IS26 carrying bla_KPC−2 mediates carbapenem resistance heterogeneity in extensively drug-resistant Klebsiella pneumoniae isolated from clinical sites

Abstract

Background

Due to the widespread and irrational use of antibiotics, the emergence and prevalence of carbapenem-resistant Klebsiella pneumoniae (K. pneumoniae) have become a major challenge in controlling bacterial infections in hospitals. The bla_KPC−2 gene located on mobile genetic elements has further complicated the control of resistant bacteria transmission.

Results

In this study, K. pneumoniae strains were isolated from blood cultures of patients. Using the Kirby-Bauer disk diffusion method, we found carbapenem resistance heterogeneity. The resistant subpopulation KPTA-R1 and the sensitive subpopulation KPTA-S1 were purified. Whole-genome sequencing revealed that the bla_KPC−2 gene in KPTA-R1 was located on an IncFII plasmid (pKPC-R), within a composite transposon (PCTs) formed by two direct repeats of IS26 elements. The structure was identified as IS26-RecA-ISKpn27-bla_KPC−2-ISKpn6-IS26. However, in KPTA-S1, a similar plasmid, pAR-S, lacked this segment. Sequence comparison analysis indicates that the deletion of this bla_KPC−2 encoding sequence in this IncFII plasmid is associated with transposition activity mediated by IS26. Multi-sequence comparison of the plasmids showed that the IS26 transposon facilitated the sequence polymorphism of these plasmids.

Conclusion

This study reveals the key role of IS26-mediated transposition activity, through homologous recombination, in the emergence of carbapenem resistance heterogeneity in clinical K. pneumoniae strains carrying bla_KPC−2. IS26 is able to promote the evolution of resistance in the IncFII plasmid, and through copy-in cointegration or targeted conservative cointegration may result in the acquisition or loss of antibiotic resistance, which may affect clinical care and pose a public health risk.

Introduction

The transposition activity mediated by mobile genetic elements (MGE) is particularly noteworthy because it enables bacteria to efficiently acquire novel antibiotic resistance genes (ARGs), thereby granting them new levels of resistance [1]. This phenomenon, which aids bacterial evolution and rapid adaptation, is of great significance, especially when these bacteria pose a threat to the health of living organisms.

Klebsiella pneumoniae (K. pneumoniae) is a common Gram-negative pathogen that frequently causes nosocomial infections in immunocompromised patients [2]. Carbapenem-resistant K. pneumoniae (CRKP) has emerged as a life-threatening pathogen in multiple countries [3]. According to data from the China Antimicrobial Surveillance Network (CHINET, http://www.chinets.com/), the resistance rate of K. pneumoniae to carbapenems has risen alarmingly, from 2.9% in 2005 to 30% in 2023. The increasing prevalence of CRKP has significantly limited the available treatment options for K. pneumoniae infections [4]. One of the primary mechanisms of carbapenem resistance in China is the presence of the carbapenemase gene bla_KPC, which is typically located within a mobile transposon on conjugative plasmids [5]. This gene has been reported to be carried by various transposons, including Tn4401, Tn1721, and IS26, which facilitate the horizontal transfer and dissemination of resistance genes [6].

Antibiotic resistance heterogeneity refers to the phenomenon in which individuals within the same bacterial population exhibit varying responses to antibiotics. This variability is one of the key factors that complicate clinical infection treatment, potentially leading to therapeutic failure or the evolution of resistant strains [7]. While the Kirby-Bauer disk diffusion method and the epsilometer test can be used to detect resistance heterogeneity, there remains a pressing need for more sensitive and rapid detection methods to enhance clinical monitoring [8]. Achieving efficient and accurate identification and management of bacterial resistance heterogeneity in clinical practice remains a critical challenge that must be addressed.

IS6 family has been shown to play an important role in the rearrangement and spread of multiple antibiotic resistance. Among them, the IS26 exhibits high activity in the horizontal transmission of antibiotic resistance genes [9]. The DDE transposase it encodes can integrate into DNA through two mechanisms. In the copy-in mechanism, IS26 promotes the spread of resistance genes by replicating the target DNA and inserting it into a new genomic location [10]. This mechanism not only acts on external sites of the target DNA but also on sites within the same DNA molecule, causing DNA sequence inversion. This inversion effect contributes to the evolution and polymorphism of resistant plasmids [11]. The target DNA sequence flanked by two IS26 elements can form pseudo-compound transposons (PCTs). These PCTs serve as carriers for gene transfer, and their movement unit is called the translocatable unit (TU). The formation of TU often relies on homologous recombination, through which IS26 carries gene sequences that are deleted internally and transfers them to new locations [12].

In this study, we report a clinically extensively drug-resistant K. pneumoniae strain exhibiting carbapenem resistance heterogeneity. Sequence analysis identified the absence of the bla_KPC−2 gene in the IncFII plasmid of KPTA-S1. Further investigation suggested that IS26-mediated transposition activity could facilitate the loss of bla_KPC−2 in the IncFII plasmid through homologous recombination. Furthermore, our analysis highlights the important role of IS26 in the evolution of IncFII plasmids. This provides a case study and insights into the crucial role of the transposition activity of mobile genetic elements in bacterial resistance.

Materials and methods

Infection history

On June 17, 2022, a 33-year-old pregnant woman presented with vomiting and chest pain lasting for half an hour. The patient underwent a cesarean section the same day and was subsequently diagnosed with disseminated intravascular coagulation (DIC), infectious shock, multi-organ failure, and an abdominal infection. Anti-infective therapy with meropenem (MEM) was initiated on June 19, along with correction of coagulation disorders, anti-shock measures, and organ support.

On June 24, Gram-negative bacilli were observed in an alveolar lavage smear, and tigecycline (TGC) was added to the treatment regimen for suspected lung infection. Blood cultures on June 29 identified multi-drug-resistant K. pneumoniae. On July 1, MEM was discontinued, and treatment was switched to polymyxin B combined with TGC. On July 23, polymyxin B was stopped, and the treatment was further adjusted to fosfomycin combined with TGC. On July 25, K. pneumoniae (KPTA-R1 and KPTA-S1) (susceptible to cotrimoxazole) was isolated from the blood culture. On July 26, the previous antimicrobials were discontinued, and the treatment was switched to ceftazidime/avibactam for infection control. By July 29, blood cultures were negative, and on August 4, 2022, the patient discontinued treatment. The treatment timeline is shown in Fig. 1.

Fig. 1

Course of Treatment. PCT, procalcitonin

Bacterial isolation and identification

KPTA-R1 was isolated from the patient’s blood. The blood sample was cultured on a blood agar plate medium (5% sheep blood, LB basal medium) (Autobio, Zhengzhou, China). The strains obtained from the culture were purified and identified by the automated microbial mass spectrometry detection system Autof ms 1000 (Autobio, Zhengzhou, China) according to the procedures of the manual.

Antimicrobial susceptibility assay

Antimicrobial susceptibility was tested using BD Phoenix M50 fully automated microbial drug sensitivity analyzer (BD, Phoenix, AZ, USA), and MIC values were determined according to the manufacturer’s instructions.

The Kirby-Bauer (K-B) disc diffusion test verified the drug sensitivity results. Three single colonies were resuspended in a normal saline solution, and the turbidity was 0.46 ~ 0.54. The bacterial solution was uniformly spread on a Mueller-Hinton agar medium (Autobio, Zhengzhou, China), and a antibiotic susceptibility disk (Thermo Fisher, Waltham, MA, USA) was pasted on the agar plate inoculated with the bacteria to be tested and incubated at 35 °C for 24 h. Results were observed and determined by the size of the antibacterial zone according to the instructions.

Isolation strains that developed carbapenem resistance heterogeneity were tested for MIC values against imipenem (IPM) and MEM using the agar diffusion and epsilometer test (E-test method). Pick 3 ~ 5 colonies to be tested, and use sterile saline to configure the bacterial suspension with 0.5 McCloud turbidity. Spread the bacterial solution evenly on the Mueller-Hinton agar medium and dry it at room temperature for 3 ~ 5 min after spreading. Use sterile tweezers to pick up the E-test paper (BIO-KONT, Wenzhou, China), stick it to the medium, and incubate it at 35 °C for 24 h. At the end of incubation, observe the inhibition circle around the test paper on the medium and read the result according to the instructions. The E-test experiment was conducted with three replicates.

The drug sensitivity results were analysed according to the 2021 CLSI M100 Executive Standard for Antimicrobial Drug Sensitivity Testing [13]. For tigecycline, the resistant breakpoint was interpreted according to the FDA criteria (https://www.fda.gov/drugs/development-resources/antibacterial-susceptibility-test-interpretive-criteria).

Isolation and purification of subtype strains

The K. pneumoniae isolate obtained from blood culture was inoculated onto chocolate agar plates (Autobio, Zhengzhou, China) containing vancomycin using a three-zone streaking method, placing an IPM susceptibility disk in the first zone. After 24 h of incubation, colonies from the edge of the inhibition zone surrounding the IPM susceptibility disk were selected. The same method was used to streak them onto new chocolate agar plates, with a new IPM susceptibility disk placed in the first zone. After another 24 h of incubation, 10 single colonies from the third plate were picked and streaked onto fresh chocolate agar plates. This process was repeated four times, and finally, a single colony from the edge of the inhibition zone around the IPM susceptibility disk was selected for bacterial identification and K-B disc diffusion test. The results revealed that the strain was sensitive to both IPM and MEM, and it was named KPTA-S1. KPTA-R1 was obtained by selecting a single colony and verifying its resistance to IPM and MEM.

Growth curve measurements

Individual bacterial colonies were isolated, and bacterial suspensions were prepared to match the 0.5 McFarland turbidity standard. A 60 µL aliquot of the bacterial suspension was transferred to 6 mL of LB liquid medium (without antibiotics) and incubated at 35 °C. Every 2 h, 100 µL of the bacterial suspension was transferred to a microtiter plate, with four wells allocated per strain, and the optical density (OD) was measured at 600 nm. Growth curves were generated based on OD measurements taken at 2, 4, 6, 8, 10, 12, and 24 h. All experiments were performed in triplicate.

Monitoring the stability of carbapenem resistance

A single bacterial colony was inoculated into LB liquid medium and incubated overnight at 37 °C with shaking at 200 rpm. Each day, 20 µL of the bacterial suspension was transferred into 2 mL of LB liquid medium for bacterial cultivation, and the resulting bacterial suspension was diluted 1:100 before being inoculated onto LB agar plates. Antibiotic susceptibility testing for IPM and MEM was performed using the K-B disc diffusion method for 10 consecutive days. The specific procedure for the K-B disc diffusion method is described above.

Conjugation test

The broth mating method was used in the conjugation test. In this method, rifampicin-resistant Escherichia coli EC600 was used as the recipient, and the KPTA-R1 strain was used as the donor. The donor and recipient strains were inoculated in 3 mL LB broth (Binhe, Hangzhou, China) and incubated overnight at 37 °C. The donor and recipient bacteria were mixed 1:1 and incubated at 37 °C for 16 to 18 h. The mixture was then applied to LB plates with a concentration of 4 µg/mL meropenem and 2.5 µg/mL rifampicin resistance and incubated at 37 °C for 24 h. Finally, single colonies were identified using an automated microbial mass spectrometry detection system Autof ms 1000, and MIC values for meropenem were determined using a BD Phoenix M50 fully automated microbial drug sensitivity analyzer.

PCR verification of the deletion sequence on pKPC-R

The primers used were primer-F (5’-GGGCCGTCGACAGCCGGGGCCGCA-3’) and primer-R (5’-GGTACAGATACGCCCAGC-3’). The template was genomic DNA prepared from an overnight culture of KPTA-R1/KPTA-S1. Following the manufacturer’s instructions, PCR was performed using Taq DNA polymerase (Vazyme Biotech Co., Ltd.) in a 25 µL reaction mixture under the following conditions: 10 ng template DNA, 65 °C annealing temperature, 2 min extension time, and 35 cycles. The PCR products were separated by 1.0% agarose gel electrophoresis and sequenced (Beijing Tsingke Biotech Co., Ltd.).

Whole genome sequencing of bacteria and DNA sequence analysis

PacBio sequencing and analysis were conducted by BIOZERON Co. Ltd., (Shanghai, China) [14]. For DNA sequence analysis, the genome was assembled using Canu [15]. Gene prediction for the assembled genome was performed with Prodigal (v 2.6.3).

The resistance genes of the strain were annotated in the Comprehensive Antibiotic Resistance Database (https://card.mcmaster.ca/) using BLAST [16]. Multilocus sequence typing (MLST) and capsular serotyping of strains KPTA-R1 and KPTA-S1 were analyzed using Pathogenwatch (v18.3.0) (https://pathogen.watch/). Plasmid typing for the two strains was carried out using the PlasmidFinder (v 2.0.1) platform (https://cge.food.dtu.dk/services/PlasmidFinder/) [17], where plasmid genome sequences were uploaded, and the Enterobacteriaceae database was selected with a minimum coverage of 60%. Virulence gene annotation was performed using the Virulence Factors of Pathogenic Bacteria database (http://www.mgc.ac.cn/VFs/links.htm) [18]. The mobile genetic elements of the two strains were annotated using the Mobile Element Finder tool (https://cge.food.dtu.dk/services/MobileElementFinder/). The nucleotide sequences of IS26 and the TIRs were obtained from ISfinder (https://isfinder.biotoul.fr/) and compared and localized in pKPC-R and pAR-S using SnapGene (v 4.1.8). The sequence differences between the pKPC-R and pAR-S plasmids were identified by performing sequence comparison using SnapGene. The origin of DNA transfer (oriT) and conjugation-associated proteins on the plasmids was identified using Orifinder (https://tool-mml.sjtu.edu.cn/oriTfinder/oriTfinder.html) [19]. Plasmid profiles were generated using the BLAST Ring Image Generator (v 0.95) [20]. Genetic evolutionary trees were generated using MEGA 11: multiple sequence alignment was performed using CLUSTALW, and evolutionary trees were built using the Neighbor-Joining Method.

The genome sequences of the strains used in this study were downloaded from the National Center for Biotechnology Information: K. pneumoniae strain ARLG-7683 plasmid pARLG-7683-2 (CP139221.1), K. pneumoniae strain 2,014,042,281 plasmid p42281-KPC (MT810369.1), K. pneumoniae strain 49,088 plasmid p49088-279.2 (CP089000.1), K. pneumoniae strain 33,367 plasmid p33367 KPC2 (CP099415.1), K. pneumoniae strain A1708 plasmid pA1708-KPC (MT810354.1), K. pneumoniae strain L388 plasmid pKPC-L388 (CP029225.1), K. pneumoniae strain 675,920 plasmid p675920-1 (MF133495.1), K. pneumoniae strain CRKP41 plasmid unnamed1 (CP107313.1).

Result

Genomic characterization and antimicrobial resistance of isolated strains

The K. pneumoniae strain isolated from the patient’s blood culture showed resistance heterogeneity to IPM and MEM in the K-B disk diffusion assay (Fig. 2A-1). The strain resistant to IPM and MEM was named KPTA-R1, while the sensitive strain was named KPTA-S1. To further confirm the antimicrobial susceptibility of these strains, the minimum inhibitory concentration (MIC) of IPM and MEM was determined using the E-test method. The results showed that the MIC values for KPTA-R1 were 32 µg/mL for both IPM and MEM (Fig. 2A-2) while KPTA-S1 exhibited MIC values of 0.5 µg/mL for IPM and 1 µg/mL for MEM (Figs. 2A-3). Complete antimicrobial susceptibility results are presented in Table 1. Except for IPM, MEM, and ceftazidime, the antimicrobial susceptibility profiles of KPTA-S1 and KPTA-R1 were identical (Table 1).

Fig. 2

Isolation and Characterization of Subtype Strains. (A-1) Clinical isolates exhibited resistance heterogeneity to meropenem (MEM) and imipenem (IPM). (A-2) Zone patterns of KPTA-R1 in the MIC gradient test strips for IPM and MEM. (A-3) Zone patterns of KPTA-S1 in the MIC gradient test strips for IPM and MEM. (B) Growth curve measurement of KPTA-R1 and KPTA-S1. (C) Monitoring of carbapenem resistance stability in KPTA-R1 and KPTA-S1. The results from days 1, 3, 6, 8, and 10 of the 10-day passaging are shown. The IPM and MEM zones are highlighted with red boxes

To assess the growth capacity of the subtypes, the growth curves of KPTA-R1 and KPTA-S1 were analyzed. The results showed that the two subtypes behaved similarly (Fig. 2B). Furthermore, after 10 generations of passaging KPTA-R1 and KPTA-S1, no changes in sensitivity to IPM and MEM were observed, indicating that the carbapenem resistance heterogeneity was stable (Fig. 2C).

Table 1 Drug sensitivity test results of clinically isolated strains

The whole genome sequence of both isolates was obtained, and the genomic characteristics of KPTA-R1 and KPTA-S1 were summarized in Table 2. The KPTA-R1 chromosome is 5,501,168 bp in length and contained four plasmids, including three circular plasmids and one linear plasmid. The KPTA-S1 chromosome is slightly shorter at 5,501,115 bp and carries five circular plasmids. Average nucleotide identity (ANI) analysis showed a 99.993% similarity between KPTA-R1 and KPTA-S1. MLST typing identified ST11 and capsule type KL64 in both strains.

KPTA-R1 contained four plasmids: a 111,465-bp plasmid (pTA2-R); an IncHI1B virulence plasmid (pVir-R, 219,497 bp) carrying iucA, B,C, D, iutA, rmpA/rmpA2; a linear plasmid (pKPC-R, 82,981 bp) carrying bla_KPC−2, bla_TEM−1, and rmtB; and a circular plasmid (pTA3-R, 81,332 bp) carrying tet(A), sul2, qnrS1, and bla_LAP−2. These plasmids contribute virulence and drug-resistance genes to KPTA-R1.

KPTA-S1 carrieed five plasmids. Compared with KPTA-R1, pVir-S was consistent with pVir-R, pTA2-S with pTA2-R, and pTA3-S with pTA3-R, showing no mutations or deletions. Additionally, KPTA-S1 contained pAR-S (143,613 bp), encoding bla_SHV−25, bla_CTX−M−65, bla_TEM−1, rmtB, and fosA, as well as pTA5-S, the smallest plasmid encoding phage-related genes (Table 2).

Table 2 Genomic characterisation of isolated strains and their plasmids

IS26-mediated deletion of bla _KPC−2 on IncFII plasmid

The bla_KPC−2 gene carried on plasmids is a common contributor to carbapenem resistance in K. pneumoniae. The pKPC-R plasmid (82,981 bp) in the KPTA-R1 strain encodes bla_KPC−2, while the pAR-S plasmid (143,613 bp) in the KPTA-S1 strain has the highest similarity to pKPC-R. Both plasmids carry the RepA replication initiation gene and belong to the IncFII plasmid family. The BLAST comparison between pKPC-R and pAR-S shows a percent identity of 99.99% and a query cover of 53%, indicating a high degree of homology between the two plasmids.

Plasmid sequence alignment and annotation analysis were performed on pKPC-R and pAR-S, and the results are shown in Fig. 3A. The alignment indicates that ~ 70kbp of the DNA sequence between pKPC-R and pAR-S are highly conserved. pAR-S is a more complete IncFII plasmid encoding multiple IS26 transposase genes, which form four novel resistance transposon regions: IS26-RecA-Tn3-bla_SHV−25-IS26, IS26-bla_CTX−M−65-IS903B-IS26, and IS26-rmtB-bla_TEM−1-RecA-IS26-fosA3-IS26, IS26-mer-genecluster-IS5075-IS1-IS26 (Fig. 3B).

Fig. 3

Comparative Sequence Analysis of pKPC-R and pAR-S. (A) Ring comparison of plasmids pKPC-R and pAR-S. Black regions denote resistance genes; red indicates transposases; rosy marks relaxase, orange represents the origin of transfer; blue highlights the type IV secretion system and type IV coupling protein. (B) IS26 transposons in plasmids pKPC-R and pAR-S contain resistance genes. The red background highlights resistance genes and mercury resistance protein domain; arrows indicate the direction of transposable elements gene encoding

Compared to pKPC-R, pAR-S lacked the DNA fragment encoding the bla_KPC−2 gene. In pKPC-R, bla_KPC−2 was located within a composite transposon (PCTs) formed by two directly oriented IS26 elements, with the structure: IS26-RecA-ISKpn27-bla_KPC−2-ISKpn6-IS26. Each IS26 element contained a 14 bp terminal inverted repeat (TIRs) at both ends (Fig. 4A). As shown in Fig. 4B, sequence alignment with pAR-S revealed that the deletion breakpoint of the IS26-flanking DNA fragment in pKPC-R is located at position 340 within the two IS26 elements, and a single copy of IS26 was found at the corresponding site in pAR-S. PCR amplification and sequencing analysis further confirmed this sequence alignment. A pair of primers was designed at both ends of the deletion junction in pAR-S, and a ~ 7,000 bp fragment was amplified from KPTA-R1. Subsequent sequencing verified that the fragment precisely corresponded to the deleted sequence (Fig. 4C). These findings indicate that in this IncFII plasmids, the loss of the bla_KPC−2 gene can be caused by IS26-mediated transposition activity through homologous recombination.

Fig. 4

IS26-mediated Sequence Variations Between Plasmids Via Homologous Recombination. (A) Compared to pKPC-R, the deleted sequence structure in pAR-S. This fragment is part of a composite transposon (PCTs) formed by two directly repeated IS26 elements. Triangles represent recombinase proteins, a: Transcriptional Repressor Protein KorC, c: Antirestriction Protein, d: Replication Protein. The IS26 elements are flanked by 14 bp terminal inverted repeats (TIRs). (B) A schematic diagram showing the sequence differences between pKPC-R and pAR-S mediated by IS26 through homologous recombination. The IS26 element is indicated by a blue box, and the arrows indicate the direction of IS26 encoding. (C) PCR validation of the deleted sequence in pKPC-R. Primers were designed upstream and downstream of the deletion site in pAR-S. The whole genome of KPTA-R1 was used as a template, and a ~ 7,000 bp band was successfully amplified

Evolutionary relationship between IS26 and IncFII plasmids

As described in the above results, the IncFII-type plasmids (pKPC-R and pAR-S) reported in this study encode multiple antibiotic resistance genes, all of which are associated with IS26-associated transposable elements (Fig. 3B), which may allow for a great evolutionary potential for the carriage of resistance genes by this IncFII plasmid. To further illustrate, After performing BLAST analysis on pKPC-R and pAR-S, eight plasmids with high similarity were identified. We constructed a phylogenetic tree for these 10 plasmids, and all nodes were supported by 100% bootstrap values. These plasmids were all classified as IncFII-type and were isolated from ST11 and KL64 K. pneumoniae strains collected over different years and from various countries. Phylogenetic analysis further confirmed that pAR-S and pKPC-R are distinct evolutionary variants of this IncFII plasmid (Fig. 5A).

Fig. 5

Evolutionary Relationship Between IS26 and IncFII Plasmids. (A) Genetic evolutionary tree was constructed by comparing pAR-S and pKPC-R with 8 similar plasmids. (B) Sequence comparison of pAR-S and pKPC-R with the 8 similar plasmids used to generate a circular diagram. The color annotations are consistent with those in Fig. 3A

Comparative circular mapping revealed that all 10 IncFII plasmids share a common conserved genomic region, and consistently carries the bla_TEM−1 and rmtB resistance genes. Beyond this conserved region, multiple MGEs were identified, with the highest number being IS26. We also observed various gene deletions around these MGEs (Fig. 5B), which are associated with the different numbers of resistance genes carried by these plasmids. These findings suggest that IS26 may play a key role in driving the evolution of these IncFII plasmids and is closely linked to their resistance functions.

Discussion

K. pneumoniae has become one of the most important pathogens causing hospital-acquired infections in recent years [21]. The World Health Organization (WHO) has identified extended spectrum β-lactamase (ESBL)-producing and carbapenem-resistant K. pneumoniae (CRKP) as a severe public health threat. In this study, we isolated an extensively drug-resistant CRKP (XDR-CRKP) strain from blood cultures of hospital-acquired infections, with carbapenem resistance heterogeneity observed during subsequent in clinical treatment [22]. After isolation and purification, we found that the resistant subtypes were almost resistant to most classes of antibiotics, while the susceptible strain lost resistance to carbapenem antibiotics. It is important to note that the XDR-CRKP strain exhibiting carbapenem resistance heterogeneity was not isolated from the early stages of infection. Initially, we defined this heterogeneity as a stable heteroresistance (HR) (carbapenem resistance remained stable in the subtypes after 10 passages) [23]. However, the definition of HR is currently unclear in the field. It is generally defined as the discovery of resistant subpopulations within sensitive strains, and there is a lack of attention to the outcomes of HR under clinical antibiotic pressure [24, 25]. In this study, the XDR-CRKP strain was isolated at the mid-treatment stage, after exposure to clinical antibiotic pressure. The sensitive subtype KPTA-S1 was obtained from the resistant phenotype strain, and we termed this heterogeneity as carbapenem resistance heterogeneity. So far, there is no standard method to detect this heterogeneity. Commonly used laboratory methods include E-test, K-B disc diffusion test, and phenotypic amplification phenomenon (PAP) test. The PAP test is considered a more reliable method, but due to its high cost and complicated procedures, it is difficult to implement in clinical settings. Therefore, new monitoring methods need to be developed for the clinical detection of HR/resistance heterogeneity [26].

To explore the mechanisms behind this resistance heterogeneity, we conducted an in-depth analysis of the bla_KPC−2-carrying pKPC-R plasmid and its similar plasmid, pAR-S. After confirming the high similarity between the two plasmids, the study highlighted the differences between pKPC-R and pAR-S plasmids. In the pKPC-R plasmid, the bla_KPC−2 gene is located within a composite transposon IS26, which contains insertion sequences ISKpn27 and ISKpn6. In contrast, at the matching sequence location in pAR-S, only a single copy of IS26 is annotated. Our analysis suggests that the deletion of the bla_KPC−2-carrying DNA sequence in pAR-S may be mediated by IS26 through a homologous recombination mechanism, which could be the cause of the observed carbapenem resistance heterogeneity between strains [27].

It has been reported that IncFII plasmids are the predominant carriers of bla_KPC−2. These IncFII plasmids tend to stably persist in ST11 K. pneumoniae strains. The host preference for specific plasmids in ST11 strains may explain why ST11 has become the dominant strain carrying bla_KPC−2 [28]. The IncFII plasmid described in this study encodes up to six resistance genes, as well as a mercury resistance domain. The study confirmed that the IncFII plasmid harbors multiple IS26 transposons, and various gene deletion sites observed in the plasmid correspond to the locations of IS26 (Fig. 5B). This suggests that IS26 has an impact on the evolution of this plasmid. The effect of IS26 on MDR plasmids, promoting the acquisition and loss of resistance, may complicate the clinical interpretation of resistance phenotypes and thus impact treatment decisions [29, 30]. IS26 has been reported to mediate the transfer of abundant antibiotic resistance genes (ARGs) between compatible plasmids, facilitating plasmid fusion and evolution [31]. Yang et al. reported that in three ST11 CR-HvKP strains, IS26 and IS903B in an IncI1 plasmid facilitated the formation of a novel fusion plasmid between two plasmids through homologous recombination, leading to the acquisition of new virulence elements by the strain [32]. This suggests that transposition events mediated by mobile genetic elements promote the polymorphism and functional evolution (both resistance and virulence) of plasmids. Additionally, IS26 has been shown to mediate ARGs transfer between incompatible plasmids [33].

Notably, the IncFII plasmids (pKPC-R and pAR-S) studied here, though containing a complete conjugative transfer region (Relaxase, oriT, Type IV Coupling Protein, Type IV Secretion System) [34], failed to successfully transfer in conjugation experiments using these plasmids as donors. Furthermore, in addition to K. pneumoniae, IncFII plasmids, which are highly homologous to pKPC-R, were also found in Escherichia coli, Citrobacter werkmanii, Enterobacter cloacae, and Enterobacter hormaechei. This suggests that such plasmids can be transmitted and colonized by Enterobacteriaceae bacteria, which are now prevalent in different countries and regions [35]. Given that this IncFII plasmid carries multiple resistance genes and mobile genetic elements, it should be given attention and monitored. Additionally, it was observed that both subtypes carried an IncHIB-type virulence plasmid, pVir-R/pVir-S, encoding the aerobactin-related toxin protein genes (iucA-D, iutA) and RmpA/RmpA2. Notably, pVir-R appears to be a novel plasmid, as BLAST analysis in public databases did not reveal similar sequences in bacterial chromosomes, plasmids, or phages.

In conclusion, our study reports a phenomenon of carbapenem resistance heterogeneity in a clinical isolate of K. pneumoniae. It was found to be related to the differences in the IncFII plasmids carried by KPTA-R1 and KPTA-S1. In this IncFII plasmid, the loss of bla_KPC−2 can be mediated by the transposition activity of IS26 through homologous recombination. Additionally, the presence of multiple IS26 elements on this plasmid plays a significant role in the polymorphism of the plasmid sequence and the evolution of its resistance functions.