-stage drug discovery efforts involve iterative medicinal chemistry synthesis and testing of molecules, gradually improving both the ligand’s shape and charge properties until a molecule is found that binds tight enough to the target. To make matters worse, for many important target classes (e.g. GPCRs, ion channels), 3D-shape information is usually unavailable for the proteins, so not only is the starting shape of the ligand unknown, but also the bound shape the chemists are seeking is unknown. This all leads to an expensive, time-consuming process in which the solution is found after a large amount of trial-and-error, if at all.
The Conformetrix proprietary technology platform allows the 3D-shapes of free ligands to be precisely measured from experimental data, meaning that medicinal chemists no longer need to work in the dark. In cases where protein 3D-shapes are available, comparison between free and bound ligand 3D-shapes immediately highlights the exact areas of the molecule that the medicinal chemist needs to address to redesign the drug molecule to be the right shape. In cases where there is no protein 3D-shapes, comparison between multiple ligand 3D-shapes also provides clear direction and decision-making for the medicinal chemist. Moreover, other problems such as cell permeability, metabolic liabilities and off-target effects can all be overcome using ligand 3D-structures.
Conformational analysis
measuring drug 3D-shapes
The quantified 3D-shapes (conformations) that a molecule or part of a molecule can adopt is termed its ‘conformetrics’. Measurement of a free drug molecule’s conformetrics is far from trivial because in most cases drugs are flexible molecules and, with typically 5-13 rotatable bonds, display very complex conformational envelopes.
The Conformetrix proprietary technology platform allows a molecule’s conformetrics to be precisely measured directly from experimental data. The method was developed by C4XD founders Charles Blundell and Andrew Almond to address the inability of existing experimental and computational chemistry methods to determine this information. It uses Nuclear Magnetic Resonance (NMR) data to provide detailed, atomic resolution information for the molecule in its near-natural state (i.e. solution). The NMR data is used in the C4XD proprietary software suite which interprets it in a novel, patented algorithm to determine the molecule’s dynamic 3D-shape (conformetrics). The output is an accurate description of the complete dynamic behaviour of the molecule in 3D-space, providing medicinal chemists with the vital information and insight they need to make key drug discovery design decisions.
Using Conformetrix, typical drug molecule dynamic 3D-structures can be solved in 1-2 weeks, integrating well with the medicinal chemistry design-make-test cycle. Our in-house database contains several hundred dynamic 3D-structures, a few examples of which are shown in the lower figure. All classes and kinds of drug molecules and natural ligands (e.g., cofactors, peptides) can be analysed, meaning that every C4XD programme is structurally enabled, irrespective of the target class and availability of target 3D-information.
Cozaar
Cydectin
Epivir
Gleevec
Lipitor
Neupro
Norvir
Paxil
Prinivil
Simpanorm
If you would like to know more about how we measure solution state free ligand 3D structures, please read our article ‘Quantification of free ligand conformational preferences by NMR and their relationship to the bioactive conformation’ in Bioorganic & Medicinal Chemistry.
Conformational design - measuring drug 3D-shapes
Typically, the more a drug molecule changes shape upon binding to its target, the weaker its interaction will be, and the weaker its functional biological (disease-modifying) effect. Hits and early-stage molecules in a drug discovery pipeline therefore often have many shapes in solution, only one or a few of which are the right shape to bind to the target. ‘Conformational design’ aims to make rational, precision changes to molecules so that they are the right shape for binding with the target (the bioactive conformation) for as much time as possible. C4XD are world-leaders in conformational design for drug discovery.
The ability to measure drug dynamic 3D-shapes using the Conformetrix proprietary technology greatly enables conformational design. When the desired bioactive conformation is known from protein X-ray crystallography data, simple comparison between the set of free drug dynamic shapes and the bioactive conformation immediately highlights and prioritises areas for redesign of the molecule. This is usefully done by displaying the measured conformetrics for a single rotatable bond with a circular histogram. Bonds where conformational adjustment is required to improve the binding affinity by training the free ligand’s conformational envelope towards the bioactive conformation are immediately obvious and provide clear rationales for medicinal chemistry decision making. These graphs further help distinguish different kinds of conformational design (modal, torsional or librational selection) and even indicate where molecules are too rigid and need to be made more flexible.
In some cases, the bioactive conformation is not known from protein X-ray crystallography data. In these cases, comparison of the local bond conformetrics of ligands from the same series with different affinities immediately identifies the conformational-affinity trend to follow. When two or more ligands with different scaffolds but overlapping pharmacophores are known, comparison of their global shapes allows the bioactive conformation to be rapidly and reliably determined, even without any protein 3D-shape data. Many natural ligands (e.g. peptides) have high affinity and in these cases, they tend to have a small number of 3D-shapes, and this means they can be used to predict the binding pharmacophore with surprising accuracy. In this 3D-context, scaffold-hopping between series, and even from peptide to drug, becomes a relatively trivial exercise.
Off-target effects can also be specifically and rationally removed from a series using conformational design. In our Orexin-1 programme, we were able to precisely measure the shapes required for binding to Ox-1 over the very similar Ox-2 target and thereby design highly selective (1000x fold) Ox-1 antagonists, even in the absence of protein 3D-data. Breakthroughs like this lead to safer drug molecules with much lower side-effect risk profiles, providing better and cleaner molecules for later drug development.
Free drug 3D-shape also has an impact on a range of other areas in drug discovery and development, including ADME, solid-state polymorphism and catalyst development. For example, polar surface area as a predictor for permeability is well known to be very dependent on the quality of conformational data used.
If you would like to know more about the importance of solution state free ligand 3D structures and how they can be used to transform the medicinal chemistry workflow please read our article ‘Measurement, Interpretation and Use of Free Ligand Solution Conformations in Drug Discovery’ in Progress in Medicinal Chemistry.
