Pycnodysostosis (Pycno) is a rare sclerosing skeletal dysplasia inherited as an autosomal recessive trait. Features include short stature, osteosclerosis, and abnormalities of cranial and facial bones. Previously, we determined by positional cloning and mutation analysis that the Pycno gene was the cysteine lysosomal protease, cathepsin K (Cat K). To date, Cat K mutations on 21 Pycno families were determined by exonic amplification and sequencing. Thirteen novel mutations were identified, including three nonsense, a stop codon (X330W), and nine missense mutations. Missense mutations have occurred in the signal peptide, pro region, and mature enzyme portion of the Cat K prepropeptide. There was no apparent phenotype-genotype correlation. Constructs for six of the missense mutations (G79E, G146R, Y212C, A277E, A277V and R312G) were created by site-directed mutagenesis and expressed in the Pichia system. Mutant protein was detected by western blot and enzyme activity was assayed. All expressed mutants had immunologically detectable protein but only the Y212C mutant had residual activity. After quantitating the Y212C protein using the cysteine cathepsin inhibitor, E64, kinetic studies were performed with the fluorescent substrate Z-Leu-Arg-MCA. These studies revealed that the Km for Y212C was not significantly different from wild type enzyme while the kcat was significantly reduced to approximately 40% of the wild type value. The pH profile of the Y212C enzyme was broad with an optimum at about 6.0, similar to that of the wild type enzyme. Stability assays revealed marked instability of the Y212C protein compared to wild type Cat K. Previous molecular modeling had predicted that the G79E, G146R, and A277E mutant proteins would be stable but enzymatically inactive, findings confirmed by the expression data. In contrast, modeling of the Y212C mutation had predicted that this mutation might interfere with the C170-C210 disulfide bridge on the enzyme's surface, potentially destabilizing the molecule. The in vitro expression data for Y212C were consistent with that conclusion, since substrate binding affinity and the pH optimum were maintained and the proteolytic efficiency was only moderately reduced, but the protein was unstable. To further validate the Cat K model, a predicted revertant of the G146R mutant (S143D/G146R) was expressed and partially restored enzymatic activity at low pH. Thus, molecular modeling and mutant expression studies provide insight into the molecular basis of the Cat K deficiency caused by these Pycno mutations.