Fragment Libraries and X-Ray Crystal Structures

Fragment Libraries: Collections of Versatile Compounds

There are several reasons why fragment libraries have become increasingly popular in the recent years. With just a few hundred to a few thousand compounds, it is possible to cover a large portion of the chemical space, providing good starting points for drug design. After a fragment has been confirmed as a hit through biophysical assays, it can then be decorated with additional functionalities to enhance its potency. Due to the small size of the fragments, they allow for considerable creative flexibility during the compound extension.

Basics of Fragments and High-Throughput X-Ray Screening

Fragments are defined by the rule of three: no more than three hydrogen bond acceptors and donors each, no more than three rotatable bonds, a molecular weight of less than 300, and a logP no greater than 3. This ensures that the compound will not violate the rule of five during lead optimization when additional functionalities are added.

During a high-throughput crystallographic fragment screening, target structures are exposed to fragments in an attempt to crystallize complexes that contain a binder. With the right capacity, hundreds of crystal structures of a single target can be collected, each containing or or more fragments, sometimes even at different binding sites. This is where the decision-making process begins, determining which fragments are of particular interest and should be further optimized in the subsequent steps.

Parameters to Consider During Fragment Assessment

To decide which fragment will be further developed, there are a number of parameters that can be considered. Here is a selection of perspectives that can be crucial for the process:

The most important property of fragments in the context of fragment-based drug design (FBDD) is that they can form high-quality interactions with the target as a unit. Therefore, it is crucial to understand whether all the heavy atoms truly contribute positively to the binding affinity.
Do other fragments with a similar structure exhibit the same binding mode as the fragment in question? A consistent pattern can indicate that the fragment has a preferred binding mode to the target, which subsequently simplifies structure-based approaches.
Are you looking at an artificial or a native binding mode? In some cases, binding pockets may form through the assembly of homo- or heteromers. These binding sites are not always found in a cellular context but are instead a result of the crystallization or soaking process.
Is the ligand easy to evolve? If a large part of the ligand is solvent-exposed, it can become very challenging to gain additional affinity by adding new groups. Therefore, it is crucial that the fragment has sufficient contact with the target structure, while still allowing enough space for further decoration. Lipophilic ligand efficiency (LLE) is a valuable numerical parameter that can be used to assess this. An LLE of 0.3 to 0.5 can be considered a good range for a fragment binder.
The simplest and fastest method to elucidate the structure-activity relationships is to test a few analogs containing the fragment. It can be helpful to examine a diverse set of compounds in order to efficiently gain insights with minimal resources. In this process, the number of accessible, expanded versions of the initial fragment plays a crucial role.1

Commercial Availability of Molecular Fragment Motifs

While some substructures appear relatively often in small molecules, there are fewer mature analogs available for other fragments. One reason for this imbalance is that some fragments are more privileged than others: For example, there are many compounds that contain a hinge binder motif, as this target class has been extensively studied.

A substructure search in a web browser can highlight the difference: While nearly 2,000 compounds are found that contain a certain fragment as a substructure, only about 30 compounds are found for another fragment.
Chemical Spaces can bridge this gap. Unlike compound catalogs, which only contain structures that have been 'actively' considered or are already in stock, Chemical Spaces reveal all possibilities that are synthetically accessible. While traditional FBDD typically deals with thousands of compounds containing a specific fragment, Chemical Spaces can provide access to millions or even billions of relevant entries.

BioSolveIT software for fragment libraries:

infiniSee: Chemical Space navigation platform with a graphical user interface.
infiniSee retrieves relevant chemistry from ultra-large Chemical Spaces containing billions or even trillions of compounds based on their similarity to a query compound. Results are synthetically accessible per design in one or two steps and, in the case of our partners' Chemical Spaces, can be ordered directly to your table.
In the context of FBDD, infiniSee's Motif Matcher Mode is a valuable source. It can perform exact substructure and maximum common substructure searches to retrieve close relatives to a fragment of interest. Furthermore, it is also possible to define growing points at the fragment to retrieve only structures that contain additional heavy atoms in desired areas.
infiniSee xREAL: Exclusive platform to screen Enamine's largest compound catalog featuring trillions of compounds.
infiniSee xREAL contains all features of infiniSee and supports all three Chemical Space exploration search modes.

Command-line tools for fragment libraries:

SpaceMACS: Performs maximum common substructure searches, as well as exact substructure mining. Algorithm behind the Motif Matcher Mode.
SpaceMACS can extract up to 1,000,000 molecules for a single query compound per run from a Chemical Space. Additionally, it is also possible to use an SDF list with multiple query fragments to collect all relevant related compounds for each fragment. This enables the creation of project-specific substance libraries of commercially available extended molecules.

Notably, SMARTS support is also featured, allowing the definition of chemical patterns to make searches even more versatile.

Structure-Based Fragment Growing

The inclusion of the 3D target structure can significantly accelerate the entire process. For instance, sublibraries generated with infiniSee or SpaceMACS can subsequently be docked into the target structure to assess which fragments are the most promising candidates for acquisition or synthesis. This approach also provides insights into which fragments have the potential for substantial growth, with the prospect of achieving reasonable affinity values.

For a more creative and efficient approach, FastGrow (also featured in SeeSAR's Inspirator Mode) can be utilized. FastGrow scans through a 3D structure database to identify entries that best complement the target's binding site. Hundreds of thousands of structures and their conformations can be searched for the most relevant results within seconds on standard hardware. By manipulating the binding fragment in the complex and specifying the growing vector, you can efficiently screen numerous structures in the command line, seamlessly integrating the entire workflow into ongoing processes.

BioSolveIT software for fragment growing:

SeeSAR: Visual, drug design dashboard for computational and medicinal chemists.
The Inspirator Mode includes the FastGrow feature, which searches databases for optimal extensions for a fragment. Multiple options are available, making it versatile for various scenarios.

Command-line tools for docking:

FastGrow: Pocket exploration algorithm to screen thousands of molecular extensions for the best fits.

Excited for more drug discovery solutions?

Keep on exploring!