Sulfamoyl Heteroarylcarboxylic Acids as Promising Metallo-β-Lactamase Inhibitors for Controlling Bacterial Carbapenem Resistance

Abstract

Production of metallo-β-lactamases (MBLs), which hydrolyze carbapenems, is a cause of carbapenem resistance in Enterobacteriaceae Development of effective inhibitors for MBLs is one approach to restore carbapenem efficacy in carbapenem-resistant Enterobacteriaceae (CRE). We report here that sulfamoyl heteroarylcarboxylic acids (SHCs) can competitively inhibit the globally spreading and clinically relevant MBLs (i.e., IMP-, NDM-, and VIM-type MBLs) at nanomolar to micromolar orders of magnitude. Addition of SHCs restored meropenem efficacy against 17/19 IMP-type and 7/14 NDM-type MBL-producing Enterobacteriaceae to satisfactory clinical levels. SHCs were also effective against IMP-type MBL-producing Acinetobacter spp. and engineered Escherichia coli strains overproducing individual minor MBLs (i.e., TMB-2, SPM-1, DIM-1, SIM-1, and KHM-1). However, SHCs were less effective against MBL-producing Pseudomonas aeruginosa Combination therapy with meropenem and SHCs successfully cured mice infected with IMP-1-producing E. coli and dually NDM-1/VIM-1-producing Klebsiella pneumoniae clinical isolates. X-ray crystallographic analyses revealed the inhibition mode of SHCs against MBLs; the sulfamoyl group of SHCs coordinated to two zinc ions, and the carboxylate group coordinated to one zinc ion and bound to positively charged amino acids Lys224/Arg228 conserved in MBLs. Preclinical testing revealed that the SHCs showed low toxicity in cell lines and mice and high stability in human liver microsomes. Our results indicate that SHCs are promising lead compounds for inhibitors of MBLs to combat MBL-producing CRE.IMPORTANCE Carbapenem antibiotics are the last resort for control of severe infectious diseases, bloodstream infections, and pneumonia caused by Gram-negative bacteria, including Enterobacteriaceae However, carbapenem-resistant Enterobacteriaceae (CRE) strains have spread globally and are a critical concern in clinical settings because CRE infections are recognized as a leading cause of increased mortality among hospitalized patients. Most CRE produce certain kinds of serine carbapenemases (e.g., KPC- and GES-type β-lactamases) or metallo-β-lactamases (MBLs), which can hydrolyze carbapenems. Although effective MBL inhibitors are expected to restore carbapenem efficacy against MBL-producing CRE, no MBL inhibitor is currently clinically available. Here, we synthesized 2,5-diethyl-1-methyl-4-sulfamoylpyrrole-3-carboxylic acid (SPC), which is a potent inhibitor of MBLs. SPC is a remarkable lead compound for clinically useful MBL inhibitors and can potentially provide a considerable benefit to patients receiving treatment for lethal infectious diseases caused by MBL-producing CRE.

Keywords: CRE; carbapenems; metallo-β-lactamase; sulfamoyl heteroarylcarboxylic acids.

FIG 1

SFC inactivates B1 MBLs. (A) Summary plot of all 22,671 compounds; values are demonstrated as residual ratios. The dashed line indicates a temporary cutoff value (residual ratio, 0.6) for initial selection of an effective IMP-1 enzyme inhibitor. (B) Representative results of the susceptibility test for IMP-1-producing E. coli cells (E. coli DH5α/pBC-IMP-1). The MIC value of MPM was determined in the presence of 10 μg/ml of the tested compounds. The chemical structure of the 243-G03 compound, 2,5-dimethyl-4-sulfamoylfuran-3-carboxylic acid (SFC), is shown. (C) Inhibition of subclass B1 (IMP-1, NDM-1, and VIM-2), B2 (SFH-1), and B3 (L1 and SMB-1) MBLs by SFC. Data represent the means ± standard deviations (SD) of results from three replicate experiments. (D) Heat maps obtained from checkerboard analyses of IMP-1-producing E. coli strains (E. coli DH5α/pBC-IMP-1, MPM MIC = 1.0 μg/ml), NDM-1-producing (E. coli DH5α/pBC-NDM-1, MPM MIC = 64 μg/ml), and VIM-2-producing (E. coli DH5α/pBC-VIM-2, CAZ MIC = 16 μg/ml).

FIG 2

SFC rescues the activity of MPM. (A to D) Plots showing the MPM MIC values for 19 IMP-type MBL-producing Enterobacteriaceae (A), 8 Acinetobacter spp. (B), 17 P. aeruginosa strains (C), and 14 NDM-type MBL-producing Enterobacteriaceae (including 1 dually NDM/VIM-producing K. pneumoniae strain) (D). MIC values greater than or equal to those represented by the dense dotted lines indicate the “resistant” (R) criteria according to CLSI guidelines, while those less than or equal to those represented by the spaced dotted lines indicate the “susceptible” (S) criteria. Solid lines represent the MIC₅₀ values. (E) Time-kill curves of IMP-1-producing E. coli NUBL-24 strain in the presence of MPM or SFC alone or their combination during a 24-h incubation. Data represent the means ± SD of results from three independent experiments. The symbol “#” indicates the detection limit (200 CFU/ml). (F) Representative cell morphology images of E. coli NUBL-24 after exposure to MPM and/or SFC. Scale bars = 5 μm. Arrows indicate round cells. (G) Cytotoxicity of SFC and staurosporine in HeLa cells. Abs, absorbance. (H) Dose-response mortality curve of intravenously injected SFC in mice (n = at least 5 mice per dose). (I) Mouse survival curves from evaluations of the therapeutic effect of MPM (0.2 or 0.4 mg/kg of body weight) or SFC (100 mg/kg) alone or in combination. Mice were intraperitoneally infected with IMP-1-producing E. coli NUBL-24 (5 × 10⁶ CFU; n = 10 mice per group). Statistical analyses of Kaplan-Meier survival curves were performed with a log rank test using the SigmaPlot 13 suite (Hulinks). *, P = 0.01; **, P < 0.001.

FIG 3

Mode of MBL inhibition by SFC. (A) Schematic representation of the overall structure of IMP-1 in complex with SFC and interactions between IMP-1 and SFC. The |F_o| − |F_c| omit map of SFC, which was contoured at 3.0σ (red mesh), is shown. SFC is illustrated in green (carbon), ochre (sulfur), red (oxygen), and blue (nitrogen) sticks. The amino acids of IMP-1 are represented by silver sticks. Zinc ions are illustrated as orange spheres. Black and orange dashed lines indicate hydrogen and coordination bonds, respectively. (B) Interactions between NDM-1 and SFC. The amino acids of NDM-1 are represented by cyan sticks. The SFC molecule is shown as described for panel A. (C) Interactions between VIM-2 and SFC. The amino acids of VIM-2 are illustrated using deep-green-colored sticks. The SFC molecule is shown as described for panel A. (D) Surface representation of IMP-1 (shown in transparent gray). The SFC molecule is shown as described for panel A. Trp64 and His263 are represented by silver sticks. (E) Surface representation of NDM-1 (shown in transparent cyan). The SFC molecule is shown as described for panel A. His263 is represented by cyan sticks. (F) Summary of the binding mode between subclass B1 MBLs (IMP-1, NDM-1, and VIM-2) and SFC.

FIG 4

In vitro evaluation of synthesized SHCs and mode of NDM-1/VIM-2 inhibition by SPC. (A and B) Chemical structures of 10 SHCs (A) and their evaluations (B). K_i values represent the means of results from three replicates. E. coli DH5α/pBC-IMP-1, E. coli DH5α/pBC-NDM-1, and E. coli DH5α/pBC-VIM-2 were used for the determination of MIC values. The colors of the highlighted rows correspond to the colors highlighting chemical structures in panel A. (C) Interactions between NDM-1 and SPC. The |F_o| − |F_c| omit map of SPC, which was contoured at 3.0σ (yellow mesh), is shown. SPC is illustrated using yellow (carbon), ochre (sulfur), red (oxygen), and blue (nitrogen) sticks. The amino acids of NDM-1 are represented by cyan sticks. Zinc ions are illustrated as orange spheres. Black and orange dashed lines indicate hydrogen and coordination bonds, respectively. The distances between SPC and the amino acids of L3 loop are illustrated as blue dashed lines together with numbers (Å). (D) Interactions between VIM-2 and SPC. The amino acids of VIM-2 are represented by deep-green-colored sticks. Other architecture is colored as described for panel C. (E) Accumulation assay of SFC and SPC incorporated in E. coli K-12 MG1655. Data represent the means ± SD of results from three independent experiments. Statistical significance was determined by Welch’s t test. *, P < 0.01.

FIG 5

SPC inactivates clinically relevant MBLs. (A and B) Heat maps obtained from checkerboard analyses of the IMP-1-producing E. coli NUBL-24 strain by the use of SFC (A) and SPC (B). (C) Summary of the reduction in MPM MIC values (fold changes) for 19 IMP-type MBL-producing Enterobacteriaceae isolates in the presence of 4 μg/ml SPC and SFC. (D and E) Heat maps obtained from checkerboard analyses of the dually NDM-1/VIM-1-producing K. pneumoniae MS5674 strain using SFC (D) and SPC (E). (F) Summary of the reduction in MPM MIC values (fold changes) for 14 NDM-type MBL-producing Enterobacteriaceae isolates in the presence of 4 μg/ml SPC and SFC. The statistical analyses whose results are indicated in panels C and F were performed by the use of the Wilcoxon signed-rank test and the JMP Pro 14 suite (SAS). *, P < 0.01. (G) Cytotoxicity of SPC and staurosporine in HeLa cells. (H and I) Mouse survival curves from evaluations of the therapeutic effect of MPM alone (0.8 mg/kg) or MPM-SPC (0.8 and 10 mg/kg, respectively) on IMP-1-producing E. coli NUBL-24 (1 × 10⁷ CFU (H) and MPM alone (4 mg/kg) or MPM-SPC (4 and 10 mg/kg, respectively) on dually NDM-1/VIM-1-producing K. pneumoniae MS5674 (1 × 10⁷ CFU) (I). n = 10 mice per group. Statistical analyses of Kaplan-Meier survival curves were performed with a log rank test by the use of the SigmaPlot 13 suite (Hulinks). *, P < 0.01; **, P < 0.001.