Novel C5α-substituted carbapenems enhance Mycobacterium abscessus killing via selective target binding and reduced hydrolysis by BlaMab

Abstract

Mycobacterium abscessus (Mab) poses significant clinical challenges and underscores the urgent need for safer and more effective treatments, including β-lactams. Among currently available carbapenems, imipenem is widely used to treat Mab infections by combining with other antibiotics. Commercial carbapenems share a common scaffold with C2 modifications, whereas this study focuses on novel carbapenem candidates with C5α modifications. We evaluated their antibacterial activity against Mab ATCC 19977 and clinical isolates, as well as their acylation of peptidoglycan target receptors (L,D-transpeptidases [LDTs] and penicillin-binding proteins [PBPs]) and the β-lactamase enzyme Bla_Mab. In vitro studies of two C5α-modified carbapenems, JDB/NA-1-157 and JDB/NA-1-208, revealed distinct antibacterial effects. JDB/NA-1-157 demonstrated potent bacterial killing with low minimum inhibitory concentrations (MICs; 0.125-8 mg/L) and near-complete eradication within 5 days, surpassing the efficacy of the standard-of-care regimen (amikacin + clarithromycin + imipenem). In contrast, JDB/NA-1-208 exhibited poor bacterial killing, with high MICs (16-256 mg/L) and limited efficacy in time-kill studies. However, JDB/NA-1-208 showed synergistic killing when combined with other β-lactams. Mechanistically, JDB/NA-1-208 is not a substrate for Bla_Mab, while JDB/NA-1-157 is, albeit with low catalytic efficiency. This is supported by the observation that the addition of avibactam did not enhance synergy with JDB/NA-1-157. The substantial bacterial killing effect of JDB/NA-1-157 is attributed to its high binding affinity for PBP-B, PBP-lipo, PonA2, D,D-carboxypeptidase, and LDT1-2. These findings highlight the potential of novel C5α-modified carbapenems, particularly JDB/NA-1-157, as promising therapeutic candidates for Mab infections.

Keywords: L,D-transpeptidases; Mycobacterium abscessus; carbapenems; penicillin-binding proteins.

Fig 1

Chemical structures of (A) JDB/NA-1-157 and (B) JDB/NA-1-208.

Fig 2

Thermal stability of imipenem (IPM), meropenem (MEM), JDB/NA-1-157, and JDB/NA-1-208 was evaluated at 100 mg/L in 7H9 broth at 30°C over 24 h. JDB/NA-1-208 demonstrated the highest thermal stability, with approximately 10% degradation observed after 24 h. JDB/NA-1-157 exhibited 16% degradation over the same period.

Fig 3

MIC distribution of JDB/NA-1-157 and JDB/NA-1-208 as single agent or in combination against 18 clinical isolates and Mab ATCC 19977. (A) JDB/NA-1-157 (red) demonstrated MIC values ranging from 0.125 to 8.0 mg/L, while JDB/NA-1-208 (blue) exhibited significantly higher MIC values, ranging from 16 to >128 mg/L. (B) JDB/NA-1-157, when combined with avibactam (AVI) 4 mg/L, durlobactam (DUR) 1 mg/L, cefuroxime (CXM) 2 mg/L, or amoxicillin (AMX) 4 mg/L + AVI 4 mg/L, resulted in a slight reduction in MIC values, though they remained similar to monotherapy. JDB/NA-1-208, in combination with CXM 2 mg/L or AMX 4 mg/L + AVI 4 mg/L, showed a decrease in MIC for some isolates compared to single JDB/NA-1-208.

Fig 4

Time-kill curves of JDB/NA-1-157 as a single agent (A); in combination with β-lactamases, cephalosporins, or penicillin (B); JDB/NA-1-208 as a single agent (C); and in combination with β-lactamases, cephalosporins, or penicillin (D) against Mab ATCC 19977. (E) Time-kill curves without JDB/NA-1-157 or JDB/NA-1-208, demonstrating the effects of avibactam (AVI) 4 mg/L, durlobactam (DUR) 12 mg/L, ceftaroline (CFT) 8 mg/L, cefuroxime (CXM) 8 mg/L, amoxicillin (AMX) 2 mg/L, and AMX 2 mg/L + AVI 4 mg/L. Standard-of-care regimens, including amikacin (AMK) 12 mg/L + clarithromycin (CLM) 0.3 mg/L + cefoxitin (FOX) 7 mg/L or AMK 12 mg/L + CLM 0.3 mg/L + imipenem (IPM) 12 mg/L, were used as comparators to assess the efficacy of JDB/NA-1-157 and JDB/NA-1-208.

Fig 5

The efficacy of JDB/NA-1-157 and JDB/NA-1-208, either as a single agent or in combination with β-lactamases, cephalosporins, or penicillin, was evaluated using time-kill assay against selected clinical isolates Mab122 (A and B), Mab 686 (C and D), and UH01 (E and F).

Fig 6

Binding of (A) JDB/NA-1-157 and (B) JDB/NA-1-208 to Bla_Mab, LDT1–5, D,D-carboxypeptidase, PBP B, and PBP-lipo. The intensities of the PBP-lipo complex bands were quantified using GelAnalyzer software and normalized to the intensity of PBP-lipo without β-lactams (PBP-lipo Apo), which was set to 1.

Fig 7

Mechanism of LDT1 acylation by JDB/NA-1-157 and JDB/NA-1-208, followed by the subsequent elimination of the substituent, resulting in an observed Δm = +86 (±5) Da adduct.

Fig 8

BlaMab with IPM (blue) and MEM (green) as acyl-enzyme complexes (A) and the Michaelis–Menten (MM) complex of JDB/NA-1-157 (B) show W105 in two conformations (t-105 “up” and m-95 “down”). In the MM complex, JDB/NA-1-157 adopts conformations with the carbonyl group oriented toward the oxyanion hole, stabilized by hydrophobic interactions between the indole ring of W105 and a methyl group at C5α. In the acyl-enzyme (C and D), the JDB/NA-1-157 core adopts a restrained conformation, with π-methyl stacking interactions with W105 (C) and with the hydroxyethyl group interacting with F237 and N132 (D). The methyl group maintains its interaction with W105, regardless of its rotamer conformation. Hydrogen bonds with S130, K232, T233, and the oxyanion hole carbonyl are preserved. The JDB/NA-1-157 tail appears more flexible to accommodate conformational changes in the active site.

Fig 9

LDT2 and JDB/NA-1-157 acyl-enzyme complex, with JDB/NA-1-157 covalently bound to C351. The carbonyl group forms hydrogen bonds with C351 and G350, while the hydroxyethyl group forms hydrogen bonds with G329. The new C5α methyl group interacts with M300 and Y331, and the tail interacts with S328. Y305, located at the entrance of the active site, may help stabilize the JDB/NA-1-157 compound in a productive conformation through stacking interactions with the pyrrolidine ring.