Aggrecanases are now believed to be the principal proteinases responsible for aggrecan degradation in osteoarthritis. Given their potential as a drug target, we solved crystal structures of the two most active human aggrecanase isoforms, ADAMTS4 and ADAMTS5, each in complex with bound inhibitor and one wherein the enzyme is in apo form. These structures show that the unliganded and inhibitor-bound enzymes exhibit two essentially different catalytic-site configurations: an autoinhibited, nonbinding, closed form and an open, binding form. On this basis, we propose that mature aggrecanases exist as an ensemble of at least two isomers, only one of which is proteolytically active.Keywords: protein structure; enzymes; metalloproteins; aggrecanases Supplemental material: see www.proteinscience.org Osteoarthritis (OA) is a progressive disease that results in degradation of articular cartilage and chronic pain. The extracellular matrix is composed of two major components, aggrecan and collagen. Aggrecan is a large multidomain proteoglycan that provides cartilage with compressibility and elasticity by swelling and hydrating the collagen network (Vertel and Ratcliffe 2000). Loss of aggrecan is considered a critical early event in OA, occurring initially at the joint surface and progressing to the deeper zones. This is followed by degradation of collagen fibrils and mechanical failure of the tissue (Nagase and Kashiwagi 2003). Aggrecanase-1 (ADAMTS4) and aggrecanase-2 (ADAMTS5), members of the ADAMTS (a disintegrin and metalloprotease with thrombospondin motifs) gene family, cleave aggrecan at a unique site termed the ''aggrecanase site Tortorella et al. 1999). ADAMTS4 and ADAMTS5 are expressed in human normal and OA cartilage (Yamanishi et al. 2002) and in OA synovium, and contribute to the structural damage that characterizes human OA (Powell et al. 2007;Song et al. 2007). However, there is no consensus in the literature as to which aggrecanase is the most important in human OA. In mice, ADAMTS5 (but not ADAMTS4) is responsible for disease progression in a surgically induced model of OA (Glasson et al. 2004(Glasson et al. , 2005. ADAMTS4/ADAMTS5 double knockout mice are physiologically normal (Majumdar et al. 2007) and also protected from developing OA. Given the normal phenotype of the double knockout mice, dual inhibition Article published online ahead of print. Article and publication date are at http://www.proteinscience.org/cgi

The reaction between bovine heart cytochrome oxidase and dioxygen was investigated at room temperature following photolysis of the fully reduced CO-bound enzyme. Time-resolved optical absorption difference spectra were collected by a gated multichannel analyzer in the visible region (lambda = 460-720 nm) from 50 ns to 50 ms after photolysis. Singular value decomposition (SVD) analysis indicated the presence of at least seven intermediates. Multiexponential fitting gave the following apparent lifetimes: 1.2 microseconds, 10 microseconds, 25 microseconds, 32 microseconds, 86 microseconds, and 1.3 ms. On the basis of the SVD results and a double difference map, a sequential kinetic mechanism is proposed from which the spectra and time-dependent populations of the reaction intermediates were determined. The ferrous-oxy complex (compound A), with a peak at 595 nm and a trough at 612 nm versus the reduced enzyme, reaches a maximum concentration approximately 30 microseconds after photolysis. It decays to a 1:6 mixture of peroxy species (a3(3+)-O(-)-O-) in which cytochrome a is reduced and oxidized. Cytochrome a3 in both species has a peak at 606 nm versus its oxidized form. The peroxy species decay to a ferryl intermediate, with a peak at 578 nm versus the oxidized enzyme, followed by electron redistribution between CuA and cytochrome a. The two ferryl species reach a maximum concentration approximately 310 microseconds after photolysis. The excellent agreement between the experimental and theoretical spectra of the intermediates provides unequivocal evidence for the presence of peroxy and ferryl species during dioxygen reduction by cytochrome oxidase at room temperature.

No abstract

Intramolecular electron transfer and conformational changes in cytochrome c oxidase were studied at room temperature following the photodissociation of CO bound to mixed-valence enzyme (cytochrome a3(2+)-CO CuB+ cytochrome a3+ CuA2+) and fully reduced enzyme. Time-resolved optical absorption difference spectra were collected in the Soret region on time scales of nanoseconds to milliseconds using a gated optical spectrometric multichannel analyzer. A global exponential fitting procedure combined with a singular value decomposition method was used to analyze the transient difference spectra at various times following CO photolysis. The analysis shows that at least two processes, with apparent lifetimes of 1.4 microseconds and 11.1 ms, are present following the photodissociation of CO bound to the fully reduced enzyme. These are attributed to a conformational change and CO recombination at the cytochrome a3 site, respectively. Global analysis of the mixed-valence CO complex transient difference spectra showed the presence of five intermediates with apparent lifetimes of 1.0 microseconds, 5.2 microseconds, 83.7 microseconds, 10.5 ms, and 25.3 ms. The data on a microsecond time scale are consistent with a mechanism involving a conformational change at cytochrome a3, followed by electron transfer from cytochrome a3 to cytochrome a with subsequent electron transfer to CuA. One of the two processes on a millisecond time scale is attributed to CO recombination and the other to a structural rearrangement or heme-heme electron transfer. On the basis of this mechanism, the kinetics and the absorption spectra of the intermediates involved in the conformational and electron transfer dynamics of the mixed-valence enzyme were determined.

The photolysis intermediates of partially and fully reduced CO-bound cytochrome oxidase derivatives were investigated. A gated optical spectrometric multichannel analyzer was used to collect visible and near-infrared transient difference spectra on time scales from nanoseconds to milliseconds. The spectra were analyzed by a singular value decomposition method combined with a global exponential fitting procedure. Global analysis of the mixed-valence CO complex transient difference spectra shows that five intermediates are present with apparent lifetimes of 1.4 microseconds, 4.8 microseconds, 76.7 microseconds, 10.6 ms, and 21.6 ms. The data were fitted to a kinetic model involving a sequential pathway with accompanying equilibria. On the basis of this mechanism, the absorption spectra of the intermediates were determined. The first step, also present in the fully reduced enzyme, is attributed to a conformational change at cytochrome a3. The spectral changes associated with the second step are similar to those expected for 1:1 electron transfer from cytochrome a3 to cytochrome a, except for a higher absorbance between 480 and 550 nm. A comparison of the experimental spectral change associated with this step, (a2+ minus a3+) minus (a3(2+) minus a3(3+), and the calculated spectral change, (a2+ CuA+ minus a3+ CuA2+) minus a3(2+) CuB+ minus a3(3+) CuB2+), allowed extraction of the absorbance spectrum of CuA2+ in the 480-550 nm region. The spectral change associated with the third step is consistent with the oxidation of cytochrome a. A decrease in the 830 nm band on the same time scale indicates that the electron acceptor is CuA. The data also suggest that the redox state of CuB significantly affects the absorption spectrum of oxidized cytochrome a3 in the visible region. The two processes on a millisecond time scale are attributed to CO recombination and intramolecular electron transfer.