Three CRKP (strains SP1023 and F1025 from Patient A, and strain SP1030 from Patient B) were isolated from two patients admitted to the neurology intensive care unit (NICU) of a tertiary hospital in Hangzhou City in 2022. Species identification was determined by MALDI-TOF MS (Bruker Daltonik GmbH, Bremen, Germany). This study was approved by the Ethics Committee of The Second Affiliated Hospital of Zhejiang University School of Medicine.

Patient A, a 38-year-old male, had poorly controlled hypertension for several years. He underwent three consecutive intracranial hematoma evacuations at a local hospital due to cerebellar hemorrhage. Blood cultures revealed carbapenem-resistant Acinetobacter baumannii (CRAB) and a combination of cefoperazone/sulbactam (1,1, 2 g IV every 6 h) and polymyxin B (750,000 IU IV every 12 h) was administered (Figure 1A). The patient was in a coma and was transferred to the NICU of our hospital for further treatment. Upon admission, the patient received supportive treatments, including mechanical ventilation, blood transfusion, fluid replacement, and appropriate medications. The original antimicrobial therapy regimen was continued. The subsequent sputum culture revealed the growth of Klebsiella aerogenes and cefoperazone/sulbactam- and carbapenem-resistant A. baumannii. Therefore, cefoperazone/sulbactam was replaced with tigecycline (100 mg IV every 12 h) on Day 7. Nine days later (Day 16), both organisms were cleared. However, CRKP (strain SP1023) exhibiting ceftazidime/avibactam and polymyxin B resistance was isolated from the sputum sample (Figure 1A). On Day 18, fecal screening for CRE yielded pandrug-resistant K. pneumoniae (strain F1025), which was resistant to ceftazidime/avibactam, polymyxin B, and tigecycline (Figure 1A). In addition to CRKP, Pseudomonas aeruginosa and carbapenem-resistant A. baumannii (CRAB) were detected in the sputum sample on Day 32. Meropenem was used (1 g IV every 12 h) instead of the previous antimicrobials, and only CRKP was cleared after 9 days of treatment. Tigecycline was used again, and amikacin (400 mg nasogastric feeding every 12 h) was added to the treatment regimen 5 days later. However, P. aeruginosa that developed carbapenem resistance and CRAB persisted in the patient’s respiratory tract. Due to his extremely poor condition and acute exacerbation of chronic renal failure, the patient died of multiple organ failure on Day 68.

Patient B, a 40-year-old male with spontaneous intracerebral hemorrhage, underwent surgical evacuation of the intracranial hematoma at a local hospital. Six days later, he remained in a coma and was transferred to the NICU of our hospital. The CT scans showed postoperative hematoma in the surgical area and scattered infiltrates in both lungs. The patient received supportive treatments to reduce intracranial pressure, sedation, pain management, enteral nutrition, endotracheal intubation, and mechanical ventilation. Additionally, piperacillin/tazobactam (8:1, 4.5 g IV every 6 h) was administered for antimicrobial therapy throughout the hospitalization period (Figure 1B). On the second day of admission, CRAB was isolated from the sputum sample and persisted until discharge. On Day 17 and Day 19, CRKP (strain SP1030) was detected in the sputum culture that was resistant to ceftazidime/avibactam and colistin (Figure 1B). After 8 days of mechanical ventilation, the patient was weaned off the ventilator (Day 15). Since the patient’s vital signs stabilized and his mental function recovered, he was discharged for further treatment at a rehabilitation hospital on Day 20.

The minimal inhibitory concentrations (MICs) of 18 antimicrobial agents, including imipenem, meropenem, ertapenem, ceftazidime/avibactam, ceftazidime, cefotaxime, cefepime, piperacillin/tazobactam, cefoperazone/sulbactam, cefmetazole, aztreonam, ciprofloxacin, amikacin, chloramphenicol, fosfomycin, tetracycline, tigecycline, and colistin, were determined using the broth microdilution method (Clinical and Laboratory Standards Institute, 2018) and interpreted according to the Clinical and Laboratory Standards Institute (CLSI) guidelines (Clinical and Laboratory Standards Institute, 2021). Tigecycline susceptibility was interpreted using breakpoints recommended by the US Food and Drug Administration.² Escherichia coli ATCC 25922, K. pneumoniae 700603, and Pseudomonas aeruginosa ATCC 27583 were used as the quality control strains in parallel.

To investigate the evolution and genetic relatedness of these CRKP, the genomic DNA of three CRKP (strains SP1023, F1025, and SP1030) was subjected to WGS by both the short-read Illumina NovaSeq 6000 platform and the hybrid long-read Oxford Nanopore PromethION 48 platform. The complete genome was assembled by Flye assembler v2.9.2 (Kolmogorov et al., 2020) and polished by Pilon v1.24 (Walker et al., 2014). The antimicrobial resistance genes and the plasmid types for the assembly scaffolds were identified by ResFinder 4.1 and PlasmidFinder 2.0, respectively, at the Center for Genomic Epidemiology.³ The sequence types and virulence factors were identified using Kleborate v0.3.0 (Lam et al., 2021). A pairwise comparison of genomes and variant callings for single-nucleotide polymorphisms (SNPs) was conducted using Snippy v4.4.5 with default settings. The plasmids encoding bla_KPC-14 and virulence-associated genes were annotated by the RAST server (Overbeek et al., 2014) and BLASTN program. The comparison of plasmids was visualized and annotated by BRIG v0.95 (Alikhan et al., 2011).

The transferability of bla_KPC-14 genes was estimated by conjugation experiments with filter mating methods (Cai et al., 2008). Rifampin-resistant E. coli EC600 was used as the recipient strain. The putative transconjugants grown on selective media supplemented with 8 mg/L CZA or 30 mg/L tetracycline were identified by MALDI-TOF MS and screened for the presence of bla_KPC-14 genes. The conjugation frequency equaled the number of transconjugants divided by the number of recipients.

The G. mellonella (wax moth larvae) infection model was used to confirm the hypervirulent phenotype of the CRKP strains as previously described (McLaughlin et al., 2014). Overnight cultures of K. pneumoniae were diluted in sterile phosphate-buffered saline to obtain a concentration of 10⁸ CFU/mL. Wax moth larvae weighing 250–300 mg (Tianjin Huiyude Biotech Company, Tianjin, China) were injected with 10 μL bacterial suspension and incubated for 48 h at 35°C. The survival rate of G. mellonella was recorded at 12 h, 24 h, 36 h, and 48 h. ST11 K. pneumoniae FJ8 without virulence factors and the hypervirulent K. pneumoniae 4 were used as the negative and positive controls, respectively (Gu et al., 2018). All experiments were performed in triplicate. Kaplan–Meier survival curves were plotted using Prism 9.

The complete genome of the chromosome and plasmids for K. pneumoniae SP1023, F1025, and SP1030 were downloaded with BioSample accession numbers SAMN36464959, SAMN36465167, and SAMN36465169, respectively.

As Table 1 illustrated, three CRKP shared a high-level resistance to CZA with MIC values of >64/4 mg/L and showed resistance or decreased susceptibility to meropenem and ertapenem. These strains also shared an extensive drug resistance (XDR) profile (Magiorakos et al., 2012) to ceftazidime, cefotaxime, cefepime, aztreonam, ciprofloxacin, amikacin, chloramphenicol, fosfomycin, and tetracycline but retained susceptibility to imipenem. More worrisome, resistance to colistin was observed in these CRKP, and K. pneumoniae F1024 exhibited additional resistance to another clinically important antibiotic, tigecycline, which was interpreted as resistance to almost all the antimicrobial agents frequently used in clinical settings.

All three CRKP were identified as the ST11 type, which was the predominant ST type of CRKP in China (Liu et al., 2022a). Moreover, these three strains all belonged to the K64 serotype. Genome-based phylogenetic analysis suggested that these CRKP were closely related within 27 SNPs (with strain F1025 as the reference genome), indicating that these XDR strains originated from the same clone.

Whole-genome analysis demonstrated that three CRKP exhibited a similar carriage profile of β-lactamase genes, including bla_KPC-14, bla_SHV-11, bla_TEM-1, and bla_LAP-2, while K. pneumoniae F1025 and SP1030 additionally expressed SHV-12 (Table 2). However, the amplification of the bla_KPC gene failed both in A.baumannii and P. aeruginosa isolated from Patient A. Multiple antimicrobial resistance genes were also identified in three CRKP, conferring resistance to quinolones (qnrS1), phenicols (catA2), sulfonamides (sul1 and sul2), tetracyclines [tet(A)], fosfomycin (fosA and fosA3), and aminoglycosides (aadA2b and rmtB). Detailed analysis showed three CRKP shared that mutations for type 1 Tet(A) variants (I5R, V55M, I75V, T84A, S201A, F202S, and V203F) and the insertion of ISKpn26 elements at position 75 in the acrR gene, both of which contributed to the elevated MICs of tigecycline (Chiu et al., 2017). Insertional inactivation by ISKpn18 was also detected in ramR, another tigecycline resistance determinant gene, further driving the generation of resistance to this antibiotic in K. pneumoniae SP1025. The mgrB genes was interrupted by ISKpn26 at the same position (nucleotide 75) in three CRKP, thus accounting for the resistance to colistin.

The bla_KPC-14 gene could be conjugated into E. coli EC600 from K. pneumoniae F1025 and SP1030 with similar conjugation efficiencies of 2.1 × 10⁻⁵ and 3.5 × 10⁻⁵, respectively, but conjugation failed in strain SP1023, suggesting a different location of the bla_KPC-14 gene among these strains. Elevated CZA and several β-lactam MIC values were observed in E. coli transconjugants for K. pneumoniae F1025 and SP1030, confirming the functionality of bla_KPC-14 (Table 1). Notably, these E. coli transconjugants also showed resistance to tetracycline and slightly higher tigecycline MIC values, indicating the possible cotransfer of tet(A) and bla_KPC-14 genes. Therefore, selective media containing tetracycline were used to screen for putative transconjugants. Each of the three CRKP was able to transfer the tet(A) gene to the recipient E. coli EC600 with a similar efficiency at approximately 10⁻⁵; however, E. coli transconjugants with different donors displayed heterogeneity in antimicrobial susceptibility profiles. Unlike the E. coli transconjugants F1025-TE and SP1030-TE (with K. pneumoniae F1025 and SP1030 as donors, respectively), which displayed similar profiles to those of their counterparts for bla_KPC-14, the E. coli transconjugant SP1023-TE (with K. pneumoniae SP1023 as the donor) showed decreased susceptibility to tetracycline and tigecycline but retained the same MIC values for CZA and other β-lactams as the recipient strain. The amplification of bla_KPC-14 was also carried out in E. coli transconjugants F1025-TE and SP1030-TE, but failed in SP1023-TE, further supporting the dissimilar bla_KPC-14 gene locations and plasmid carriage of the three CRKP.

To clarify the location and genetic platforms of bla_KPC-14 genes, hybrid long-read sequencing of all three strains was performed. This yielded the complete genome for three CRKP with similar sizes of approximately 6 Mbp, consisting of one chromosome and varied numbers of plasmids (Table 3). A similar IncFII/IncR hybrid plasmid encoding bla_KPC-14 was harbored by all three CRKP, designated as plasmids pSP1023-KPC, pF1025-KPC, and pSP1030-KPC for K. pneumoniae SP1023, F1025, and SP1030, respectively. These plasmids carried a variety of additional antimicrobial resistance determinants, including the bla_TEM-1, fosA3, and rmtB genes, with sizes ranging from 112,363–112,364 bp. Blast analysis demonstrated that the backbone of this plasmid showed the highest similarity to plasmid pCRKP66R-3 with 100% identity (100% coverage, GenBank accession CP063835), which harbored the bla_KPC-2 gene. These plasmids also displayed 100% identity to two other bla_KPC-2-encoding plasmids named plasmid pC76-KPC (86% coverage, GenBank accession CP080299) and pCT77-KPC (87% coverage, GenBank accession CP080305), both of which were previously discovered in CRKP in our hospital (Figure 2) (Chen et al., 2021). The bla_KPC-14 gene was in the non-Tn4401 element as an NTE_KPC-Ib-like transposon with a genetic array of “IS26-ΔTn3-ISKpn8-bla_KPC-14-ISKpn6-korC-klcA-rep-orf-IS26,” which was commonly and uniquely identified in ST11 KPC-producing CRKP (Yang et al., 2021). Consistent with many IncFIIK/IncR bla_KPC-harboring plasmids widely disseminated in China, this plasmid was nonconjugative, which was probably due to the absence of relaxase, thus explaining the failure of the single transfer of bla_KPC-14 in the conjugation experiment.

The type 1 Tet(A) variant colocalized with the catA2, bla_LAP-2, qnrS1, and sul2 genes on unknown Inc-type plasmids with sizes of 84,754 bp, 101,767 bp, and 101,768 bp for K. pneumoniae SP1023, F1025, and SP1030, respectively (Figure 2A). This backbone was also observed in tet(A)-harboring plasmids in K. pneumoniae C76 and K. pneumoniae CT77 with 100% identity (91% coverage) but integrated with a 4,956 bp DNA fragment (based on K. pneumoniae SP1030) containing the bla_LAP-2 and qnrS1 genes. Unlike plasmid pSP1023-tetA in K. pneumoniae SP1023, the bla_KPC-14 locus of 8,248 bp was inserted into the tet(A)-carrying plasmids in K. pneumoniae F1025 and SP1030, which were further designated as plasmid pF1025-KPC-tetA and pSP1030-KPC-tetA, respectively. Another resistance locus harboring the bla_SHV-12 gene was also observed in these two plasmids but was absent in plasmid pSP1023-tetA. The structural discrepancy in these tet(A)-carrying plasmids echoed the differences in the genetic and phenotypic profiles of the transconjugants we noticed. These observations indicated that the plasmids carrying both tet(A) and bla_KPC-14 genes possibly evolved from those with a single occurrence of the tet(A) gene, such as pSP1023-tetA, which could also be further traced back to previously identified plasmids. The bla_KPC-14 genes were flanked by an NTE_KPC-Ib-like transposon similar to that on the IncFII/IncR plasmids in our study, indicating that these nonconjugative plasmids might be the source of the bla_KPC-14-carrying fragments for plasmids pF1025-KPC-tetA and pSP1030-KPC-tetA. As previously reported, the diversity of mobile elements and transposons in NTE_KPC-I elements actively promoted the transposition of the bla_KPC genes to various genetic locations (Yang et al., 2021); thus, we speculated that the mobilization event mediated the integration of the bla_KPC-14 locus into the backbones of tet(A)-carrying plasmids, thus creating the binary carriage profile of bla_KPC-14-harboring plasmids in K. pneumoniae F1025 and SP1030. The integrated plasmids retained the fully functional conjugative genes homologous to the previous tet(A)-carrying plasmids (Figure 2), further facilitating the cotransfer and dissemination of KPC-14 and the type 1 Tet(A) variant among Enterobacterales.

The hypervirulent phenotype of three CRKP was observed in the G. mellonella infection model (Figure 3). At 48 h post-infection, the survival rates of larvae infected by strains SP1023, F1025, and SP1030 were 12.5, 29.2, and 16.7%, respectively, which were lower than that of larvae infected by the negative control strain K. pneumoniae FJ8 at 79.2%. WGS analysis showed that these strains expressed the same virulence factor profile specific to hypervirulent K. pneumoniae (hvKP), including RmpA2, aerobactin, and yersiniabactin (Table 2); thus, the CRKP in our study were classified as CR-hvKP. Long-read sequencing revealed that the rmpA2 and the iutAiucABCD gene cluster (aerobactin siderophore) were located on the IncHI1B/IncFIB-type pLVPK-like plasmid with sizes ranging from 204,770 to 204,778 bp (Table 3, Figure 2C). These plasmids showed high identity (>99%) to a variety of other hypervirulence plasmids of K. pneumoniae in the NCBI database, indicating the wide dissemination of these plasmids conferring hypervirulent phenotypes.