RESUMO
Despite its genome sequencing more than two decades ago, the majority of the genes of Mycobacterium tuberculosis remain functionally uncharacterized. Patatins are one such class of proteins that, despite undergoing an expansion in this pathogenic species compared to their non-pathogenic cousins, remain largely unstudied. Recent advances in protein structure prediction using machine learning tools such as AlphaFold2 have provided high-confidence predicted structures for all M. tuberculosis proteins. Here we present detailed analyses of the patatin family of M. tuberculosis using AlphaFold-predicted structures, providing insights into likely modes of regulation, membrane interaction and substrate binding. Regulatory domains within this family of proteins include cyclic nucleotide binding, lid-like domains and other helical domains. Using structural homologues, we identified the likely membrane localization mechanisms and substrate-binding sites. These analyses reveal diversity in their regulatory capacity, mechanisms of membrane binding and likely length of fatty acid substrates. Together, this analysis suggests unique roles for the eight predicted patatins of M. tuberculosis.