Phospholipid hydroperoxide glutathione peroxidase (GPx4) is a moonlighting selenoprotein, which has been implicated in anti-oxidative defense, sperm development, and cerebral embryogenesis. Among GPx-isoforms, GPx4 is unique because of its capability to reduce complex lipid hydroperoxides and its tendency toward polymerization, but the structural basis for these properties remained unclear. To address this, we solved the crystal structure of the catalytically active U46C mutant of human GPx4 to 1.55 A resolution. X-ray data indicated a monomeric protein consisting of four alpha-helices and seven beta-strands. GPx4 lacks a surface exposed loop domain, which appears to limit the accessibility of the active site of other GPx-isoforms, and these data may explain the broad substrate specificity of GPx4. The catalytic triad (C46, Q81, and W136) is localized at a flat impression of the protein surface extending into a surface exposed patch of basic amino acids (K48, K135, and R152) that also contains polar T139. Multiple mutations of the catalytic triad indicated its functional importance. Like the wild-type enzyme, the U46C mutant exhibits a strong tendency toward protein polymerization, which was prevented by reductants. Site-directed mutagenesis suggested involvement of the catalytic C46 and surface exposed C10 and C66 in polymer formation. In GPx4 crystals, these residues contact adjacent protein monomers.

specific, the phenomena cross-reactivity and polyspecificity are observed quite frequently (see Kramer et al., 1997 [this issue of Cell]). As described previously, we suggest discriminating between the terms crossreactivity and polyspecificity; the term cross-reactivity

The 9E10 antibody epitope (EQKLISEEDL) derives from a protein sequence in the human proto-oncogen p62(c-myc) and is widely used as a protein fusion tag. This myc-tag is a powerful tool in protein localization, immunochemistry, ELISA or protein purification. Here, we characterize the myc-tag epitope by substitutional analysis and length variation using peptide spot synthesis on cellulose. The key amino acids of this interaction are the core residues LISE. The shortest peptide with a strong binding signal is KLISEEDL. Dissociation constants of selected peptide variants to the antibody 9E10 were determined. scFv constructs with the shortest possible myc-tags were successfully detected by Western blot and ELISA, giving a signal comparable to that of the original myc-tag.

A new method is presented that uses parallel peptide array synthesis on cellulose membranes to characterize protease/peptide inhibitor interactions. A peptide comprising P5-P4' of the third domain of turkey ovomucoid inhibitor was investigated for both binding to and inhibition of porcine pancreatic elastase. Binding was studied directly on the cellulose membrane, while inhibition was measured by an assay in microtiter plates with punched out peptide spots. The importance of each residue for binding or inhibition was determined by substitutional analyses, exchanging every original amino acid with all other 19 coded amino acids. Seven hundred eighty individual peptides were investigated for binding behavior to porcine pancreatic elastase, and 320 individual peptides were measured in inhibition experiments. The results provide new insights into the interaction between the ovomucoid derived peptide and subsites in the active site of elastase. Combining these data with length analysis we designed new peptides in a step-wise fashion which in the end not only inhibited elastase 400 times more strongly than the original peptide, but are highly specific for the enzyme. In addition, the optimized inhibitor peptide was protected against exopeptidase attack by substituting D-amino acids at both termini.