Design, synthesis and evaluation of a novel series of spiroketals based on the structure of the antibacterial gyrase inhibitor novobiocin

Molecular modelling has been used in conjunction with crystallographic and biological data in an attempt to design compounds that mimic the structure and activity of the coumarin antibiotic novobiocin 2. Calculations on four conformations of a 6,6-spiroketal system 8–11 suggest that whilst 9 should be the lowest energy, 8 should also be readily accessible and this conformation overlays well with the coumarin and sugar rings of novobiocin. Incorporation of key hydroxy and carbamate groups onto the cyclohexane and a hydrogen bond acceptor onto the aromatic ring suggest 12a as the initial target. The crystal structure of a model compound 24 shows the conformation illustrated in 9 and thus supports the molecular modelling. However, the crystal structure of novobiocin bound to a 24 kD fragment of gyrase B, the target of the coumarin antibiotics, reveals that it is the lactone carbonyl rather than the 4-oxy group which is responsible for a key interaction with an Arg residue and consequently the carboxy group in 12a is misplaced as a corresponding H-bond acceptor. Subsequently, compounds incorporating an extra aromatic (as in 13a) or heteroaromatic ring (as in 45) bearing an H-bond acceptor, along with the extra dimethyl groups on the cyclohexane ring, have been designed. Models of these structures overlay convincingly with novobiocin bound to 24 kD gyrase B. Versatile synthetic routes have been developed allowing these compounds and the underlying hypotheses to be tested. Unfortunately, none of the compounds demonstrates significant enzyme or antibacterial activity, which we attribute to a combination of features, including the lack of a replacement for the noviose methoxy group and the failure to achieve a good stacking (charge transfer) interaction with Arg-76.