A series of Co complexes with bis­(pyridylmonoimine)-based ligands with different degrees of structural flexibility have been prepared and studied for the electrocatalytic CO2 reduction reaction to CO. First, electrochemical kinetic studies of the structurally rigid [Co­(L-L)] complex show that it undergoes a reductive dimerization upon reduction to the CoI complex. This dimerization is facilitated by the planar geometry of the [Co­(L-L)] complex. The dimer structure dissociates after reduction of the ligand, forming a monomer species that is active for CO2 reduction. The reductive dimerization can be sterically prevented either by adding the strong axially coordinating ligand such as triphenylphosphine (PPh3) or by distorting the square planarity of the Co geometry by modulating the flexibility of the ligand scaffold. The more flexible [Co­(L-R-L)] complexes prevent catalyst dimerization and operate with more positive catalytic onset potentials for CO2 reduction compared to the more rigid [Co­(L-L)] complex but operate with lower overall activity in the presence of a proton source. CO-binding and inhibition studies provide evidence that the lower activity for CO2 reduction of the more flexible [Co­(L-R-L)] complexes compared to [Co­(L-L)] is due to CO poisoning because of the stronger binding affinity of the CO product to the flexible [Co­(L-R-L)] complexes. This highlights an important trade-off in catalyst design for this class of molecular electrocatalysts: Co bis­(pyridylmonoimine) complexes with higher degrees of structural flexibility prevent dimerization and shift the onset of CO2 reduction catalysis to more positive potentials but decrease the maximum activity due to CO product inhibition.

Here we report the testing and application of a simple, structure-aware framework to design target-specific screening libraries for drug development. Our approach combines advances in generative artificial intelligence (AI) with conventional molecular docking to rapidly explore chemical space conditioned on the unique physiochemical properties of the active site of a biomolecular target. As a proof-ofconcept, we used our framework to construct a focused library for cyclin-dependent kinase type-2 (CDK2). We then used it to rapidly generate a library specific to the active site of the main protease (M pro ) of the SARS-CoV-2 virus, which causes COVID-19. By comparing approved and experimental drugs to compounds in our library, we also identified six drugs, namely, Naratriptan, Etryptamine, Panobinostat, Procainamide, Sertraline, and Lidamidine, as possible SARS-CoV-2 M pro targeting compounds and, as such, potential drug repurposing candidates. To complement the open-science COVID-19 drug discovery initiatives, we make our SARS-CoV-2 M pro library fully accessible to the research community (https://github.com/atfrank/SARS-CoV-2).

Here, we report the implementation and application of a simple, structure-aware framework to generate target-specific screening libraries. Our approach combines advances in generative artificial intelligence (AI) with conventional molecular docking to explore chemical space conditioned on the unique physicochemical properties of the active site of a biomolecular target. As a demonstration, we used our framework, which we refer to as sample-and-dock, to construct focused libraries for cyclin-dependent kinase type-2 (CDK2) and the active site of the main protease (Mpro) of the SARS-CoV-2 virus. We envision that the sample-and-dock framework could be used to generate theoretical maps of the chemical space specific to a given target and so provide information about its molecular recognition characteristics.

RNAs can recognize small-molecule ligands. However, the extent to which the molecules that they recognize differ from those recognized by proteins remains an open question. Cheminformatics analysis of experimentally validated RNA binders strongly suggests that RNA binders occupy a specific region of chemical space. However, less than 100 validated small molecule ligands are currently known. Here, we demonstrate how structure-based approaches could be used to navigate vast regions of the chemical space specific to ligand binding sites in five highly-structured RNAs. Our method involves using generative-AI to design target- and site-specific virtual libraries and then analyzing them using similar cheminformatics approaches as those used to assess experimentally validated RNA binders. Despite employing a completely orthogonal strategy, our results essentially reproduce the trends observed by analyzing the experimentally validated RNA binders. Large-scale generation of target and site-specific libraries may therefore prove to be helpful in simultaneously mapping the regions of chemical space unique to RNA and generating libraries that could be mined to identify novel RNA binders.

No abstract