Search

IMP dehydrogenase (IMPDH) is an essential enzyme that catalyzes the first step unique to GTP synthesis. To provide a basis for the evaluation of IMPDH inhibitors as antimicrobial agents, we have expressed and characterized IMPDH from the pathogenic bacterium Streptococcus pyogenes. Our results show that the biochemical and kinetic characteristics of S. pyogenes IMPDH are similar to other bacterial IMPDH enzymes. However, the lack of sensitivity to mycophenolic acid and the Km for NAD (1180 microM) exemplify some of the differences between the bacterial and mammalian IMPDH enzymes, making it an attractive target for antimicrobial agents. To evaluate the basis for these differences, we determined the crystal structure of the bacterial enzyme at 1.9 A with substrate bound in the catalytic site. The structure was determined using selenomethionine-substituted protein and multiwavelength anomalous (MAD) analysis of data obtained with synchrotron radiation from the undulator beamline (19ID) of the Structural Biology Center at Argonne's Advanced Photon Source. S. pyogenes IMPDH is a tetramer with its four subunits related by a crystallographic 4-fold axis. The protein is composed of two domains: a TIM barrel domain that embodies the catalytic framework and a cystathione beta-synthase (CBS) dimer domain of so far unknown function. Using information provided by sequence alignments and the crystal structure, we prepared several site-specific mutants to examine the role of various active site regions in catalysis. These variants implicate the active site flap as an essential catalytic element and indicate there are significant differences in the catalytic environment of bacterial and mammalian IMPDH enzymes. Comparison of the structure of bacterial IMPDH with the known partial structures from eukaryotic organisms will provide an explanation of their distinct properties and contribute to the design of specific bacterial IMPDH inhibitors.

Past difficulty in growing good crystals of cholera toxin has prevented the study of the crystal structure of this important protein. We have determined that failure of cholera toxin to crystallize well has been due to its heterogeneity. We have now succeeded in overcoming the problem by isolating a single isoelectric variant of this oligomeric protein (one A subunit and five B subunits). Cholera toxin purified by our procedure readily forms large single crystals. The crystal form (space group P2(1), a = 73.0 A, b = 92.2 A, c = 60.6 A, beta = 106.4 degrees, one molecule in the asymmetric unit) has been described previously [Sigler et al. (1977) Science (Washington, D.C.) 197, 1277-1278]. We have recorded data from native crystals of cholera toxin to 3.0-A resolution with our electronic area detectors. With these data, we have found the orientation of a 5-fold symmetry axis within these crystals, perpendicular to the screw dyad of the crystal. We are now determining the crystal structure of cholera toxin by a combination of multiple heavy-atom isomorphous replacement and density modification techniques, making use of rotational 5-fold averaging of the B subunits.

We have shown by kinetic and magnetic resonance measurements that a spin-labeled substrate analogue, spiro[doxyl-2,3'-5' alpha-androstan]-17'beta-ol, binds at the substrate site of crystalline delta 5-3-ketosteroid isomerase (steroid delta-isomerase; EC 5.3.3.1) of Pseudomonas testosteroni. The spin-labeled steroid is a linear competitive inhibitor with a Ki value (25 +/- 5 microM) that is consistent with dissociation constants obtained by direct binding measurements based on changes in the electron paramagnetic resonance spectrum of the nitroxide, longitudinal relaxation rates of water protons, and longitudinal and transverse relaxation rates of carbon-bound protons of the isomerase. These binding studies yield a stoichiometry for the nitroxide of 1 per subunit of the enzyme. Measurements of the longitudinal relaxation rates of water protons indicate that the 3-doxyl portion of the spin-label is highly immobilized yet is exposed to solvent. Paramagnetic effects of the nitroxide on T1 defined distances to several previously assigned [Benisek, W. F., & Ogez, J. R. (1982) Biochemistry 21, 5816-5825] and newly assigned protons of the enzyme. These distances were then used to locate (with an accuracy of +/- 2 A) the nitroxide moiety at a unique position in a partially refined 2.5-A resolution X-ray structure of native isomerase. Three of five additional proton resonance peaks, attributed to ring-shielded methyl groups, could be assigned to specific residues on the basis of distances from the spin-label in the X-ray structure. The remaining portion of the spin-labeled steroid was then docked into the X-ray structure in a hydrophobic cavity of the enzyme. This position of the steroid is consistent with the steroid binding site previously proposed [Westbrook, E. M., Piro, O. E., & Sigler, P. B. (1984) J. Biol. Chem. 259, 9096-9103]. However, the rotational orientation of this steroid about its long axis could not be unambiguously established. If we assume that steroid substrates and the spin-labeled inhibitor bind to the same site, but with reversal of the 3- and 17-positions, then the phenolic hydroxyl of Tyr-55 is optimally positioned to function as the general acid that protonates the 3-keto group of the substrate, facilitated by the negative end of the dipole of a 10-residue alpha-helix, the only helix in the molecule.(ABSTRACT TRUNCATED AT 400 WORDS)