Many drug candidates contain a basic functional group that results in lysosomotropism--the accumulation of drug in the acidic lysosomes of a cell. When evaluating inhibitors of lysosomal enzymes, such as the cathepsins, this physical property can have a dramatic impact on the functional selectivity of the test compounds. A basic P3 substituent in cathepsin K inhibitors provides a means of achieving potent and selective enzyme inhibition. To evaluate the whole-cell selectivity of the basic cathepsin K inhibitor L-006235, we identified the irreversible pan-selective cathepsin probe BIL-DMK and used it to design whole-cell enzyme-occupancy assays. These cell-based assays showed a dramatic reduction in selectivity against cathepsins B, L, and S relative to the selectivities observed in enzyme assays. Two-photon confocal fluorescence microscopy showed punctated subcellular localization of L-006235, which colocalized with BODIPY-labelled Lysotracker, consistent with compound lysosomotropism. To address this potential problem, a series of potent cathepsin K inhibitors was developed by replacing the P2--P3 amide bond with a metabolically stable trifluoroethylamine moiety. X-ray crystallography has identified the binding of this functional group to active-site residues in cathepsin K. This modification resulted in increased potency and selectivity that allowed the removal of the basic P3 substituent. The resulting nonbasic inhibitor L-873724 is a 0.2 nM inhibitor of cathepsin K with cathepsin B, L, and S potencies that were not shifted between purified enzyme and whole-cell assays; thus indicating that this compound is not lysosomotropic. L-873724 exhibits excellent pharmacokinetics and is orally active in a monkey model of osteoporosis at 3 mg kg(-1) q.d.