Outer membrane vesicles mediating horizontal transfer of the epidemic bla_OXA-232 carbapenemase gene among Enterobacterales

ABSTRACT

OXA-232 is one of the most common OXA-48-like carbapenemase derivatives and is widely disseminated in nosocomial settings across countries. The bla_OXA-232 gene is located on a 6-kb non-conjugative ColKP3-type plasmid, while the dissemination of bla_OXA-232 into different Enterobacterales species and the polyclonal dissemination of OXA-232-producing K. pneumoniae revealed the horizontal transfer of bla_OXA-232. However, it’s still unclear how this non-conjugative ColKP3 plasmid could facilitate the mobilization of bla_OXA-232. Here, we observed the in vivo intraspecies transfer of bla_OXA-232 during a nosocomial outbreak of OXA-232-producing K. pneumoniae. We demonstrated the presence of ColKP3 OXA-232 plasmid in the outer membrane vesicles (OMVs) derived from clinical isolates, and OMVs could facilitate the horizontal transfer of bla_OXA-232 among Enterobacterales. In contrast, for the most prevalent carbapenemase genes, including bla_KPC-2 and bla_NDM-1, though the presence of carbapenemase genes and plasmid backbones in the vesicular lumen was observed, OMVs couldn’t promote effective transformation, probably due to the low copy number of plasmids in clinical isolates and the low number of plasmids loaded into vesicles. Conjugation assay revealed that the epidemic IncX3 NDM-1 and IncFII(pHN7A8)/IncR KPC-2 plasmids were conjugative and could be horizontally transferred via independent conjugation or with the help of a co-existent conjugative plasmid. For the large-size and low-copy number conjugative plasmids carrying carbapenemase genes, OMVs-mediated gene exchange may only serve as an alternative pathway for horizontal transfer. In conclusion, diverse mobilization strategies were employed by plasmids harbouring carbapenemase genes, and plasmids display a proper choice of mobility pathway due to their individual properties.

KEYWORDS:

• OXA-232
• outer membrane vesicles
• horizontal gene transfer
• carbapenemase genes
• plasmid mobility

Introduction

The emergence and global dissemination of carbapenem-resistant Klebsiella pneumoniae (CRKP) pose a significant therapeutic challenge to public health [1]. The production of carbapenemases is the most crucial cause of carbapenem resistance in K. pneumoniae [2]. Carbapenemases could efficiently hydrolyze carbapenems and other commonly used β-lactams and are associated with multi- or pan-drug resistance [3]. Following KPC and NDM, the OXA-48 group is one of the most common carbapenemases, mainly distributed within specific regions [4]. OXA-232 is one of the most common OXA-48-like carbapenemase derivatives and was described in 2013 from E. coli and K. pneumoniae [5]. To date, K. pneumoniae and E. coli remain the most common bacteria associated with bla_OXA-232, and bla_OXA-232 is endemic in many countries, including India [4], United States [6], France [7], Poland [8], China [9], and South Korea [10]. For K. pneumoniae, several nosocomial outbreaks have been reported with different sequence types (STs), including ST14, ST15, ST16, ST23, ST101, ST231, and ST437 [4]. The dissemination of bla_OXA-232 into different Enterobacterales species and the polyclonal dissemination of OXA-232-producing K. pneumoniae revealed the horizontal transfer of bla_OXA-232.

The bla_OXA-232 gene is located on a 6-kb ColKP3-type non-conjugative plasmid, which has been identified in almost all OXA-232-producing isolates [4,5]. The genetic environment surrounding bla_OXA-232 is a truncated Tn2013 transposon containing a large deletion on the 5′ end of ISEcp1 [5]. This deletion disrupted the ISEcp1 transposase activity and abolished its ability to mobilize bla_OXA-232, indicating the significant role of the ColKP3 plasmid on the prevalence of bla_OXA-232. As a small-size plasmid, the ColKP3 plasmid doesn’t possess an intact type IV secretion system (T4SS) for conjugation. It’s still unclear how this non-conjugative plasmid could facilitate the horizontal transfer of bla_OXA-232.

Except for conjugation, bacteria have developed several efficient and complex processes of horizontal gene transfer, including transformation and transduction. For natural transformation, bacteria must be in a state of competence in order to acquire extracellular DNA, while the Enterobacterales species, such as K. pneumoniae and E. coli, are not naturally transformable and couldn’t uptake extracellular DNA directly [11]. The transduction involves bacteriophages that transfer DNA to bacterial cells through infection, and the infection is limited to host specificity [12], nor could it explain the wide dissemination of bla_OXA-232 among Enterobacterales species. In other words, there might be other mobilization pathways that could facilitate the horizontal transfer of bla_OXA-232.

Outer Membrane Vesicles (OMVs) are spherical nanostructures, 50–250 nm in diameter, naturally produced by Gram-negative bacteria [13]. OMVs originate from the outer membrane and lipopolysaccharide, peptidoglycan, phospholipids, genetic materials (DNA and RNA), and periplasmic and cytoplasmic protein components were loaded into the vesicular lumen during their biogenesis. OMVs represent a unique bacterial secretion pathway that selects and protects its cargo, allowing bacteria to act on and interact with their environment over extended distances without direct contact [13]. A novel horizontal gene transfer mechanism mediated by OMVs has been recently identified, which has been reported in various bacterial species, including K. pneumoniae, E. coli, and Acinetobacter baumannii [14,15]. OMVs derived from K. pneumoniae have been identified as vectors for the horizontal transfer of antibiotic resistance and virulence genes [16,17]. OMVs could protect luminal genes or plasmids against DNases and function as an effective horizontal gene transfer system [18].

Here, we observed the in vivo intraspecies transfer of bla_OXA-232 during a nosocomial outbreak of OXA-232-producing K. pneumoniae. We demonstrated the presence of the bla_OXA-232-harbouring ColKP3-type plasmid in OMVs derived from clinical K. pneumoniae isolates, and ColKP3 plasmid-packaged OMVs could facilitate the horizontal transfer of bla_OXA-232 among Enterobacterales. For the first time, we determined the horizontal transfer pathway for the carbapenemase gene bla_OXA-232, since it was reported in 2013. In contrast, we found that OMVs couldn’t promote an effective transformation of the most prevalent plasmid-encoded carbapenemase genes, including bla_KPC-2 and bla_NDM-1, probably due to the low number of plasmids loaded into vesicles, while these two genes could be transferred through independent plasmid conjugation or with the help of a co-existent conjugative plasmid. We highlighted that diverse mobilization strategies were employed by plasmids harbouring carbapenemase genes, and plasmids display a proper choice of mobility pathway due to their individual properties.

Materials and methods

Bacterial strains for transformation and conjugation

For transformation and conjugal transfer, we inserted the hygromycin phosphotransferase gene (hph) into E. coli MG1655 and deleted its hsdR, the major subunit of type I restriction-modification (RM) system, constructing a hygromycin-resistant and type I RM system deficient E. coli MG1655 hph ΔhsdR as the recipient strain. To construct a model K. pneumoniae strain, one clinical K. pneumoniae isolate TU37 was employed [19]. The virulence plasmid pVir in TU37 was cured with the sodium dodecyl sulphate (SDS)-based plasmid curing approach, and the key gene wza for capsule synthesis was deleted, constructing a plasmid- and capsule-free E. coli-like mutant strain K. pneumoniae TU37-vf Δwza.

Clinical strains

Clinical samples were routinely collected from inpatients and sent to the Clinical Microbiology section for bacterial culture and identification. Clinical isolates were identified by matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS). K. pneumoniae isolates displaying intermediate or resistance to meropenem through the disk-diffusion method were selected for carbapenemase detection with colloidal gold immunoassay [20]. Briefly, a 1-mL loopful of bacteria was mixed with five drops of extraction buffer. Next, 100 μL of the mixture was transferred into the Carba-5 cassette, and the results were evaluated after incubation for 15 min. A total of 46 nonduplicate OXA-232-producing K. pneumoniae strains were isolated from 16 patients, who were hospitalized in the Surgery and Neurosurgery intensive care unit (SICU and NICU).

Antimicrobial susceptibility testing

Antimicrobial susceptibility testing was performed with the reference broth microdilution method. The results were interpreted according to the Clinical and Laboratory Standards Institute (CLSI) 2022 guideline, with the exception of tigecycline and colistin, where the breakpoint was defined by the European Committee on Antimicrobial Susceptibility Testing (EUCAST, version 12.0, http://www.eucast.org/).

Whole genome sequencing and bioinformatic analysis

All 46 bla_OXA-232-positive K. pneumoniae isolates were sequenced using Illumina NovaSeq 6000 platform. The raw reads were trimmed and filtered to remove low-quality sequences and adaptors by Trimmomatic (version 0.36). De novo assembly was conducted by Canu 2.0. Antimicrobial resistance genes and plasmid replicon analysis were performed using ResFinder and PlasmidFinder tools via the CGE server. Virulence genes were identified using the online virulence factor database VFDB, and multilocus sequence typing (MLST) and capsular (K) type were determined using Kleborate 1.0.0. Easyfig tools were used to visualize the bla_OXA-232-carrying ColKP3 plasmids and genetic context comparisons. Pan-genome analysis was done using Roary v3.11.2. Genes in at least 95% of isolates were considered core genes. Core genome single-nucleotide polymorphisms (cgSNPs) were extracted using SNP-sites from core gene alignment. cgSNPs were filtered using VCF tools v0.1.17. A maximum likelihood phylogenomic tree of these genomes was constructed by RAxML-NG v1.1.0. The phylogenetic tree was displayed and annotated using iTOL v5.

OMV isolation

OMVs were isolated by modification of the technique described previously for the recovery of S. aureus vesicles from culture supernatants [21]. Overnight-cultured cells of K. pneumoniae clinical strains were inoculated into fresh LB with 1:1000 dilution and grown for 16 h at 37°C with shaking at 180 rpm. Cell cultures were centrifuged and the supernatants were filtered through a 0.22-μm pore-size filter. The filtrate was then concentrated by ultrafiltration using a 100-kDa hollow-fiber membrane (Millipore). Then, the crude OMVs were obtained by ultracentrifugation of the above filtrate at 174,900×g at 4°C for 4 h in a SW 32Ti rotor (Beckman Coulter). The pellets were suspended in 4.4 ml 50% OptiPrep solution (Axis-Shield) and applied to the bottom of 4 ml 40% and 1.6 ml 10% OptiPrep solutions. After centrifugation at 200,000×g for 5 h at 4°C in a SW 41Ti rotor, the visible band between 10% and 40% gradients was collected.

Nanoparticle tracking analysis (NTA)

Nanoparticles in the isolated OMV suspensions were analyzed using a NanoSight NS300 (Malvern) instrument equipped with a 488-nm laser and a sCMOS camera. The platform was cleaned with sterile PBS to ensure no residue was present and to remove any potential air bubbles before analyzing samples. All samples were diluted in PBS to a final volume of 2 ml and injected into the sample chamber with sterile syringes, yielding particle concentrations of 10⁸ particles per millilitre in accordance with the manufacturer’s recommendations.

Transmission electron microscopy (TEM) imaging

Generally, 10 μl of OMV solution was dropped onto a copper grid, let to sit for 7 min and wiped dry. A drop of uranyl acetate (3%) was added to the sample to stain for 10 s. The grid was then washed with PBS twice and dried at room temperature for about 5 min before being put on the microscope. Electron micrographs were recorded with a HT7700 microscope (Hitachi) at 100 kV acceleration voltage.

OMV DNA extraction

OMV samples were treated with DNase to hydrolyze surface-associated DNA and free DNA in suspension. Reactions were stopped by the heat treatment of the mixtures for 2 min at 85°C. DNase- treated vesicles were then lysed with 0.2% Triton X-100 solution for 30 min at 4°C, and OMV DNA was purified with a QIAamp DNA Mini Kit (QIAGEN). The OMV DNA was quantified using a Nanodrop ND-1000 instrument.

OMV-mediated transformation

Transformation assay was performed according to a method described previously [22]. BCA Protein Assay Kit (Beyotime Biotechnology, China) was used to determine the protein concentration of OMVs samples according to the manufacturer’s protocols. Overnight-cultured cells of K. pneumoniae TU37-vf Δwza or E. coli MG1655 hph ΔhsdR were inoculated into fresh LB with 1:100 dilution and grown for 3 h at 37°C with shaking at 180 rpm. Cells were pelleted and resuspended in PBS with 0.4% glucose. This cell suspension was mixed with purified OMVs obtained from K. pneumoniae clinical strains. The suspensions were incubated statically for 3 h at 37°C. Three millilitres of LB broth were then added to the suspension, and the incubation continued at 37°C overnight with shaking (180 rpm). The broth culture grown overnight was pelleted and resuspended in 200 μl of LB broth. Transformants were selected on LB agar plates with 8 µg/ml piperacillin and 4 µg/ml tazobactam. PCR assay was used to determine the presence of carbapenemase genes as key molecular markers in successful transformants. Transformation frequency was calculated as the number of transformants (cfu/mL) in the total cell count (cfu/mL).

Plasmid copy number determination

The copy number of plasmid carrying carbapenemase genes was measured with carbapenemase genes relative to an internal K. pneumoniae housekeeping gene, rpoB, as previously described [23]. Standard curves were generated for both the target and the endogenous control (rpoB) using 10-fold dilutions of template DNA at known concentrations.

Construction of genetically modified strains

To construct markerless mutants, we generated vectors for each mutation using pCONJ5H, in which hygromycin phosphotransferase gene (hph) was employed to replace the resistance genes in pCONJ working vectors [24]. Each target gene’s upstream and downstream flanking regions (approximately 500 bp) were amplified by PCR with the relevant primers and assembled into pCONJ5H, with NEBuilder HiFi DNA Assembly Cloning Kit (New England Biolabs). K. pneumoniae mutants were then generated by conjugation. A single colony of the transconjugant was used to inoculate liquid LB with hygromycin. Bacteria were collected by centrifugation, washed and diluted 5-fold into 5 ml LB, and then grown for 4 h at 37°C. 5 μl of the culture was then plated on salt-free LB agar containing 10% (w/v) sucrose to select for loss of integrated plasmid; colonies were analyzed by PCR to identify mutants and confirmed by sequencing.

Conjugation assay

For conjugal transfer, E. coli MG1655 hph ΔhsdR was employed as the recipient strain. Three millilitres of overnight cultures of each donor and recipient strains were mixed together. The mixture was spotted on LB agar plate and then incubated for mating at 37°C for 4 h. Bacteria were scraped from the plate and resuspended in 1 ml of PBS, and 200 µl was plated on LB agar plates containing 200 µg/ml hygromycin together with 8 µg/ml piperacillin and 4 µg/ml tazobactam for selecting an E. coli transconjugant carrying bla_OXA-232. For selecting E. coli transconjugant carrying bla_KPC-2 or bla_NDM-1, 2 µg/ml meropenem was employed. Conjugation efficiency was calculated by dividing the number of transconjugants (cfu/mL) by the number of recipient cells (cfu/mL).

Results

In vivo intraspecies transfer of bla_OXA-232 during a nosocomial outbreak of OXA-232-producing K. pneumoniae

A nosocomial outbreak of OXA-232-producing K. pneumoniae was observed in our hospital during the period of December 2022 to April 2023 (Figure 1). A total of 46 OXA-232-producing K. pneumoniae strains were isolated from 16 patients, who were hospitalized in the Surgery and Neurosurgery intensive care unit (SICU and NICU) (Figure 1). None of the patients had a recent medical history or international travel. Most isolates were obtained from sputum (58.7%) and bronchoalveolar lavage (15.2%). Antimicrobial susceptibility testing revealed that all isolates were resistant to cefepime, ceftazidime, aztreonam, piperacillin-tazobactam, levofloxacin, and ciprofloxacin. Most of them were susceptible to ceftazidime-avibactam, tigecycline, and polymyxin B, whereas the majority of isolates showed intermediate or resistant meropenem, but were susceptible to imipenem (Table S1).

Figure 1. Timeline of the isolation of OXA-232-producing K. pneumoniae during the period of December 2022 to April 2023. A. Timeline of the isolation of OXA-232-producing K. pneumoniae. Coloured rectangles represent the presence of the patient in the corresponding ward. Only patients with isolation of at least two strains were displayed. The yellow and red rectangles represent the detection of ST15 and ST147 OXA-232-producing K. pneumoniae strains, respectively. SICU, surgery intensive care unit; NICU, neurosurgery intensive care unit. B. Linear alignment of bla_OXA-232-harbouring ColKP3 plasmids from ST15 and ST147 strains isolated from patient P1. Genes were denoted by arrows, and were coloured based on their functional classification. Shading denotes regions of homology (nucleotide identity ≥99%).

All OXA-232-producing K. pneumoniae isolates were sent for genomic sequencing and two different clones were identified, with ST15 (n = 43) as the dominant clone, followed by ST147 (n = 3). All isolates carried bla_OXA-232 and ESBLs genes, with bla_CTX-M-15 in ST15 and bla_CTX-M-19 in ST147 strains, and no other carbapenemase genes were detected (Figure S1). The bla_OXA-232 gene was harboured by the 6-kb ColKP3 plasmid in all isolates. The ST147 OXA-232-producing K. pneumoniae strains were isolated after two months of clonal outbreak of ST15 strains, and the first ST147 strain was isolated from patient P1 with successive isolation of ST15 strains (Figure 1). Both ST15 and ST147 strains were recovered from the respiratory tract of P1. During the nosocomial outbreak, no ST147 OXA-232-producing K. pneumoniae was identified in other departments within this hospital. The ColKP3 plasmid sequences displayed 100% coverage and >99% identity between ST15 and ST147 strains (Figure 1). The presence of OXA-232-positive ST15 and ST147 K. pneumoniae in the same patient suggests the horizontal transfer of bla_OXA-232-harbouring ColKP3 plasmid.

The wide dissemination of bla_OXA-232-harbouring ColKP3 plasmid in clinical K. pneumoniae with various genetic backgrounds

We then collected whole genome sequences of clinical OXA-232-positive K. pneumoniae strains from GenBank and performed the core genome single-nucleotide polymorphisms (cgSNPs)-based phylogenetic analysis. These reference strains were isolated from various parts of the world from 2013 to 2023 (Figure 2). Though they displayed various genetic backgrounds and belonged to different STs, the bla_OXA-232 gene was harboured by the 6-kb ColKP3 plasmid in all strains, highlighting the horizontal transfer of OXA-232 ColKP3 plasmid in K. pneumoniae. In addition to bla_OXA-232, several antibiotic-resistance genes have also been identified in these strains, including carbapenemase gene bla_NDM-1, ESBLs gene bla_CTX-M-15, aminoglycoside resistance genes rmtF and aac(6′)-Ib, sulphonamide resistance genes sul1 and sul2.

Figure 2. Phylogenetic analysis of OXA-232-producing K. pneumoniae strains. The phylogenetic tree was constructed with core genome sequences of 8 strains collected in this study (ST15 and ST147), and other reference strains retrieved from the GenBank database. The filled block of colour indicates the presence of plasmids and the antimicrobial resistance genes.

OMVs mediating horizontal transfer of the epidemic bla_OXA-232-harbouring ColKP3 plasmid

To determine the pathway for the horizontal transfer of bla_OXA-232-harbouring ColKP3 plasmid, outer membrane vesicles (OMVs) were purified from ST15 and ST147 OXA-232-producing K. pneumoniae strains. Transmission electron microscope (TEM) showed that these purified OMVs were oval and spherical (Figure 3). Nanoparticle tracking analysis (NTA) showed that the purified OMVs displayed a median size of 208 nm and a concentration of 5 × 10¹¹ particles/ml (Figure 3). No bacteria were observed under the microscope, and no contaminated bacteria were detected through LB agar plating. No significant differences in morphology and the number of OMVs isolated from ST15 and ST147 strains were observed, and OMVs from the ST15 strain were employed for transmission of bla_OXA-232.

Figure 3. The distribution of plasmids harbouring carbapenemase genes in outer membrane vesicles (OMVs) derived from carbapenemase-producing K. pneumoniae. (A, D, G) Transmission electron microscopy (TEM) imaging of OMVs from OXA-232-Kp, KPC-2-Kp, and NDM-1-Kp. (B, E, H) Nanoparticle tracking analysis (NTA) of OMVs from OXA-232-Kp, KPC-2-Kp, and NDM-1-Kp. (C, F, I) PCR screening for carbapenemase genes and the corresponding plasmid backbones. For each gene, genomic DNA was employed as the positive control and was placed at the left lane. OMVs DNA was also used as the template and the PCR products were run at the right lane. (J) Plasmid copy number of bla_OXA-232 ColKP3 (pOXA-232), bla_KPC-2 IncFII(pHN7A8)/IncR (pKPC-2), and bla_NDM-1 IncX3 (pNDM-1) plasmids in K. pneumoniae. (K) The number of plasmids per nanogram of vesicle protein. (L) The number of plasmids per OMV. The number of OMV was determined with NTA analysis. Error bars signify standard deviations. ****, P < 0.0001.

DNase treatment and PCR analysis proved the presence of bla_OXA-232 gene and the ColKP3 plasmid backbone in the vesicular lumen (Figure 3). The OMV-mediated transformation assay was performed using K. pneumoniae TU37-vf Δwza and E. coli MG1655 hph ΔhsdR as the recipient strains, with different amounts of OMVs. The transformation assay displayed that OMVs could efficiently promote the mobilization of bla_OXA-232, and the number of transformants grew at increasing OMVs amounts added to the recipient bacterial cells (Table 1). The recipient cells were positive for the bla_OXA-232 gene, and agarose gel electrophoresis demonstrated that transformants have acquired the ColKP3 plasmid (Figure 4). Antimicrobial susceptibility testing revealed that transformants displayed significantly increased resistance to carbapenems, piperacillin and piperacillin-tazobactam (Table 2). No transformants were obtained when free plasmid or triton-lysed OMVs were incubated with recipient strains, proving that the transfer was mediated exclusively by the intact vesicles (Figure 4).

Figure 4. OMVs-mediated transformation assay of bla_OXA-232. (A) Protocol of OMVs-mediated transformation assay. (B) Transformants were selected on LB agar plates with 8 µg/ml piperacillin and 4 µg/ml tazobactam. Recipient cells were treated with free plasmid pOXA-232, OMVs and triton-lysed OMVs. Eco, E. coli MG1655 hph ΔhsdR; Kpn, K. pneumoniae TU37-vf Δwza. (C) PCR screening for positive transformants after OMVs mediated transformation assay using E. coli MG1655 hph ΔhsdR and K. pneumoniae TU37-vf Δwza as the recipient strains. (D) Agarose gel electrophoresis of bla_OXA-232-harbouring ColKP3 plasmids extracted from E. coli MG1655 hph ΔhsdR (Ec) and K. pneumoniae TU37-vf Δwza (Kp) transformants.

Table 1. Outer membrane vesicles (OMVs) mediated transformation of plasmids-harbouring carbapenemase genes.

OMVs donor strain^a	Recipient strain	Amount of OMVs (μg)	Transformation frequency^b
OXA-232-Kp
	K. pneumoniae TU37-vf Δwza	100	4.47 × 10^−7
		50	1.72 × 10^−7
		20	0.56 × 10^−7
	E. coli MG1655 hph ΔhsdR	100	2.31 × 10^−7
		50	1.12 × 10^−7
		20	/
KPC-2-Kp
	K. pneumoniae TU37-vf Δwza	100	/
	E. coli MG1655 hph ΔhsdR	100	/
NDM-1-Kp
	K. pneumoniae TU37-vf Δwza	100	/
	E. coli MG1655 hph ΔhsdR	100	/

^a Carbapenemase-producing clinical K. pneumoniae strains, with 6.1-kb bla_OXA-232 ColKP3 plasmid in OXA-232-Kp, 178-kb bla_KPC-2 IncFII(pHN7A8)/IncR plasmid in KPC-2-Kp, and 54-kb bla_NDM-1 IncX3 plasmid in NDM-1-Kp.

^b /,no positive transformants.

Table 2. Antibiotic susceptibility profiles of donor, recipient and OMVs-mediated bla_OXA-232 transformants.

Antibiotics^a	MIC (μg/ml)^b
	OXA-232-Kp	Eco	T-Eco	Kpn	T-Kpn
PIP	≥128	1	≥128	4	≥128
TZP	≥128	1	64	1	≥128
ATM	≥64	≤0.06	≤0.06	0.5	0.5
CAZ	≥64	0.125	0.125	0.25	0.5
FEP	≥32	≤0.06	0.125	0.25	0.5
IPM	1	≤0.06	0.5	0.125	1
MEM	2	≤0.06	0.25	0.125	1
ETP	32	≤0.06	0.5	0.125	4
LEV	≥8	≤0.25	≤0.25	≤0.25	≤0.25
CIP	≥4	≤0.25	≤0.25	≤0.25	≤0.25
AMK	4	≤1	≤1	≤1	≤1
TGC	≤0.25	≤0.25	≤0.25	≤0.25	≤0.25
COL	16	≤0.25	≤0.25	≤0.25	≤0.25
CZA	0.5	≤0.25	≤0.25	≤0.25	≤0.25

^a PIP, piperacillin; TZP, piperacillin-tazobactam; ATM, aztreonam; CAZ, ceftazidime; FEP, cefepime; IPM, imipenem; MEM, meropenem; ETP, ertapenem; LVX, levofloxacin; CIP, ciprofloxacin; AMK, amikacin; TGC, tigecycline; COL, polymycin B; CZA, ceftazidime-avibactam.

^b OXA-232-Kp, ST15 OXA-232-producing K. pneumoniae; Eco, E. coli MG1655 hph ΔhsdR; Kpn, K. pneumoniae TU37-vf Δwza; T, transformant.

OMVs couldn’t promote an effective transformation of the bla_KPC-2 and bla_NDM-1 genes

To further determine the role of OMVs in the dissemination of carbapenemase genes, the OMV-mediated transformation assay was performed for the most prevalent carbapenemase genes, bla_KPC-2 and bla_NDM-1. Two reference clinical K. pneumoniae strains, CR-hvKP4 and Kp1902226, which carried the epidemic 178-kb bla_KPC-2-harbouring IncFII(pHN7A8)/IncR plasmid and 54-kb bla_NDM-1-harbouring IncX3 plasmid, respectively, were employed in this study (Table S2). Even with different genetic backgrounds, no significant difference was observed in the size and the amount of OMVs purified from different strains (Figure 3). DNase treatment and PCR analysis proved the presence of the carbapenemase genes and plasmid backbones in the vesicular lumen, however, OMVs couldn’t promote an effective transformation for these two carbapenemase genes (Table 1). The bla_KPC-2 and bla_NDM-1-harbouring plasmids displayed significantly lower plasmid copy numbers in K. pneumoniae when compared to the bla_OXA-232-harbouring ColKP3 plasmid (Figure 3). Furthermore, the number of plasmids loaded into vesicles as measured per vesicle protein and the average loading of plasmids per vesicle all decreased significantly for CR-hvKP4 and Kp1902226 (Figure 3).

Plasmid conjugation facilitated the horizontal transfer of the bla_KPC-2 and bla_NDM-1 genes

The epidemic IncFII(pHN7A8)/IncR KPC-2 plasmids share a conserved backbone and we can see large deletions and inversions of the tra gene cluster encoding for the F-type type IV secretion system (T4SS) (Figure S2). The reference KPC-2 plasmid pKPC-KP4 in this study displayed a complete inversion of the tra gene cluster. Conjugation assay demonstrated that pKPC-KP4 was conjugative (Table S3). pKPC-KP4 could still be transferred into E. coli MG1655 hph ΔhsdR via conjugation after deletion of one of the major tra genes, traC, while deletion of traU on another co-existent conjugative plasmid p3 could totally abolish the conjugal transfer of pKPC-KP4 (Table S3). In contrast, the epidemic IncX3 NDM-1 plasmids also share a conserved backbone, including the gene cluster for P-type T4SS. Deletion of one of the key vir genes, virD4, could completely abolish its conjugal transfer (Table S3).

Discussion

Carbapenems are the last resort antibiotics for treating infections due to Gram-negative bacilli, and rising carbapenem resistance is of particular concern for K. pneumoniae [1]. The production of carbapenemases is the most crucial cause of carbapenem resistance in K. pneumoniae [2]. Carbapenemase genes are mostly plasmid-encoded, the mobility of plasmid could facilitate the wide dissemination of carbapenemase genes among Enterobacterales species [3]. OXA-232 is one of the most common OXA-48-like carbapenemase derivatives and is widely disseminated in nosocomial settings across countries [4]. The bla_OXA-232 gene is located on a 6-kb non-conjugative ColKP3 plasmid, while the dissemination of bla_OXA-232 into different Enterobacterales species and the polyclonal dissemination of OXA-232-producing K. pneumoniae revealed the horizontal transfer of bla_OXA-232 [7,25]. However, it’s still unclear how this non-conjugative ColKP3 plasmid could facilitate the mobilization of bla_OXA-232. Here, we observed the in vivo intraspecies transfer of bla_OXA-232 during a nosocomial outbreak of OXA-232-producing K. pneumoniae. We demonstrated the presence of bla_OXA-232-harbouring ColKP3 plasmid in OMVs derived from clinical K. pneumoniae isolates, and OMVs could facilitate the horizontal transfer of bla_OXA-232 among Enterobacterales. For the first time, we determined the horizontal transfer pathway for bla_OXA-232, since it was reported in 2013.

For the detailed mobilization pathway of bla_OXA-232 mediated by OMVs, DNase treatment and PCR analysis proved the presence of bla_OXA-232 gene and the ColKP3 plasmid backbone in the vesicular lumen (Figure 3). OMVs could facilitate the horizontal transfer of bla_OXA-232 into K. pneumoniae TU37-vf Δwza and E. coli MG1655 hph ΔhsdR. The recipient cells were positive for the bla_OXA-232 gene, and agarose gel electrophoresis demonstrated that transformants have acquired the ColKP3 plasmid (Figure 4). In other words, OMVs-mediated mobilization of bla_OXA-232 was achieved by horizontal transfer of the bla_OXA-232-harbouring ColKP3 plasmid. The 6-kb bla_OXA-232-harbouring ColKP3 plasmid could be efficiently loaded into OMVs and OMVs then merge with the outer membrane of the recipient cell, which promotes the migration of the ColKP3 plasmid into the cytoplasm and facilitates the horizontal transfer of bla_OXA-232. No transformants were obtained when free plasmid or triton-lysed OMVs were incubated with recipient strains, proving that the transfer was mediated exclusively by the intact vesicles (Figure 4). Since the OXA-232 plasmid could not be transferred via natural transformation and conjugation, OMVs could be a major pathway for horizontal transfer of bla_OXA-232.

To further determine the role of OMVs in the dissemination of carbapenemase genes, the OMV-mediated transformation assay was performed for the most prevalent carbapenemase genes, including bla_KPC-2 and bla_NDM-1. In China, IncFII(pHN7A8)/IncR and IncX3 appear to be the most common type of plasmids carrying bla_KPC-2 and bla_NDM-1 [26,27]. Though the presence of carbapenemase genes and plasmid backbones in the vesicular lumen was observed, OMVs couldn’t promote an effective transformation for bla_KPC-2 and bla_NDM-1, which might be attributed to the low copy number of plasmids in clinical isolates and the low number of plasmids loaded into vesicles (Figure 3). Previous studies revealed that plasmid properties, such as copy number and size, could significantly affect the loading of the plasmid into vesicles and modulate vesicle-mediated horizontal gene transfer [15,28]. The number of plasmids loaded into vesicles is positively correlated with the plasmid copy number in a bacterium, and the plasmid size inversely affected the number of plasmid copies in vesicles [28]. Since bla_KPC-2 and bla_NDM-1 plasmids are usually in large-size and low-copy number, they could not be efficiently loaded into OMVs and could not be transferred via OMVs.

The bla_KPC-2-harbouring IncFII(pHN7A8)/IncR and bla_NDM-1-harbouring IncX3 plasmids possessed intact F-type (traM-traI) and P-type T4SS (virB1-virB11), respectively (Figure S2), and were conjugative through conjugation assay (Table S3). The bla_NDM-1-harbouring IncX3 plasmid was an independent conjugative plasmid and displayed high conjugation frequency. In contrast, a complete inversion of the tra gene cluster encoding for the F-type T4SS was observed on bla_KPC-2-harbouring IncFII(pHN7A8)/IncR plasmid pKPC-KP4, resulting in a non-functional T4SS (Table S3). However, the existence of a helper plasmid facilitated the conjugal transfer of pKPC-KP4. The pKPC-KP4 plasmid could be horizontally transferred via conjugation with the help of a co-existent IncI1 conjugative plasmid within the same bacterial host (Table S3). IncI1 plasmids are prevalent in K. pneumoniae and have also been reported to facilitate the transmission of pLVPK-like virulence plasmid [29]. The co-existent helper plasmid could compensate for the non-functional T4SS of IncFII(pHN7A8)/IncR KPC-2 plasmid and promote its horizontal transfer.

Till now, whether OMVs could facilitate the mobilization of bla_KPC-2 and bla_NDM-1 genes remains a controversial issue, and different results have been reported by several independent research groups [17,30,31]. Even for the positive report of OMVs-mediated transmission of bla_NDM-1, transformation frequency was much lower than plasmid conjugation efficiency [15], probably due to the low number of plasmids loaded into vesicles. It’s still unclear how plasmid DNA was wrapped into OMVs and the mechanism by which plasmids were transported into cytoplasm after OMVs merge with the outer membrane of the recipient cell. According to the phenotype assays, this study suggested that OMVs-mediated genetic material exchange may only serve as an alternative pathway for horizontal transfer of the large-size and low-copy number conjugative plasmids harbouring carbapenemase genes.

Here, we demonstrated that OMVs-mediated transformation could facilitate the mobilization of bla_OXA-232 among Enterobacterales. Without the functional conjugation machinery and surrounding mobile genetic elements, the wide dissemination of bla_OXA-232 among Enterobacterales has also been established with OMVs-mediated transformation. In contrast, for bla_KPC-2 or bla_NDM-1-harbouring conjugative plasmids, conjugation remains the dominant horizontal transfer mechanism. Non-conjugative plasmids tend to be poorly adapted and are frequently lost in the long term [32]. However, the ColKP3 plasmid seems to have evolved to become efficient at plasmid mobilization by employing OMVs. Plasmids harbouring carbapenemase genes display diverse mobilization strategies and manage to establish the high prevalence of carbapenemase genes among bacterial populations. Understanding the transmission mechanism is of great importance to delay or arrest the propagation of OXA-232 and other prevalent carbapenemases within K. pneumoniae and to other species.

Ethical approval

All procedures performed in this study involving human participants were in accordance with the ethical standards of the Institutional Review Board Ethics Committee of Renji Hospital. For this type of retrospective study, formal consent is not required.

Nucleotide sequence accession numbers

Genome sequences of all OXA-232-producing K. pneumoniae strains have been deposited in the NCBI database under BioProject accession numbers PRJNA1001133.

Transparency declarations

The authors declare that they have no competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

We thank the authority of CR-hvKP4 by Prof. Rong Zhang from the Second Affiliated Hospital of Zhejiang University School of Medicine. We thank the authority of Kp1902226 by Prof. Xing Wang from Shanghai Children’s Medical Center.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This study was supported by National Natural Science Foundation of China (82272374), Shanghai Pujiang Program (22PJ1409600), and a research fund from Renji Hospital for young scholars (RJTJ22-MS-018). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.