Bacterial strains

The K. pneumoniae 0692 strain used in this study, harboring two carbapenemases, NDM-1 and KPC-2, was isolated from the First People’s Hospital of Yunnan Province, China. It was identified simultaneously using the VITEK 2 Compact and VITEK MS systems (bioMérieux, Lyon, France).

Bacterial drug-resistant gene detection

Plasmid DNA was extracted using a plasmid extraction kit (Tiangen, Beijing, China). The blaNDM and blaKPC plasmids were identified using polymerase chain reaction (PCR) and sequencing.

DNA extraction and PCR

The extraction of genomic DNA from K. pneumoniae 0692 isolates was performed with a bacterial genomic DNA extraction kit (Tiangen, Beijing, China). Table 1 lists the specific primers. Set up the PCR reaction system and amplification programme according to the Master Mix instructions (TSINGKE, China).

Table 1 List of primers used in this studyWhole-genome sequencing of K. pneumoniae 0692 and phages

The genomic DNA of the clinically isolated K. pneumoniae 0692 strain was extracted using the Bacterial Genomic DNA Extraction Kit (Tiangen, China) and sequenced using the PacBio Sequel II/PacBio Sequel IIe and Illumina NovaSeq PE150 sequencing platforms. Similarly, phage genomic DNA was extracted using the Phage λ Genomic DNA Extraction Kit (Yuanye Biology, China) and sequenced using Illumina NovaSeq sequencing platforms following the manufacturer’s instructions. The sequencing results were compared with NCBI blast (https://blast.ncbi.nlm.nih.gov/Blast.cgi).

Determination of the minimum inhibitory concentrations (MICs)

The MICs of antimicrobial agents were determined using the microdilution susceptibility testing method. The following antibiotics were tested: imipenem, meropenem, ertapenem, tigecycline, ceftazidime/avibactam, cefoxitin, amikacin, and aztreonam. MIC results were interpreted as specified by Clinical Laboratory Standards Institute (CLSI) guidelines [14], except for tigecycline, which followed the interpretation defined by the US Food and Drug Administration (susceptible: MIC ≤ 2 mg/L; resistant: MIC ≥ 8 mg/L). Escherichia coli ATCC 25922 and Pseudomonas aeruginosa ATCC 27853 were used as quality controls.

Isolation and purification of phages

Phage isolation and purification were performed using the double-layer plate method, as previously described [15]. In brief, 100 μL of K. pneumoniae 0692 microbial suspension from an overnight culture and 100 μL of untreated sewage filtered through a 0.22 μm membrane filter (Thermo Fisher Scientific, USA) were mixed with 10 mL of fresh LB medium (Coolaber, China). The mixture was incubated in a shaking incubator at 160 rpm and 37 °C, and the resulting culture solution was filtered through a 0.22 μm filter to retain the filtrate. The lower plate of LB solid medium (2% agar) was prepared, and the upper plate was prepared by mixing 100 μL of overnight cultured K. pneumoniae 0692 and 100 μL of culture filtrate in semi-solid LB medium (0.5% agar), then spreading the plate and incubating at 37 °C overnight. Transparent round phage spots on top of the background of the host bacterium were observed. Single phage spots were picked and mixed into the K. pneumoniae 0692 microbial suspension to prepare double-layer plates for further purification. At least three rounds of purification were performed to efficiently lyse K. pneumoniae 0692. Plates with 30–300 phage spots were selected for counting. Titer (potency) = (number of phage spots × dilution × 10) PFU/mL.

Isolation and purification of OMVs

OMVs were prepared through density gradient ultracentrifugation in the absence or presence of 2 μg/mL tigecycline (MedChemExpress, Monmouth Junction, NJ, USA) and phage, following previously established protocols with minor adjustments [16]. Briefly, single colonies grown on LB agar plates were inoculated into 20 mL of LB broth medium (OD600 = 0.4). Subsequently, 1 mL of the microbial suspension was inoculated into 300 mL of antibiotic (2 μg/mL tigecycline) or phage-containing (1 mL of phage [1.42 × 1010 PFU/mL] solution added to 300 mL of LB medium) solutions and then incubated at 37 °C with shaking at 150 rpm for 16 h. Cultures were taken at 8000 rpm for 30 min at 4 °C to remove bacteria and cellular debris. The supernatant after centrifugation was collected and filtered using a 0.22 μm filter to remove fine debris. After the supernatant was ultracentrifuged at 25,000 rpm for 50 min at 4 °C in a P100AT2 rotor (Hitachi CP-100WX, Japan), the pellet was resuspended in 8 mL of phosphate-buffered saline (PBS) (Servicebio, China). OptiPrep™ (Sigma-Aldrich, Germany) was used to prepare 10%, 20%, and 30% density gradient solutions, which were distributed into discrete 10–30% density gradient layers. The resuspended OMV samples were placed on the density gradient and centrifuged at 4 °C and 25,000 rpm for 2 h. The OMV layer was collected and purified by washing with a tenfold volume of PBS solution at 25,000 rpm and 4 °C for 30 min. The resulting pellet was resuspended with 5 mL of PBS to obtain purified OMVs. The purified OMV solution was passed through a 0.22 μm membrane filter and filtered again. External DNA and proteins were removed using 2U DNase I (Biosharp, China) and 100 μg/mL Protease K (Biosharp, China). Purified OMV solution was inoculated onto Columbia blood plates to verify sterility.

Transmission electron microscopy

Purified phages and OMVs were observed through transmission electron microscopy (TEM). Samples were prepared by applying 10 μL of phage or OMVs suspended in PBS onto a carbon-coated copper grid for 5 min. After drying, the samples were stained with 2% uranyl acetate (Bio-Rad, Hercules, CA, USA) for 2 min, and the excess liquid was blotted. Observations were performed using a JEM-1400 Plus TEM (JEOL, Tokyo, Japan), operating at 80 kV.

Protein quantification of OMVs

Approximately 50 μL of OMVs were treated with 2U of DNase I (Biosharp, China) and 100 μg/mL of Protease K (Biosharp, China), followed by the addition of 50 μL of RIPA lysate (Beyotime, China). The mixture was then incubated at 4 °C for 30 min and subsequently centrifuges at 10,000×g at 4 °C for 15 min. A bicinchoninic acid (BCA) protein assay kit (Biosharp, China) was used to determine the protein concentration of OMVs samples, according to the manufacturer’s protocols.

Mass spectrometry assay of OMVs

The mass spectrometry experimental analysis process involved protein extraction, peptide enzymolysis, liquid chromatography-tandem mass spectrometry, and database retrieval. Briefly, SDT buffer (4% SDS, 100 mM Tris–HCl, pH 7.6; Bio-Rad, USA) was directly added to the different samples. DTT (with a final concentration of 40 mM; Bio-Rad, USA) was added to each sample individually and mixed at 600 rpm for 1.5 h (37 °C). After the samples were cooled to room temperature, iodoacetamide (Bio-Rad, USA) was added at a final concentration of 20 mM to block the reduced cysteine residues. Subsequently, the samples were incubated for 30 min in the dark, then transferred to filters (MicroCon units, 10 kDa). The filters were washed thrice with 100 μL UA buffer, then twice with 100 μL 25 mM NH4HCO3 (Sigma, Germany) buffer. Finally, trypsin was added to the samples (at a trypsin: protein (wt/wt) ratio of 1:50; Promega, Madison, WI, USA) and incubated at 37 °C for 15–18 h (overnight). The resulting peptides were collected as the filtrate. The peptides from each sample were analyzed using an Orbitrap™ Astral™ mass spectrometer (Thermo Scientific, USA) connected to a Vanquish Neo system liquid chromatograph (Thermo Scientific, USA) in data-independent acquisition (DIA) mode. DIA data were analyzed using DIA-NN 1.8.1 software.