Finding a new chemical series can help a stalled drug discovery project escape from a toxicity or patent issue. Many projects also look for alternative series as a project backup strategy.
There are different tools and tricks to finding a new chemical series, but they all involve setting up a ‘screen’, i.e., a way of assessing a collection of molecules to find novel ‘hit’ compounds which show some activity at the biological target of interest. The differences lie in the input data available for setting up the screen, and the type of screen you choose – virtual, biochemical or biophysical. Virtual screening offers a highly cost effective method of identifying new chemical series, whether the starting point is ligand or protein data.
When the protein and the ligand(s) are both known
The ideal scenario is when the protein and the ligand(s) are both known. The two sets of information can be brought together and used as a virtual screening seed to find new chemistry. For example, you can perform a ligand-based virtual screen that identifies a compound with the same electrostatics and shape as the first series, but is a new chemical entity. This approach can provide an ideal back-up series for a project.
Without doubt, if you know the protein-ligand interaction, then using it to carry out a virtual screen is absolutely the best method for finding a new chemical series. Many projects do not have such complete starting knowledge, but it is often possible to deduce a protein-ligand interaction from analyzing either the ligand or the protein data.
Starting from the protein
When the protein is known, the focus is the binding pocket. The goal is to find molecules (hits) which bind into the pocket in a way which activates or deactivates the biological target, triggering the desired pharmacological effect.
Docking is the process of looking for a ligand that has complementary shape, electrostatics and H-bonding patterns to interact with the binding pocket: the overall dipole must also match. Running a High-Throughput Docking (HTD) experiment on large collections of virtual compounds, available from in house repositories or commercially available, is a very effective virtual screening strategy.
To ensure this approach is successful, an accurate preparation of the protein preparation is vital. If a ligand is available in a bound state, then it may be possible to determine whether the protein is in an ‘active’ or ‘inactive’ state. The residues must be in positions consistent with the binding mode, and any waters that are important for binding must also be considered.
The result of a HTD screen is a score that can be used to prioritize output for purchase. The Lead Finder™ docking algorithm implemented in Flare is equipped with a dedicated Virtual Screening scoring function optimized to provide maximum efficiency in virtual screening experiments, where the important outcome is an optimal discrimination of active and inactive compounds.
The ligand-only route
In the absence of a protein structure, ligand-based methods are a very effective way to find hits which match the electrostatic and shape properties of an active query molecules, but are novel chemical entities. Cresset’s XED force field is used to analyze the electrostatics, hydrophobicity and shape of the ligands to produce an alignment hypothesis that is consistent with the data and which can be used in a ligand-based virtual screen with Blaze™ or Ignite™.
Virtual screening focuses your biological screening
Running biochemical or biophysical screens involves the selection and purchase of a library of compounds to screen against your assays. Typical screens include a hundred thousand compounds, which is an expensive and not very thorough way of exploring chemical space. The number and diversity of purchasable chemical space is orders of magnitude in excess of this.
A screen can be carried out against millions of virtual screening compounds, using a large computational cluster to compare tens of millions of compounds in a few days. Conformational sampling of all available commercial compounds can be carried out and these can be docked and scored. The goal is to cut down the field of 30 million commercially available compounds to tell you which are the best 1,000 or so compounds to focus on. A wet screen would still be necessary, but only to screen those compounds that are most likely to be active.
This streamlined method is a very cost effective way of finding new chemical series. Adding in silico work enables a short cut through the process, making it far more streamlined and cost effective.
By applying the proper experiment to your project, you can rapidly assess new molecule designs for their fit to the protein active site and/or high enrichments in virtual screening experiments.
