Dear Editor,
Tuberculosis (TB), an infectious disease caused by the bacillus Mycobacterium tuberculosis, leads to substantial mortality worldwide1. The emergence of multidrug-resistant (MDR) and extensively drug-resistant (XDR) strains of M. tuberculosis has challenged conventional anti-TB therapy and threatens global disease control of TB2,3. The development of new anti-TB drugs is urgently required4. β-lactams are effective antibiotics widely used to treat bacterial infections; however, so far no effective anti-TB antibiotics have emerged from this class of drugs. Previous studies have indicated that an important reason for the lack of effective β-lactam anti-TB antibiotics is the presence of a chromosomally-encoded class A (Ambler) β-lactamase, BlaC, in M. tuberculosis5.
A recent study by Hugonnet et al.6 has demonstrated that a combination of meropenem-clavulanate was effective against 13 XDR strains of M. tuberculosis. Meropenem is a member of the carbapenem class of β-lactams and contains a bicyclic nucleus, a pyrroline ring and a β-lactam ring, and clavulanate is a β-lactamase inhibitor. Both are FDA-approved drugs and could potentially be used to treat TB, providing a novel clinical treatment strategy6. However, the bactericidal mechanism of this drug combination in treating TB is unclear. Hugonnet et al. suggest that one reason for the efficacy of the meropenem-clavulanate combination against XDR strains is that meropenem is poorly hydrolyzed by BlaC6.
Carbapenems have been reported to target L,D-transpeptidases that generate the unusual 3→3 transpeptide linkages that are the predominant transpeptide linkages (80%) in the peptidoglycan layer of nonreplicating M. tuberculosis7,8. L,D-transpeptidase type 2 (LdtMt2) is the main L,D-transpeptidase in M. tuberculosis and is critical for cell wall synthesis, virulence and amoxicillin tolerance of M. tuberculosis9. To kill M. tuberculosis effectively, the combination of meropenem-clavulanate not only inactivates β-lactamase but also prevents the formation of transpeptide linkages in the peptidoglycan layer9. As LdtMt2 is the enzyme that generates 80% of the transpeptide linkages in M. tuberculosis, the inactivation of LdtMt2 is likely to be the major mechanism underlying the effectiveness of meropenem-clavulanate against M. tuberculosis. Here, we report the crystal structures of an N-terminal-truncated LdtMt2 (LdtMt2-ΔN54, residues 55-408), a trypsin-degraded fragment of LdtMt2 (LdtMt2-ΔN139, residues 140-408) and the complex of LdtMt2-ΔN139 and meropenem, at 2.5, 1.8 and 1.4 Å resolutions, respectively (Supplementary information, Figure S1 and Table S1). These structures reveal the structural mechanism through which meropenem may act to inhibit LdtMt2.
Our apo LdtMt2-ΔN54 structure (Figure 1A) shows the linear arrangement of the two N-terminal β-barrel domains (residues 60-148 and 149-250) and the C-terminal YkuD domain (residues 251-408). The two N-terminal β-barrel domains, both of which adopt an IgG-like fold, contain one three-stranded and one four-stranded sheet, respectively. We suggest that these two IgG-like domains act as a spacer arm for the YkuD catalytic domain. The YkuD domain is characterized by a β-sheet structure similar to that in LdtBs from Bacillus subtilis10 and Ldtfm from Enterococcus faecium11 (Supplementary information, Figure S1) with a root-mean-square deviation of Cα superposition of 1.6 Å and 1.8 Å and sequence identities of 20% and 24%, respectively. As in LdtBs and Ldtfm10,11, the YkuD domain in LdtMt2 contains a catalytic triad Cys354-His336-Ser337, and it is likely that residue Cys354 is directly involved in enzyme activity and is the target site for carbapenems. LdtMt2 also contains two additional segments, A (residues 300-323) and B (residues 379-408), which are not observed in LdtBs and Ldtfm (Supplementary information, Figure S1).
Co-crystallization of LdtMt2-ΔN54 and meropenem failed, but co-crystallization of LdtMt2-ΔN139 and meropenem was successful. The asymmetric unit of LdtMt2- ΔN139-meropenem co-crystals contains two LdtMt2-meropenem complexes. Meropenem adopts different conformations in the two complexes, hereafter referred to as State I (Figure 1B and 1D) and II (Figure 1C and 1E). In both states, meropenem is covalently linked to the catalytic residue Cys354 via thioester formation. The carbapenem nucleus of meropenem appears clearly, displays well-defined electron densities and shows the same conformation in both State I and II (Figure 1B and 1D). The planar arrangement of atoms C2, C3, N4, C5, and C31 of meropenem in complex with LdtMt2 results in the appearance of a double bond between C3 and N4 (Figure 1F), replacing the C2/C3 double bond observed in meropenem12,13. The C3/N4 double bond was also observed in meropenem covalently linked to BlaC6 and other β-lactamases13. This tautomerism of the carbapenem nucleus is accompanied by the covalent linking of meropenem to LdtMt2. As tautomerization of the carbapenem nucleus has been reported to cause slow turnover in carbapenem hydrolyzation by BlaC6 and other β-lactamases13, we propose that tautomerization helps to stablize the LdtMt2-meropenem adduct.
Despite these similarities, the two states are distinct from each other. In State I, atom C7 of meropenem (Figure 1F and Supplementary information, Figure S2) is close to atom N4 at a distance of 2.87 Ã…. This conformation is very similar to that reported previously for meropenem in complex with BlaC6, and likely represents the initial step during the action of meropenem on LdtMt2. In State II, the carbapenem nucleus has flipped approximately 180 degrees around the single bond connecting atoms C5 and C6 (Supplementary information, Figure S3), resulting in a larger C7-N4 distance of 3.21 Ã….
Interestingly, by flipping the carbapenem nucleus in State II, residue Tyr308 of LdtMt2 adopts a different rotamer conformation by rotating approximately 90 degrees around the Cα/Cβ bond (Supplementary information, Figure S3). We suggest that rotating of the Tyr308 side chain in LdtMt2 and flipping of the carbapenem nucleus in meropenem are the second step following the β-lactam ring opening in the first step during the action of meropenem on LdtMt2, and result in three additional hydrogen bonds (Tyr318OH-MerOH62, Tyr318OH-MerO32, and Tyr308OH-MerO32) between the enzyme and meropenem in State II (Figure 1E and Supplementary information, Figure S3). In contrast, only one hydrogen bond (Gly353N-MerO7) is observed in State I (Figure 1D). Residues Tyr308 and Tyr318 that form hydrogen bonds with meropenem are conserved in LdtMt2 and its homologs, but not in other L,D-transpeptidases (Supplementary information, Figure S1). Mutant protein Y318F or Y318A hydrolyzes meropenem at a faster rate than the wild-type protein (Supplementary information, Figure S4), supporting that the additional hydrogen bonds are important for stabilization of the enzyme-drug adduct and indicating that the flipping of meropenem and the resulting additional hydrogen bonds are involved in the mechanism underlying the inhibition of LdtMt2 by meropenem (Figure 1F).
As a main peptidoglycan-crosslinking L,D-transpeptidase, LdtMt2 is an important target in the development of drugs against XDR M. tuberculosis9,14. Our analyses of the LdtMt2-meropenem complex structure reveal a two-step mechanism of drug actions (Figure 1F top). The first step involves opening the β-lactam ring of meropenem and formation of a thioester between meropenem and LdtMt2, accompanied by the simultaneous tautomerization of the C2/C3 double bond to C3/N4 double bond (State I). These events of the first step are also observed in the complex of meropenem and BlaC6 (Figure 1F bottom). During the second step, the C5/C6 single bond of meropenem rotates approximately 180 degrees, and this rotation induces localized structural changes in segment A (including residues Tyr308 and Tyr318) of LdtMt2, leading to the formation of several hydrogen bonds (State II), stabilizing the structure of the LdtMt2-meropenem adduct, and resulting in a slow turnover in drug hydrolysis. This mechanism may be shared by carbapenems, such as meropenem and imipenem, which have been reported to inhibit LdtMt2.
Furthermore, we explored the role of tautomerization in stabilizing the complex of LdtMt2 and antibiotics. When we performed mass spectrometry assays with cephems such as cepholotin and cefuroxime, a subclass of β-lactams that do not inhibit LdtMt2, we found that cephems also form covalent adducts with LdtMt2 (Supplementary information, Table S2 and Figure S6). We thus propose that the mechanism of LdtMt2 action on cephems is the same as that of LdtMt1 action on cephems14 and involves a tautomerization of the cephem nucleus (a β-lactam ring and a dihydrothiazine ring) (Supplementary information, Figure S5A). As cephems cannot inhibit LdtMt2, we suggest that tautomerization alone is not sufficient to stabilize the adduct of LdtMt2 and cephems. To stabilize the adduct of LdtMt2 and carbapenem (or any other potential inhibitor of LdtMt2), tautomerization is necessary but not sufficient, and the conformational change in the second step likely also plays an important role. The results reported here, together with the reported structure of the BlaC-meropenem complex15, reveal the structural basis for the stabilization of enzyme-carbapenem adducts and may suggest new strategies for the design of antibiotics derived from β-lactams to fight against XDR M. tuberculosis.
Accession codes
Coordinates and structure factor files of LdtMt2-ΔN54, LdtMt2-ΔN139, and the LdtMt2-ΔN139-meropenem complex have been deposited in the Protein Data Bank with accession codes 3 VYN, 3 VYO and 3 VYP.
References
Dye C, BG Williams . Science 2010; 328:856–861.
WHO global tuberculosis control report 2010. Summary. Cent Eur J Public Health 2010; 18:237.
Gandhi NR, Nunn P, Dheda K, et al. Lancet 2010; 375: 1830–1843.
Koul A, Arnoult E, Lounis N, et al. Nature 2011; 469:483–490.
Cole ST, Brosch R, Parkhill J, et al. Nature 1998; 393:537–544.
Hugonnet JE, Tremblay LW, Boshoff HI, et al. Science 2009; 323:1215–1218.
Lavollay M, Arthur M, Fourgeaud M, et al. J Bacteriol 2008; 190:4360–4366.
Wietzerbin J, Das BC, Petit JF, et al. Biochemistry 1974; 13:3471–3476.
Gupta R, Lavollay M, Mainardi JL, et al. Nat Med 2010; 16:466–469.
Bielnicki J, Devedjiev Y, Derewenda U, et al. Proteins 2006; 62:144–151.
Biarrotte-Sorin S, Hugonnet JE, Delfosse V, et al. J Mol Biol 2006; 359:533–538.
Drawz SM, RA Bonomo . Clin Microbiol Rev 2010; 23:160–201.
Therrien C, RC Levesque . FEMS Microbiol Rev 2000; 24:251–262.
Dubée V, Triboulet S, Mainardi JL, et al. Antimicrob Agents Chemother 2012; 56:4189–4195.
Tremblay LW, JE Hugonnet, JS Blanchard . Biochemistry 2008; 47:5312–5316.
Acknowledgements
We are grateful for the assistance of P Wu in mass spectrometry analysis and M Fan in crystallographic data collection. We thank the staff at beamline 17A, KEK, Photon Factory, Japan, and beamline 17U, Shanghai Synchrotron Radiation Facility, for assistance with data collection. This work was funded by National Basic Research Program of China (973 Program, 2011CB910300, 2011CB911103 and 2013CB911500) and the Chinese Academy of Sciences (KSCX2-EW-J-3)
Author information
Authors and Affiliations
Corresponding authors
Additional information
( Supplementary information is linked to the online version of the paper on the Cell Research website.)
Supplementary information
Supplementary Information, Figure S1
Structure superposition (A) and sequence alignment (B) of LdtMt2 (green), LdtBs (cyan), and Ldtfm (gray) show that segments A and B are unique in LdtMt2 and its homologs. (PDF 839 kb)
Supplementary Information, Figure S2
Molecular structures of the β-lactam antibiotics described in this manuscript. (PDF 263 kb)
Supplementary Information, Figure S3
Stereo view of meropenem molecules in State I (cyan) and II (gray). (PDF 259 kb)
Supplementary Information, Figure S4
Comparison of the rates of meropenem turnover by LdtMt2 and its mutant proteins. (PDF 271 kb)
Supplementary Information, Figure S5
Proposed mechanisms of cephem and clavulanate action on LdtMt2. (PDF 450 kb)
Supplementary Information, Figure S6
Raw mass spectrometry data for LdtMt2 in complex with different beta-lactams. (PDF 628 kb)
Supplementary Information, Table S1
Data collection, phasing and refinement statistics. (PDF 109 kb)
Supplementary Information, Table S2
Detection of LdtMt2-β-lactam adducts using mass spectrometry. (PDF 88 kb)
Supplementary information, Data S1
Materials and Methods (PDF 155 kb)
Rights and permissions
About this article
Cite this article
Li, WJ., Li, DF., Hu, YL. et al. Crystal structure of L,D-transpeptidase LdtMt2 in complex with meropenem reveals the mechanism of carbapenem against Mycobacterium tuberculosis. Cell Res 23, 728–731 (2013). https://doi.org/10.1038/cr.2013.53
Published:
Issue Date:
DOI: https://doi.org/10.1038/cr.2013.53
This article is cited by
-
Molecular insight on the non-covalent interactions between carbapenems and l,d-transpeptidase 2 from Mycobacterium tuberculosis: ONIOM study
Journal of Computer-Aided Molecular Design (2018)
-
Structural insight into the inactivation of Mycobacterium tuberculosis non-classical transpeptidase LdtMt2 by biapenem and tebipenem
BMC Biochemistry (2017)
-
Non-classical transpeptidases yield insight into new antibacterials
Nature Chemical Biology (2017)