Phage-antibiotic synergy suppresses resistance emergence of Klebsiella pneumoniae by altering the evolutionary fitness

Abstract

Phage-antibiotic synergy (PAS) represents a superior treatment strategy for pathogen infections with less probability of resistance development. Here, we aim to understand the molecular mechanism by which PAS suppresses resistance in terms of population evolution. A novel hypervirulent Klebsiella pneumoniae (KP) phage H5 was genetically and structurally characterized. The combination of H5 and ceftazidime (CAZ) showed a robust synergistic effect in suppressing resistance emergence. Single-cell Raman analysis showed that the phage-CAZ combination suppressed bacterial metabolic activities, contrasting with the upregulation observed with phage alone. The altered population evolutionary trajectory was found to be responsible for the contrasting metabolic activities under different selective pressures, resulting in pleiotropic effects. A pre-existing wcaJ point mutation (wcaJ_G949A) was exclusively selected by H5, conferring a fitness advantage and up-regulated activity of carbohydrate metabolism, but also causing a trade-off between phage resistance and collateral sensitivity to CAZ. The wcaJ point mutation was counter-selected by H5-CAZ, inducing various mutations in galU that imposed evolutionary disadvantages with higher fitness costs, and suppressed carbohydrate metabolic activity. H5 and H5-CAZ treatments resulted in opposite effects on the transcriptional activity of the phosphotransferase system and the ascorbate and aldarate metabolism pathway, suggesting potential targets for phage resistance suppression. Our study reveals a novel mechanism of resistance suppression by PAS, highlighting how the complexity of bacterial adaptation to selective pressures drives treatment outcomes.

Importance: Phage-antibiotic synergy (PAS) has been recently proposed as a superior strategy for the treatment of multidrug-resistant pathogens to effectively reduce bacterial load and slow down both phage and antibiotic resistance. However, the underlying mechanisms of resistance suppression by PAS have been poorly and rarely been studied. In this study, we tried to understand how PAS suppresses the emergence of resistance using a hypervirulent Klebsiella pneumoniae (KP) strain and a novel phage H5 in combination with ceftazidime (CAZ) as a model. Our study reveals a novel mechanism by which PAS drives altered evolutionary trajectory of bacterial populations, leading to suppressed emergence of resistance. The findings advance our understanding of how PAS suppresses the emergence of resistance, and are imperative for optimizing the efficacy of phage-antibiotic therapy to further improve clinical outcomes.

Keywords: evolutionary trajectory; hypervirulent Klebsiella pneumoniae; metabolic activity; phage-antibiotic synergy.

Fig 1

Three-dimensional structural analysis of H5 phage by cryo-EM. Negative staining TEM (A) and cryo-EM images (B) of the H5 phage. (C) Fivefold surface view of corresponding 3D reconstructions. The surfaces are colored by radius. The color key indicates the scale (in angstroms) of the radial color scheme. (D) A cut-away view reveals the internal density of the inside content of the H5 phage. (E) The Fourier shell correlation curves for gold standard reconstructions of the full reconstructed phage. (F) A model of the asymmetric unit of H5 phage. (G) The atomic model a of the H5 phage capsid gp10 protein (rainbow) with T7 phage gp10 (gray) superimposed. (H) Atomic model of the entire H5 phage capsid in the view as shown in (B).

Fig 2

Effect of antibiotic class on PAS. Bacteria in the log phase were inoculated into a 96-well plate coated with H5 and antibiotics, and the OD₆₀₀ was measured every 10 min for a total of 24 h with shaking. The synergistic effect of H5 was estimated with (A) ceftazidime and meropenem; (B) ciprofloxacin and levofloxacin; and (C) chloramphenicol and kanamycin. Synograms (t = 24 h) represent the mean percentage reduction of each treatment from three biological replicates: reduction (%) = [(OD_{growth control} − OD_treatment)/OD_{growth control}] × 100.

Fig 3

Physiological changes of bacteria in response to the phage H5 and CAZ synergy. (A) Effect of combinatorial treatment on bacterial regrowth. Effect of CAZ concentrations on cell size (B) and plaque size (C). The number of cells or plagues counted for each concentration is indicated. UMAP representation of phenotypic responses of individual cells to different CAZ and phage treatments based on the whole Raman fingerprint region (D) and C–D region (E). (F) Single-cell metabolic activity under different treatments was revealed by single-cell Raman with D₂O labeling. Each point is a measurement of a single cell, and box plots show the quartiles of each evolved population. Statistical analysis was performed using Mann-Whitney U tests or one-way ANOVA followed by Tukey HSD tests. ***, P < 0.001.

Fig 4

Effect of PAS on a bacterial evolutionary trajectory. (A) The phage adsorption rate and EOP of WT and mutants. (B) Phage resistance of 43816ΔwcaJ and different complements indicated by EOP. (C) Phage resistance of 43816ΔgalU and different complements indicated by EOP. (D) The effect of point mutation G949A on the 3D structure of WcaJ. Predicted binding conformation of UDP-D-glucose at the substrate-binding site of the modeled WcaJ (E) and WcaJ_G949A (F) structure. The side chains of residues connected with UDP-D-glucose with hydrogen bonds are indicated in the enlarged views. (G) The susceptibility of WT and mutants to CAZ. (H) The susceptibility of 43816ΔwcaJ and different complements to CAZ. (I) The susceptibility of 43816ΔgalU and different complements to CAZ. Susceptibility was defined as the ratio of the number of viable cells in the mutants compared to WT strains under CAZ (0.25 mg/L) treatment. A phage that produces an equal number of plaques on WT has an EOP of 1.0 (dotted line). A “bd” at the lower, dashed line indicates that the EOP was below the limit of detection (~10⁻⁸). Statistical analysis was performed using one-way ANOVA tests. ^**, P < 0.01 and ***, P < 0.001.

Fig 5

Impact of PAS on fitness cost and biological processes of K. pneumonia ATCC 43816. The area under the growth curves (A) and competitiveness (B) of WT (ATCC 438186) and two mutants (P13 and CP6) and the competition index was defined as the ratio of the number of viable cells in the mutants to the WT in mixed culture. The relative intensity of major biomolecules in mutants (C) and gene knockout strains (43816ΔwcaJ and 43816ΔgalU) (D). Each point is a measurement of a single cell, and box plots show the quartiles of each evolved population. Statistical analysis was performed using one-way ANOVA tests. ^*, P < 0.05; ^**, P < 0.01; and ^***, P < 0.001.