Human dipeptidyl-peptidase III (hDPP III) is a zinc-dependent hydrolase cleaving dipeptides off the N-termini of various bioactive peptides. Thus, the enzyme is likely involved in a number of physiological processes such as nociception and is also implicated in several forms of cancer. We present high-resolution crystal structures of hDPP III in complex with opioid peptides (Met-and Leu-enkephalin, endomorphin-2) as well as with angiotensin-II and the peptide inhibitor IVYPW. These structures confirm the previously reported large conformational change of the enzyme upon ligand binding and show that the structure of the closed conformation is independent of the nature of the bound peptide. The overall peptide-binding mode is also conserved ensuring the correct positioning of the scissile peptide bond with respect to the catalytic zinc ion. The structure of the angiotensin-II complex shows, how longer peptides are accommodated in the binding cleft of hDPP III. Differences in the binding modes allow a distinction between real substrates and inhibitory peptides or “slow” substrates. The latter displace a zinc bound water molecule necessitating the energetically much less favoured anhydride mechanism as opposed to the favoured promoted-water mechanism. The structural data also form the necessary framework for the design of specific hDPP III inhibitors.

The genome of the yeast Saccharomyces cerevisiae encodes four flavodoxin-like proteins, namely Lot6p, Pst2p, Rfs1p and Ycp4p. Thus far only Lot6p was characterized in detail demonstrating that the enzyme possesses NAD(P)H:quinone oxidoreductase activity. In the present study, we heterologously expressed PST2 in Escherichia coli and purified the produced protein to conduct a detailed biochemical and structural characterization. Determination of the three-dimensional structure by X-ray crystallography revealed that Pst2p adopts the flavodoxin-like fold and forms tetramers independent of cofactor binding. The lack of electron density for FMN indicated weak binding, which was confirmed by further biochemical analysis yielding a dissociation constant of 20±1μM. The redox potential of FMN bound to Pst2p was determined to -89±3mV and is thus 119mV more positive than that of free FMN indicating that reduced FMN binds ca. five orders of magnitude tighter to Pst2p than oxidized FMN. Due to this rather positive redox potential Pst2p is unable to reduce free FMN or azo dyes as reported for other members of the flavodoxin-like protein family. On the other hand, Pst2p efficiently catalyzes the NAD(P)H dependent two-electron reduction of natural and artificial quinones. The kinetic mechanism follows a ping-pong bi-bi reaction scheme. In vivo experiments with a PST2 knock out and overexpressing strain demonstrated that Pst2p enables yeast cells to cope with quinone-induced damage suggesting a role of the enzyme in managing oxidative stress.

(Crabtree & Moorehouse, 1986) and well formed prismatic crystals obtained by diffusion of ether into a solution in dichloromethane (cod = 1,5-cyclooctadiene, py = pyridine). Lattice parameters were derived from the setting angles for 25 reflections with 7.5 < 0 < 9.7 °. Data collected using a crystal ca 0.25 × 0.25 × 0.25 mm, on an Enraf-Nonius CAD-4 diffractometer, monochromated Mo Ka radiation in the 0--20 mode, with dO = (0.8 + 0.35tan0) ° and a maximum scan time of 1 min. A total of 8119 measured reflections for 2 < 0 < 25 ° and h 0---, 13, k 0---,22, l -23-+23, 7737 unique reflections, Rint = 0.017, and 5438 observed reflections with IF21 > 2o'(F2), where or(F2) = {ora(/)+(0.041/)2}m/Lp, were used in the refinement. Two reference reflections remeasured every hour showed 14% decay, for which a correction was made. An absorption correction was applied using DIFABS (Walker & Stuart, 1983), 0108-2701/92/040735-03503.00 after isotropic refinement, and gave maximum and minimum corrections of 1.14 and 0.81. Lorentz and polarization corrections were made. The structure was solved by routine heavy-atom methods and non-H atoms refined anisotropically by full-matrix least squares. H atoms were held fixed at calculated positions with Uis o = 1.3 Ueq for the atoms to which they are bonded. Solvent CH2C12 at half occupancy and disordered, modelled with two C1 atoms disordered over three sites and H atoms omitted. With a weighting scheme of w = 1/or2(F), Yw(IFoL-IF~I) 2 minimized, the final residuals were R = 0.050 and wR = 0.070, 5438 observed reflections, 503 variables, S = 2.0, (A/Or)max = 0.01, Apmax = 1.9, Apmin =

No abstract

No abstract