The nature of the sugars attached to many drugs (e.g., the antibiotics erythromycin and vancomycin) can influence their biological activity, but researchers have few means of manipulating glycosylation. In cells, the first step in the glycosylation pathway is the priming of a sugar with a nucleotide diphosphate, which facilitates its attachment to the substrate. However, most of the enzymes involved in this process work with a restricted repertoire of sugars. To broaden this repertoire, scientists at the Memorial Sloan–Kettering Cancer Center (New York) and Brookhaven National Laboratory (Brookhaven, NY) have created an enzyme that is more promiscuous in its preference for sugars (Nat. Struct. Biol. 8, 545–549, 2001). The researchers first solved the three-dimensional structure of the Ep enzyme, a nucleotidylyltransferase from Salmonella. They then engineered Ep to create a variant capable of liaising with sugars not naturally targeted by the wild-type enzyme. Dimitar Nikolov, senior author, says: “Our aim is to use completely in vitro steps with engineered enzymes, the final result of which will be libraries of modified or randomized compounds based on naturally glycosylated compounds.” Some of these variants could have superior properties as medicines.