Identification of Potent ADCK3 Inhibitors through Structure-Based Virtual Screening

ADCK3 is a member of the UbiB family of atypical protein
kinases
in humans, with homologues in archaea, bacteria, and eukaryotes. In
lieu of protein kinase activity, ADCK3 plays a role in the biosynthesis
of coenzyme Q10 (CoQ10), and inactivating mutations can cause a CoQ10
deficiency and ataxia. However, the exact functions of ADCK3 are still
unclear, and small-molecule inhibitors could be useful as chemical
probes to elucidate its molecular mechanisms. In this study, we applied
structure-based virtual screening (VS) to discover a novel chemical
series of ADCK3 inhibitors. Through extensive structural analysis
of the active-site residues, we developed a pharmacophore model and
applied it to a large-scale VS. Out of ∼170,000 compounds virtually
screened, 800 top-ranking candidate compounds were selected and tested
in both ADCK3 and p38 biochemical assays for hit validation. In total,
129 compounds were confirmed as ADCK3 inhibitors, and among them,
114 compounds are selective against p38, which was used as a counter-target.
Molecular dynamics (MD) simulations were then conducted to predict
the binding modes of the most potent compounds within the ADCK3 active
site. Through metadynamics analysis, we successfully detected the
key amino acid residues that govern intermolecular interactions. The
findings provided in this study can serve as a promising starting
point for drug development.