INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

UDP-N-acetylglucosamine enolpyruvyl transferase (MurA) catalyzes the first committed step in bacterial cell wall biosynthesis (9, 20). The enzyme transfers an enolpyruvyl group from phosphoenolpyruvate (PEP) to UDP-N-acetylglucosamine (UDPAG) to form UDP-N-acetylglucosamine enolpyruvate (Fig. 1A). This is a precursor to UDP-N-acetylmuramate, an essential building block for the bacterial cell wall. MurA is inhibited by the antibiotic fosfomycin (Fig. 1B) (10), and because of its importance in peptidoglycan biosynthesis, it is of interest as a target for the design of novel antibacterial agents.

View larger version (12K): [in this window] [in a new window]   FIG. 1.   (A) The enolpyruvyl transfer reaction catalyzed by MurA. (B) Inactivation of MurA by fosfomycin as a result of the covalent linkage between Cys-115 of MurA and fosfomycin.

MurA (sometimes called MurZ) from Escherichia coli has been studied in great depth. There is one copy of the murA gene in E. coli (16), and this has been shown to be essential by gene deletion experiments (3). The E. coli murA gene has been overexpressed, and the MurA enzyme has been purified and kinetically characterized (16). The K_ms of UDPAG and PEP were reported to be 15 and 0.4 µM, respectively, while the k_cat was 3.8 s¹ (11). The crystal structure of the enzyme, either ligand free or complexed with UDPAG and fosfomycin, has been solved (21, 22). This shows the enzyme to be made up of two domains, with the active site located in a deep cavity between them, and identifies catalytically and structurally important residues. One catalytically important residue (Cys-115) has been investigated in detail. It was shown that this residue is the site of alkylation by fosfomycin, a finding that has been confirmed by kinetics and mutagenesis studies (Fig. 1B) (15, 27). Mutations in the cysteine residue (C115S) inactivated the enzyme, confirming that this is an important residue in catalysis (27). Another residue (Asp-305) is indicated by the structural data as being the general base involved in the deprotonation of the substrate UDPAG (22).

The importance of MurA as an antibacterial target led us to investigate the MurA enzymes from gram-positive pathogens. While there is only one murA gene in E. coli, we have identified in gram-positive bacteria two murA genes, which have arisen by a possible gene duplication. To investigate if the two genes present in the gram-positive bacteria both encode active enzymes, we performed an in-depth study of the two genes from the gram-positive bacterium Streptococcus pneumoniae. Gene deletion experiments were used to investigate the essentiality of the murA genes. The genes were then overexpressed, the enzymes that they encode were purified, and the biochemical activities of the enzymes were investigated.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

General materials. S. pneumoniae R6 is a nonencapsulated strain (13). pJS3 (2) was the kind gift of Sanford Lacks (Brookhaven National Laboratory). Biochemical reagents were from Sigma. Enzymes and reagents used in cloning were from Gibco BRL (Gaithersburg, Md.). DNA preparations were performed with assorted kits from Qiagen (Valencia, Calif.). Column chromatography was performed on an AKTA FPLC system (Pharmacia) with Pharmacia Hi-Load columns. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analyses were performed on Novex (San Diego, Ca) 4 to 12% bis-Tris gels. The N-terminal amino acid sequence of the purified enzyme was deduced by automated Edman degradation with a Hewlett-Packard model G1000A sequencer. The molecular weight of the enzyme was determined by matrix-assisted laser-desorption ionization-time of flight (MALDI-TOF) mass spectrometry using a Voyager RP Biospectrometer (PerSeptive Biosystems, Framingham, Mass.).

Phylogenetic analysis. Homologous protein sequences were retrieved from public and proprietary genomic sequence databases using the software BLASTP and TBLASTN (1). The partial genome of S. pneumoniae type 4 from The Institute for Genome Research was used as the source of MurA1 and MurA2 described here. The proteins were aligned using the program CLUSTALW version 1.7 (23) with the BLOSUM62 (8) similarity matrix and gap opening and extension penalties of 10.0 and 0.05, respectively. The multiple sequence alignments were refined manually using the program SEQLAB of the Wisconsin Package, version 9.0 (Genetics Computer Group, Madison, Wis.).

Generation of S. pneumoniae allelic replacement mutants. S. pneumoniae 0100993 chromosomal DNA was prepared as follows. Four milliliters of S. pneumoniae culture with an optical density at 650 nm of 0.6 to 0.7 was spun down, resuspended in 100 µl of 1 M Tris-HCl (pH 8)-0.25 M EDTA-5% sodium deoxycholate-40% glucose, and incubated at 37°C for 20 min; 10 µl of 10% (wt/vol) SDS was added, and the suspension was incubated at 37°C for a further 10 min. The mixture was phenol-chloroform extracted three times, followed by a single chloroform extraction. The DNA was precipitated with isopropanol, washed twice in 70% ethanol, and resuspended in water. Chromosomal DNA fragments flanking murA1 and murA2 were PCR amplified from chromosomal DNA and purified as previously described (19). These were then used to make antibiotic-resistant gene-targeting constructs as described in Results, which were used to transform S. pneumoniae R6 competent cells prepared according to published protocols. A total of 10⁶ cells were incubated with 500 ng of allelic replacement cassette at 30°C for 30 min and then transferred to 37°C for 90 min to allow expression of antibiotic resistance (7). Cells were plated in AGCH medium (12) containing erythromycin (1 µg/ml) or chloramphenicol (2.5 µg/ml) as appropriate and incubated at 37°C for 36 h in 5% CO₂. Transformation efficiencies were 10³ transformants per µg of control DNA.

Whole-cell antimicrobial activity. Whole-cell antimicrobial activity was determined by broth microdilution. Fosfomycin was dissolved in dimethyl sulfoxide and diluted 1:10 in water to produce a 256-µg/ml stock solution. Using a 96-well microtiter plate, a Microlab AT Plus 2 (Hamilton Co., Reno, Nev.) serially diluted 50 µl of the stock solution into cation-adjusted Mueller-Hinton broth (Becton Dickinson, Cockeysville, Md.). After the fosfomycin was diluted, a 50-µl aliquot of the test isolate (~10⁶ CFU/ml) was added to each well of the microtiter plate. The final test concentrations ranged from 0.06 to 64 µg/ml. Inoculated plates were incubated at 35°C in ambient air for 18 to 24 h. The MIC was determined as the lowest concentration of fosfomycin that inhibited visible growth.

Cloning and overexpression. The murA1 and murA2 genes were amplified by PCR from the genomic DNA of S. pneumoniae 100993. The murA1 gene (1,281 bp) was amplified in two fragments: fragment A has NcoI and EcoRI restriction ends and was amplified using the oligonucleotides 5'-GGAAACGACCATGGATAAAATTGTGGTTCAAGGTGG-3' and 5'-CGCCTTTTGCAACTGTCATCAAGGCTG-3'. Fragment B has EcoRI and HindIII restriction ends and was amplified using the oligonucleotides 5'-GGGTGTTGAAGTAATTGAAGAAGACG-3' and 5'-GACCGATTCCTAGAGCCAGAGTACC-3'. Following the restriction digestion, the fragments were ligated into pET28a(+) (Novagen, Madison, Wis.), and the new construct was used to transform E. coli BL21(DE3) competent cells (Stratagene, La Jolla, Calif.). The expression of murA1 was induced with 1 mM isopropylthio--D-galactoside (IPTG) at 18°C for 24 h.

Enzyme purification. E. coli BL21(DE3) cells carrying the plasmid containing the murA1 or murA2 gene were grown, respectively, in 1 or 4 liters of LB medium containing kanamycin (0.1 mg/ml) and 2% (wt/vol) glucose. Gene expression was induced using the induction conditions described above. The harvested cells were resuspended in 100 mM Tris (pH 8.0)-5 mM dithiothreitol (DTT) and disrupted by sonication. For the MurA1 enzyme, the cell lysate was fractionated by ammonium sulfate precipitation (20, 40, and 70%), and the 70% ammonium sulfate supernatant was further purified on a Hi-Load (16/10) Q Sepharose column, equilibrated in 100 mM Tris (pH 8.0)-5 mM DTT. Protein was eluted with a gradient of 0 to 1 M KCl over 200 ml. For the MurA2 enzyme, the cell lysate was brought to 1.5 M ammonium sulfate, and the supernatant of this solution was purified on a Hi-Load (16/10) phenyl-Sepharose column, equilibrated in 100 mM Tris (pH 8)-5 mM DTT-1.5 M ammonium sulfate. Protein was eluted with a gradient of 1.5 to 0 M ammonium sulfate over 200 ml. MurA2 was further purified on a Hi-Load (16/10) Q Sepharose column under the same conditions as used for MurA1. MurA1 and MurA2 were identified by SDS-PAGE analyses and activity assays throughout the purification.

Enzyme assays. The activity of MurA1 and MurA2 was assayed by measuring the release of P_i from the UDPAG and PEP reaction, using a malachite green assay (14) in 50 mM HEPES (pH 7.5) at room temperature. A typical assay of 200 µl contained substrates and enzyme (MurA1 at 200 nM or MurA2 at 50 nM). For kinetic measurements, the assay was performed with a five-by-five array of various substrates on a half-area 96-well microtiter plate (Costar 3696; Corning Inc., Corning, N.Y.) with a SpectraMax Plus plate reader (Molecular Devices Corp., Sunnyvale, Calif.). Actual concentrations of the substrates are listed in the footnotes to Table 1. Enzyme activity during purification was assayed with both UDPAG and PEP fixed at 1 mM.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Sequence and structural comparisons of UDP-N-acetylglucosamine enolpyruvyl transferases. Database searches revealed that the gene encoding UDP-N-acetylglucosamine enolpyruvyl transferase exists widely in nearly all bacteria. However, in low-G+C gram-positive bacteria (Bacillus subtilis, S. pneumoniae, S. pyogenes, Staphylococcus aureus, and Clostridium acetobutylicum), two complete genes encoding UDP-N-acetylglucosamine enolpyruvyl transferase were found (Fig. 2). One of these genes (murA1) was more closely related to the murA genes found in gram-negative bacteria than the second gene copy (murA2). The extent of identical amino acids among MurA1 proteins ranged from 32 to 96% (calculated over ungapped alignments), while that for MurA2 homologs ranged from 53 to 82%. The genus Mycoplasma lacked any open reading frames (ORFs) with significant homology to UDP-N-acetylglucosamine enolpyruvyl transferase, as would be expected for an organism that contains no peptidoglycan. Both genomes were thoroughly searched at the protein (BLASTP) and nucleotide (TBLASTN) levels using B. subtilis MurA1 and MurA2 (1).

View larger version (46K): [in this window] [in a new window]   FIG. 2.   Phylogenetic tree showing divergence of UDP-N-acetylglucosamine enolpyruvyl transferase isoforms (MurA1 and MurA2). Tree was constructed by the NJ method as implemented by the program NEIGHBOR of the PHYLIP 3.57c package (see Materials and Methods). The scale bar represents 0.1 expected amino acid replacements per site as estimated by the program PROTDIST using the Dayhoff PAM substitution matrix. Numbers at the branching points represent the percent occurrence in 1,000 ML puzzling steps and 1,000 random bootstrap replications of MP and NJ methods, respectively. ML and MP analyses were done with the programs PUZZLE version 4.0 (23) and PAUP* (24), respectively. Nodes where values by all three methods were 90% or greater are labeled with an asterisk. Values less than 50% are not shown or are indicated by a dash. Major taxonomic groups of bacteria (18) are indicated with abbreviations given for gram-positive bacteria (GP) and Cytophaga-Flexibacter-Bacteroides (CFB) group.

View larger version (89K): [in this window] [in a new window]   FIG. 3.   Multiple sequence alignment of UDP-N-acetylglucosamine enolpyruvyl transferases (MurA1 and MurA2) from B. subtilis (Bs), S. aureus (Sa), S. pneumoniae (Sp), and E. coli (Ec). Shown are the locations of  helices (h) and  sheets (s) in the determined structure (St) of E. coli MurA (22). Highly conserved residues (from CLUSTALW output) are boldface, while residues of E. coli MurA involved in ligand interactions are in italics and boxed.

Essentiality. We first attempted to inactivate each S. pneumoniae murA gene by allelic replacement. Regions up- and downstream of each gene ranging from 406 to 555 bp were amplified by PCR and used to make constructs in which they flanked an antibiotic resistance gene. In the case of murA1, the constitutively expressed erythromycin resistance gene cassette ermAM from pAM1 was used (17), while the chloramphenicol resistance cat gene from pJS3 was used in the murA2 construct (2). To minimize the potential polar effects of gene replacement, the PCR primers were chosen so that flanking genes and potential promoters would remain intact in the deletion mutant. In addition, the cassettes were designed so that the resistance gene was inserted in the same orientation as the target gene, to ensure transcription of the downstream region. This strategy was designed to generate mutants which were nonpolar (since there is no structural or transcriptional disruption of adjacent genes), stable in the absence of selective pressure (because there are no insertion site sequence duplications to allow spontaneous excision), and null (given that there can be no residual target gene activity).

Biochemical characterization of MurA1 and MurA2. MurA1 and MurA2 from S. pneumoniae were investigated further at the protein level. The murA1 and murA2 genes were overexpressed, and the enzymes that they encode, MurA1 and MurA2, respectively, were purified and kinetically characterized. Overexpression of the murA1 and murA2 genes in E. coli accounted for ca. 40 and 30% of total cell protein, and under the selected induction conditions, more than 50 and 40% of the MurA1 and MurA2 enzymes, respectively, remained soluble.

View larger version (40K): [in this window] [in a new window]   FIG. 4.   SDS-PAGE analysis of purified MurA1 and MurA2. The experiment was performed on a Novex 4 to 12% bis-Tris gel with morpholineethanesulfonic acid running buffer. Lane 1, 10-kDa molecular mass ladder (with selected molecular masses [in kilodaltons] marked); lane 2, purified MurA1 after the Hi-Load Q column; lane 3, purified MurA2 after the Hi-Load Q column.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

We have shown by phylogenetic analysis that two distinct classes of UDP-N-acetylglucosamine enolpyruvyl transferase exist (Fig. 2). The first class of transferases, MurA1, occurs throughout all bacteria except gram-positive Mycoplasma spp. The second type of transferase, MurA2, exists as a duplicate gene copy in only the low-G+C gram-positive bacteria. A combination of sequence analysis, gene knockout experiments, and direct biochemical characterization was used to analyze the two copies of murA found in these gram-positive organisms. Sequence analysis indicates both genes are complete and that the enzymes they encode contain the important catalytic residues previously identified in the E. coli MurA. Structurally, and probably functionally, MurA1 and MurA2 are highly similar. Gene deletion experiments further indicate that both genes must encode active enzymes since removal of either still results in a viable organism although this activity is clearly essential for viability. This suggests that the two enzymes have the same function and that one enzyme can substitute for the other.

To demonstrate conclusively that these two genes encode active enzymes, we overexpressed the two genes and purified the two proteins from S. pneumoniae. Both MurA enzymes are active and catalyze the reaction between UDPAG and PEP to give UDP-N-acetylglucosamine enolpyruvate and P_i. Both enzymes from S. pneumoniae appear to have higher K_ms for their substrates and lower k_cats than the E. coli enzyme, indicating that at a kinetic level MurA1 more closely resembles MurA2 than its gram-negative counterpart. They are also both inhibited by the antibiotic fosfomycin. Fosfomycin possesses antibacterial activity against both gram-positive and gram-negative organisms. The majority of the data generated on the mode of action of this antibiotic has been focused on MurA from E. coli (for example, see references 11, 15, 16, and 26). In this study we have demonstrated that fosfomycin inhibits both MurA enzymes from S. pneumoniae, thus providing the first confirmatory evidence that MurA is also the antibacterial target of fosfomycin in this organism.

The existence of two active copies of murA gene in these gram-positive organisms raises the question of how the two copies arose. The phylogenetic tree suggests that murA2 might have evolved from a duplication of the murA1 gene in gram-positive bacteria, although the statistical support for this scenario, or any alternative, is weak. The branching of Deinococcus radiodurans and Thermotoga maritima, both not considered G+C gram-positive bacteria (18), among the low G+C gram-positive bacteria might be taken as negative evidence against the scenario of murA2 evolution by gene duplication within the gram-positive bacteria. However, there is considerable debate about the phylogenetic placement of these bacterial taxa (Cavalier-Smith [5] proposed that T. maritima might be a gram-positive bacterium) and the clustering of Deinococcus/Thermus and Thermotogales among gram-positive bacteria has been observed in other protein phylogenies such as glutamine synthetase (4).

The phylogenetic and structural analysis does support the notion that MurA1 and MurA2 proteins are closely related and are likely to have evolved somewhere in the lineage leading to the low-G+C gram-positive bacteria. MurA1 and MurA2 proteins might have evolved from an even more ancient source as suggested by Psi-BLAST homology searches (1), which reveal significant similarities between MurA1/MurA2 and 5-enolpyruvylshikimate-3-phosphate synthase, a key enzyme in the shikimate pathway.

It is also unclear why a duplication of the enolpyruvyl transfer function has arisen and whether MurA1 and MurA2 carry out identical functions at a physiological level within the cell. The two enzymes catalyze the same reaction, which is the first committed step in peptidoglycan biosynthesis. The higher k_cat/K_m(UDPAG) and k_cat/K_m(PEP) values for MurA2 from S. pneumoniae (Table 1) indicate that MurA2 is a more efficient enzyme in utilizing both substrates under the assay conditions used here, but this may not be true in vivo. Both enzymes appear to be less efficient than their E. coli counterparts (11). It would be interesting to investigate the relative levels of expression of the two murA genes within the cell, as this might shed more light on roles of their gene products in the peptidoglycan biosynthetic pathway. It is possible that although both MurA enzymes are active, only one murA gene is expressed and its expression product is utilized in S. pneumoniae.

It is interesting that the other genes encoding enzymes involved in peptidoglycan biosynthesis in gram-positive bacteria appear to be present as single copies. However, there is a further example in peptidoglycan biosynthesis of an enzyme activity that appears to have been duplicated: D-Ala-D-Ala ligase in gram-negative bacteria, which is encoded by the ddlA and ddlB genes. In this case again, the two purified enzymes show similar kinetic parameters and susceptibilities to inhibitors, and it is unclear why a duplication in activity has occurred (28).

MurA is potentially an important antibacterial target. The presence of two active copies of the enzyme in gram-positive pathogens illustrates the need for a MurA-directed antibacterial to inhibit both forms of the enzyme. This is clearly possible, as illustrated by fosfomycin, which inhibits both MurA1 and MurA2 and thus has antibacterial activity against S. pneumoniae. This observation and the fact that we have shown the enolpyruvyl transferase activity to be essential in this organism further validate the MurA enzyme as an antibacterial target.