Jean-Emmanuel Hugonnet

A Novel Peptidoglycan Cross-linking Enzyme for a β-Lactam-resistant Transpeptidation Pathway

The β-lactam antibiotics remain the most commonly used to treat severe infections. Because of structural similarity between the β-lactam ring and the d-alanyl4-d-alanine5 extremity of bacterial cell wall precursors, the drugs act as suicide substrates of the dd-transpeptidases that catalyze the last cross-linking step of cell wall assembly. Here, we show that this mechanism of action can be defeated by a novel type of transpeptidase identified for the first time by reverse genetics in aβ-lactam-resistant mutant of Enterococcus faecium. The enzyme, Ldtfm, catalyzes in vitro the cross-linking of peptidoglycan subunits in a β-lactam-insensitive ld-transpeptidation reaction. The specificity of Ldtfm for the l-lysyl3-d-alanine4 peptide bond of tetrapeptide donors accounts for resistance because the substrate does not mimic β-lactams in contrast to d-alanyl4-d-alanine5 in the pentapeptide donors required for dd-transpeptidation. Ldtfm homologues are encountered sporadically among taxonomically distant bacteria, indicating that ld-transpeptidase-mediated resistance may emerge in various pathogens.

Unexpected Inhibition of Peptidoglycan LD-Transpeptidase from Enterococcus faecium by the β-Lactam Imipenem

The β-lactam antibiotics mimic the d-alanyl4-d-alanine5 extremity of peptidoglycan precursors and act as “suicide” substrates of the dd-transpeptidases that catalyze the last cross-linking step of peptidoglycan synthesis. We have previously shown that bypass of the dd-transpeptidases by the ld-transpeptidase of Enterococcus faecium (Ldtfm) leads to high level resistance to ampicillin. Ldtfm is specific for the l-lysyl3-d-alanine4 bond of peptidoglycan precursors containing a tetrapeptide stem lacking d-alanine5. This specificity was proposed to account for resistance, because the substrate of Ldtfm does not mimic β-lactams in contrast to the d-alanyl4-d-alanine5 extremity of pentapeptide stems used by the dd-transpeptidases. Here, we unexpectedly show that imipenem, a β-lactam of the carbapenem class, totally inhibited Ldtfm at a low drug concentration that was sufficient to inhibit growth of the bacteria. Peptidoglycan cross-linking was also inhibited, indicating that Ldtfm is the in vivo target of imipenem. Stoichiometric and covalent modification of Ldtfm by imipenem was detected by mass spectrometry. The modification was mapped into the trypsin fragment of Ldtfm containing the catalytic Cys residue, and the Cys to Ala substitution prevented imipenem binding. The mass increment matched the mass of imipenem, indicating that inactivation of Ldtfm is likely to involve rupture of the β-lactam ring and acylation of the catalytic Cys residue. Thus, the spectrum of activity of β-lactams is not restricted to transpeptidases of the dd-specificity, as previously thought. Combination therapy with imipenem and ampicillin could therefore be active against E. faecium strains having the dual capacity to manufacture peptidoglycan with transpeptidases of the ld- and dd-specificities.

Role of Class A Penicillin-Binding Proteins in PBP5-Mediated β-Lactam Resistance in Enterococcus faecalis

Peptidoglycan polymerization complexes contain multimodular penicillin-binding proteins (PBP) of classes A and B that associate a conserved C-terminal transpeptidase module to an N-terminal glycosyltransferase or morphogenesis module, respectively. In Enterococcus faecalis, class B PBP5 mediates intrinsic resistance to the cephalosporin class of beta-lactam antibiotics, such as ceftriaxone. To identify the glycosyltransferase partner(s) of PBP5, combinations of deletions were introduced in all three class A PBP genes of E. faecalis JH2-2 (ponA, pbpF, and pbpZ). Among mutants with single or double deletions, only JH2-2 DeltaponA DeltapbpF was susceptible to ceftriaxone. Ceftriaxone resistance was restored by heterologous expression of pbpF from Enterococcus faecium but not by mgt encoding the monofunctional glycosyltransferase of Staphylococcus aureus. Thus, PBP5 partners essential for peptidoglycan polymerization in the presence of beta-lactams formed a subset of the class A PBPs of E. faecalis, and heterospecific complementation was observed with an ortholog from E. faecium. Site-directed mutagenesis of pbpF confirmed that the catalytic serine residue of the transpeptidase module was not required for resistance. None of the three class A PBP genes was essential for viability, although deletion of the three genes led to an increase in the generation time and to a decrease in peptidoglycan cross-linking. As the E. faecalis chromosome does not contain any additional glycosyltransferase-related genes, these observations indicate that glycan chain polymerization in the triple mutant is performed by a novel type of glycosyltransferase. The latter enzyme was not inhibited by moenomycin, since deletion of the three class A PBP genes led to high-level resistance to this glycosyltransferase inhibitor.

Inactivation of Mycobacterium tuberculosis l,d-Transpeptidase LdtMt1 by Carbapenems and Cephalosporins

ABSTRACT The structure of Mycobacterium tuberculosis peptidoglycan is atypical since it contains a majority of 3→3 cross-links synthesized by l,d-transpeptidases that replace 4→3 cross-links formed by the d,d-transpeptidase activity of classical penicillin-binding proteins. Carbapenems inactivate these l,d-transpeptidases, and meropenem combined with clavulanic acid is bactericidal against extensively drug-resistant M. tuberculosis. Here, we used mass spectrometry and stopped-flow fluorimetry to investigate the kinetics and mechanisms of inactivation of the prototypic M. tuberculosis l,d-transpeptidase LdtMt1 by carbapenems (meropenem, doripenem, imipenem, and ertapenem) and cephalosporins (cefotaxime, cephalothin, and ceftriaxone). Inactivation proceeded through noncovalent drug binding and acylation of the catalytic Cys of LdtMt1, which was eventually followed by hydrolysis of the resulting acylenzyme. Meropenem rapidly inhibited LdtMt1, with a binding rate constant of 0.08 μM−1 min−1. The enzyme was unable to recover from this initial binding step since the dissociation rate constant of the noncovalent complex was low (<0.1 min−1) in comparison to the acylation rate constant (3.1 min−1). The covalent adduct resulting from enzyme acylation was stable, with a hydrolysis rate constant of 1.0 × 10−3 min−1. Variations in the carbapenem side chains affected both the binding and acylation steps, ertapenem being the most efficient LdtMt1 inactivator. Cephalosporins also formed covalent adducts with LdtMt1, although the acylation reaction was 7- to 1,000-fold slower and led to elimination of one of the drug side chains. Comparison of kinetic constants for drug binding, acylation, and acylenzyme hydrolysis indicates that carbapenems and cephems can both be tailored to optimize peptidoglycan synthesis inhibition in M. tuberculosis.

Specificity of L,D-Transpeptidases from Gram-positive Bacteria Producing Different Peptidoglycan Chemotypes

We report here the first direct assessment of the specificity of a class of peptidoglycan cross-linking enzymes, the l,d-transpeptidases, for the highly diverse structure of peptidoglycan precursors of Gram-positive bacteria. The lone functionally characterized member of this new family of active site cysteine peptidases, Ldtfm from Enterococcus faecium, was previously shown to bypass the d,d-transpeptidase activity of the classical penicillin-binding proteins leading to high level cross-resistance to glycopeptide and β-lactam antibiotics. Ldtfm homologues from Bacillus subtilis (LdtBs) and E. faecalis (Ldtfs) were found here to cross-link their cognate disaccharide-peptide subunits containing meso-diaminopimelic acid (mesoDAP3) and l-Lys3-l-Ala-l-Ala at the third position of the stem peptide, respectively, instead of l-Lys3-d-iAsn in E. faecium. Ldtfs differed from Ldtfm and LdtBs by its capacity to hydrolyze the l-Lys3-d-Ala4 bond of tetrapeptide (l,d-carboxypeptidase activity) and pentapeptide (l,d-endopeptidase activity) stems, in addition to the common cross-linking activity. The three enzymes were specific for their cognate acyl acceptors in the cross-linking reaction. In contrast to Ldtfs, which was also specific for its cognate acyl donor, Ldtfm tolerated substitution of l-Lys3-d-iAsn by l-Lys3-l-Ala-l-Ala. Likewise, LdtBs tolerated substitution of mesoDAP3 by l-Lys3-d-iAsn and l-Lys3-l-Ala-l-Ala in the acyl donor. Thus, diversification of the structure of peptidoglycan precursors associated with speciation has led to a parallel evolution of the substrate specificity of the l,d-transpeptidases affecting mainly the recognition of the acyl acceptor. Blocking the assembly of the side chain could therefore be used to combat antibiotic resistance involving l,d-transpeptidases.

The CroRS Two-Component Regulatory System Is Required for Intrinsic β-Lactam Resistance in Enterococcus faecalis

Enterococcus faecalis produces a specific penicillin-binding protein (PBP5) that mediates high-level resistance to the cephalosporin class of beta-lactam antibiotics. Deletion of a locus encoding a previously uncharacterized two-component regulatory system of E. faecalis (croRS) led to a 4,000-fold reduction in the MIC of the expanded-spectrum cephalosporin ceftriaxone. The cytoplasmic domain of the sensor kinase (CroS) was purified and shown to catalyze ATP-dependent autophosphorylation followed by transfer of the phosphate to the mated response regulator (CroR). The croR and croS genes were cotranscribed from a promoter (croRp) located in the rrnC-croR intergenic region. A putative seryl-tRNA synthetase gene (serS) located immediately downstream from croS did not appear to be a target of CroRS regulation or to play a role in ceftriaxone resistance. A plasmid-borne croRp-lacZ fusion was trans-activated by the CroRS system in response to the presence of ceftriaxone in the culture medium. The fusion was also induced by representatives of other classes of beta-lactam antibiotics and by inhibitors of early and late steps of peptidoglycan synthesis. The croRS null mutant produced PBP5, and expression of an additional copy of pbp5 under the control of a heterologous promoter did not restore ceftriaxone resistance. Deletion of croRS was not associated with any defect in the synthesis of the nucleotide precursor UDP-MurNAc-pentapeptide or of the D-Ala(4)-->L-Ala-L-Ala-Lys(3) peptidoglycan cross-bridge. Thus, the croRS mutant was susceptible to ceftriaxone despite the production of PBP5 and the synthesis of wild-type peptidoglycan precursors. These observations constitute the first description of regulatory genes essential for PBP5-mediated beta-lactam resistance in enterococci.

Aslfm, the D-aspartate ligase responsible for the addition of D-aspartic acid onto the peptidoglycan precursor of Enterococcus faecium.

d-Aspartate ligase has remained the last unidentified peptide bond-forming enzyme in the peptidoglycan assembly pathway of Gram-positive bacteria. Here we show that a two-gene cluster of Enterococcus faecium encodes aspartate racemase (Racfm) and ligase (Aslfm) for incorporation of d-Asp into the side chain of the peptidoglycan precursor. Aslfm was identified as a new member of the ATP-grasp protein superfamily, which includes a diverse set of enzymes catalyzing ATP-dependent carboxylate-amine ligation reactions. Aslfm specifically ligated the β-carboxylate of d-Asp to the ϵ-amino group of l-Lys in the nucleotide precursor UDP-N-acetylmuramyl-pentapeptide. d-iso-Asparagine was not a substrate of Aslfm, indicating that the presence of this amino acid in the peptidoglycan of E. faecium results from amidation of the α-carboxyl of d-Asp after its addition to the precursor. Heterospecific expression of the genes encoding Racfm and Aslfm in Enterococcus faecalis led to production of stem peptides substituted by d-Asp instead of l-Ala2, providing evidence for the in vivo specificity and function of these enzymes. Strikingly, sequencing of the cross-bridges revealed that substitution of l-Ala2 by d-Asp is tolerated by the d,d-transpeptidase activity of the penicillin-binding proteins both in the acceptor and in the donor substrates. The Aslfm ligase appears as an attractive target for the development of narrow spectrum antibiotics active against multiresistant E. faecium.

In Vitro Cross-Linking of Mycobacterium tuberculosis Peptidoglycan by l,d-Transpeptidases and Inactivation of These Enzymes by Carbapenems

ABSTRACT The Mycobacterium tuberculosis peptidoglycan is cross-linked mainly by l,d-transpeptidases (LDTs), which are efficiently inactivated by a single β-lactam class, the carbapenems. Development of carbapenems for tuberculosis treatment has recently raised considerable interest since these drugs, in association with the β-lactamase inhibitor clavulanic acid, are uniformly active against extensively drug-resistant M. tuberculosis and kill both exponentially growing and dormant forms of the bacilli. We have purified the five l,d-transpeptidase paralogues of M. tuberculosis (Mt1 to -5) and compared their activities with those of peptidoglycan fragments and carbapenems. The five LDTs were functional in vitro since they were active in assays of peptidoglycan cross-linking (Mt5), β-lactam acylation (Mt3), or both (Mt1, Mt2, and Mt4). Mt3 was the only LDT that was inactive in the cross-linking assay, suggesting that this enzyme might be involved in other cellular functions such as the anchoring of proteins to peptidoglycan, as shown in Escherichia coli. Inactivation of LDTs by carbapenems is a two-step reaction comprising reversible formation of a tetrahedral intermediate, the oxyanion, followed by irreversible rupture of the β-lactam ring that leads to formation of a stable acyl enzyme. Determination of the rate constants for these two steps revealed important differences (up to 460-fold) between carbapenems, which affected the velocity of oxyanion and acyl enzyme formation. Imipenem inactivated LDTs more rapidly than ertapenem, and both drugs were more efficient than meropenem and doripenem, indicating that modification of the carbapenem side chain could be used to optimize their antimycobacterial activity.

β-Lactamase inhibition by avibactam in Mycobacterium abscessus

OBJECTIVES Two β-lactams, cefoxitin and imipenem, are part of the reference treatment for pulmonary infections with Mycobacterium abscessus. M. abscessus has recently been shown to produce a broad-spectrum β-lactamase, BlaMab, indicating that the combination of β-lactams with a BlaMab inhibitor may improve treatment efficacy. The objectives of this study were to evaluate the impact of BlaMab production on the efficacy of β-lactams in vitro and to assess the benefit of BlaMab inhibition on the activity of β-lactams intracellularly and in an animal model. METHODS We analysed the mechanism and kinetics of BlaMab inactivation by avibactam, a non-β-lactam β-lactamase inhibitor currently in Phase III of development, in combination with ceftazidime for the treatment of serious infections due to Gram-negative bacteria. We then deleted the gene encoding BlaMab to assess the extent of BlaMab inhibition by avibactam based on a comparison of the impact of chemical and genetic inactivation. Finally, the efficacy of amoxicillin in combination with avibactam was evaluated in cultured human macrophages and in a zebrafish model of M. abscessus infection. RESULTS We showed that avibactam efficiently inactivated BlaMab via the reversible formation of a covalent adduct. An inhibition of BlaMab by avibactam was observed in both infected macrophages and zebrafish. CONCLUSIONS Our data identify avibactam as the first efficient inhibitor of BlaMab and strongly suggest that β-lactamase inhibition should be evaluated to provide improved therapeutic options for M. abscessus infections.

Characterization of CrgA, a New Partner of the Mycobacterium tuberculosis Peptidoglycan Polymerization Complexes

The role(s) in cell division of the Mycobacterium tuberculosis Rv0011c gene product, a homolog of the Streptomyces CrgA protein that is responsible for coordinating growth and cytokinesis in sporogenic aerial hyphae, is largely unknown. We show that an enhanced cyan fluorescent protein-M. tuberculosis CrgA (ECFP-CrgA(MT)) fusion protein is localized to the cell membrane, midcell, and cell pole regions in Mycobacterium smegmatis. Furthermore, the ECFP-CrgA(MT) fusion protein colocalized with FtsZ-enhanced yellow fluorescent protein (EYFP) in M. smegmatis. Bacterial two-hybrid assays indicated strong interactions of M. tuberculosis CrgA with FtsZ, FtsQ, and the class B penicillin-binding proteins, FtsI (PBPB) and PBPA. The midcell localization of CrgA(MT) was severely compromised under conditions of FtsZ depletion, which indicated that CrgA localizes to the midcell region after assembly of the FtsZ ring. M. tuberculosis cells with reduced CrgA levels were elongated and grew more slowly than wild-type cells, which indicated defects in cell division, whereas CrgA overproduction did not show growth defects. A M. smegmatis ΔcrgA strain exhibited a bulged cell morphology, elongated cells with a chain-like phenotype, cells with polar bulbous structures, and a modest growth defect. FtsZ and FtsI levels were not affected in cells producing altered levels of CrgA. Septal and membrane localization of GFP-FtsI was enhanced by CrgA overproduction and was diminished in a ΔcrgA strain, which indicates that one role of CrgA is to promote and/or stabilize FtsI localization. Overall, these data indicate that CrgA is a novel member of the cell division complex in mycobacteria and possibly facilitates septum formation.