Covalent Drug Design

The term "warhead" in covalent drug design refers to functional groups capable of forming a covalent bond with a target structure. This typically involves a reaction between the warhead and a nucleophilic side chain of an amino acid residue. The most common target is cysteine due to its high reactivity, but there are warheads designed to preferentially bind other residues like lysine or histidine. Serine and tyrosine are also potential candidates for forming covalent bonds; however, due to the lower nucleophilicity of oxygen, these bonds are more likely to be reversible covalent interactions.

Warheads exhibit varying levels of reactivity. For instance, α-chloroacetamides are relatively more reactive than α,β-unsaturated ketones, making them capable of binding to cysteine residues on non-target structures. To avoid side effects, it is crucial to select a warhead that forms covalent bonds selectively with the intended target structure. In recent years, Michael acceptors (α,β-unsaturated carbonyl compounds and their derivatives) have gained popularity due to their tunable reactivity, which can be controlled through the addition of electron-withdrawing groups.

Designing covalent drugs can be approached using structure-based or ligand-based methods. In the structure-based context, this involves designing new, more potent ligands, optimizing them, and screening for novel scaffolds and bioactive compounds for a target structure of interest.

During de novo design, one starts from a known structure and modifies it to fine-tune and improve its parameters. 3D methods can identify and exploit previously unused potential interactions in the target binding site, enabling the introduction of new groups to the molecule to enhance affinity. Additionally, modifications around the warhead can be evaluated to influence its reactivity while maintaining potential interaction patterns.

Virtual screening is another structure-based method that can be extended to covalent drugs. In this approach, a virtual substance library is docked to a target, and the generated poses are scored and ranked. Under the premise that bioactive substances are more likely to bind to the target given their stronger predicted interactions, this method can be used to pre-filter potential candidates for in vitro confirmation. Covalent docking also enables the prediction of binding modes for drugs containing a covalent warhead within the same context. It is a proven method for cost-effectively identifying a preliminary selection of potential binders.

When engaging in computer-aided drug design for covalent drugs, one quickly realizes how cumbersome and clunky the process within the software can be.

At BioSolveIT, we have focused intensely on simplifying this process without compromising the quality of the results. Additionally, user convenience was a critical aspect to ensure that even non-software experts can achieve their goals.

In this context, we have developed radically simple covalent docking, which automatically identifies warheads and translates them into their target-bound form, ready to be docked into the binding pocket. This feature covers over 35 of the most common warheads in medicinal chemistry. Users can conveniently select their target residue of interest and seamlessly initiate covalent docking or virtual screening. A range of docking parameters ensures full control over the results and allows fine-tuning, further enhanced with pharmacophore constraints to deliver even better outcomes.

Moreover, Chemical Spaces are an excellent source of compounds containing warheads that are not found in standard catalogs. Their key advantage, beyond the novel intellectual property they represent, is their vast size. With billions or even trillions of entries, the number of compounds and specific warheads of interest increases significantly. Within our platforms infiniSee and infiniSee xREAL, it is possible to perform substructure searches for desired warheads. The command-line tool SpaceMACS further extends the search capabilities with SMARTS queries, enabling even more targeted searches for specific motifs. BioSolveIT platforms for mining of covalent drug candidates:

• infiniSee: Chemical Space navigation platform with a graphical user interface.
• infiniSee xREAL: Exclusive platform to screen Enamine's largest compound catalog featuring trillions of compounds.

Command-line tools for mining of covalent drug candidates:

• SpaceMACS: Performs maximum common substructure searches, as well as exact substructure mining. Algorithm behind the Motif Matcher Mode.