Crystal structure of PurE (BA0288) from <i>Bacillus anthracis</i> at 1.8 Å resolution

摘要

As part of a structural genomics program, Structural Proteomics in Europe (SPINE), we are determining protein structures from the causative agent of anthrax, Bacillus anthracis, a Gram-positive spore-forming bacterium. Among initial candidates for crystallographic analysis are enzymes involved in nucleotide biosynthesis. The BA0288 gene (www.tigr.org) of B. anthracis encodes a protein with 57% amino acid sequence identity to the Escherichia coli 5′-phosphoribosyl-5-aminoimidazole carboxylase (PurE).1 PurE proteins are highly conserved and are designated as Class I or Class II according to their enzymatic activity. Class I enzymes, found in yeast, plants, and prokaryotes, catalyze the second of a two-step conversion of 5-aminoimidazole ribonucleotide (AIR), via the intermediate N5-carboxyaminoimidazole ribonucleotide (N5-CAIR), to 4-carboxy-5-aminoimidazole ribonucleotide (CAIR).2, 3 The conversion of AIR to N5-CAIR is catalyzed by N5-CAIR synthetase (PurK) in the presence of ATP and bicarbonate. Class II enzymes from higher eukaryotes catalyze the conversion of AIR to CAIR directly, in the presence of bicarbonate or CO2.4 Therefore, Class I PurE proteins function as phosphoribosylaminoimidazole mutases, while Class II enzymes are carboxylases. This difference in activity could be exploited to provide potential targets for antibacterial therapies.5 The structure of PurE was determined to 1.8 Å resolution by molecular replacement using the coordinate set for the Escherichia coli orthologue (PDB code 1QCZ6) as a search model. Data collection, refinement and model-building statistics are summarized in Table I. The refined model consists of eight protein molecules (residues −7–161) for chains A–H and a total of 1219 water molecules. Residues 156–162 are not clearly defined in the electron-density maps for most subunits and are assumed to be disordered, although complete backbone chains can be traced for subunits B, E, and F. The Matthews' coefficient (Vm) for the crystals is 2.4 Å3/Da, and the estimated solvent content is 48.5%. The Ramachandran plot produced by PROCHECK 3.47 shows that 93.9% of residues are in the most favored regions with 6.1% in additional allowed regions. All the observed X-ray data in the resolution range 50.0–1.8 Å were used in the structure solution and refinement. While the Rsym value is high (60%) in the outer resolution shell, the value of I/σ(I) is 2.5, indicating these data are significant. The maximum likelihood program REFMAC5,8 which we used for refinement, is designed to handle such weak data in a robust manner and accord them appropriate weights. This is confirmed by the value of Rfree = 0.288 in the outer shell, which should be compared to a theoretical R-factor value for a random atom model of 0.586. For this B. anthracis PurE structure, the inclusion of all data in the refinement and map calculation considerably aided the building of the rather poorly ordered N-terminal region of two of the eight chains (C and G). The residues (SHHHHHHMKS) in these regions correspond to seven from the purification tag plus three from PurE itself. The B. anthracis PurE monomer [Fig. 1(a)] consists of a five-stranded, parallel β-sheet with a β2β1β3β4β5 strand topology together with six α-helices in a ‘flavodoxin-like’ fold.9 The strands and helices α1–α5 form a compact (βα)5 globular domain with helices α1 and α5 packing against one face of the β-sheet and helices α2, α3, and α4 packing onto the opposite face. Helix α6 protrudes away from the monomer, as illustrated in Figure 1(a). The PurE octamer is constructed as a dimer of tetramers, which are related by a two-fold symmetry axis perpendicular to the four-fold symmetry axis. Viewed down the four-fold axis, the assembly has the shape of a square from which the corners have been cut [Fig. 1(b)]. The protruding helix α6 makes numerous contacts with helix α2 and strand β2 of the neighboring subunit, thus contributing extensively to the contacts between adjacent mon