Decarboxylation of the Catalytic Lysine Residue by the C5α-Methyl-Substituted Carbapenem NA-1-157 Leads to Potent Inhibition of the OXA-58 Carbapenemase

Abstract

Antibiotic resistance in bacteria is a major global health concern. The wide spread of carbapenemases, bacterial enzymes that degrade the last-resort carbapenem antibiotics, is responsible for multidrug resistance in bacterial pathogens and has further significantly exacerbated this problem. Acinetobacter baumannii is one of the leading nosocomial pathogens due to the acquisition and wide dissemination of carbapenem-hydrolyzing class D β-lactamases, which have dramatically diminished available therapeutic options. Thus, new antibiotics that are active against multidrug-resistantA. baumannii and carbapenemase inhibitors are urgently needed. Here we report characterization of the interaction of the C5α-methyl-substituted carbapenem NA-1-157 with one of the clinically important class D carbapenemases, OXA-58. Antibiotic susceptibility testing shows that the compound is more potent than commercial carbapenems against OXA-58-producingA. baumannii, with a clinically sensitive MIC value of 1 μg/mL. Kinetic studies demonstrate that NA-1-157 is a very poor substrate of the enzyme due mainly to a significantly reduced deacylation rate. Mass spectrometry analysis shows that inhibition of OXA-58 by NA-1-157 proceeds through both the classical acyl-enzyme intermediate and a reversible covalent species. Time-resolved X-ray crystallographic studies reveal that upon acylation of the enzyme, the compound causes progressive decarboxylation of the catalytic lysine residue, thus severely impairing deacylation. Overall, this study demonstrates that the carbapenem NA-1-157 is highly resistant to degradation by the OXA-58 carbapenemase.

Keywords: Acinetobacter; OXA-58 carbapenemase; X-ray structure; antibiotic resistance; inhibitor.

Figure 1.. Kinetics of the interaction of OXA-58 with NA-1–157.

(A) Representative progress curve under steady-state conditions. (B) Time course of a single turnover reaction for determination of k₂. The inset shows the fast phase of the reaction. (C) Recovery of enzyme activity to determine k₃. (D) Time dependent inactivation of OXA-58 in the presence of NA-1–157. € Progress curves to determine k_off in the presence (light gray) and absence (black) of NA-1–157. (F) The k_inter values plotted versus concentration of NA-1–157 to determine k_NA-1–157 and K_I. The “x” on the y-axis indicates the expected starting point of the reaction. The dashed and solid black lines show the best fit of the data.

Figure 2.. Structure of apo-OXA-58.

(A) Stereoview of the active site of apo-OXA-58. The 2F_o-F_c electron density (blue mesh, 1σ) is shown for the Ser83 and Lys86 side chains. Elongated residual F_o-F_c (green mesh, 3.5σ), and 2F_o-F_c density for Lys86 indicates its carboxylation state. These maps were calculated prior to refinement. (B) Stereoview of the refined apo-OXA-58 structure showing the interactions (black lines) of Lys86^CO2.

Figure 3.. The 2.5 h OXA-58-NA-1–157 complex.

(A) Structures of NA-1–157 and meropenem. (B) Residual F_o-F_c electron density (green mesh, 3.5σ) in the OXA-58 active site. (C) The refined OXA-58-NA-1–157 2.5 h complex. The hydrogen bonding interactions are shown. (D) Active site of the OXA-58-NA-1–157 complex showing the final 2F_o-F_c density (blue mesh, 1σ). (E) Superposition of the OXA-58-NA-1–157 complex (blue) onto the apo-OXA-58 structure (yellow).

Figure 4.. Comparison of the NA-1–157 and meropenem complexes of OXA-58.

(A) Superposition of the OXA-58-meropenem complex (7vvi, brown) onto the 2.5 h OXA-58-NA-1–157 complex (blue). (B) Close-up view of panel (A). Close contacts are shown as red dashed lines. Magenta sticks indicate residues of the OXA-58-meropenem complex.

Figure 5.. Structures of OXA-58-NA-1–157 complexes at different time points after soaking.

(A) 1.5 min and (B) 7.5 min 2F_o-F_c density for the Lys86 side chain. The K86^CO2 in the apo-OXA-58 structure is shown as pink sticks. (C) Monomer C at 10 min showing two partially occupied CO₂ molecules and also a partially occupied water (Wat3). (D) Monomer B at 10 min showing the presence of two fully occupied Wat1 and Wat3 molecules. Hydrogen-bonding interactions are shown as black dashed lines. In all panels, the 2F_o-F_c density is contoured at 1σ.

Figure 6.. Solvent accessible surfaces in the active site.

(A) Apo-OXA-58, with the hydrophobic cap residues Val132 and Leu170 in pink and Ser83 in red. (B) Panel (A) rotated 90° about the vertical axis, showing a cutaway view into the lysine pocket. (C) The 20 min OXA-58-NA-1–157 complex, showing a hole in the hydrophobic cap. (D) Panel (C) rotated 90° about the vertical axis, showing the channel linking the lysine pocket with the active site. (E) Stereoview of the superposition of the 20 min OXA-58-NA-1–157 complex (blue) onto apo-OXA-58 (yellow). C–O and C–C nonbonded contact distances are 3.22 and 3.4 Å, respectively. Close nonbonded contacts between NA-1–157 and residues in apo-OXA-58 are shown as dashed red lines.

Figure 7.. Molecular docking and catalytic mechanism.

(A) Stereoview of the pre-acylation tetrahedral transition state models for OXA-58 with meropenem and NA-1–157. Hydrogen bonds are shown as black dashed lines. The Ser130-N4 contact distances are shown as green and red dashed lines for meropenem and NA-1–157, respectively. (B) Mechanism of carbapenem hydrolysis by OXA-58.

Scheme 1.

Minimal Reaction Pathway for the Interaction of OXA-58 with NA-1–157