Roughly 70 percent of current lead compounds in modern drug discovery derive directly from the natural products, many of which are glycosylated bacterial metabolites. While it is known that the sugar ligands of these pharmaceutically important metabolites often define their corresponding biological activity, efficient methods to alter these essential carbohydrate ligands are still lacking. This proposal outlines a stepwise approach to accomplish this goal while also providing invaluable mechanistic and structural information on two critical, but poorly understood, enzyme classes; namely, nucleotidylytransferases and glycosyltransferases. Specifically, the proposed studies are designed to exploit structure/function-based protein engineering to generate a promiscuous in vitro nucleotidylyltransferase/glycosyltransferase systems which will provide a library of potentially new bioactive metabolites. The model system selected includes the Salmonella rmlA-encoded alpha-D-glucose-1-phosphate thymidylyltansferase (Ep) and three glycosyltransferases, which act ponerthronolide B (EryBV, EryCIII and MegGT). Among the many advantages of the presented model, it has been shown that the MegGT-catalyzed addition of a single sugar (megosamine) to erthromycin leads to a metabolite (megalomicin) with remarkably different biological activity. Thus, the selected model promises varied biological activity from metabolites anticipated from the proposed studies. The specific goals include: 1) structural and mechanistic studies on Ep3 EryBV, EryCIII and MegGT, 2) structure-based Ep engineering for the construction of nucleotide sugar libraries and 3) the use of this library, in conjunction with EryBV, EryCIII and MegGT, to generate libraries of glycosylated erythronolides with potentially new bioactivity.
Showing the most recent 10 out of 56 publications
