| One person dies every 15 seconds because of tuberculosis. Isoniazid (INH) is the frontline drug used to combat this disease caused by Mycobacterium tuberculosis. Isoniazid's proposed target is InhA, an enoyl reductase enzyme in fatty acid synthesis--II pathway. To inhibit InhA, isoniazid requires activation by catalase-peroxidase (KatG). Mutations in the katG gene are the single largest determinant of isoniazid resistance, thus complicating the TB epidemic by the rising tide of multi-drug-resistant strains. Our research focuses on designing new drugs that do not require activation.;Similar MM-GBSA calculations performed with InhA (M.tuberculosis enzyme) reveal that the binding energy is correlated to the extent of binding loop ordering, with loop ordering accompanied by stronger affinity. This is a critical observation since the Fabl binding loop becomes ordered by triclosan binding, while that of InhA does not. We have performed simulations to test this hypothesis and determine how the inhibitor can be modified to optimize interaction with the binding loop and obtain a strong binding affinity to InhA, as is seen for Fabl. Obtaining such lead compounds is a critical step toward effective and affordable treatment of drug-resistant TB.;Designing new inhibitors is a challenge due to a lack of detailed structural information about the interaction of the enzyme with its substrate and the acyl carrier protein that delivers it. The same problem is encountered in InhA homolog E.coli enzyme, Fabl. Firstly, we employed X-ray crystallography. Although the experiment revealed the relative orientation of the enzymes, a dynamic interface resulted in poor electron density for the substrate and side-chains in the protein:protein interface. Starting with this data, we successfully employed molecular modeling and MD simulations to predict the structure of the productive Fabl-ACP complex. The resulting model is in excellent agreement with kinetic studies on wild-type and mutant Fabls. Importantly, the binding mode of the substrate differs from that of current inhibitors. Secondly, we complemented this structural study by applying the MM-PB(GB)SA to the binding by Fabl of a series of triclosan analogs that span 450,000 orders of magnitude in affinity. The same compounds were screened experimentally. We obtained a 98% correlation between calculated and measured relative binding affinities. |
