Latent human fibroblast collagenase (HFC) can be activated by a variety of seemingly disparate means. In addition to the well-characterized activation by trypsin and organomercurial compounds, the enzyme can be activated to various extents by surfactants such as sodium dodecyl sulfate, by chaotropic ions such as SCN-, by disulfide compounds such as oxidized glutathione, by sulfhydryl alkylating agents such as N-ethylmaleimide, and by oxidants such as NaOCl. The underlying basis for these activations is the modification, exposure, or proteolytic release of the Cys73 residue from its habitat in the latent enzyme where it is thought to be complexed to the active-site zinc atom. This residue is not accessible for reaction with small molar excesses of dithionitrobenzoate in native, latent HFC. However, on addition of EDTA, this residue becomes fully exposed and is quantitatively labeled. All modes of activation of latent HFC are believed to involve the dissociation of Cys73 from the active-site zinc atom and its replacement by water, with the concomitant exposure of the active site. This is thought to be the primary event that precedes the well-known autolytic cleavages that are observed following the appearance of collagenase activity. The dissociation of Cys73 from the zinc atom in the latent enzyme "switches" the role of the zinc from a noncatalytic to a catalytic one. This "cysteine switch" mechanism of regulation may be applicable to the entire collagenase gene family.Human fibroblast collagenase (HFC) is a member of a homologous class of proteinases called the matrix metalloproteinases (MMPs) that are collectively capable of catabolizing the major components of the extracellular matrix (1, 2). HFC and the collagenase from neutrophils are the only two human enzymes capable of catalyzing the breakdown of interstitial collagens in tissue at an appreciable rate, while the remaining MMPs efficiently catalyze the destruction of other matrix macromolecules. Therefore, a knowledge of the factors that control the activity of these enzymes is critical to an understanding of connective tissue breakdown in vivo. Although the behavior of all of the MMPs has not yet been completely characterized, these enzymes appear to share the characteristic that they are synthesized and secreted in an inactive form and are subsequently activated in situ. The biochemical basis for this latency and the physiological mechanisms of activation of the latent form of these enzymes have not yet been delineated. HFC is the best-characterized member of the MMP family and can now be prepared in sizable quantities (1). Thus, it is a good candidate for detailed studies of the activation of the latent form.The sequence of latent HFC has been elucidated from a cDNA clone (3). It has a molecular mass of 52 kDa, and a I)D lllkill .----.

We have prepared a series of achiral aminoacetonitriles, bearing tri-ring benzamide moieties and an aminocyclohexanecarboxylate residue at P2. This combination of binding elements resulted in sub-250 pM, reversible, selective, and orally bioavailable cathepsin K inhibitors. Lead compounds displayed single digit nanomolar inhibition in vitro (of rabbit osteoclast-mediated degradation of bovine bone). The best compound in this series, 39n (CRA-013783/L-006235), was orally bioavailable in rats, with a terminal half-life of over 3 h. 39n was dosed orally in ovariectomized rhesus monkeys once per day for 7 days. Collagen breakdown products were reduced by up to 76% dose-dependently. Plasma concentrations of 39n above the bone resorption IC50 after 24 h indicated a correlation between functional cellular and in vivo assays. Inhibition of collagen breakdown by cathepsin K inhibitors suggests this mechanism of action may be useful in osteoporosis and other indications involving bone resorption.

Protein kinases c-Abl, b-Raf, and p38alpha are recognized as important targets for therapeutic intervention. c-Abl and b-Raf are major targets of marketed oncology drugs Imatinib (Gleevec) and Sorafenib (Nexavar), respectively, and BIRB-796 is a p38alpha inhibitor that reached Phase II clinical trials. A shared feature of these drugs is the fact that they bind to the DFG-out forms of their kinase targets. Although the discovery of this class of kinase inhibitors has increased the level of emphasis on the design of DFG-out inhibitors, the structural determinants for their binding and stabilization of the DFG-out conformation remain unclear. To improve our understanding of these determinants, we determined cocrystal structures of Imatinib and Sorafenib with p38alpha. We also conducted a detailed analysis of Imatinib and Sorafenib binding to p38alpha in comparison with BIRB-796, including binding kinetics, binding interactions, the solvent accessible surface area (SASA) of the ligands, and stabilization of key structural elements of the protein upon ligand binding. Our results yield an improved understanding of the structural requirements for stabilizing the DFG-out form and a rationale for understanding the genesis of ligand selectivity among DFG-out inhibitors of protein kinases.

Plasma kallikrein is a serine protease that has many important functions, including modulation of blood pressure, complement activation, and mediation and maintenance of inflammatory responses. Although plasma kallikrein has been purified for 40 years, its structure has not been elucidated. In this report, we described two systems (Pichia pastoris and baculovirus/Sf9 cells) for expression of the protease domain of plasma kallikrein, along with the purification and high resolution crystal structures of the two recombinant forms. In the Pichia pastoris system, the protease domain was expressed as a heterogeneously glycosylated zymogen that was activated by limited trypsin digestion and treated with endoglycosidase H deglycosidase to reduce heterogeneity from the glycosylation. The resulting protein was chromatographically resolved into four components, one of which was crystallized. In the baculovirus/Sf9 system, homogeneous, crystallizable, and nonglycosylated protein was expressed after mutagenizing three asparagines (the glycosylation sites) to glutamates. When assayed against the peptide substrates, pefachrome-PK and oxidized insulin B chain, both forms of the protease domain were found to have catalytic activity similar to that of the full-length protein. Crystallization and x-ray crystal structure determination of both forms have yielded the first three-dimensional views of the catalytic domain of plasma kallikrein. The structures, determined at 1.85 Å for the endoglycosidase H-deglycosylated protease domain produced from P. pastoris and at 1.40 Å for the mutagenically deglycosylated form produced from Sf9 cells, show that the protease domain adopts a typical chymotrypsin-like serine protease conformation. The structural information provides insights into the biochemical and enzymatic properties of plasma kallikrein and paves the way for structure-based design of protease inhibitors that are selective either for or against plasma kallikrein.

No abstract