Structure analysis and molecular model of the selenoenzyme glutathione peroxidase at 2.8 A resolution.

Abstract The three-dimensional structure of bovine erythrocyte glutathione peroxidase, a tetrameric enzyme containing 4 gram atoms of selenium per mole ( M r = 84,000), has been determined at 2.8 A resolution using the multiple isomorphous replacement method. By correlation calculations in Patterson space the tetramers were shown to exhibit molecular [222] symmetry, proving the monomers to be identical or at least very similar. The monomer consists of a single polypeptide chain of 178 amino acid residues. Its shape is nearly spherical with a radius of r ≈ 19 A . A tentative sequence corresponding to a partially refined model ( R = 0.38) is given. Each subunit is built up from a central core of two parallel and two anti-parallel strands of pleated sheet surrounded by four α-helices. One of the helices runs antiparallel to the neighbouring β-strands giving rise to a βαβ substructure, an architecture that has been found in several other proteins e.g. flavodoxin, thioredoxin, rhodanese and dehydrogenases. A comparison of the glutathione peroxidase subunit structure with thioredoxin-S 2 revealed large regions of structural resemblance. The central four-stranded β structure together with two parallel α-helices resembles nearly 80% of the thioredoxin fold. The active sites of glutathione peroxidase are located in flat depressions on the molecular surface. Probably each active centre is built up by segments from two subunits. The catalytically active selenocysteines were found at the N-terminal ends of long α-helices and are surrounded by an accumulation of aromatic side-chains. A difference Fourier map between oxidized and substrate-reduced glutathione peroxidase as well as heavy-atom binding led to the conclusion that the two-electron redox-cycle involves a reversible transition of the active-site selenium from a selenenic acid (RSeOH) to a seleninic acid (RSeOOH).