In this study, a total of 363 strains of Klebsiella pneumoniae (KP) were selected from multiple hospitals located in Ningbo, Zhejiang Province, China, from January 2019 to December 2021. These included 150 strains of carbapenem-resistant Klebsiella pneumoniae (CRKP) and 213 strains of carbapenem-sensitive Klebsiella pneumoniae (CSKP). This study was approved by the Ethics Committee of Ningbo Medical Center LiHuiLi Hospital, Ningbo University (KY2023SL347-01). The specimens were obtained from various sources, including sputum, throat swabs, wound secretions, blood, and urine.

We used the VITEK 2 Compact automated system (bioMérieux, France) for bacterial strain identification and antimicrobial susceptibility testing. Data on drug susceptibility can be found in Supplementary Table 1. Carbapenem resistance was determined using the broth microdilution method. The quality control strains used for antimicrobial susceptibility testing were Escherichia coli ATCC 25922 and Pseudomonas aeruginosa ATCC 27853 (purchased from the National Center for Clinical Laboratories, Ministry of Health). The definition of carbapenem-resistant Klebsiella pneumoniae (CRKP) in this study was based on the MIC breakpoints specified by the Clinical and Laboratory Standards Institute (CLSI) in 2023, which indicate intermediate or resistant levels to one or more carbapenem drugs.

Genomic DNA was extracted from the samples using a boiling method (Liao et al., 2020b). Subsequently, polymerase chain reaction (PCR) was employed to amplify various genes related to capsule polysaccharide-related genes (rmpA, rmpA2, magA), iron carrier-associated virulence genes (iucA, iutA, entB, aerobactin, iroB, ybts, kfu), fimbriae-related genes (fimH, markD, alls), lipopolysaccharide-related genes (wabG), as well as inner membrane transport protein (peg-344) and capsular typing genes (K1, K2, K5, K20, K54, K57). The primer sequences and annealing temperatures for each gene were provided in Supplementary Table 2. The PCR reaction consisted of 12.5 μL of 2X Taq MasterMix, 1.0 μL of forward and reverse primers each, 2.0 μL of DNA template, and 8.5 μL of ddH₂O, with a total volume of 25 μL. The PCR reaction conditions were as follows: initial denaturation at 94°C for 3 min, denaturation at 94°C for 30 s, annealing at the respective temperatures for 30 s, extension at 72°C for 1 min, with a total of 30 cycles, and a final extension at 72°C for 10 min. The PCR products were subjected to agarose gel electrophoresis on a 1.0% agarose gel at a voltage of 110 V for 30 min. Gel imaging and analysis were performed using the GELDOC XR gel imaging analysis system (Bio-Rad, USA), and the presence or absence of the target genes was determined based on the position of the marker bands. For positive bands, sequencing analysis was performed after gel purification.

The strains were inoculated onto Columbia blood agar plates using the streaking method in three zones. The petri dishes were then placed in a constant temperature incubator and cultured at 37°C with 5% CO₂ for 18–24 h. A sterile disposable inoculation loop was used to pick up bacteria from a single colony and streaked upward. If the length of the mucoviscosity string observed during streaking was greater than 5 mm, and this phenomenon could be observed repeatedly more than 2 times, the string test result was considered positive, indicating that the strain was hypermucoviscosity Klebsiella pneumoniae (hmKP). Conversely, if the string test result was negative, the strain was considered non-hypermucoviscosity Klebsiella pneumoniae (n-hmKP).

Genomic DNA of Klebsiella pneumoniae was first extracted using a boiling method. PCR was then used to amplify 7 housekeeping gene fragments (gapA, infB, mdh, pgi, phoE, rpoB, tonB). The primer sequences and annealing temperatures for each gene can be found in Supplementary Table 1. The PCR reaction system consisted of 12.5 μL of 2X Taq MasterMix, 1.0 μL of forward and reverse primers each, 2.0 μL of DNA template, and 8.5 μL of ddH₂O, with a total volume of 25 μL. The PCR reaction conditions were as follows: initial denaturation at 94°C for 5 min, denaturation at 94°C for 30 s, annealing at the respective temperatures for 45 s, extension at 72°C for 45 s, with a total of 35 cycles, and a final extension at 72°C for 5 min. The amplification products were sent to Shanghai Biotechnologies Corporation for sequencing. The sequencing results were uploaded to http://bigsdb.pasteur.fr/klebsiella/ and then compared with the database on the website to obtain the allele number and MLST type.

In this study, strains carrying the same virulence genes were selected from CSKP and CRKP, respectively, and their relative expression differences were analyzed using RT-qPCR. Details of the strains are shown in Supplementary Table 3. First, the Column Bacterial Total RNA Extraction Purification Kit (Bioteke, Shanghai) was used to extract RNA from the tested strains. Then, the One Step RT-qPCR Kit (Dye method) (Sangon Biotech, Shanghai) was used for RT-qPCR. The primer sequences can be found in Supplementary Table 1. Each gene was replicated three times. For data analysis, the 16S rRNA gene was used as the internal reference, and the 2^–△△CT method was used for calculation. Finally, GraphPad Prism 9.5 software (GraphPad Software, San Diego, CA, USA) was used for data visualization and analysis of relative gene expression levels.

We used Galleria mellonella larvae as an infection model (Ménard et al., 2021; Asai et al., 2023). The strains were derived from RT-qPCR consistently, as detailed in Supplementary Table 3. Galleria mellonella larvae from Huiyude (Tianjin) weighing 250–350 milligrams, pale yellowish white in color, and exhibiting good activity and responsive behavior were selected as experimental subjects. Firstly, pilot experiments were conducted using control strains with pre-set concentrations of 10⁷ CFU/mL, 10⁶ CFU/mL, 10⁵ CFU/mL of injection solutions to determine the optimal concentration by observing larval mortality. The criteria for determining larval mortality were blackening of the body and no response to touch (Figure 7A shows survival and Figure 7B shows death). Ten larvae were injected for each strain, with the injection site being the rear lateral side of the second-last abdominal proleg, and care was taken to prevent any obvious leakage during the injection process. The injected larvae were placed in sterile disposable culture dishes and incubated in a lightproof incubator (Shanghai Yiheng) at 37°C. The survival of larvae was recorded every 6 h, and a survival curve was plotted. A saline control group was set up (10 larvae injected with 10 μL of physiological saline). The negative control strain was Klebsiella quasipneumoniae ATCC 700603, and the positive control strain was NTUH-K2044 (hvKP) (purchased from the National Center for Clinical Medicine Examination, Ministry of Health).

In this study, we retrospectively analyzed the data of clinically isolated bacteria using WHONET 5.6 software. For data processing, we used SPSS 25.0 software (IBM, USA). The comparison of sample rates was analyzed using the Pearson’s chi-square 2 × 2 (2-sided) test or Fisher’s exact test (2-sided). Survival analysis was performed using the Log Rank (Mantel-Cox) test, with a P-value < 0.05 considered statistically significant. For the evaluation of virulence gene correlation, calculations, and graphical analyses were conducted using GraphPad Prism 9.5 software (GraphPad Software, San Diego, CA, USA).

In this study, a total of 150 carbapenem-susceptible Klebsiella pneumoniae (CSKP) strains and 213 carbapenem-resistant Klebsiella pneumoniae (CRKP) strains were collected. The specimens comprised urine, sputum, bile, drainage fluid, puncture fluid, and blood. We found significant differences in the types of specimens between the two groups (detailed in Figure 1).

A total of 363 KP strains were examined for virulence genes. The positive rates of the 5 virulence genes (rmpA, rmpA2, peg-344, iucA, and iroB) in KP were 30.9, 34.2, 32.0, 41.9, and 27.3%, respectively. Further comparing the detection rates of virulence genes in carbapenem-susceptible KP (CSKP) and carbapenem-resistant KP (CRKP), we found that CSKP had significantly higher positive rates of genes rmpA, peg-344, and iroB compared to CRKP (P < 0.01), at rates of 44.1, 42.7, and 44.6%, respectively. However, there was no significant difference in the detection rates of genes rmpA2 and iucA between the two groups (P > 0.05). Additionally, the detection rate of gene ybts in CSKP was 47.9%, significantly lower than that in CRKP (P < 0.01). Furthermore, CSKP and CRKP also showed significant differences in the detection rates of other virulence genes including magA, aerobactin, alls, kfu, and entB (P < 0.01) (Table 1).

For capsule genotypes, we found the number of K1, K2, K5, K20, K54, and K57 types were (35)9.6%, (21) 5.8%, (27)7.4%, (8) 2.2%, (8) 2.2%, and (12)3.3%, respectively. Furthermore, comparing the capsule genotypes between CSKP and CRKP, we found the detection rates of K1, K2, and K54 types in CSKP were significantly higher than in CRKP (Table 1).

We analyzed the distribution and correlation of virulence genes (Figure 2) and found that in CSKP, the genes rmpA, rmpA2, iucA, iroB, peg-344, and aerobactin tended to coexist. Additionally, the iutA gene showed a highly positive correlation with rmpA, rmpA2, iucA, iroB, and aerobactin. The entB gene showed a positive correlation with the wabG gene, and the K1 gene showed a positive correlation with the magA and allS genes. However, the K non-typeable gene showed a negative correlation with the rmpA, iroB, and aerobactin genes. In CRKP, the correlations among the genes were significantly reduced, the positive correlation was only found between the iutA and iucA genes, the fimH and rmkD, wabG genes, and between the entB and wabG genes.

Among the 363 strains of KP, 99 strains (27.3%) were identified as hypermucoviscous positive (hmKP). Among them, 90 strains (42.3%) in the CSKP group exhibited a high mucoviscosity phenotype, while only 9 strains (6.0%) in the CRKP group showed this phenotype, indicating a significant difference between the two groups (P < 0.01). We defined hvKP based on the detection of the rmpA, rmpA2, peg-344, iucA, and iroB genes. A total of 65 strains (17.9%) were determined as hvKP. Among them, the detection rate of CS-hvKP (30.0%) was significantly higher than that of CR-hvKP (0.7%) (Figure 3). At the same time, we divided the strains into sputum, blood, urine, and others sources (mostly sterile body fluids). The positive rate of hmKP was 22.2% in sputum, 36.2% in blood, 17.9% in urine and 35.7% in others isolates. The positive rate of blood isolates was similar to others isolates and higher than that of sputum and urine isolates. The detection of hvKP in each group also showed acacia results. The positive rate of hvKP was 13.5% in sputum, 21.3% in blood, 12.8% in urine and 25.0% in others isolates. The positive rate of blood isolates was similar to others isolates and higher than that of sputum and urine isolates (Figure 3).

Among the 363 strains of KP, the sequence type (ST) was successfully determined for 358. A total of 71 different ST types were detected. For CSKP, 58 different ST types were identified, with ST23 being the most common sequence type (n = 27, 12.7%), followed by ST65 (n = 11, 5.2%), ST268 (n = 10, 4.7%), and ST1764 (n = 10, 4.7%). On the other hand, for CRKP, 19 different ST types were identified, with ST11 being the most common sequence type (n = 75, 50.0%), followed by ST437 (n = 19, 15.2%) and ST268 (n = 10, 4.7%) (Figure 4).

Through examination of virulence characteristics and the sequence type (Figure 5), we observed that 98.5% of hvKP strains in this study were CSKP, while 1.5% were CRKP. Among hvKP strains, 65.65% were identified as hmKP. Additionally, the positive rate of capsule genes in hvKP was 95.38%, Among them, the detection rates of K1, K2, K5, K20, K54, and K57 were 47.69, 16.92, 6.15, 6.15, 4.62, and 13.85%, respectively.

Among hmKp isolates, the most common ST type was ST23 (n = 26, 26.3%), followed by ST412 (n = 7, 7.1%) and ST65 (n = 5, 5.1%). Among hvKP isolates, the most common ST type was also ST23 (n = 27, 41.5%), followed by ST412 (n = 7, 10.8%) and ST65 (n = 5, 7.7%). The ST types of these two groups showed a high degree of similarity. According to the classification of capsule serotypes, there were differences in the distribution of ST types among different capsule types. ST23 was primarily associated with K1 type (77.1%), while K2 type showed diversity, with ST65 being the predominant type (23.8%). K20 type was mainly associated with ST268 (62.5%) and ST368 (25.0%), while K54 type was mainly associated with ST29 (50.0%), and K57 type was primarily associated with ST412 (83.3%). More details can be found in Figure 5.

Based on the screening of virulence genes in this study, we selected genes rmpA, rmpA2, and iutA, which showed high positive rate in both CSKP and CRKP for RT-qPCR experiments (for detailed information on experimental strains, please refer to Supplementary Table 3). The results (Figure 6) showed significant differences in the expression of the same virulence genes among different strains. In CSKP, the expression level of the rmpA gene was significantly higher than that in CRKP (P < 0.01), followed by rmpA2 (P = 0.015), while there was no significant difference in the expression of iutA (P = 0.200).

We studied the impact of different KP strains on larval survival rates in the model of Galleria mellonella larvae. Control strains included the negative control strain ATCC 700603, the positive control strain NTUH-K2044 (representing hvKP), and larvae injected with physiological saline. The survival rates of abovementioned controls were 90.0, 50.0, and 100% respectively. We also selected strains positive or negative for genes rmpA, rmpA2, and iutA for injection experiments (Supplementary Table 3). Compared to the control groups, we found significant differences in larval survival rates after injection with different strains expressing virulence genes. Particularly, the larval survival rate after injection with rmpA-positive strain was 66.7%, significantly lower than rmpA-negative strain (90.0%) (P = 0.005) (Figure 7C). The larval survival rate after injection with rmpA2-positive strain was 55.8%, significantly lower than rmpA2-negative strain (89.0%) (P < 0.001) (Figure 7D). The larval survival rate after injection with iutA-positive strain was 62.5%, significantly lower than iutA-negative strain (93.3%) (P < 0.001) (Figure 7E). In hvKP strains, the larval survival rate after injection with CS-hvKP strain (13.3%) was significantly lower than CR-hvKP strain (70.0%) (P = 0.003) (Figure 7F).