Abstract

-lactam antibiotics, including penicillins and cephalosporins, inhibit penicillin-binding proteins (PBPs), which are essential for bacterial cell wall biogenesis. Pathogenic bacteria have evolved efficient antibiotic resistance mechanisms that, in Gram-positive bacteria, include mutations to PBPs that enable them to avoid -lactam inhibition¹. Lactivicin (LTV; Compound 1) contains separate cycloserine and -lactone rings and is the only known natural PBP inhibitor that does not contain a -lactam^{2, 3, 4}. Here we show that LTV and a more potent analog, phenoxyacetyl-LTV (PLTV; Compound 2), are active against clinically isolated, penicillin-resistant Streptococcus pneumoniae strains. Crystallographic analyses of S. pneumoniae PBP1b reveal that LTV and PLTV inhibition involves opening of both monocyclic cycloserine and -lactone rings. In PBP1b complexes, the ring-derived atoms from LTV and PLTV show a notable structural convergence with those derived from a complexed cephalosporin (cefotaxime; Compound 3). The structures imply that derivatives of LTV will be useful in the search for new antibiotics with activity against -lactam–resistant bacteria.

Introduction

S. pneumoniae, the causative agent of pneumonia, meningitis, otitis media and bacteremia⁵, is predominantly treated with -lactams. However, there is a clear relationship between the appearance of penicillin-resistant S. pneumoniae strains and heavy clinical use of -lactams⁶. Resistance in S. pneumoniae involves mutations that result in PBPs incapable of efficiently recognizing -lactams¹.

Efforts to identify new types of -lactam antibiotics with better activity spectra or resistance profiles against pathogenic bacteria such as S. pneumoniae have resulted in generation of antibiotics including the cephalosporins, cephamycins, carbapenems and monobactams⁷ (Fig. 1a). The isolation of LTV (Fig. 1b), which has a unique ring structure comprising a functionalized L-cycloserinyl ring linked to a -lactone ring, was important as LTV was the first—and so far, only—natural PBP inhibitor without a -lactam^{2, 3, 4} (Fig. 1). LTV is highly active against Gram-positive bacteria and, to a lesser extent, Gram-negative pathogens⁸. Binding experiments have identified PBPs as the targets of LTV², although the mechanism of PBP inhibition is unknown. Here we use microbiological, kinetic and crystallographic techniques to determine how lactivicins inhibit PBPs from drug-sensitive and drug-resistant pneumococcal strains.

Figure 1: Structures and mechanisms of action of -lactams and lactivicins.

(a) Structures of -lactam antibiotics. The conserved -lactam ring is highlighted in green. (b) Structures of epimeric LTV and PLTV. (c) Outline mechanism for -lactams showing formation of the hydrolytically stable acyl-enzyme complex (for cefotaxime: R, aminothiazolemethoxyoxime; X, OCOCH₃). (d) Proposed mechanism of PBP acylation by LTV and PLTV, involving formation of a stable acyl-enzyme complex whose structure is closely analogous to that formed by cephalosporins.

Full size image (81 KB)

Initially, LTV and PLTV were synthesized (see Methods and Supplementary Methods online) and assayed in liquid cultures for activity against the noninfectious pneumococcal strain R6 (ref. 9). Both were found to be bacteriolytic agents (Fig. 2a). Highly -lactam–resistant isolates were then tested for susceptibility to LTV and PLTV (Table 1). For penicillin G (Compound 4), the minimal inhibitory concentration (MIC) values for drug-resistant strains were 150- to 600-fold higher than those for the drug-sensitive strain, consistent with modification of PBPs to avoid -lactam recognition. In contrast, for LTV/PLTV, the MICs for resistant strains were only between 5- and 10-fold higher than those for the drug-sensitive R6 strain, possibly reflecting a difference in the inhibition mechanism for the lactivicins and the -lactams. Thus, although the PBPs in drug-resistant strains are mutated to avoid -lactam inhibition¹, they can still recognize and be inhibited by lactivicins.

Figure 2: Kinetic and microbiological analyses of LTV and PLTV activity.

(a) Analysis of growth curves for drug-sensitive S. pneumoniae strain R6, in the absence or presence of various amounts of PLTV. The decrease of optical density (OD) values to zero after introduction of PLTV reveals that the antibiotic is bacteriolytic. (b) k₂/K values for penicillin G, LTV and PLTV, measured for both PBP2x-R6 (drug-sensitive laboratory strain) and PBP2x-5204 (highly drug-resistant, clinical pneumococcal strain). See Methods for conditions.

Full size image (36 KB)

Table 1: MIC values for -lactam–sensitive and -lactam–resistant strains toward penicillin G, LTV and PLTV

Full table

Because PBP2x is a resistance determinant in pneumococcal strains¹⁰, kinetic analyses for lactivicins were carried out on PBP2x from sensitive (R6) and -lactam-resistant (5204) strains (Fig. 2b). PLTV had a half-maximal inhibitory concentration (IC₅₀) of 7.9 M against PBP2x-5204, compared to 150 M for LTV. Second-order rate constants (k₂/K) are used to characterize the acylation efficiency of PBPs by -lactams. The k₂/K value of penicillin G for PBP2x-5204 was over 1,000-fold lower than that for PBP2x-R6 (ref. 11). LTV and PLTV also had lower k₂/K values against PBP2x-5204, and PLTV was apparently the more efficient inhibitor. Caution should be exerted in comparing the k₂/K data for different types of inhibitors, although it is notable that the difference in apparent k₂/K values for the two PBPs was markedly lower for PLTV than for LTV and penicillin G.

To investigate the mechanism of PBP inhibition by LTV and PLTV, we initially incubated a soluble form of pneumococcal PBP1b (PBP1b*) with the inhibitors and conducted mass spectrometric analyses. The mass increments between PBP1b* (51,416 Da) and the inhibitor complexes were 272 Da and 364 Da for LTV and PLTV respectively, consistent with the formation of covalent adducts. The complexes were stable for more than 24 h, well beyond the lifetime of the bacterial division cycle.

We then solved crystal structures of S. pneumoniae PBP1b* complexed to LTV or PLTV (see Supplementary Table 1 online). PBP1b*-inhibitor structures were only obtained after PBP1b* crystals complexed with the pseudosubstrate N-benzoyl-D-alanylmercaptoacetic acid (S2d, Compound 5) were washed and then incubated with LTV or PLTV¹². Structures were solved to 2.6-Å (LTV) and 2.4-Å resolution (PLTV) by molecular replacement using the PBP1b structure as a search model (see Methods). The overall folds of the complexes were similar to the reported PBP1b* structure¹² (r.m.s. deviation between the LTV and PLTV complex structures for all backbone atoms (NCCO) was 0.91 Å). Briefly, PBP1b* is a three-domain protein, comprising a central transpeptidase domain flanked by a C-terminal -strand–rich region and an N-terminal domain that links the transpeptidase and glycosyltransferase domains (latter domain mostly absent in the structure). When complexed to LTV or PLTV, the conformations of the active site are very similar to that observed in an apparently 'open' form, as when PBP1b* is complexed with relatively large -lactam inhibitors such as cefotaxime¹². These conformations contrast with the apo form of PBP1b*, where the active site is unavailable for inhibitor binding because the entry of the catalytic cleft is closed by the 3/4 loop¹².

The structures reveal that the atoms derived from the L-cycloserine and -lactone rings of LTV and PLTV are nearly superimposable and that both rings of the inhibitors are open. LTV and PLTV are covalently linked to the nucleophilic O of Ser460 by an ester and are present in high occupancy (Fig. 3a,b). The carboxylate group of LTV and PLTV to the cleaved cycloserine nitrogen interacts with the O of Thr652 and Thr654, part of the conserved SXXK motif, and with the O of Ser516, part of the conserved SXN motif (Fig. 4 and Supplementary Fig. 1 online). The carbonyl of the cleaved cycloserine is located in the oxyanion hole and interacts with backbone nitrogen atoms from Ser460 and Thr654. In the PLTV structure, the carboxyethyl moiety resulting from the cleavage of the -lactone is positioned to form hydrogen bonds with the hydroxyl group of Tyr498 (Fig. 4 and Supplementary Fig. 1). The amide carbonyl of the LTV and PLTV side chain is positioned to form hydrogen bonds with Asn518 N2, which lies on the loop between 4 and 5 (conserved SXN motif), and with the backbone carbonyl of Thr654. The hydrophobic phenoxyacetyl side chain of PLTV, which might have a role in stabilizing surrounding loops, folds back over the ring-derived inhibitor atoms and sterically blocks the entry to the catalytic cleft (Figs. 3 and 4), consistent with the kinetic data revealing that PLTV is a more potent inhibitor than LTV.

Figure 3: Stereoviews from crystal structures of the PBP1b*-LTV and PBP1b*-PLTV complexes.

(a) F_o – F_c map (green) contoured at 2.4 , generated before inclusion of LTV in the model. (b) F_o – F_c map (green), contoured at 2.4 , for the PLTV molecule. Selected active site residues are shown as sticks; LTV and PLTV backbones are shown in blue. Notably, ester link is present between the Ser460 side chain and the carbonyl of the LTV/PLTV cycloserine ring, and both cycloserine and lactone rings of the inhibitors are open. Figure was prepared using PyMOL (http://pymol.sourceforge.net/).

Full size image (96 KB)

Figure 4: Stereoview overlay of the active site complexes of PBP1b*-cefotaxime, PBP1b*-LTV and PBP1b*-PLTV.

Atoms derived from the opened -lactam and lactone rings in LTV and PLTV show structural convergence to those derived from the reaction of cefotaxime with PBP1b*. In total, PLTV makes eight hydrogen bonds to PBP1b*, whereas LTV makes seven. LTV (green), PLTV (blue) and cefotaxime (yellow) are shown as sticks. Residues involved in hydrogen bonds (gray dashed lines) with PLTV are shown as white sticks (Ser516 not shown). Figure was prepared using PyMOL.

Full size image (64 KB)

Although the lactivicin and cephalosporin ring systems are very different (Fig. 1), the backbone structures of the LTV and PLTV complexes reveal convergence with the backbone of cefotaxime, in which only the -lactam is opened (Figs. 1 and 4 and Supplementary Fig. 1). Structural analogy is observed for the C3, C4 and C6 carbon and N5 nitrogen atoms of the cefotaxime dihydrothiazine ring and the O1 and N2 cycloserine atoms and most atoms from the cleaved LTV and PLTV lactone ring. The C4 cefotaxime carboxylate and the C6 carboxylate groups of LTV and PLTV occupy closely related positions, and the LTV- and PLTV-derived carboxyethyl moieties point in the same direction as the C3 methylene of cefotaxime, suggesting that substitutions on lactivicins at this position will improve activity as observed for cephalosporins.

In contrast to the ring-derived atoms, the acetamido side chains of LTV and PLTV occupy different positions from that of cefotaxime. Although the oxygen atoms of the cefotaxime oximino group and the PLTV phenoxy group occupy the same region, the absence of atoms analogous to the sulfur and C-2 of the cephem nucleus in the lactivicins seems to enable the hydrophobic side chain of PLTV to fold and approach the position occupied by the sulfur of the dihydrothiazine ring in the cephalosporin-PBP complex. Such a side chain conformation implies that the potency of the lactivicins could be further improved by substitution to enable occupation of the typical -lactam side chain binding pocket.

The combined results imply that the stable acyl-enzyme complex from the lactivicins is formed by a process in which initial nucleophilic attack of Ser460 O onto the carbonyl of the L-cycloserine releases the cycloserine nitrogen lone pair, thus enabling opening of the neighboring -lactone ring (Fig. 1d). Thermodynamically favorable isomerization of the double bond then leads to the crystallographically observed complex. A study on the hydrolysis rates of lactams found that -lactams are approximately 100 times less reactive than analogous -lactams¹³. However, the unique bicyclic ring structure of the lactivicins enables them to fulfill the requirements to act as potent PBP inhibitors. In addition to fitting within the PBP1b* active site (D.S.F., unpublished modeling data), the lactivicins are sufficiently activated to (i) react rapidly with the active site nucleophile, (ii) react irreversibly and (iii) form an acyl-enzyme complex that is stable to hydrolysis. In the case of the lactivicins, the potential deficiencies of direct -lactam analogs of -lactams¹⁴ with respect to (i) and (ii) are overcome by the electron-withdrawing groups on the lactam nitrogen and in the case of (ii) by 'secondary' opening of the -lactone ring. This secondary ring opening has some precedent in the mechanism of serine -lactamase inhibition by clavulanic acid (Compound 6)^{15, 16}. In the case of (iii), as in the -lactams, hydrolysis is probably prevented by the presence of the lactam/cycloserine skeleton, as with the natural substrate (the stem peptide), the C-terminal portion exits the active site and allows a 'hydrolytic' water molecule to enter.

The lower difference in MIC and apparent k₂/K values between the drug-sensitive and drug-resistant S. pneumoniae strains treated with PLTV compared to -lactams (Table 1) may reflect a different mechanistic pathway by which the various bicyclic ring systems of the lactivicins and the -lactams react with the PBP active sites to achieve formation of a stable acyl-enzyme complex. Thus, lactivicins and their analogs may represent a good starting point for the identification of new drugs that target -lactam-resistant pathogens.

Acknowledgments

We thank the ESRF ID14 and ID23 beamline staff for help with data collection, D. Lemaire (IBS) for mass spectrometric assays, J. Croizé (Hopital Universitaire de Grenoble) for drug-resistant pneumococcal strains, R. Carapito for the purification of PBP2x (5204) and O. Dideberg for support. We thank J.M. Frère (ULG) for useful discussions. The work was funded by the European Commission LSHM-CT-2004-512138 (EUR-INTAFAR).

Competing interests statement:

The authors declare no competing financial interests.

Received 30 March 2007; Accepted 6 July 2007; Published online 5 August 2007.

References

1. Macheboeuf, P., Contreras-Martel, C., Job, V., Dideberg, O. & Dessen, A. Penicillin Binding Proteins: Key players in bacterial cell cycle and drug resistance processes. FEMS Microbiol. Rev. 30, 673–691 (2006). | Article | PubMed | ISI | ChemPort |
2. Nozaki, Y. et al. Binding of a non--lactam antibiotic to penicillin-binding proteins. Nature 325, 179–180 (1987). | Article | PubMed | ISI | ChemPort |
3. Nozaki, Y., Katayama, N., Harada, S., Ono, H. & Okazaki, H. Lactivicin, a naturally occuring non--lactam antibiotic having -lactam-like action: Biological activities and mode of action. J. Antibiot. (Tokyo) 42, 84–93 (1989). | PubMed | ChemPort |
4. Harada, S. et al. Structure of lactivicin, an antibiotic having a new nucleus and similar biological activities to -lactam antibiotics. Tetrahedr. Lett. 27, 6229–6232 (1986). | ISI | ChemPort |
5. Schuchat, A. et al. Bacterial meningitis in the United States in 1995. Active Surveillance Team. N. Engl. J. Med. 337, 970–976 (1997). | Article | PubMed | ISI | ChemPort |
6. Bronzwaer, S.L. et al. A European study on the relationship between antimicrobial use and antimicrobial resistance. Emerg. Infect. Dis. 8, 278–282 (2002). | PubMed | ISI |
7. Page, M.I. The Chemistry of -Lactams (ed. Page, M.I.) (Blackie, Glasgow, 1992).
8. Harada, S. et al. Chemistry of a new antibiotic: Lactivicin. Tetrahedron 44, 6589–6606 (1988). | Article | ISI | ChemPort |
9. Inui, T., Oshida, T., Endo, T. & Matsushita, T. Potent bacteriolytic activity of ritipenem associated with a characteristic profile of affinities for penicillin-binding proteins of Haemophilus influenzae. Antimicrob. Agents Chemother. 43, 2534–2537 (1999). | PubMed | ISI | ChemPort |
10. Hakenbeck, R. et al. Penicillin-binding proteins in beta-lactam-resistant Streptococcus pneumoniae. Microb. Drug Resist. 5, 91–99 (1999). | PubMed | ISI | ChemPort |
11. Chesnel, L. et al. The structural modifications induced by the M339F substitution in PBP2x from Streptococcus pneumoniae further decreases the susceptibility to -lactams of resistant strains. J. Biol. Chem. 278, 44448–44456 (2003). | Article | PubMed | ISI | ChemPort |
12. Macheboeuf, P. et al. Active site restructuring regulates ligand recognition in class A penicillin-binding proteins. Proc. Natl. Acad. Sci. USA 102, 577–582 (2005). | Article | PubMed | ChemPort |
13. Imming, P., Klar, B. & Dix, D. Hydrolytic stability versus ring size in lactams: Implications for the development of lactam antibiotics and other serine protease inhibitors. J. Med. Chem. 43, 4328–4331 (2000). | Article | PubMed | ISI | ChemPort |
14. Wilmouth, R.C. et al. Mechanistic insights into the inhibition of serine proteases by monocyclic lactams. Biochemistry 38, 7989–7998 (1999). | Article | PubMed | ISI | ChemPort |
15. Brown, R.P.A., Aplin, R.T. & Schofield, C.J. Inhibition of TEM-2 -lactamase from Escherichia coli by clavulanic acid: Observation of intermediates by electrospray ionization mass spectrometry. Biochemistry 35, 12421–12432 (1996). | Article | PubMed | ISI | ChemPort |
16. Imtiaz, U. et al. Inactivation of class-A -lactamases by clavulanic acid—the role of arginine-244 in a proposed nonconcerted sequence of events. J. Am. Chem. Soc. 115, 4435–4442 (1993). | Article | ISI | ChemPort |
17. Llinás, A. et al. Inactivation of bacterial DD-peptidase by -sultams. Biochemistry 44, 7738–7746 (2005). | PubMed |
18. Lakaye, B. et al. Synthesis, purification and kinetic properties of fluorescein-labelled penicillins. Biochem. J. 300, 141–145 (1994). | PubMed | ISI | ChemPort |
19. Carapito, R., Chesnel, L., Vernet, T. & Zapun, A. Pneumococcal -lactam resistance due to a conformational change in penicillin-binding protein 2x. J. Biol. Chem. 281, 1771–1777 (2006). | PubMed | ISI | ChemPort |
20. Natsugari, H., Kawano, Y., Morimoto, A. & Yoshioka, K. Antibiotic 2-(3-oxo-2-isoxazolidinyl)-5-oxo-2-tetrahydrofuran-carboxylates. US Patent 4,851,422 (1989).
21. Ghuysen, J.M., Frère, J.M., Leyh-Bouille, M., Nguyen-Disteche, M. & Coyette, J. Active-site-serine D-alanyl-D-alanine-cleaving-peptidase-catalysed acyl-transfer reactions. Procedures for studying the penicillin-binding proteins of bacterial plasma membranes. Biochem. J. 235, 159–165 (1986). | PubMed | ISI | ChemPort |
22. Adam, M., Damblon, C., Plaitin, B., Christiaens, L. & Frère, J.M. Chromogenic depsipeptide substrates for beta-lactamases and penicillin-sensitive DD-peptidases. Biochem. J. 270, 525–529 (1990). | PubMed | ISI | ChemPort |
23. De Meester, F. et al. Automated analysis of enzyme inactivation phenomena. Application to beta-lactamases and DD-peptidases. Biochem. Pharmacol. 36, 2393–2403 (1987). | Article | PubMed | ChemPort |
24. Kabsch, W. Automatic processing of rotation diffraction data from crystals of initially unknown symmetry and cell constants. J. Appl. Cryst. 26, 795–800 (1993). | Article | ISI | ChemPort |
25. Navaza, J. Implementation of molecular replacement in AMoRe. Acta Crystallogr. Sect. D Biol. Crystallogr. 57, 1367–1372 (2001). | Article | ISI | ChemPort |
26. Emsley, P. & Cowtan, K. Coot: model-building tools for molecular graphics. Acta Crystallogr. Sect. D Biol. Crystallogr. 60, 2126–2132 (2004). | Article | ISI | ChemPort |
27. Collaborative Computational Project 4. The CCP4 suite: programs for protein crystallography. Acta Crystallogr. Sect. D Biol. Crystallogr. 50, 760–763 (1994). | ISI |
28. Brünger, A. Crystallography & NMR system: A new software for macromolecular structure determination. Acta Crystallogr. Sect. D. Biol. Crystallogr. 54, 905–921 (1998). | Article |

1. Institut de Biologie Structurale Jean-Pierre Ebel Commissariat à l'énergie atomique — Centre National de La Recherche Scientifique – Université Joseph Fourier, 41 rue Jules Horowitz, F-38027 Grenoble, France.
2. Chemistry Research Laboratory, 12 Mansfield Road, Oxford, OX1 3TA, UK.
3. Centre de Recherches du Cyclotron, B30, Université de Liège, Sart-Tilman, B-4000, Liège, Belgium.
4. Centre d'Ingénierie des Protéines, Institut de Chimie, B6a, Université de Liège, Sart-Tilman, B-4000, Liège, Belgium.
5. These authors contributed equally to this work.

Correspondence to: Andréa Dessen¹ e-mail: andrea.dessen@ibs.fr

Correspondence to: Christopher J Schofield² e-mail: christopher.schofield@chem.ox.ac.uk

MORE ARTICLES LIKE THIS

These links to content published by NPG are automatically generated.

NPG Journals

by Subject Area

• Chemistry

  • Chemistry
  • Drug discovery
  • Biotechnology
  • Materials
  • Methods & Protocols

• Clinical Practice & Research

  • Cancer
  • Cardiovascular medicine
  • Dentistry
  • Endocrinology
  • Gastroenterology & Hepatology
  • Methods & Protocols
  • Pathology & Pathobiology
  • Urology

• Earth & Environment

  • Earth sciences
  • Evolution & Ecology

• Life sciences

  • Biotechnology
  • Cancer
  • Development
  • Drug discovery
  • Evolution & Ecology
  • Genetics
  • Immunology
  • Medical research
  • Methods & Protocols
  • Microbiology
  • Molecular cell biology
  • Neuroscience
  • Pharmacology
  • Systems biology

• Physical sciences

  • Physics
  • Materials

by A - Z Index