An approach to the generation of ultra-large chemical libraries of readily accessible (“REAL”) compounds is described. The strategy is based on the use of two- or three-step three-component reaction sequences and available starting materials with pre-validated chemical reactivity. After the preliminary parallel experiments, the methods with at least ∼80% synthesis success rate (such as acylation – deprotection – acylation of monoprotected diamines or amide formation – click reaction with functionalized azides) can be selected and used to generate the target chemical space. It is shown that by using only on the two aforementioned reaction sequences, a nearly 29-billion compound library is easily obtained. According to the predicted physico-chemical descriptor values, the generated chemical space contains large fractions of both drug-like and “beyond rule-of-five” members, whereas the strictest lead-likeness criteria (the so-called Churcher's rules) are met by the lesser part, which still exceeds 22 million.
The nonstructural protein 3 (NSP3) of the severe acute respiratory syndrome-coronavirus-2 (SARS-CoV-2) contains a conserved macrodomain enzyme (Mac1) that is critical for pathogenesis and lethality. While small-molecule inhibitors of Mac1 have great therapeutic potential, at the outset of the COVID-19 pandemic, there were no well-validated inhibitors for this protein nor, indeed, the macrodomain enzyme family, making this target a pharmacological orphan. Here, we report the structure-based discovery and development of several different chemical scaffolds exhibiting low- to sub-micromolar affinity for Mac1 through iterations of computer-aided design, structural characterization by ultra-high-resolution protein crystallography, and binding evaluation. Potent scaffolds were designed with in silico fragment linkage and by ultra-large library docking of over 450 million molecules. Both techniques leverage the computational exploration of tangible chemical space and are applicable to other pharmacological orphans. Overall, 160 ligands in 119 different scaffolds were discovered, and 153 Mac1-ligand complex crystal structures were determined, typically to 1 Å resolution or better. Our analyses discovered selective and cell-permeable molecules, unexpected ligand-mediated conformational changes within the active site, and key inhibitor motifs that will template future drug development against Mac1.
Deep generative neural networks have been used increasingly in computational chemistry for de novo design of molecules with desired properties. Many deep learning approaches employ reinforcement learning for optimizing the target properties of the generated molecules. However, the success of this approach is often hampered by the problem of sparse rewards as the majority of the generated molecules are expectedly predicted as inactives. We propose several technical innovations to address this problem and improve the balance between exploration and exploitation modes in reinforcement learning. In a proof-of-concept study, we demonstrate the application of the deep generative recurrent neural network architecture enhanced by several proposed technical tricks to design inhibitors of the epidermal growth factor (EGFR) and further experimentally validate their potency. The proposed technical solutions are expected to substantially improve the success rate of finding novel bioactive compounds for specific biological targets using generative and reinforcement learning approaches.