The rapid emergence and spread of drug-resistant Mycobacterium tuberculosis infections has driven the urgent need for new antibiotics. Bedaquiline is the first new anti-tubercular compound with a novel mechanism of action licensed for use in humans in over 40 years. Bedaquiline works through the inhibition of subunit c (atpE) of F0-subunit of the ATP synthase to kill both replicating and dormant strains. To guide the effective use of Bedaquiline, and help minimize the development of clinical resistance, we built a predictive tool which could preemptively identify likely resistance mutations in patient samples.
Using the high-resolution X-ray structure of Bedaquiline bound to its target atpE, we have used a structural approach to characterize how variants lead to resistance. Computational tools were used to calculate the biophysical effects of the variants on the formation of the ATP synthase complex and drug binding, revealing a set of rules that explain how variants were likely to reduce drug binding while maintaining protein function. Combined with structural changes associated with the mutation, these were used to build a novel predictive tool that accurately identified likely resistance mutations (95% accuracy). This tool was used to evaluate circulating variants present in the Asia-Pacific region in the absence of Bedaquiline treatment, highlighting their susceptibility to Bedaquiline therapy. The knowledge accumulated during this study will now be used to explore how structural changes alter fitness and transmission dynamics, influencing the likelihood of resistance spreading.