Exploring synthetically accessible alternatives to P2Y12 antagonists using electrostatics and shape

Poster presented at the 5th RSC / SCI Symposium on GPCRs in Medicinal Chemistry.


Clopidogrel is a member of the thienopyridine class of ADP-induced platelet aggregation inhibitors. The mechanism of action requires oxidative activation, resulting in opening of the thiophene ring to generate the active antithrombotic agent which is an irreversible antagonist of the ADP receptor P2Y12 (a GPCR). The mechanism of inactivation probably involves formation of a disulfide bond between the thiol group of the active metabolite and a cysteine residue in the receptor.

Since this compound acts through an irreversible mechanism by which it is covalently bound, both slow onset and effect duration are an issue which can lead to problems, particularly for patients requiring bypass surgery.

This poster shows how I have used Cresset’s field based technology to explore potential binding interactions of recently developed antagonists. By applying our electrostatics to probe these examples in the context of new protein structures, we reveal some less intuitive interaction patterns. These insights relate all the current antagonists and suggest binding modes for Clopidogrel and Elinogrel.

Spark fragment replacement results
Spark fragment replacement results.

View the poster.

Torch V10.3 delivers intuitive visualization and analysis tools for multiple parameter optimization

Cambridge, UK – 30th June 2014 – Cresset, innovative provider of computational chemistry software and services, announces the release of Torch V10.3. This version includes significant new science enabling medicinal chemists to visualize and optimize multiple activities to make the best molecular design decisions.

“An important part of modern medicinal chemistry is to understand how improving one property is likely to affect other properties,” says Dr Tim Cheeseright, Director of Products at Cresset. “This new release of Torch enables chemists to analyse and visualize their project data in an intuitive way while balancing multiple activity goals.”

Highlights of this release:

  • Find and understand structure-selectivity relationships using the new support for multiple activities in the innovative Activity Miner module
  • Visualize molecular properties as a simple, visually intuitive profile using the radial plots. Quickly see how each compound compares to a corporate or project profiles
  • Understand how physical properties change across a dataset of compounds using interactive scatter plots
  • Create publication ready pictures to communicate designs or results across a team or through publication. The enhanced visualization of proteins enables the creation of simple pictures that communicate the important results
  • Find new chemical intellectual property using Cresset’s Blaze directly from your desktop. Submit or retrieve the results from in-house or cloud hosted ligand-based virtual screening experiments directly from Torch.

Dr Robert Scoffin, Cresset’s CEO, says, “We spend a lot of time listening to our customers, and many of these new features have been developed in response to specific customer input and feedback. We are confident that Torch V10.3 will significantly improve the ability of medicinal chemists to make the best use of their project data to make informed design decisions.”


All targets are not created equal

The factors that determine whether a protein is druggable are complex. The question is whether it is possible to find an inhibitor that will be able to conform to the requirements of the protein’s location, shape and lipophilicity.

The first challenge to consider is the target location. This determines the properties required of the compound for it to reach the location at sufficient concentration for inhibition. This has to be balanced against the specific features required of the compound to fit the binding site within the protein in order to achieve inhibition.

Sometimes, these two parameters are diametrically opposed to each other and thus the target is not easily druggable. However, in many cases innovative solutions can often be found once there is wide enough understanding of the specific system.

The serine proteases are a case in point. Many of them require a highly basic arginine-like moiety such as benzamidine to access a key binding site location: the S1 pocket. This same benzamidine moiety disallows penetration through membranes. One solution for this, rather than defeat, is a pro-drug – a membrane soluble benzamidine analogue such as etixilate.

A more informative gauge of drugability of a target is the extent of the requirement for lipophilicity. This is the bane of medicinal chemistry, as almost inevitably lipophilicity increases during lead expansion (as presented by Mike Hann at CDDD Verona). Getting the drug to the target rather than to lipid membrane compartments and other targets is very difficult. Off target effects and toxicity both increase with lipophilicity. Lipid binding GPCRs such as prostaglandins and other receptors such as the cholesterol binding / transporter proteins are very challenging drug targets.

Cresset consultants work across a range of protein targets, as illustrated below. The chart classifies Cresset’s recent consulting projects according to target class. Proteases and protein-protein interactions are the most challenging targets.

Distribution of target classes for recent Cresset consulting projects
Figure 1: The distribution of target classes for recent Cresset consulting projects.


Receptors are found inside cells or on cell membranes. When activated by a ligand, the receptor performs a specific biochemical action.

Receptors can act as an intermediary between a ligand and a protein across a cell membrane. The ligand activates the receptor so that it can activate the protein, forming a ligand-receptor-protein complex, as in the case of GPCRs.

G-protein coupled receptors (GPCRs) are one of the most important receptor families of proteins. About 40% of all available drugs act on GPCRs.

Conversely, nuclear hormone receptors act inside the nucleus of the cell and are involved in the regulation of gene expression which in turn is affected by small molecule hormones such as steroids and synthetic molecules such as Bis phenol A (BPA) – an endocrine disruptor leaching from polycarbonate containers.

Figure 2: A highly lipophilic agonist ligand (slogP 6.1) of the PPARgamma nucleur hormone receptor PDB:4JL4 showing hydrophobic surface (brown) and vdW interaction hotspots (in yellow).

Enzymes and proteases

Enzyme inhibitors are an important group of drugs. They are effective in therapeutic areas as diverse as cancer, depression, cardiovascular disease and pain.

Protease enzymes break down proteins and peptides. AIDS is treated with HIV protease inhibitors. Over 700 human proteases have been identified so far. However, it is notoriously difficult to develop drugs for protease targets due to selectivity issues and property space.

Ion channels

Ion channels control the flow of ions such as sodium, calcium or potassium across a cell membrane. They play an important role in regulating many physiological processes, so can be very useful targets.

Protein-protein targets

Protein-protein interactions are a medial step in many biological processes. Designing a compound that will inhibit these interactions is a promising way to intercept mechanisms that have an adverse effect on a disease condition.

The challenges of designing suitable inhibitors relates mainly to the large surface area of the proteins. The computing power required to model the proteins means that identifying suitable binding sites can be a slow process.

However, there is a high therapeutic potential in designing suitable small molecules for protein-protein targets.

Working towards innovative solutions

The type of protein target has a fundamental impact on the nature of a drug discovery project.

Cresset’s computational methods are not dependent on knowledge of the target – we can take a ligand based approach to drug design. However, if there is structural information available for the target then we use it.

Cresset consultants are always keen to take on new challenges. Contact us to discuss how we can work with you to arrive at innovative solutions.

Scaffold hopping in medicinal chemistry

Scaffold hopping in medicinal chemistry

Scaffold hopping in medicinal chemistry

This book is part of the series ‘methods and principles in medicinal chemistry’.

The first section serves as an introduction to the topic by describing the concept of scaffolds, their discovery, diversity and representation, and their importance for finding new chemical entities. The following parts contain a general description as well as case studies of the most common tools and methods for scaffold hopping, whether topological, shape-based or structure-based. Part two, chapter 13 explores Cresset’s ‘XED force field and Spark‘ and was co-authored by Dr Andy Vinter and Dr Martin Slater. The final part contains three fully documented real-world examples of successful drug development projects by scaffold hopping that illustrate the benefits of the approach for medicinal chemistry.

While most of the case studies are taken from medicinal chemistry, both chemical and structural biologists will also benefit greatly from the insights presented here.

Case study 2 ‘Bioisosteric Replacements for the Neurokinin 1 receptor (NK1R)’ by Francesca Perruccio (pages 259-278) cites Spark, Cresset’s dedicated scaffold hopping too.

Now available as an eBook from various online vendors, the book is scheduled for physical publication in December 2013.

Author information

Nathan Brown is the Head of the In Silico Medicinal Chemistry group in the Cancer Research UK Cancer Therapeutics Unit at the Institute of Cancer Research in London (UK). At the ICR, Dr. Brown and his group support the entire drug discovery portfolio together with developing new computational methodologies to enhance the drug design work. Nathan Brown conducted his doctoral research in Sheffield with Professor Peter Willett focusing on evolutionary algorithms and graph theory applied to challenges in chemoinformatics.

After a two-year Marie Curie Fellowship in Amsterdam in collaboration with Professor Johann Gasteiger in Erlangen, he joined the Novartis Institutes for BioMedical Research in Basel for a three year Presidential Fellowship in Basel working with Professors Peter Willett and Karl-Heinz Altmann.

His work has led to the pioneering work on mulitobjective design in addition to a variety of discoveries and method development in bioisosteric identification and replacement, scaffold hopping, molecular descriptors and statistical modeling. Nathan continues to pursue his research in all aspects of medicinal chemistry.

Torch V10.2 released

Cresset are delighted to announce the release of a new version of Torch, the intuitive molecule design suite for medicinal chemists. This version comes with many new features and significant interface enhancements.  It also includes the new Activity Miner module for SAR interpretation and understanding.

Activity Miner

Activity Miner is designed to give you rapid access to the important points in the structure activity relationships of your projects. It uses the concept of ‘activity cliffs’ or ‘disparity’ to identify where a small change in the molecule has caused a large change in the measured activity. Uniquely, Activity Miner works using multiple metrics for deciding if a change is significant or not. See the Activity Miner page for more information or see it in action in these short videos.

Activity Miner's Activity View

Torch Interface Improvements

The main Torch interface has received significant improvements, such as:

  • the new Fullscreen mode
  • the inclusion of additional columns into the results table, including Ligand Efficiency (LE) and Lipophilic Ligand Efficiency (LLE) data for every molecule
  • an option to import molecules from or update the data for molecules from csv files.

The new Fullscreen mode gives the option to display the 3D window across the whole of one monitor. We use it to discuss molecule designs or protein-ligand interactions in meetings but you could equally use it for creating stunning pictures for papers or for browsing the results of an alignment experiment. You can now ‘rock’ or ‘spin’ the 3D display at the same time for sensational full screen presentations. This short video shows it in action:

In this new version of Torch any activity data that is associated with your molecules is automatically detected and used to calculate activity based properties such as the ligand efficiency and the lipophilic ligand efficiency. You have always had the ability to filter the molecules in the table but with these new properties it is easier than ever to focus your thoughts on the compounds with the best possible match of physical and activity criteria.

Other improvements to the results table include a ‘notes’ field associated with every molecule so that you can record your thinking when you are designing a new compound. Simply type the notes into the molecule editor and they will be automatically transferred to the main table when you save the molecule.

Creating Molecule Notes in Torch

The ways to load molecules into Torch continue to expand and improve. In previous versions we introduced reading molecules from smiles and the ability to download and process pdb files. In this version you have support for reading both molecules and data from csv files, making the loading and updating of molecule associated data easier than ever. Simply open the csv file, choose to load molecules or data and specify which column contains the molecule smiles or the matching field. See it in action in this short video:

Download a FREE Evaluation

You can try all these features with a free evaluation of Torch, or contact us to find out more.

Using scaffold hopping to search nearby chemical space

How Can Computational Chemistry Help Find New Drugs from Old?

In this series of blogs, Dr Robert Scoffin, CEO of Cresset, explores how computational chemistry is being applied to the field of reprofiling to help find new drugs from old.

As R&D budgets are cut and blockbusters come off patent, drug discovery is becoming more conservative.  Many companies are taking existing drugs or drug-like compounds as the starting point for the discovery of new chemical entities (NCEs) because they know that this is the most likely root to success.  In my last post I discussed Using Pharmacophores to Find New Targets for Existing Drugs.  This month I’m taking a look at the technique of scaffold hopping that makes changes to existing drugs.

Scaffold hopping involves a computational search of the chemical space around existing compounds by making structural changes to known drugs.  The aim is to find a new compound that not only retains the activity of the parent compound, but also shows some improvement, either in terms of improved activity or a reduction of side effects.  The fact that the resulting compound will have a similar chemistry to a known active significantly increases the likelihood of activity and, therefore, return on R&D spend.

This approach is well validated.  Bayer’s Levitra was created by altering Pfizer’s Viagra molecule.  The small chemical change involved a substitution of the nitrogen positions in a fused ring, which was not covered by Pfizer’s patent.  Computational analysis of the field patterns of the two compounds shows that this structural change had only a subtle effect on the molecule’s field pattern: computational analysis shows that Levitra and Viagra share a very similar pharmacophore template.

Scaffold hopping using Spark is an ideal approach to finding NCEs from existing compounds.  Spark uses Cresset’s field technology to find biologically equivalent replacements for key moieties.  The activity profile of the new compound is compared to the starting point or to a pharmacophore.  It is an extremely effective method of exploring new chemical space.  Our recent poster: Rapid Technique for New Scaffold Generation highlights the power of scaffold hopping to both explore new chemical space and to provide enhanced protection for NCEs.

Viagra Levitra Unknown

Above left: Sildenafil (Viagra) from pdb code 1udt; Above center: Vardenafil (Levitra) found by scaffold hopping in Spark; Above right: Potential PDE5 active that mimics both Sildenafil and Vardenafil but lies outside of both patent landscapes.

When I left a trial copy of Spark with a customer at a major pharma company he used it to run a scaffold hopping experiment on some of his current leads.  The results blew him away.  Yes,Spark returned some bioisosteres that he had been expecting and had already thought of.  It also returned some that weren’t viable for patent or ADME reasons.  But it also came up with some gold – new, bioisosteric structures that he had not thought of, that were completely ‘out of the box’.  Needless to say, he bought the software.

In the next post I’ll talk about a recent collaboration between Cresset and RedX that focuses on discovering NCEs by switching the oxidation state of known drugs.

Previous posts in this series:

Dr Robert Scoffin,

Torch Brings the Next Generation of Molecule Design to the Desktop

Protein’s eye view gives clear insights into the causes of biological activity

Welwyn Garden City, UK – 25 October 2012 – Cresset, innovative provider of software and services, announces the release of Torch, a complete desktop molecular design and 3D SAR tool for medicinal chemists.

Torch uses molecular fields to show the binding patterns of your compounds. This protein’s eye view gives clear insights into the causes of biological activity.

Working with Torch, medicinal chemists can see how to optimize the shape and electrostatic properties of their series, and rapidly identify the best next molecule to synthesize.

Cresset CEO, Dr Robert Scoffin said, “We believe that molecular fields are the most intuitive way of assessing biological activity. We would like every chemist to access the power of fields so that they can truly understand the mechanisms by which their leads interact with targets. Torch is such an easy to use desktop tool that it brings these insights within the reach of every medicinal chemist.”

torchV10 screenshot

Torch is a complete desktop molecular design and Structure Activity Relationship (SAR) tool that brings the power of fields within the reach of every chemist.

Medicinal chemists, research managers, academics and computational chemists who want to find out more about Torch can register for a free introductory webinar.

Torch Released

We are delighted to announce the release of Torch, the next generation of FieldAlign, an intuitive 3D molecule designer. Torch brings new capabilities to medicinal chemistry desktops while maintaining our focus on easy to use, easy to learn software that produces meaningful results within minutes.

Existing FieldAlign users can upgrade to Torch for free. If you have not already received your new license file then please let us know. If you’re new to our molecule design software you can get a free one month evaluation here.

In common with Forge and Spark, Torch uses the next generation of our proprietary XED force field to describe molecules and their interaction from the protein’s view point. Using Torch gives medicinal chemists a unique insight into how molecules relate to each other and to the target protein, enabling a deep understanding of structure activity relationships. However, understanding is only half of the equation. Torch brings a powerful molecule editor and sketcher that chemists use to design their next molecule. The editor provides immediate feedback on each design giving a rapid progression of idea to synthesis.

Specific additions in Torch are:

  • New molecular mechanics force field (XED3).
  • New protein importer to read and process pdb files into protein and reference molecule.
  • Download pdb files directly into Torch and view protein-ligand interactions.
  • Display and use Forge QSAR models in the scoring of molecules.
  • Improved molecule reading to give fewer clicks and more intuitive operation.
  • Measure distances, angles and torsions, automatically display intramolecular H-bonds and atomic clashes.

Rapid Technique for New Scaffold Generation

Scaffold hopping remains a central task in medicinal chemistry for generating and protecting intellectual property. We present Spark, an application for rapidly generating reasonable yet novel scaffold replacements.

Our technique uses the molecular interaction fields of the parent molecule and assesses replacements in the context in which they will be synthesized. This enables the differing steric and electronic effects of potential new scaffolds to be used. An added bonus of our method is that replacements for terminal substituents can be considered alongside more central moieties enabling its use in growing fragments and lead optimization as well as lead generation.

View the poster, presented at EFMC-ISMC 2012, here…

Computational Tools for Medicinal Chemists – Increasing the Dimensions of Drug Discovery

For an insight into the following, view this presentation from Rob Scoffin (Cresset’s CEO) which was made at ACS Fall 2012:

– Building Desktop Tools – A History
– About Cresset BioMolecular Discovery
– Applications in Medicinal Chemistry
– PDE-5 Examples
Computational Tools for Medicinal Chemists