We report here the crystallographic and biophysical analysis of a soluble, catalytically active fragment of the Escherichia coli type I signal peptidase (SPase ⌬2-75) in complex with arylomycin A 2 . The 2.5-Å resolution structure revealed that the inhibitor is positioned with its COOH-terminal carboxylate oxygen (O45) within hydrogen bonding distance of all the functional groups in the catalytic center of the enzyme (Ser 90 O-␥, Lys 145 N-, and Ser 88 O-␥) and that it makes ␤-sheet type interactions with the ␤-strands that line each side of the binding site. Ligand binding studies, calorimetry, fluorescence spectroscopy, and stopped-flow kinetics were also used to analyze the binding mode of this unique non-covalently bound inhibitor. The crystal structure was solved in the space group P4 3 2 1 2. A detailed comparison is made to the previously published acyl-enzyme inhibitor complex structure (space group: P2 1 2 1 2) and the apo-enzyme structure (space group: P4 1 2 1 2). Together this work provides insights into the binding of pre-protein substrates to signal peptidase and will prove helpful in the development of novel antibiotics.Type I signal (leader) peptidase (SPase, 1 EC 3.4.21.89) is the membrane-bound serine endopeptidase that catalyzes the cleavage of the amino-terminal signal (or leader) peptide from secretory proteins and some membrane proteins (for recent reviews, see Refs. 1-3). Evolutionarily, SPase belongs to the protease clan SF and the protease family S26 (4). The Escherichia coli SPase has served as the model Gram-negative SPase and is the most thoroughly characterized SPase to date. It has been cloned (5), sequenced (6), overexpressed (7), purified (6,8,9), and kinetically (10), and structurally (11, 12) characterized. E. coli SPase (323 amino acids, 35,988 Da, pI 6.9) contains two amino-terminal transmembrane segments (residues 4 -28 and 58 -76), a small cytoplasmic region (residues 29 -58), and a carboxyl-terminal periplasmic catalytic region (residues 77-323). A catalytically active fragment of SPase (SPase ⌬2-75) corresponding to the periplasmic region (lacking the two transmembrane segments and the cytoplasmic domain) has been cloned, purified, characterized (13,14), and crystallized (15). Interestingly, the ⌬2-75 construct required detergent or lipid for optimal activity (14) and crystallization (15).The crystal structure of ⌬2-75 has been solved in complex with a ␤-lactam-type inhibitor as well in the apo-form (11, 12). The structures of E. coli SPase ⌬2-75 revealed that the periplasmic region of bacterial signal peptidase has a unique, mostly ␤-structure protein fold made of several coiled ␤-sheets and contains an Src homology 3-like barrel. The periplasmic region of SPase is made up of two domains. Domain I contains the catalytic residues and all of the conserved regions of sequence. It also contains an unusually large exposed hydrophobic surface that is consistent with a membrane association surface and possibly the detergent/lipid requirement of the ⌬2-75 deletion construct. The ...

Ceftobiprole exhibited tight binding to PBP2a in methicillin-resistant Staphylococcus aureus, PBP2x in penicillin-resistant Streptococcus pneumoniae, and PBP3 and other essential penicillin-binding proteins in methicillin-susceptible S. aureus, Escherichia coli, and Pseudomonas aeruginosa. Ceftobiprole also bound well to PBP2 in the latter organisms, contributing to the broad-spectrum antibacterial activity against gram-negative and gram-positive bacteria.

The kinetics of beta-lactam hydrolysis by wild-type AmpC beta-lactamase from Escherichia coli and three mutant proteins created by substitution of tyrosine 150 have been examined. The catalytic efficiency was decreased 10- to 1000-fold according to the substrate and mutant being studied. The effect of the mutation was much stronger with rapidly hydrolyzed substrates (e.g., cephalothin) than it was with slowly hydrolyzed substrates (e.g., ceftriaxone). With the latter substrates, the mutagenesis had a much stronger effect on apparent affinity than it did on rates of catalysis. Indeed, the enzyme appeared to be more reactive toward certain of the slowly hydrolyzed substances (e.g., methicillin, aztreonam, and ceftriaxone). These observations were not compatible with an obligatory role of tyrosine 150 in catalysis. The analysis of the effects of the mutation on activity was complicated by the observation of at least two, kinetically distinct, forms of the enzymes. It appeared that mutation of tyrosine 150 influenced the kinetic properties of one state and that this residue is involved in the partitioning of the enzyme between the different reactive states.

The interactions of ceftobiprole with purified ␤-lactamases from molecular classes A, B, C, and D were determined and compared with those of benzylpenicillin, cephaloridine, cefepime, and ceftazidime. Enzymes were selected from functional groups 1, 2a, 2b, 2be, 2d, 2e, and 3 to represent ␤-lactamases from organisms within the antibacterial spectrum of ceftobiprole. Ceftobiprole was refractory to hydrolysis by the common staphylococcal PC1 ␤-lactamase, the class A TEM-1 ␤-lactamase, and the class C AmpC ␤-lactamase but was labile to hydrolysis by class B, class D, and class A extended-spectrum ␤-lactamases. Cefepime and ceftazidime followed similar patterns. In most cases, the hydrolytic stability of a substrate correlated with the MIC for the producing organism. Ceftobiprole and cefepime generally had lower MICs than ceftazidime for AmpCproducing organisms, particularly AmpC-overexpressing Enterobacter cloacae organisms. However, all three cephalosporins were hydrolyzed very slowly by AmpC cephalosporinases, suggesting that factors other than ␤-lactamase stability contribute to lower ceftobiprole and cefepime MICs against many members of the family Enterobacteriaceae.

N-(5'-Phosphoribosyl)anthranilate isomerase-indole-3-glycerol-phosphate synthase from Escherichia coli is a monomeric bifunctional enzyme ofMr 49,500 that catalyzes two sequential reactions in the biosynthesis of tryptophan. The three-dimensional structure ofthe enzyme has been determined at 2.8-A resolution by x-ray crystallography. The two catalytic activities reside on distinct functional domains of similar folding, that of an eightfold parallel ,8-barrel with a-helices on the outside connecting the f8-strands. Both active sites were located with an iodinated substrate analogue and found to be in depressions on the surface of the domains created by the outward-curving loops between the carboxyl termini of the ,8-sheet strands and the subsequent a-helices. They do not face each other, making "channeling" of the substrate between active sites virtually impossible. Despite the structural similarity of the two domains, no significant sequence homology was found when topologically equivalent residues were compared. MATERIALS AND METHODS Crystals of PRAI-IGPS were obtained as described (14,15). They belong to space group P41 with unit cell parameters a = b = 104.7 A, c = 68.0 A, and one molecule per asymmetric unit (67% solvent content). Screening of heavy atom derivatives was done by soaking crystals in a buffered solution of the appropriate heavy atom compound in 3 M ammonium sulfate (Table 1). For potentially useful derivatives, a low-resolution diffractometer data set was collected, while oscillation data sets to 2.5-A resolution were collected for the three best derivatives. One derivative was prepared by diffusing an iodinated form of the substrate analogue N-(5'-phosphoribit-1-yl)anthranilate (rCdRP) into the crystals. 5-Iodo-rCdRP was prepared as described for rCdRP (16) except that 5-iodoanthranilate (Fluka) was used in place of anthranilate as starting material and was dissolved in dioxane instead of in 50% aqueous ethanol.The crystals were mounted in thin-walled glass capillaries. Diffractometer data were collected as described elsewhere (17) with the following modifications: after crystal alignment, a single test reflection was used to determine the direction of the c-axis of the crystal; several crystals were required per data set due to strong radiation sensitivity; a maximum intensity loss of about 30% was accepted; and intensities were measured by c-scan over 0.8-1.5°, depending on mosaic spread.Oscillation data were collected and processed as described elsewhere (18) with the following modifications: a cooling device kept the crystal at about 40C; data were collected over 900; two sheets of film per film pack were used; crystal orientation was checked by still photographs at the beginning and end of data collection; and the films were digitized, using a 3.0 OD scale.The oscillation and diffractometer data sets (Table 1) were not merged at this point. Heavy atom sites were accepted only if found with both sets of data. They were located by difference Patterson and difference Fourier techniqu...