Notes
Thousands of nuclear and cytosolic proteins are subjected to post-translational modification at serine or threonine residues by O-linked ß-N-acetylglucosamine (O-GlcNAc). This unique type of glycosylation, known as O-GlcNAcylation, does not result in complex glycans but is rather cycled on and off proteins by the two enzymes O-GlcNAc transferase (OGT) and O-GlcNAcase (OGA), respectively. Notably, these enzymes work in an interplay with phosphorylation regulating a wide variety of biological processes, including gene expression, cell cycle progression, epigenetics, and stress response. O-GlcNAcylation is thus critical in maintaining cell homeostasis, and its dysregulation is linked with severe human pathologies like cancer, diabetes, and cardiovascular and neurodegenerative diseases. Hence, the key enzymes OGT and OGA have both been proposed as potential therapeutic targets for these conditions. For instance, an increase in O-GlcNAcylation levels and OGT expression can be observed in many human cancers, and different studies have shown that OGT inhibition decreases cancer cell proliferation. Although the available data are promising, the exact role of O-GlcNAcyation in the pathogenesis of tumors and other conditions is far from being well understood. One of the main obstacles to studying this critical post-translational modification is the limited availability of potent and cell-permeable OGT inhibitors. Namely, most inhibitors are structurally related to the glycosyl donor uridine diphosphate N-acetylglucosamine (UDP-GlcNAc) or the enzymatic reaction byproduct, UDP, so they are usually not cell-permeable nor selective. Only in recent years the first series of potent small molecules OGT inhibitors, known as OSMIs, has been developed, but the compounds are not suitable for in vivo studies. Our work within the scope of this dissertation focused on filling this gap and providing the field with new chemical tools to study OGT's role in different biological systems. In particular, we employed four different strategies to design novel OGT inhibitors. We began by developing two new libraries of fragment-like inhibitors based on the 2-hydroxyquinoline uridine mimetic scaffold discovered by our group in a previous study. The compounds were active in vitro at micromolar concentrations, and the best hit showed an IC50 value of 13 µM. Since the libraries were screened with two different biochemical assays, our data also pointed out that special care should be taken when testing UDP mimics with the commercial UDP Glo? assay, as false positives may occur. Our second approach consisted in performing virtual screening of an extensive library of drug-like molecules. This allowed us to identify novel uridine mimetic chemotypes and select the most promising commercially available hits for in vitro screening. Notably, some of the tested molecules inhibited the enzyme at micromolar concentrations, and the best hit displayed an IC50 value of 7 µM, which places it amongst the most potent OGT inhibitors reported to date. Thus, the compound was additionally explored by developing a focused library of its derivatives to study its SAR. The inhibitor was also tested in cell-based assays but unfortunately was not active at micromolar concentrations. Since the first nanomolar OGT inhibitors became available while we were working on our project, we decided to resynthesize the most potent one, OSMI-4, and use it to study O-GlcNAcylation in different human cancer cell lines. Our results revealed that the compound's metabolic stability varies between cell lines and could be improved by replacing a labile ethyl ester featured in the inhibitor's structure with more stable functional groups. We therefore designed and synthesized a library of amide and amine-based OSMI-4 derivatives and demonstrated that ester replacement does not significantly affect the compound's potency. Notably, while the cell permeability of the molecules is still being investigated, their stability to hydrolysis by plasma esterases has greatly improved. Hence, we believe this strategy will help us modulate the inhibitors' drug properties and might eventually yield the first potent chemical probe suitable for in vivo studies. Finally, we presented a new potentially druggable pocket near OGT's active site. The pocket surrounds Asp554, which is thought to be involved in the enzymatic catalysis and could thus represent a target for OGT inhibition. Our approach consisted in designing a small series of fragments using various computational tools, followed by their synthesis and in vitro screening. One fragment showed promising results by inhibiting the enzyme with an IC50 value of 1.4 mM. However, since the binding site of these compounds still needs to be experimentally confirmed, x-ray crystallographic data of the ligand-protein complex should be acquired to help design more potent inhibitors.