Abstract We previously reported potent ligands and inhibitors of Mycobacterium tuberculosis dethiobiotin synthetase ( Mt DTBS), a promising target for antituberculosis drug development (Schumann et al., ACS Chem Biol. 2021, 16, 2339-2347); here the unconventional origin of the fragment compound they were derived from is described for the first time. Compound 1 (9b-hydroxy-6b,7,8,9,9a,9b-hexahydrocyclopenta[3,4]cyclobuta[1,2-c]chromen-6(6a H )-one), identified by in silico fragment screen, was subsequently shown by surface plasmon resonance to have dose-responsive binding ( K D 0.6 mM). Clear electron density was revealed in the DAPA substrate binding pocket, when 1 was soaked into Mt DTBS crystals, but the density was inconsistent with the structure of 1 . Here we show the lactone of 1 hydrolyses to carboxylic acid 2 under basic conditions, including those of the crystallography soak, with subsequent ring-opening of the component cyclobutane ring to form cyclopentylacetic acid 3 . Crystals soaked directly with authentic 3 produced electron density that matched that of crystals soaked with presumed 1 , confirming the identity of the bound ligand. The synthetic utility of fortuitously formed 3 enabled subsequent compound development into nanomolar inhibitors. Our findings represent an example of chemical modification within drug discovery assays and demonstrate the value of high-resolution structural data in the fragment hit validation process. Synopsis A molecule flagged in an in silico docking screen against Mt DTBS, was inadvertently hydrolysed in the crystal conditions used for hit validation. The resulting fragment-sized molecule bound to the DAPA substrate binding pocket of the target enzyme ( Mt DTBS) with millimolar affinity, as measured by surface plasmon resonance, but was later modified to a highly potent (nanomolar) ligand and promising lead for the development of novel tuberculosis treatments. Graphical Abstract