Search

The essential bacterial cell division protein FtsZ (filamentation temperature-sensitive protein Z) is a distant homologue to the eukaryotic cytoskeletal protein tubulin. We have examined the GTP hydrolytic activity of Escherichia coli FtsZ using a real-time fluorescence assay that monitors phosphate production. The GTPase activity shows a dramatic, nonlinear dependence on FtsZ concentration, with activity only observed at enzyme concentrations greater than 1 microM. At 5 microM FtsZ, we have determined a K(m) of 82 microM GTP and a V(max) of 490 nmol of P(i) min(-1) (mg of protein)(-1). Hydrolysis of GTP requires Mg(2+) and other divalent cations substitute only poorly for this requirement. We have compared the concentration dependence of FtsZ GTPase activity with the oligomeric state by use of analytical ultracentrifugation and chemical cross-linking. Equilibrium analytical ultracentrifugation experiments show that FtsZ exists as 68% dimer and 13% trimer at 2 microM total protein concentration. Chemical cross-linking of FtsZ also shows that monomer, dimer, trimer, and tetramer species are present at higher (>2 microM) FtsZ concentrations. However, as shown by analytical centrifugation, GDP-bound FtsZ is significantly shifted to the monomeric state, which suggests that GTP hydrolysis regulates polymer destabilization. We also monitored the effect of nucleotide and metal ion on the secondary structure of FtsZ; nucleotide yielded no evidence of structural changes in FtsZ, but both Mg(2+) and Ca(2+) had significant effects on secondary structure. Taken together, our results support the hypothesis that Mg(2+)-dependent GTP hydrolysis by FtsZ requires oligomerization of FtsZ. On the basis of these results and structural comparisons with the alpha-beta tubulin dimer, GTP is likely hydrolyzed in a shared active site formed between two monomer subunits.

Rat liver arginase contains a dimanganese(II,II) center per subunit that is required for catalytic hydrolysis of l-arginine to form urea and l-ornithine. A recent crystallographic study has shown that the Mn2 center consists of two coordinatively inequivalent manganese(II) ions, MnA and MnB, bridged by a water (hydroxide) molecule and two aspartate residues [Kanyo et al. (1996) Nature 383, 554-557]. A conserved residue, His141, is located near the proposed substrate binding region at 4.2 A from the bridging solvent molecule. The present EPR studies reveal that there is no essential alteration of the Mn2 site upon mutation of His141 to an Asn residue, which lacks a potential acid/base residue, while the catalytic activity of the mutant enzyme is 10 times lower vs wild-type enzyme. The binding affinity of l-lysine, l-arginine (substrate), and Nomega-OH-l-arginine (type 2 binders) increases inversely with the pKa of the side chain. Binding of l-lysine is more than 10 times weaker, and the substrate Michaelis constant (Km) is >6-fold greater (weaker binding) in the His141Asn mutant than in wild-type arginase. L-Lysine and Nomega-OH-L-arginine, type 2 binders, induce extensive loss of the EPR intensity, suggesting direct coordination to the Mn2 center. From these data and the pH dependence of type 2 binders, we conclude that His141 functions as the base for deprotonation of the side-chain amino group of L-lysine and the substrate guanidinium group, -NH-C(NH2)2+ and that the unprotonated side chain of these amino acids is responsible for binding to the active site. A different class of inhibitors (type 1), including L-isoleucine, L-ornithine, and L-citrulline, suppresses enzymatic activity, producing only minor change in the zero-field splitting of the Mn2 EPR signal and no change in the EPR intensity, suggestive of minimal conformational transformation. We propose that type 1 alpha-amino acid inhibitors do not bind directly to either Mn ion, but interact with the recognition site on arginase for the alpha-aminocarboxylate groups of the substrate. A new mechanism for the arginase-catalyzed hydrolysis of L-arginine is proposed which has general relevance to all binuclear hydrolases: (1) Deprotonation of substrate l-arginine(H+) by His141 permits entry of the neutral guanidinium group into the buried Mn2 region. Binding of the substrate imino group (>C=NH), most likely to MnB, is coupled to breaking of the MnB-(mu-H2O) bond, forming a terminal aquo ligand on MnA. (2) Proton transfer from the terminal MnA-aqua ligand to the substrate Ndelta-guanidino atom forms the nucleophilic hydroxide on MnA and the cationic NdeltaH2+-guanidino leaving group. Protonation of the substrate -NdeltaH2+-group is likely assisted by hydrogen bonding to the juxtaposed anionic carboxylate group of Glu277. (3) Attack of the MnA-bound hydroxide at the electrophilic guanidino C-atom forms a tetrahedral intermediate. (4) Formation of products is initiated by cleavage of the Cepsilon-NdeltaH2+ bond, yielding urea and L-ornithine(H+).

The solution structures of the binuclear Mn centers in arginase, Mn catalase, and the Mn-substituted forms of the Fe enzymes ribonucleotide reductase and hemerythrin have been determined using X-ray absorption spectroscopy (XAS). X-ray absorption near edge structure (XANES) spectra for these proteins were compared to those obtained for Mn(II) models. The Mn model spectra show an inverse correlation between the XANES peak maximum and the root-mean-square (RMS) deviation in metal-ligand bond lengths. For these complexes, the XANES maxima appear to be more effective than the 1s --> 3d areas as an indicator of metal-site symmetry. Arginase and Mn-substituted ribonucleotide reductase have symmetric nearest neighbor environments with low RMS deviation in bond length, while Mn catalase and Mn-substituted hemerythrin appear to have a larger RMS bond length deviation. The 1s --> 3d areas for arginase and Mn-substituted ribonucleotide reductase are consistent with six coordinate Mn, while the 1s --> 3d areas for Mn catalase and Mn-substituted hemerythrin are larger, suggesting that one or both of the Mn ions are five-coordinate in these proteins. Extended x-ray absorption fine structure (EXAFS) spectra were used to determine the Mn2 core structure for the four proteins. In order to quantitate the number of histidine residues bound to the Mn2 centers, EXAFS data for the crystallographically characterized model hexakis-imidazole Mn(II) dichloride tetrahydrate were used to calibrate the Mn-imidazole multiple scattering interactions. These calibrated parameters allowed the outer shell EXAFS to be fit to give a lower limit on the number of bound histidine residues. The EXAFS spectra for Mn-substituted ribonucleotide reductase and arginase are nearly identical, with symmetric Mn-nearest neighbor environments and outer shell scattering consistent with a lower limit of one histidine per Mn2 core. In contrast, the EXAFS data for Mn catalase and Mn-substituted hemerythrin show two distinct Mn-nearest neighbor shells, modeled as Mn-O at ca. 2.1 A and Mn-N at ca. 2.3 A, and outer shell carbon scattering consistent with a lower limit of ca. 2-3 His residues per Mn2 core. Only Mn catalase shows clear evidence for Mn...Mn scattering. The observed Mn...Mn distance is 3.53 A, which is significantly longer than the approximately 3.3 A distances that are typically observed for Mn(II)2 cores with two single atom bridges, but which is typical of the distances seen in Mn(II)2 cores having one single atom bridge (e.g., aqua or hydroxo) together with one or two carboxylate bridges. The absence of EXAFS-detectable Mn...Mn interactions for the other three proteins suggests either that there are no single atom bridges in these cases or that the Mn...Mn interactions are more disordered.