Overexpression of β-lactamase genes (bla_KPC, bla_SHV) and novel CirA deficiencies contribute to decreased cefiderocol susceptibility in carbapenem-resistant Klebsiella pneumoniae before its approval in China

ABSTRACT

Cefiderocol (FDC) is an effective antibiotic that is used to treat severe infections caused by carbapenem-resistant Klebsiella pneumoniae (CRKP). The mechanisms underlying FDC resistance and molecular epidemiology in China remain unclear. We collected 477 non-duplicate CRKP clinical isolates in central China and characterized their susceptibility to FDC, virulence genes, and sequence typing. The overall FDC susceptibility rate of CRKP was 99.2% in central China, which was higher than that in North America and Europe (96.1%), with MIC_50/90 values of 1/2 mg/L. The decrease in FDC susceptibility in central China was concentrated in the ST11 CRKP-carrying virulence plasmids. Whole-genome sequencing (WGS) and quantitative reverse transcription PCR (qRT-PCR) experiments showed that serine β-lactamases, especially highly expressed KPC and SHV, substantially decreased FDC susceptibility in four FDC non-susceptible isolates (two resistant and two intermediate isolates). Notably, different CirA deficiencies, p.E450GfsTer16 and p.E133Ter, were found in both of the resistant isolates. In contrast, global WGS data indicate that the resistance mechanisms in North America and Europe were primarily associated with NDM and KPC variants, predominantly found in ST307 and ST147. Overall, FDC exhibits excellent activity against CRKP in central China, with resistance mechanisms primarily related to high KPC and SHV expression, along with deficiencies in CirA, frequently observed in ST11. This is remarkably different from the situation in North America and Europe and will directly impact the choice of clinical interventions. Additionally, the surveillance of FDC resistance in China is imperative.

INTRODUCTION

The prevalence of carbapenem-resistant Klebsiella pneumoniae (CRKP) continues to rise globally, posing a serious threat to public health worldwide (1, 2). In China, the treatment for CRKP infections commonly relies on polymyxins, tigecycline, and ceftazidime-avibactam (CZA) due to the development of resistance to carbapenems (3). However, with the widespread and frequent use of these antibiotics, their activity against CRKP is gradually diminishing, and some limitations are emerging, such as toxicity (4), ineffectiveness against metallo-β-lactamases (5), and inefficacy in monotherapy for urinary tract infections and bacteremia (6, 7). The continued emergence of CRKP resistance to these drugs and their limitations highlight the stark reality of the need for new antimicrobial agents that are safe and effective for combating CRKP in China.

Cefiderocol (FDC) is a novel siderophore cephalosporin and the first iron-chelating cephalosporin approved for clinical use in the United States and Europe. Although it has not yet been authorized in China, phase III clinical trials are currently underway. FDC shares some structural similarities with catecholate-type siderophores (e.g., salmochelin and enterobactin) allowing it to enter the periplasmic space through catecholate siderophore receptors such as the colicin I receptor (cirA). The emergence of isolates resistant to FDC due to mutations in the cirA gene has been reported in certain regions, including China and the United Arab Emirates, where FDC has not yet been approved for use (8–10). However, there have been no reports of cirA mutations leading to FDC resistance in the dominant CRKP clone ST11 in China (11). Although FDC has demonstrated structural stability against hydrolysis by extended-spectrum β-lactamases (ESBLs) and carbapenemases (12, 13), multiple FDC-resistant CRKP isolates associated with NDM and various KPC variants have been reported (14–16). Meanwhile, Sun et al. found that OmpK35 truncation is also associated with FDC resistance (17).

Although FDC is expected to be approved soon, knowledge about FDC resistance and molecular epidemiology in China remains limited. Our study aimed to investigate the resistance profile and molecular mechanisms of CRKP in Jiangxi, central China before FDC is approved for clinical use, and to compare it with FDC-approved regions, providing insights and guidance for the clinical use of FDC upon its approval in China.

RESULTS

In vitro activity of FDC against CRKP isolates in central China where FDC is not approved

From October 2018 to December 2022, we collected clinical CRKP isolates from different sources, including sputum, bronchoalveolar lavage fluid, urine, blood, secretions, and bile, from 477 patients at a tertiary hospital in Jiangxi Province, where FDC had not been used. The corresponding MICs for FDC ranged from 0.06 to 64 mg/L (Fig. 1; Table S1), with the highest proportion of isolates exhibiting an MIC of 1 mg/L (142/477, 29.8%). The K. pneumoniae isolates exhibited high FDC susceptibility (99.2%) with MIC_50/90 values of 1/2 mg/L, respectively.

FIG 1.

FDC MIC distribution profile for China isolates (red, n = 477) and combined North America and Europe isolates (blue, n = 280).

To compare the resistance profile of CRKP regions with and without clinical use of FDC, we searched for the distributions of FDC MICs for CRKP isolates in North America and Europe from 2020 to 2023 and 2021 to 2023, respectively, from the worldwide SENTRY Antimicrobial Surveillance Program. We searched for 439 CRKP strains, of which 280 were subjected to FDC MIC determination, including 55 from North America and 225 from Europe. As shown in Fig. 1, the susceptibility of CRKP to FDC in the regions of use was 96.1%, and the MIC_50/90 was 1/4 mg/L. The results indicated that isolates from North America and Europe showed higher overall resistance compared to those from China; however, FDC-resistant isolates remained rare (n = 1), suggesting that FDC are still very effective antimicrobial agents.

Molecular epidemiological characteristics of CRKP isolates

Considering the genetic differences between the CRKP isolates prevalent in the unapproved and approved regions, we analyzed the genotypic characteristics of these 477 CRKP isolates. Sequence typing analyses identified 28 distinct sequence types, including ST11 (376, 78.8%), ST15 (35, 7.3%), ST340 (11, 2.3%), and others (Fig. S1; Table S1). Analysis of FDC susceptibility among isolates of different sequence types revealed a significant difference in MIC distribution between ST11 and non-ST11 isolates. The FDC MIC distribution of ST11 and non-ST11 isolates showed a rightward shift to higher MICs for ST11 CRKP than for non-ST11 CRKP isolates (median 1 vs. 0.5 mg/L, P < 0.001), as depicted in Fig. 2A. The cumulative percentage of ST11 and non-ST11 isolates inhibited at various FDC MICs is shown in Fig. 2B, in which the cumulative curve of ST11 was significantly lower than that of the non-ST11 isolates.

FIG 2.

(A) MIC distribution profile and (B) cumulative percentage of FDC between ST11 and non-ST11 isolates. (C) MIC distribution profile and (D) cumulative percentage of FDC between pV-CRKp and CR-cKp isolates.

Furthermore, given that the ST11 isolates in China frequently carry virulence plasmids, we examined the virulence gene profiles of the isolates. Multiple virulence genes were identified (see Table S1), and this study defines CRKP-carrying virulence plasmid (pV-CRKP) as described by Gu et al. (18) as K. pneumoniae with several genetic markers located in the virulence plasmid including iucA, rmpA/rmpA2, and/or iroB. Of the 477 CRKP isolates, 298 (62.5%) were classified as pV-CRKP and the other 149 (37.5%) were classical CRKP (CR-cKp). Notably, within the ST11 CRKP, pV-CRKP accounted for 75.3%. As shown in Fig. 2C and D, MIC distribution and cumulative percentage between pV-CRKP and CR-cKp isolates exhibited that the pV-CRKP isolates have increased FDC resistance than the CR-cKp isolates.

Clinical and molecular characterizations of FDC-non-susceptible isolates

Among the 477 CRKP isolates, 4 were non-susceptible to FDC, comprising 2 FDC-resistant isolates, Kp241 (MIC = 64 mg/L) and Kp337 (MIC = 32 mg/L), and 2 intermediate isolates with MICs of 8 mg/L (Kp387 and Kp459). Four non-susceptible isolates were obtained from four male patients aged 35–89 years with different diseases. Various antibiotics were administered to treat patients before bacterial sampling. Unfortunately, all patients were unhealed and even died (Table S2). Furthermore, we applied Galleria mellonella larvae-infecting model and biofilm formation assay to assess pathogenicity. The results indicated that Kp241 and Kp459 exhibited similar virulence to the hypervirulent K. pneumoniae NTUH-K2044, whereas Kp337 and Kp387 were slightly more virulent than the low-virulence strain ATCC700603 (Fig. S2).

Taking into account the WGS data, three isolates (Kp241, Kp337, and Kp387) belonged to ST11-KL64, and one isolate (Kp459) belonged to ST15-KL19. All four strains harbored three to four plasmids, among which, the three ST11 isolates carried a virulence plasmid with the IncHI1B(pNDM−MAR)/repB replicon type and a resistance plasmid with the IncFII(pHN7A8) replicon type carrying bla_KPC-2 and ESBL genes. The three virulence plasmids, highly similar to pK2044 (NC_006625.1) with >92% query coverage and 99.39% nucleotide identity, contained virulence factors, including rmpA/rmpA2 and iucABCD, but lacked a fragment containing the virulence factor encoding the catecholamine-type iron carrier receptor salmochelin gene cluster iroBCDN (Fig. 3A and B). Kp459 did not carry a virulence plasmid but carried the virulence genes for the Kfu ABC iron transport system (kfuABC) in the chromosome and a repB(R1701)-type plasmid carrying bla_KPC-2, bla_OXA-1, and ESBL genes. A 18,168 bp fragment of this plasmid was completely identical to a fragment of its own chromosome, which carried the resistance genes bla_KPC-2, bla_OXA-1, bla_CTX-M-15 and aac(6′)-Ib-cr, flanked by the terminal insertion sequences (ISs) ISkpn27 and IS1380 (Fig. 3C). These two ISs may have mediated the transfer of the plasmid fragment into the chromosome, leading to the appearance of double copies of β-lactamase genes in the isolate.

FIG 3.

Molecular characterizations of FDC-non-susceptible isolates (A) circular and (B) linear comparison of pK2044 and virulence plasmids in Kp241, Kp337, and Kp387. (C) Schematic diagram of the 18,168 bp homologous fragment on the chromosome and plasmid of isolate Kp459. The arrows represented the genes encoding proteins (red: resistance; yellow: integrase recombinase and transposase; purple: transfer associated; gray: other functions).

Serine β-lactamases, especially highly expressed KPC and SHV, substantially decrease FDC susceptibility

All four non-susceptible isolates lacked metallo-beta-lactamases and contained only a variety of serine beta-lactamases (SBLs) (Table 1). To evaluate the contribution of SBLs to FDC resistance, the MICs were compared with and without the addition of the serine-β-lactamase inhibitor AVI. The combination of FDC and AVI resulted in an 8- to 16-fold decrease in the MIC of FDC in non-susceptible isolates and decreased the maximum MIC value from 64 to 4 mg/L. Additionally, we tested the MICs of ceftazidime and CZA, and the results also showed a significant reduction in MICs for the isolates in the presence of AVI (at least a 16-fold decrease) (Table 1). Furthermore, 100 mM of NaCl, a known bla_OXA inhibitor, was added to test the effect of class D beta-lactamase (DBL) on FDC resistance (Table 1) and found that NaCl addition did not alter the FDC MICs.

TABLE 1.

Antimicrobial activity of FDC with or without serine-β-lactamase inhibitors

Isolate	Serine-β-lactamases	MIC (mg/L)
		FDC	FDC + AVI	FDC + NaCl	CAZa	CZA
Kp241	KPC-2, SHV-12, SHV-11, CTX-M-65, TEM-1, LAP-2	64	4	64	>256	4
Kp337	KPC-2, SHV-12, SHV-11, CTX-M-65 × 2, TEM-1	32	4	32	>256	16
Kp387	KPC-2, SHV-12, SHV-11, CTX-M-3, CTX-M-65, TEM-1	8	1	8	>256	32
Kp459	KPC-2 × 2, TEM-1, SHV-28, CTX-M-15 × 2, OXA-1 × 2	8	0.5	8	256	4

^a CAZ, ceftazidime.

We randomly selected four isolates from the FDC-susceptible isolates that were positive for the resistance genes KPC, SHV, CTX, and TEM, as confirmed by WGS, and tested the relative expression of the four SBL genes in the FDC-non-susceptible and FDC-susceptible isolates using qRT-PCR. The KPC and SHV expression levels were higher in non-susceptible isolates than in susceptible isolates (P < 0.05), whereas CTX and TEM expression levels were not significantly different (Fig. 4). Overall, the results revealed that the overexpression of various class A β-lactamases, mainly KPC and SHV, in the four non-susceptible isolates collectively contributed to FDC resistance.

FIG 4.

Relative expression levels of four resistance genes in FDC-non-susceptible and FDC-susceptible groups. *P < 0.05; ***P < 0.001; ns, not significant. The expression of β-lactamases in all isolates was normalized by dividing by the mean expression level of four susceptible isolates.

Decreased susceptibility to FDC caused by novel cirA mutations

We further analyzed mutations [non-synonymous single-nucleotide polymorphisms (SNPs) and insertions/deletions] in seven siderophore receptor genes (cirA, fiu, fecA, fhuA, fepA, efeO, exbD) previously described to be associated with FDC resistance in four non-susceptible isolates. Two different cirA mutations were found in two resistant isolates (Kp241 and Kp337), with no mutations detected in other genes. As shown in Fig. 5A, Kp241 carried a nucleotide duplication (c.1348dupG), leading to a frameshift, followed by a premature stop codon (p.E450GfsTer16), and Kp337 carried a nonsense mutation (c.397G  >  T), causing a premature stop codon (p.E133Ter).

FIG 5.

Identified cirA alterations. (A) Alignment of cirA gene and amino acid sequence in wild type and mutant type in Kp241 and Kp337. Single base substitution or duplication (red) results in frameshift and a premature stop codon (asterisk, black). (B and C) Ribbon diagrams of CirA. (B) Intracellular view. (C) Side view. The extracellular space is located at the top and the periplasmic space is at the bottom of the image. The CirA amino acid deletions and alterations of Kp241 are shown in cyan, and those of Kp337 in blue. The CirA structure was modeled using Alphafold2 and visualized using PYMOL (DeLano).

CirA consists of a 22-stranded beta barrel spanning the outer membrane and a globular plug structural domain in the center of the barrel lumen. The structural domain of the plug mediates the translocation of siderophores through the barrel structure via conformational changes. In Kp241, the large amino acid changes and deletions in CirAE450GfsTer16 result in the loss of an entire β-strand, disrupting the integrity of the β-barrel structure and hindering FDC transport. In Kp337, the missing amino acid segment in CirAE133Ter is located in the plug structural domain, preventing FDC from passing through the β-barrel (Fig. 5B and C). Interestingly, CirAE450GfsTer16 is a novel mutation that has not been previously reported, while CirAE133Ter is a reported mutation found in an FDC-resistance ST512 K. pneumoniae strain 0296 in a patient in Italy who had been treated with FDC (8). This suggests that FDC may select for naturally occurring CirA-deficient strains with multiple SNPs, making them dominant strains. Additionally, the c.397G  >  T mutation may be a common genetic locus mutation, which is worth focusing on.

Mechanisms underlying reduced susceptibility to FDC based on global WGS data

We constructed a phylogenetic tree and analyzed the genes associated with FDC resistance of the four strains in this study and global non-susceptible isolates (MICs: ≥8 mg/L) uploaded to NCBI before February 2024. As shown in Fig. 6, the phylogenetic tree was divided into five clusters (ABCDE). Isolates from the FDC non-use regions were distributed in distantly related groups A and E, in which all cluster A strains were ST11 isolates. Cluster A was dominant in China, whereas isolates in the three sub-clusters (BCD), mainly ST147, ST307, and ST512, were predominant in North America and Europe. Regardless of whether FDC was used or not, the resistance mechanism was mainly β-lactamases accompanied by Ompk35/36 variants and truncation of CirA, and almost all strains with high-level FDC resistance (MICs: ≥32 mg/L) had premature truncation of CirA at various sites.

FIG 6.

Phylogenetic tree of FDC-non-susceptible strains. Leaf color block indicates strains isolated from unapproved areas (cyan) and from approved areas (red). Tip points of the phylogenetic tree are dotted by location where strains were isolated. Heatmap denotes (from left to right) MICs range (mg/L), MLST, Serotypes, CirA variant, Serine-carbapenemase, Metallo-carbapenemase, SHV, OmpK35 variant, OmpK36 variant, Virulence genes, IroBCDN, AMR plasmid types, Virulence plasmid types.

Compared to the FDC-approved regions, there were no small outbreaks of resistant isolates in the FDC-unapproved regions, and the resistance mechanisms were more diverse and dispersed. In terms of β-lactamases, NDM and KPC variants were predominant in FDC-approved regions, while KPC-2 and SHV were predominant in FDC-unapproved regions. Regarding Ompk35, the truncation site was more diverse in the FDC-unapproved regions, whereas in the FDC-approved regions, Ompk35-25% truncation sites were predominant. These results indicate that the resistance mechanisms in regions with and without approval for FDC use share similarities and intersections. However, the resistance mechanisms in regions with FDC approval related mainly to the presence of NDM and KPC variants in combination with Ompk35-25%, whereas in regions where FDC was not approved, mainly China, the overexpression of KPC-2 and SHV accompanied by novel CirA deficiencies was predominant.

DISCUSSION

While FDC, a siderophore catecholate cephalosporin, has not yet been approved for clinical use in China, it holds promise as an effective antimicrobial agent. Our results showed that the FDC susceptibility rate of CRKP in Jiangxi was 99.2%, which was higher than that in North America and Europe (96.1%). Our findings are consistent with many previous reports, including the SIDERO surveillance program, which showed that the FDC susceptibility rates of CRKP in North America and Europe ranged from 99.3% to 100% (19, 20), and the report from the China CRE Network, which showed a susceptibility rate of 99.7% (21). However, several studies have reported different susceptibility rates (9, 22), which may be related to differences in isolate collection sources, times, and regions.

Siderophores are essential for FDC resistance and are a key factor for the survival of hypervirulent K. pneumoniae under iron-deficient conditions (23). The common catecholate siderophore receptors in K. pneumoniae are CirA, Fiu, FepA, and IroN. Kumar et al. showed that CirA has strong affinity for FDC in K. pneumoniae, whereas other catecholate siderophores have weaker binding affinities (24). Increased expression of the enterobactin receptor gene (fepA) promotes FDC susceptibility, in contrast to the non-catecholate siderophore kfu, present in Kp459, and fecA (25). These results suggest that the absence of catecholate siderophore receptors, especially CirA, promotes the antibacterial activity of FDC, whereas the presence of other siderophore receptors may counteract the transport function of catecholate siderophores, thereby inhibiting FDC transport.

ST11 is a highly dominant CRKP clone in Asia, especially China, and carries bla_KPC-2 and multiple ESBL genes (11), while ST258 and ST512 are more common in North America and Europe, according to global genomic analyses (26). Our findings indicated that ST11 pV-CRKP isolates had stronger resistance to FDC than other CRKP isolates. This suggests that local epidemiological differences influence FDC susceptibility rates. Therefore, it is crucial to consider these regional epidemiological differences when interpreting FDC susceptibility data. Previous studies have shown that ST11 CRKP can easily evolve into CR-hvKP by acquiring a pk2044-like virulence plasmid. The virulence plasmids in ST11 were often accompanied by the absence of iroBCDN and the presence of the non-catecholate siderophore aerobactin (iuc) (27), a feature that was also observed in our study of all three ST11 FDC-non-susceptible isolates (Fig. 3A). The diversity of SBLs and the presence of specific virulence plasmids in ST11 may be the main reasons for the weak in vitro activity of FDC against ST11, which underscores the importance of monitoring antibacterial activity of FDC against ST11 CRKP.

We characterized the β-lactamase spectrum of four FDC-non-susceptible isolates and verified the role of serine β-lactamases, especially KPC and SHV, in FDC resistance. Previous studies have shown that metallo-β-lactamases (MBLs) contribute strongly to FDC resistance, and that the expression of a single serine β-lactamase in Escherichia coli under in vitro conditions does not have as strong an effect on FDC resistance as MBL (28), in particular weak hydrolysis of FDC by D-class β-lactamase (29). However, KPC and SHV are also associated with FDC heteroresistance (30, 31). A recent study by Liu et al. showed that multiple copies of bla_SHV-12 can cause high levels of FDC heteroresistance, suggesting that overexpression and duplication of SBLs can lead to FDC resistance (32). The four non-susceptible isolates in our study contained multiple SBLs, of which only bla_KPC and bla_SHV showed a significant increase in expression, contributing to decreased susceptibility to FDC.

The different mechanisms underlying resistance to FDC in different regions should inform region-specific strategies to tackle the spread of infections caused by FDC-resistant isolates. The emergence of multiple variants of KPC-3 prevalent in Europe and North America caused resistance to CZA and simultaneously diminished FDC susceptibility, for example, KPC-40, KPC-50 (14, 16). This may be due to the early use of CZA, as well as the structural similarities between FDC and ceftazidime. SBL inhibitors enhance the sensitivity of FDC, but it is not known whether the inhibitory effect on the KPC variant enhances susceptibility to FDC. This emphasizes the importance for regions without FDC approval to vigilantly monitor the emergence and impact of KPC-2 variants upon authorization, making antimicrobial susceptibility testing (AST) imperative before treating infections caused by KPC-producing organisms with FDC.

It should be noted that this study has several limitations. (i) This is a single-center study with some bias in the data, and follow-up studies with larger and more extensive multicenter are needed. Additionally, the number of strains from North America is small, with only 55 isolates. Therefore, conclusions regarding the North American region should be approached with caution. (ii) The definition of pV-CRKP used in our study only considered several genetic markers and did not WGS confirmation. Therefore, conclusions drawn from this definition may be biased. (iii) We were unable to determine the complete resistance mechanism of the four FDC-non-susceptible K. pneumoniae isolates.

In conclusion, CRKP exhibited higher FDC susceptibility in central China where it is not approved for clinical use than in North America and Europe. Resistance mechanisms are primarily related to high KPC and SHV expression, along with deficiencies in CirA, frequently observed in ST11. This is significantly different from regions in North America and Europe and will directly impact the choice of clinical interventions. Furthermore, the surveillance of FDC resistance in China is imperative.

MATERIALS AND METHODS

Bacterial isolates

In total, 477 non-repetitive clinical CRKP isolates were collected from a tertiary hospital in Jiangxi Province between October 2018 and December 2022. The isolates were identified using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (Bruker Daltonics). Carbapenem resistance was defined as a MIC of ≥4 mg/L for imipenem or meropenem.

ASTs

AST was initially performed using a VITEK-2 system (bioMérieux, Lyon, France) in sentinel hospitals and was further confirmed using broth microdilution method in our laboratory. The MIC of FDC was determined using iron-depleted cation-adjusted Mueller Hinton broth (ID-CAMHB), as described in the Clinical and Laboratory Standards Institute (CLSI) M100 33rd edition, employing the broth microdilution method (33). All susceptibility tests were performed in triplicate and interpreted according to the (33) guidelines. E. coli ATCC 25922 and Pseudomonas aeruginosa ATCC 27853 were used as quality control strains for AST. FDC-non-susceptible strains of K. pneumoniae were defined as strains with an MIC of FDC ≥8 mg/L. To examine the effect of serine-β-lactamase and class DBL on FDC activity against these isolates, FDC MICs were determined with or without β-lactamase inhibitors. We added 4 mg/L of avibactam (AVI) to inhibit serine-β-lactamase and 100 mM NaCl to inhibit class DBL.

Determination of virulence genes and multilocus sequence typing

All the template DNAs were extracted using the TIANamp bacterial DNA kit (Tiangen Biotech, Beijing, China). The presence of multilocus sequence typing (MLST) and virulence genes (iucA, iroB, ybt, clb, rmpA, and rmpA2) were determined using PCR, as previously described (34, 35). Positive products were sequenced by Sangon Biotechnology (Shanghai, China), and the results were analyzed using BLAST. In this study, all PCRs were performed at least three times, but sequencing was conducted on only one of the positive results. For sequence typing analyses, the sequences of seven conserved housekeeping genes of K. pneumoniae (gapA, infB, mdh, pgi, phoE, rpoB, and tonB) were submitted to the MLST database (https://bigsdb.pasteur.fr/klebsiella/).

Virulence assessment assay

G. mellonella larval infection and biofilm formation assays were performed to assess bacterial pathogenicity. Healthy larvae weighing 250–300 mg (purchased from the Tianjin Huiyude Biotech Company) were used in this experiment. Overnight cultures of K. pneumoniae were adjusted to a concentration of 1 × 10⁶ CFU/mL using phosphate-buffered saline (PBS). The insects were inoculated by injecting 1 × 10⁶ CFU per 10 µL aliquots into the hemocoel via the rear left proleg, followed by a recording of survival rate every 12 h for 2 days. All experiments were performed in triplicate. The K. pneumoniae strains NTUH-K2044 and ATCC700603 were used as controls for high- and low-virulence strains, respectively (36, 37).

Crystal violet staining was used to quantitatively assess biofilm formation in cells cultured in Luria-Bertani (LB) broth using a 96-well polystyrene microtiter plate, as previously described (38). Biofilm formation was quantified by measuring the optical density at 590 nm (OD590) and defining the cut-off optical density (ODc) as the mean of the control wells without bacterial suspension plus three standard deviations. According to the previous description, the specific results categorize the biofilm-forming ability into four categories: negative (OD ≤ ODc); weakly positive (ODc < OD ≤ 2 ODc); positive (2ODc < OD ≤ 4 ODc); and strongly positive (OD > 4 ODc) (39).

WGS and bioinformatic analysis

Genomic DNA from single clones in the exponential phase was extracted using the TIANamp Bacteria DNA Kit (TianGen Biotech, Beijing, China). Genomes were sequenced using the Illumina HiSeq platform (Illumina, San Diego, CA, USA) and the PacBio Sequel II system (Pacific Biosciences, Menlo Park, CA, USA). Raw reads were trimmed using Trimmomatic v0.35 and then assembled using SPAdes v.3.15.4. The hybrid assembly was conducted using Unicycler v0.5.0. The assemblies were annotated using Prokka v1.14.646 (40). Antimicrobial resistance genes, virulence genes, capsular serotype, and plasmid type were predicted using ResFinder, virulence factor database (VFDB), Kaptive v2.0.4 (41), and PlasmidFinder. Mutations in genes previously reported to be involved in FDC resistance were manually investigated. All available genomes of FDC-non-susceptible K. pneumoniae retrieved by the broth microdilution method from GenBank as of 1 February 2024, were included in this analysis (Table S3). Snippy v4.6.0 (https://github.com/tseemann/snippy) was applied to run core-genome SNPs calling and generate a phylogenetic tree based on the maximum-likelihood method with K. pneumoniae HS11286 (GenBank no. CP003200.1) was used as the reference. A phylogenetic tree was displayed and annotated using tvBOT (42). A comparative plasmid illustration was implemented using the BRIG or Easyfig tools. ISs, and other mobile genetic elements were predicted using the VRprofile (43).

qRT-PCR experiments

Real-time quantitative reverse transcription PCR was used to detect the expression levels of resistance genes in the FDC-non-susceptible and FDC-susceptible groups. The mean expression of the strains in the susceptible group was used as the reference value, and the 16sRNA was used as an internal reference (primers: q16sRNA_F: ATCGAGGAACGGAACCAACC, q16sRNA_R: GTGGATTCATCGCAGTTGCC, qKPC_F: GGCAGTCGGAGACAAAACC, qKPC_R: CCCTCGAGCGCGAGTCTA, qSHV_F: AGCCGCTTGAGCAAATTAAA, qSHV_R: GCTGGCCAGATCCATTTCTA, qCTX_F: TACCGCAGATAATACGCAGGTG, qCTX_R: CAGCGTAGGTTCAGTGCGATCC, qTEM_F: TCGGGGAAATGTGCG, qTEM_R: TGCCTTAATCAGTGAGGCACC). The expression level was assessed using TB Green Premix Ex Taq (Takara, Kyoto, Japan) in a LightCycler 480 system (Roche, Rotkreuz, Switzerland), with triplicate samples for each isolate, replicated thrice independently using the comparative CT method. 2^−ΔΔCT method was used for relative quantitation of the four resistance genes. The mean CT values from triplicate runs were used as input data (the CT value range for all triplicates was <0.5).

Statistical analysis

The Student’s t-test or Wilcoxon rank-sum test was performed for pairwise group comparisons of FDC MICs, expression levels of resistance genes, and biofilm formation. P < 0.05 was considered significant. All statistical analyses were implemented in GraphPad Prism 9.5.0.

ETHICS APPROVAL

Any personally identifiable information was removed from this study. This study protocol was approved by the Ethics Committee of the First Affiliated Hospital of Nanchang University.

DATA AVAILABILITY

Complete sequences of the four FDC-non-susceptible K. pneumoniae isolates have been deposited in NCBI with BioProject no. PRJNA1105644-PRJNA1105646https://www.ncbi.nlm.nih.gov/bioproject/PRJNA1105646/ and PRJNA1105649. The sequence accession numbers of all non-susceptible isolates are shown in Table S3.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/aac.00754-24.

Supplemental figures. aac.00754-24-s0001.docx.

Fig. S1, distribution of sequence types; Fig. S2, virulence assessment of the four nonsusceptible isolates.

Table S1. aac.00754-24-s0002.xlsx.

Genetic characteristics and antibiotic susceptibility of 477 CRKp isolates.

Table S2. aac.00754-24-s0003.docx.

Clinical characteristics of patients infected with FDC-nonsusceptible K. pneumoniae.

Table S3. aac.00754-24-s0004.docx.

Metadata and MICs to FDC of 40 FDC-nonsusceptible CRKP.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.