The development of TB drugs benefits greatly from validation of novel drug targets in predictive animal models. M. tuberculosis (Mtb) enzymes in central carbon metabolism are emerging as promising targets for drug development but have not been validated in animal models that recapitulate the diverse and heterogeneous environments Mtb encounters in humans. We propose to test the hypothesis that these heterogeneous environments result in airway, lung and granuloma-dependent nutritional restrictions that, at times, make Mtb dependent on both glycolysis and gluconeogenesis to establish and maintain infection in nonhuman primates and/or during paucibacillary infection in mice. We will infect NHPs with Mtb mutants of two enzymes – PFK and PEPCK - that are required for glycolysis or gluconeogenesis, respectively. We will use deletion mutants to determine whether these enzymes are required for establishment of infection and use conditional knockdown mutants to investigate the enzymes importance for growth and survival in different pathologies and for progression of disease. We will furthermore evaluate these mutants in a relapse mouse model to test the hypothesis that the nutritional requirements during long-term persistence in mice, when the bacteria cannot be cultured in vitro, are different from those encountered during active growth and high titer chronic infection and ask whether they mimic environments encountered in NHPs. This proposal builds on the experience of the multi- PI team in mycobacterial genetics, metabolism, and animal models, with the goal to dissect the carbon source requirements for Mtb in the model system that is most similar to human Mtb infection and disease.