Function of MoaB proteins in the biosynthesis of the molybdenum and tungsten cofactors

Molybdenum (Mo) and tungsten (W) enzymes catalyze important redox reactions in the global carbon, nitrogen, and sulfur cycles. Except in nitrogenases both metals are exclusively associated with a unique metal-binding pterin (MPT) that is synthesized by a conserved multistep biosynthetic pathway, which ends with the insertion and thereby biological activation of the respective element. Although the biosynthesis of Mo cofactors has been intensively studied in various systems, the biogenesis of W-containing enzymes, mostly found in archaea, is poorly understood. Here, we describe the function of the Pyrococcus furiosus MoaB protein that is homologous to bacterial (such as MogA) and eukaryotic proteins (such as Cnx1) involved in the final steps of Mo cofactor synthesis. MoaB reconstituted the function of the homologous Escherichia coli MogA protein and catalyzes the adenylylation of MPT in a Mg2+ and ATP-dependent way. At room temperature reaction velocity was similar to that of the previously described plant Cnx1G domain, but it was increased up to 20-fold at 80 degrees C. Metal and nucleotide specificity for MPT adenylylation is well conserved between W and Mo cofactor synthesis. Thermostability of MoaB is believed to rely on its hexameric structure, whereas homologous mesophilic MogA-related proteins form trimers. Comparison of P. furiosus MoaB to E. coli MoaB and MogA revealed that only MogA is able to catalyze MPT adenylylation, whereas E. coli MoaB is inactive. In summary, MogA, Cnx1G, and MoaB proteins exhibit the same adenylyl transfer activity essential for metal insertion in W or Mo cofactor maturation.     

Modification and processing of internalized signal sequences of prolipoprotein in Escherichia coli and in Bacillus subtilis

We have cloned the Escherichia coli lipoprotein structural gene (lpp) into a shuttle vector and studied its expression in both E. coli and in Bacillus subtilis. Using in vitro gene fusion techniques, the lpp gene was placed under the control of the promoter for the erythromycin-resistance (ery) gene. This fusion gene directed the synthesis of Braun's prolipoprotein which can be subsequently processed into the mature lipoprotein. In addition to the prolipoprotein, two ery-lpp hybrid proteins containing a 45- and a 22-amino acid extension preceding the NH2 terminus of prolipoprotein, respectively, are also synthesized in E. coli. The synthesis of these three proteins appears to involve the utilization of three distinct translation initiation sites. In B. subtilis, only two proteins are synthesized, the hybrid protein with a 45-amino acid extension and the prolipoprotein. In both E. coli and B. subtilis, the precursor forms of the hybrid proteins are lipid-modified, and they are processed to mature lipoprotein in vivo. These results indicate that internalized signal sequence containing the prolipoprotein modification and processing site (Leu-Ala-Glys-Cys) can function normally and permit the modification of hybrid proteins to lipid-modified precursors which can be subsequently processed by the globomycin-sensitive prolipoprotein signal peptidase.     

Molybdenum co-factor biosynthesis: the Arabidopsis thaliana cDNA cnx1 encodes a multifunctional two-domain protein homologous to a mammalian neuroprotein, the insect protein Cinnamon and three Escherichia coli proteins

The molybdenum co-factor (Moco) is an essential part of all eukaryotic molybdoenzymes. It is a molybdopterin and reveals the same principal structure in eubacteria, archaebacteria and eukaryotes. This paper reports the isolation of cnx1 , a cDNA clone of Arabidopsis thaliana which complements the Escherichia coli Moco mutant mogA . The mapping data of this cDNA correlate well with the mapping position of the A. thaliana molybdenum cofactor locus chl6 . As mutants in chl6 are known to be repairable by high concentrations of molybdate, the defective gene is very likely to be involved in the last step of Moco biosynthesis, that is, the insertion of molybdenum into molybdopterin. The protein encoded by cnx1 shows a two-domain structure: the N-terminal domain is homologous to the E. coli Moco protein MoeA, the C-terminal domain is homologous to the E. coli Moco proteins MoaB and MogA, respectively. These homologies show that part of the prokaryotic Moco biosynthetic pathway accomplished by monofunctional proteins in E. coli , is performed by a single multifunctional protein in eukaryotes. In addition Cnx1 is homologous to the eukaryotic proteins Gephyrin, a rat neuroprotein, and Cinnamon, a Drosophila protein with a function in Moco biosynthesis. These proteins also show a two-domain structure but the order of the domains is inversed as compared with Cnx1. Southern analysis indicates the existence of at least one further member, in addition to the cnx1 gene, of this novel gene family in the Arabidopsis genome.     

A wheat germ cell-free system is a novel way to screen protein folding and function

For high-throughput protein structural analysis, it is indispensable to develop a reliable protein overexpression system. Although many protein overexpression systems, such as that involving Escherichia coli cells, have been developed, the number of overexpressed proteins showing the same biological activities as those of the native proteins is limited. A novel wheat germ cell-free protein synthesis system was developed recently, and most of the proteins functioning in solution were synthesized as soluble forms. This suggests the applicability of this protein synthesis method to determination of the solution structures of functional proteins. To examine this possibility, we have synthesized two (15)N-labeled proteins and obtained (1)H-(15)N HSQC spectra for them. The structural analysis of these proteins has already progressed with an E. coli overexpression system, and (1)H-(15)N HSQC spectra for biologically active proteins have already been obtained. Comparing the spectra, we have shown that proteins synthesized with a wheat germ cell-free system have the proper protein folding and enough biological activity. This is the first experimental evidence of the applicability of the wheat germ cell-free protein synthesis system to high-throughput protein structural analysis.     

Feedback regulation of ribosomal protein synthesis in E. coli: Localization of the mRNA target sites for repressor action of ribosomal protein L1

E. coli ribosomal protein L1 is a translational repressor of the synthesis in vitro of both proteins encoded in the L11 operon (L11 and L1). L1 is shown to act at a single target site within the first 160 bases of the bicistronic mRNA, near (or at) the translation initiation site of the L11 cistron. Synthesis of L1 apparently requires translation of the preceding L11 cistron, allowing regulation of the synthesis of both proteins from a single mRNA target site. This observation suggests a sequential translation mechanism that results in the equimolar synthesis rates of the two proteins observed in vivo. It was found that the presence of 23S rRNA, but not 16S rRNA, relieves translational inhibition by L1. L1 presumably recognizes structural features of the mRNA target site that are homologous to the L1-binding site of 23S rRNA. Although previous work indicated that translationally inhibited ribosomal protein mRNA is degraded in vivo, L1 repressor action in the present in vitro system was found not to involve mRNA degradation.     

Autogenous regulation of RNA polymerase beta subunit synthesis in vitro.

The effects of Escherichia coli RNA polymerase and its subassemblies and subunits on the in vitro synthesis of beta subunit directed by DNA from a lambda transducing phage lambdadrif+-6 were investigated. This phage carries the structural gene (rpoB) for beta subunit as well as the genes for EF (translation elongation factor)-Tu, some ribosomal proteins, and stable RNAs of the E. coli chromosome. Among the RNA polymerase proteins examined, the two oligomers, holoenzyme and alpha2beta complex, repressed the synthesis of only the beta subunit but not of other proteins encoded by the phage DNA. The results indicate that the expression of at least the betabeta' (rpoBC) operon is under autogenous regulation, in which both holoenzyme and alpha2beta complex function as regulatory molecules with repressor activity.     

Mutagenesis of folylpolyglutamate synthetase indicates that dihydropteroate and tetrahydrofolate bind to the same site

The folylpolyglutamate synthetase (FPGS) enzyme of Escherichia coli differs from that of Lactobacillus casei in having dihydrofolate synthetase activity, which catalyzes the production of dihydrofolate from dihydropteroate. The present study undertook mutagenesis to identify structural elements that are directly responsible for the functional differences between the two enzymes. The amino terminal domain (residues 1-287) of the E. coli FPGS was found to bind tetrahydrofolate and dihydropteroate with the same affinity as the intact enzyme. The domain-swap chimera proteins between the E. coli and the L. casei enzymes possess both folate or pteroate binding properties and enzymatic activities of their amino terminal portion, suggesting that the N-terminal domain determines the folate substrate specificity. Recent structural studies have identified two unique folate binding sites, the omega loop in L. casei FPGS and the dihydropteroate binding loop in the E. coli enzyme. Mutants with swapped omega loops retained the activities and folate or pteroate binding properties of the rest of the enzyme. Mutating L. casei FPGS to contain an E. coli FPGS dihydropteroate binding loop did not alter its substrate specificity to using dihydropteroate as a substrate. The mutant D154A, a residue specific for the dihydropteroate binding site in E. coli FPGS, and D151A, the corresponding mutant in the L. casei enzyme, were both defective in using tetrahydrofolate as their substrate, suggesting that the binding site corresponding to the E. coli pteroate binding site is also the tetrahydrofolate binding site for both enzymes. Tetrahydrofolate diglutamate was a slightly less effective substrate than the monoglutamate with the wild-type enzyme but was a 40-fold more effective substrate with the D151A mutant. This suggests that the 5,10-methylenetetrahydrofolate binding site identified in the L. casei ternary structure may bind diglutamate and polyglutamate folate derivatives.     

Solution structure and backbone dynamics of the cysteine 103 to serine mutant of the N-terminal domain of DsbD from Neisseria meningitidis

The DsbD protein is essential for electron transfer from the cytoplasm to the periplasm of Gram-negative bacteria. Its N-terminal domain dispatches electrons coming from cytoplasmic thioredoxin (Trx), via its central transmembrane and C-terminal domains, to its periplasmic partners: DsbC, DsbE/CcmG, and DsbG. Previous structural studies described the latter proteins as Trx-like folds possessing a characteristic C-X-X-C motif able to generate a disulfide bond upon oxidation. The Escherichia coli nDsbD displays an immunoglobulin-like fold in which two cysteine residues (Cys103 and Cys109) allow a disulfide bond exchange with its biological partners.We have determined the structure in solution and the backbone dynamics of the C103S mutant of the N-terminal domain of DsbD from Neisseria meningitidis. Our results highlight significant structural changes concerning the beta-sheets and the local topology of the active site compared with the oxidized form of the E. coli nDsbD. The structure reveals a "cap loop" covering the active site, similar to the oxidized E. coli nDsbD X-ray structure. However, regions featuring enhanced mobility were observed both near to and distant from the active site, revealing a capacity of structural adjustments in the active site and in putative interaction areas with nDsbD biological partners. Results are discussed in terms of functional consequences.     

Crystal structure of AGR_C_4470p from Agrobacterium tumefaciens

We report here the crystal structure at 2.0 A resolution of the AGR_C_4470p protein from the Gram-negative bacterium Agrobacterium tumefaciens. The protein is a tightly associated dimer, each subunit of which bears strong structural homology with the two domains of the heme utilization protein ChuS from Escherichia coli and HemS from Yersinia enterocolitica. Remarkably, the organization of the AGR_C_4470p dimer is the same as that of the two domains in ChuS and HemS, providing structural evidence that these two proteins evolved by gene duplication. However, the binding site for heme, while conserved in HemS and ChuS, is not conserved in AGR_C_4470p, suggesting that it probably has a different function. This is supported by the presence of two homologs of AGR_C_4470p in E. coli, in addition to the ChuS protein.     

Crystal structures of YBHB and YBCL from Escherichia coli, two bacterial homologues to a Raf kinase inhibitor protein.

In rat and human cells, RKIP (previously known as PEBP) was characterized as an inhibitor of the MEK phosphorylation by Raf-1. In Escherichia coli, the genes ybhb and ybcl possibly encode two RKIP homologues while in the genomes of other bacteria and archaebacteria other homologous genes of RKIP have been found. The parallel between the cellular signaling mechanisms in eukaryotes and prokaryotes suggests that these bacterial proteins could be involved in the regulation of protein phosphorylation by kinases as well. We first showed that the proteins YBHB and YBCL were present in the cytoplasm and periplasm of E. coli, respectively, after which we determined their crystallographic structures. These structures verify that YBHB and YBCL belong to the same structural family as mammalian RKIP/PEBP proteins. The general fold and the anion binding site of these proteins are extremely well conserved between mammals and bacteria and suggest functional similarities. However, the bacterial proteins also exhibit some specific structural features, like a substrate binding pocket formed by the dimerization interface and the absence of cis peptide bonds. This structural variety should correspond to the recognition of multiple cellular partners.     

Piecing together the structure-function puzzle: experiences in structure-based functional annotation of hypothetical proteins

The combination of genomic sequencing with structural genomics has provided a wealth of new structures for previously uncharacterized ORFs, more commonly referred to as hypothetical proteins. This rapid growth has been the direct result of high-throughput, automated approaches in both the identification of new ORFs and the determination of high-resolution 3-D protein structures. A significant bottleneck is reached, however, at the stage of functional annotation in that the assignment of function is not readily automatable. It is often the case that the initial structural analysis at best indicates a functional family for a given hypothetical protein, but further identification of a relevant ligand or substrate is impeded by the diversity of function in a particular structural classification of proteins family, a highly selective and specific ligand-binding site, or the identification of a novel protein fold. Our approach to the functional annotation of hypothetical proteins relies on the combination of structural information with additional bioinformatics evidence garnered from operon prediction, loose functional information of additional operon members, conservation of catalytic residues, as well as cocrystallization trials and virtual ligand screening. The synthesis of all available information for each protein has permitted the functional annotation of several hypothetical proteins from Escherichia coli and each assignment has been confirmed through generally accepted biochemical methods.     

The crystal structure of Escherichia coli MoeA, a protein from the molybdopterin synthesis pathway.

MoeA is involved in synthesis of the molybdopterin cofactor, although its function is not yet clearly defined. The three-dimensional structure of the Escherichia coli protein was solved at 2.2 A resolution. The locations of highly conserved residues among the prokaryotic and eukaryotic MoeA homologs identifies a cleft in the dimer interface as the likely functional site. Of the four domains of MoeA, domain 2 displays a novel fold and domains 1 and 4 each have only one known structural homolog. Domain 3, in contrast, is structurally similar to many other proteins. The protein that resembles domain 3 most closely is MogA, another protein required for molybdopterin cofactor synthesis. The overall similarity between MoeA and MogA, and the similarities in a constellation of residues that are strongly conserved in MoeA, suggests that these proteins bind similar ligands or substrates and may have similar functions.     

Cytokinesis in bacteria.

Work on two diverse rod-shaped bacteria, Escherichia coli and Bacillus subtilis, has defined a set of about 10 conserved proteins that are important for cell division in a wide range of eubacteria. These proteins are directed to the division site by the combination of two negative regulatory systems. Nucleoid occlusion is a poorly understood mechanism whereby the nucleoid prevents division in the cylindrical part of the cell, until chromosome segregation has occurred near midcell. The Min proteins prevent division in the nucleoid-free spaces near the cell poles in a manner that is beginning to be understood in cytological and biochemical terms. The hierarchy whereby the essential division proteins assemble at the midcell division site has been worked out for both E. coli and B. subtilis. They can be divided into essentially three classes depending on their position in the hierarchy and, to a certain extent, their subcellular localization. FtsZ is a cytosolic tubulin-like protein that polymerizes into an oligomeric structure that forms the initial ring at midcell. FtsA is another cytosolic protein that is related to actin, but its precise function is unclear. The cytoplasmic proteins are linked to the membrane by putative membrane anchor proteins, such as ZipA of E. coli and possibly EzrA of B. subtilis, which have a single membrane span but a cytoplasmic C-terminal domain. The remaining proteins are either integral membrane proteins or transmembrane proteins with their major domains outside the cell. The functions of most of these proteins are unclear with the exception of at least one penicillin-binding protein, which catalyzes a key step in cell wall synthesis in the division septum.     

Construction and characterization of hybrid plasmids containing the Escherichia coli nrd region.

Recombinant plasmids containing all or part of the genetic region of Escherichia coli coding for the two subunits of ribonucleoside diphosphate reductase (proteins B1 and B2) were constructed with the aid of the multicopy plasmid pBR322. Two of these plasmids (pPS1 and pPS2) appeared to carry both a regulator and the complete structural information for the enzyme and, after transformation of E. coli, directed a 10- to 20-fold overproduction of both proteins B1 and B2. The other plasmids (pPS101 and pPS201) carried structural information for only protein B2. Cells carrying pPS1 and pPS2 showed a 5- to 500-fold increased resistance against the drug hydroxyurea. This establishes that in E. coli the inhibition of deoxyribonucleic acid synthesis by hydroxyurea is fully explained by its action on ribonucleotide reductase.     

UDPglucose pyrophosphorylase from Xanthomonas spp. Characterization of the enzyme kinetics, structure and inactivation related to oligomeric dissociation

The genes encoding for UDPglucose pyrophosphorylase in two Xanthomonas spp. were cloned and overexpressed in Escherichia coli. After purification to electrophoretic homogeneity, the recombinant proteins were characterized, and both exhibited similar structural and kinetic properties. They were identified as dimeric proteins of molecular mass 60kDa, exhibiting relatively high specific activity ( approximately 80Units/mg) for UDPglucose synthesis. Both enzymes utilized UTP or TTP as substrate with similar affinity. The purified Xanthomonas enzyme was inactivated after dilution into the assay medium. Studies of crosslinking with the bifunctional lysyl reagent bisuberate suggest that inactivation occurs by enzyme dissociation to monomers. UTP effectively protects the enzyme against inactivation, from which a dissociation constant of 15microM was calculated for the interaction substrate-enzyme. The UTP binding to the enzyme would induce conformational changes in the protein, favoring the subunits interaction to form an active dimer. This view was reinforced by protein modeling of the Xanthomonas enzyme on the basis of the prokaryotic UDPglucose pyrophosphorylase crystallographic structure. The in silico approach pointed out two main critical regions in the enzyme involved in subunit-subunit interaction: the region surrounding the catalytic-substrate binding site and the C-term.     

Initiation of DNA synthesis on the isolated strands of bacteriophage f1 replicative-form DNA.

Viral and complementary strand circular DNA molecules were isolated from intracellular bacteriophage f1 replicative-form DNA. Soluble protein extracts of Escherichia coli were used to examine the initiation of DNA synthesis on these DNA templates. The initiation of DNA synthesis on f1 viral strand DNA was catalyzed by E. coli DNA-dependent RNA polymerase, as was initiation of f1 viral strand DNA isolated from mature phage particles. The site of initiation was the same as that used in vivo. In contrast, no de novo initiation of DNA synthesis was detected on f1 complementary strand DNA. Control experiments demonstrated that the E. coli dnaB, dnaC, and dnaG initiation proteins were active under the conditions employed. The results suggest that the viral strand of the f1 replicative-form DNA molecule carries the same DNA synthesis initiation site as the viral strand packaged in mature phage, whereas the complementary strand of the replicative-form DNA molecule carries no site for de novo primer synthesis. These in vitro observations are consistent with the simple rolling circle model for f1 DNA replication in vivo proposed by Horiuchi and Zinder.     

Differences between the manganese- and the iron-containing superoxide dismutases of Escherichia coli detected through sedimentation equilibrium, hydrodynamic, and spectroscopic studies

The genome of Escherichia coli codes for two superoxide dismutases that may contain either iron (FeSOD) or manganese (MnSOD) at the active site. The crystal structures of MnSODs from two bacterial sources (but not E. coli) have been completed, and structural comparisons with the crystal structure of the FeSOD from either E. coli or Pseudomonas ovalis have been made. Despite the low degree (less than 50%) of sequence homology between the E. coli enzymes, the two proteins are suggested to be structurally homologous. Nonetheless, these enzymes exhibit absolute metal cofactor specificity in conferring enzymatic activity to the inactive apoenzyme. This observation is surprising considering the identity of the active site ligands and the similarities in their geometry and surrounding environment. Using analytical ultracentrifugation, we have determined that the solution properties of these two proteins are different. Thus dialysis of FeSOD but not of MnSOD against phosphate buffer in the presence or absence of EDTA caused dissociation of the homodimer. This dissociation appeared to be related to the loss of iron from native FeSOD. Thus, apoFeSOD but not apoMnSOD existed predominantly as a monomer at protein concentrations below 150 micrograms/mL. ApoMnSOD showed no evidence for dissociation under these conditions. Fluorescence data suggest that the tryptophan environments for the two enzymes are also different. The results of these physical measurements lead us to propose that subtle differences, perhaps at the subunit contact faces, exist in the structures of these crystallographically similar proteins.     

Unexpected sequence similarity between nucleosidases and phosphoribosyltransferases of different specificity.

Amino acid sequences of enzymes that catalyze hydrolysis or phosphorolysis of the N-glycosidic bond in nucleosides and nucleotides (nucleosidases and phosphoribosyltransferases) were explored using computer methods for database similarity search and multiple alignment. Two new families, each including bacterial and eukaryotic enzymes, were identified. Family I consists of Escherichia coli AMP hydrolase (Amn), uridine phosphorylase (Udp), purine phosphorylase (DeoD), uncharacterized proteins from E. coli and Bacteroides uniformis, and, unexpectedly, a group of plant stress-inducible proteins. It is hypothesized that these plant proteins have evolved from nucleosidases and may possess nucleosidase activity. The proteins in this new family contain 3 conserved motifs, one of which was found also in eukaryotic purine nucleosidases, where it corresponds to the nucleoside-binding site. Family II is comprised of bacterial and eukaryotic thymidine phosphorylases and anthranilate phosphoribosyltransferases, the relationship between which has not been suspected previously. Based on the known tertiary structure of E. coli thymidine phosphorylase, structural interpretation was given to the sequence conservation in this family. The highest conservation is observed in the N-terminal alpha-helical domain, whose exact function is not known. Parts of the conserved active site of thymidine phosphorylases and anthranilate phosphoribosyltransferases were delineated. A motif in the putative phosphate-binding site is conserved in family II and in other phosphoribosyltransferases. Our analysis suggests that certain enzymes of very similar specificity, e.g., uridine and thymidine phosphorylases, could have evolved independently. In contrast, enzymes catalyzing such different reactions as AMP hydrolysis and uridine phosphorolysis or thymidine phosphorolysis and phosphoribosyl anthranilate synthesis are likely to have evolved from common ancestors.