Azithromycin represses evolution of ceftazidime/avibactam resistance by translational repression of rpoS in Pseudomonas aeruginosa

Abstract

Antibiotic combinations can slow down resistance development and/or achieve synergistic therapeutic effects. In this study, we observed that a combined use of ceftazidime-avibactam (CZA) with azithromycin effectively repressed CZA resistance development in Pseudomonas aeruginosa. Transcriptome analysis revealed that subinhibitory concentrations of azithromycin reduced the expression of genes involved in stress-induced mutagenesis, including the stress response sigma factor rpoS. Interestingly, ribosome profiling revealed global redistribution of ribosomes by azithromycin, among which ribosome stalling was significantly intensified near the 5´ terminus of the rpoS mRNA. Further DNA mutational analysis revealed that azithromycin represses the translation of rpoS through its 5´-terminal rare codons, which in turn reduced its transcription. These in vitro observations have been recapitulated in vivo where azithromycin-repressed CZA resistance development when P. aeruginosa was passaged in mice. Overall, our study revealed the molecular mechanism of azithromycin-mediated repression of antibiotic resistance development, providing a promising antibiotic combination for the treatment of P. aeruginosa infections.IMPORTANCEAntibiotic resistance, a global public health challenge, demands the development of novel antibiotics and therapeutic strategies. Ceftazidime-avibactam (CZA) is a combination of a β-lactam antibiotic with a β-lactamase inhibitor that is effective against various gram-negative bacteria such as Pseudomonas aeruginosa. However, clinical CZA-resistant isolates have been reported. Here, we found that combining CZA with azithromycin can effectively suppress the development of resistance in P. aeruginosa in vitro and in vivo. Moreover, we found that azithromycin represses the translation initiation of rpoS through its 5´-terminal rare and less frequent codons, thereby subsequently reducing the mutational frequency of CZA resistance. Therefore, our work provides a promising antibiotic combination for the treatment of P. aeruginosa infections.

Keywords: Pseudomonas aeruginosa; antibiotic combination; ceftazidime-avibactam.

Fig 1

Development of CZA resistance by wild-type PA14. (A) Dynamics of stepwise development of resistance to CZA. Three parallel repeats were performed in the passaging, labeled CZA1–3. (B–E) Determination of the MICs of CZA over the passaging when combined with AZM at indicated ratios. C, CZA; A, azithromycin.

Fig 2

Mutation frequencies of resistance to CZA. (A) The experimental schematic diagram for measuring mutation frequencies. 10⁵ CFU of PA14 were incubated with indicated antibiotics for 6 hours. The bacteria were then resuspended in fresh LB medium without antibiotic and recovered for 20 hours, followed by serial dilution and plating on LB plates with or without CZA (16 mg/L). (B) The frequency of CZA resistance mutation. *, P < 0.05, **, P < 0.01, ***, P < 0.001, ns, not significant, compared to the cells treated with 1 mg/L CZA by Student’s t-test. C, CZA; A, azithromycin.

Fig 3

Azithromycin affects global gene expression. 10⁷ CFU of wild-type PA14 was grown in LB or LB containing 3 mg/L CZA, 16 mg/L AZM, or both of the drugs for 3 hours. Total RNA was extracted for RNA sequencing and analysis (A) or for qRT-PCR to determine the expression levels of genes related to stress-induced mutagenesis (B). C+A, CZA in combination with AZM. Role of RpoS in the mutation frequency of CZA resistance (C). Wild-type PA14, the ΔrpoS mutant, and the complemented strain (ΔrpoS/rpoS_WT) were grown in LB or LB containing 3 mg/L CZA alone or in combination with 16 mg/L azithromycin (C+A). The frequency of CZA-resistant (CZA^R) mutations was assessed. C+A, CZA in combination with AZM. *, P < 0.05, **, P < 0.01, ns, not significant, compared to the cells treated with 3 mg/L CZA by Student’s t-test. C, CZA; A, azithromycin.

Fig 4

Redistribution of ribosome on cellular mRNA following azithromycin treatment. (A) Volcano plot depicting the mRNA counts protected by ribosome with or without azithromycin (16 mg/L) treatment. The x-axis of the graph shows log₂-fold changes in mRNA counts (AZM-treated cells vs untreated cells). Red and blue dots represent significantly increased and decreased ribosome-associated mRNAs, respectively. (B) Ribosome distribution on the rpoS mRNA in wild-type PA14 grown in LB (blue) or LB containing azithromycin (red). The start and stop codons of rpoS were indicated by arrows.

Fig 5

Azithromycin affects the translation of rpoS. (A) Structure of the 6×His-tagged rpoS (rpoS-His). Expression of C-terminal 6×His-tagged rpoS with its native 25-nucleotide sequence of the 5´ untranslated region is driven by an exogenous arabinose-inducible promoter (P_BAD). PA14 containing the rpoS-His fusion was cultured in LB containing 0.1% L-arabinose and azithromycin at the indicated concentrations. When the OD₆₀₀ reached 1.0, the bacteria were collected, and the RpoS-His levels were determined by western blot. (B–D) PA14 containing the indicated rpoS-His fusions was cultivated to an OD₆₀₀ of 1 in the presence of 0.1% L-arabinose and azithromycin at indicated concentrations. Levels of the RpoS-His and RpoA (as loading control) were determined by western blot. (B) The first two to seven codons of rpoS in the P_BAD-rpoS-His fusion were replaced with the corresponding common codons, resulting in P_BAD-rpoS-M_2-7-His. The usage frequencies of the first two to seven codons (shown in red) are 61.8%, 61.45%, 34.61%, 34.61%, 37.08%, and 70.81%, respectively, all representing the most frequently used codons. (C) The rpoS-His fusion with an exogenous ribosome-binding site from pET28a was driven by the P_BAD promoter, resulting in P_BAD-RBS-rpoS-His. Then the first two to seven codons of rpoS were replaced with the corresponding common codons (P_BAD-RBS-rpoS-M_2-7-His). (D) The individual codons at positions 2–7 of the P_BAD-RBS-rpoS-His were replaced with the corresponding common codons (shown in red). The start codon (ATG) of rpoS is underlined.

Fig 6

Roles of the N-terminal codons of rpoS in the azithromycin-mediated repression of translation and mutation frequency. (A) In the rpoS-His with the native 5´-UTR and promoter, the third, fourth, and fifth codons were individually changed to the corresponding common ones, resulting in rpoS-M₃-His, rpoS-M₄-His, and rpoS-M₅-His. PA14 containing the indicated rpoS-His fusions was cultivated to an OD₆₀₀ of 1 in the presence of 0.1% L-arabinose and azithromycin at indicated concentrations. Levels of the RpoS-His and RpoA were determined by western blot. (B) The ΔrpoS mutant complemented with wild-type rpoS (rpoS-His) or the mutated rpoS (rpoS-M₃-His, rpoS-M₄-His, and rpoS-M₅-His) was grown in LB or LB containing 3 mg/L CZA alone or in combination with 16 mg/L azithromycin (C+A). The frequency of CZA-resistant (CZA^R) mutations was assessed. *, P < 0.05, **, P < 0.01, ns, not significant, compared to the cells treated with 1 mg/L CZA by Student’s t-test. The start codon (ATG) of rpoS is underlined. C, CZA; A, azithromycin.

Fig 7

Development of CZA resistance in vivo. (A) The in vivo passage in the neutropenic mice using the pneumonia model. At 72 and 24 hours before infection, the mice were injected with cyclophosphamide. At the time of infection, the mice were intranasally inoculated with 5 × 10⁸ CFU of the clinical P. aeruginosa isolate CI-PA41. At 2 and 10 hours post-infection, the mice were intraperitoneally injected with CZA alone or in combination with azithromycin. At 20 hours post-infection, the mice were euthanized, and bacteria in the lungs were collected, followed by serial dilution and plating on LB agar plates for CFU counting (B). All of the remaining bacteria were plated on LB plates. After 16 hours, bacteria were scraped from the plates and resuspended for the next round of passage (A). (C, D) Another portion of the bacterial suspension was serially diluted, and 15 µL of the bacteria suspension was spotted on LB plates with or without CZA to determine the resistance mutation frequencies. The limit of detection was 67 CFU/mL. ****, P < 0.0001, analysis was conducted using the one-way analysis of variance test. C, CZA; A, azithromycin.