Covalent ligands modulating protein activities/signals have attracted unprecedented attention in recent years, but the insufficient understanding of their advantages in the early days of drug discovery has hindered their rational discovery and development. This also left us inadequate knowledge on the rational design of covalent ligands, e.g., how to balance the contribution from the covalent group and the noncovalent group, respectively. In this work, we dissected the noncovalent docking from covalent docking by creating SCARs (steric-clashes alleviating receptors). We showed that the SCAR method outperformed those specifically developed but more complicated covalent docking protocols. We furthermore provided a "proof-of-principle" example by implementing this method in the first high-throughput screening and discovery of novel covalent inhibitors of S-adenosylmethionine decarboxylase. This work demonstrated that noncovalent groups play a predeterminate role in the design of covalent ligands, and would be of great value in accelerating the discovery and development of covalent ligands.

Metabolic reprograming is a hallmark of cancer, and the polyamine metabolic network is dysregulated in many cancers. Ornithine decarboxylase (ODC) is a rate‐limiting enzyme for polyamine synthesis in the polyamine metabolic network. In many cancer cells, ODC is over‐expressed, so this enzyme has been an attracting anti‐cancer drug target. In the catalysis axis (pathway), ODC converts ornithine to putrescine. Meanwhile, ODC’s activity is regulated by protein–protein interactions (PPIs), including the ODC‐OAZ1‐AZIN1 PPI axis and its monomer‐dimer equilibrium. Previous studies showed that when ODC’s activity is inhibited, the PPIs might counteract the inhibition efficiency. Therefore, we proposed that multipurpose inhibitors that can simultaneously inhibit ODC’s activity and perturb the PPIs would be very valuable as drug candidates and molecular tools. To discover multipurpose ODC inhibitors, we established a computational pipeline by combining positive screening and negative screening. We used this pipeline for the forward screening of multipurpose ligands that might inhibit ODC’s activity, block ODC‐OAZ1 interaction and enhance ODC non‐functional dimerization. With a combination of different experimental assays, we identified three multipurpose ODC inhibitors. At last, we showed that one of these inhibitors is a promising drug candidate. This work demonstrated that our computational pipeline is useful for discovering multipurpose ODC inhibitors, and multipurpose inhibitors would be very valuable. Similar with ODC, there are a lot of proteins in human proteome that act as both enzymes and PPI components. Therefore, this work is not only presenting new molecular tools for polyamine study, but also providing potential insights and protocols for discovering multipurpose inhibitors to target more important protein targets.