To determine the prevalence of tet(X)s-positive K. pneumoniae strains in animal-associated and clinical samples, 921 samples from pigs and humans in China were collected (Supplementary Table S1). Specifically, a total of 590 samples from a pig farm (56 swine nasal swabs; Guangdong province, China) and a swine slaughter house (419 swine nasal swabs, 67 swine anal swabs, and 48 skin swabs of workers; Guangdong province, China) were collected from October to November 2020. Besides, 184 human anal swab specimens including 171 hospitalized patients and 13 healthy individuals from hospital A collected in 2019 were also included. All samples were subjected to selection on brain heart infusion (BHI) plates containing tigecycline (2 mg/L). After incubation at 37°C for 16–18 h, the tigecycline-resistant colonies were selected. Then detection of the tet(X) genes was carried out by PCR using specific primers in Supplementary Table S2. Species identification was achieved by Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS). In addition, 147 clinical nonduplicate K. pneumoniae isolates were collected from bloodstream samples in four hospitals (named as B, C, D, and E) located in Guangdong province in China during the period 2008–2018, and all strains were tested for the presence of tet(X) genes.

The susceptibility to tetracycline (TET), chloramphenicol (CHL), ciprofloxacin (CIP), ceftazidime (CAZ), gentamicin (GEN), amikacin (AMK), cefotaxime (CTX), trimethoprim-sulfamethoxazole (SXT), imipenem (IMP), ampicillin (AMP), fosfomycin (FOS), and rifampin (RIF) was determined by the agar dilution method according to the Clinical and Laboratory Standards Institute (CLSI) guidelines (CLSI, 2019). MICs of MICs of tigecycline (TGC) and colistin (CT) were determined by the broth dilution method according to the guidelines of EUCAST (2021). The E. coil ATCC25922 served as quality control.

Conjugation experiments were performed to test the transferability of the tet(X4)-harboring plasmids, using a sodium azide resistant E. coli J53 as the recipient. Briefly, a culture of tet(X4)-producing isolates and the recipient strain E. coli J53 were mixed (ratio of 1:9) in BHI broth and subjected to incubation for 6–8 h. The mixture was then spread on BHI agar plates containing 100 mg/L sodium azide and 1 mg/L tigecycline to select transconjugants that had acquired the tet(X4)-harboring plasmid. Colonies that grew on selective plates after incubation for 16–24 h at 37°C were further confirmed by PCR and Sanger sequencing.

The Galleria mellonella infection model was used to test virulence and pathogenesis of tet(X4)-positive K. pneumoniae strains, which was carried out as described previously with slight modifications (Buckner et al., 2018). Overnight cultures of K. pneumoniae strains were washed with phosphate-buffered saline (PBS) and further adjusted with PBS to concentrations of 1 × 10⁶ CFU/ml. The larvae were injected with 10 uL of bacterial solution and negative control groups were inoculated with 10 uL of PBS, the hvKP4 K. pneumoniae strain was used as the hypervirulent control (Gu et al., 2018). Then the larvae were incubated at 37°C in the dark and the survival rates of the larvae were recorded. Kaplan–Meier survival curves were plotted using GraphPad Prism, and the log rank (Mantel-Cox) test was used to analyze the significant difference (p < 0.05) of the survival rates in G. mellonella infection model.

Genomic DNA of tet(X)s-positive K. pneumoniae isolates was extracted using the Qiagen Blood & Tissue kit (Qiagen, Hilden, Germany). DNA libraries were constructed with 350-bp paired-end fragments and sequenced using an Illumina HiSeq 2000 platform. In addition, the PacBio Sequel System was performed to long-read sequencing the TKPN_8 strain. The sequencing reads were assembled into contigs using SPAdes version 3.10 (Bankevich et al., 2012). Genome sequences were annotated using Prokka (version 1.13.3; Seemann, 2014). Multilocus sequence typing (MLST) of isolates was conducted using MLST v2.11 based on assembled contigs (Raisanen et al., 2020). The core genes of bacterial genomes were extracted and aligned using Roary (Page et al., 2015). RAxML v8.2.10 was used to construct a maximum likelihood phylogeny of the strains (Stamatakis, 2006), then the phylogenetic tree was visualized by iTOL (Letunic and Bork, 2016). The SNP (sequence-based) and core-genome-based MLST (cgMLST) strategies on BacWGSTdb 2.0 were used for source tracking bacterial pathogens, and the phylogenetic tree was generated and visualized by Grapetree (Feng et al., 2021). Capsule (K) loci was identified using Kaptive (Wick et al., 2018). To obtain the complete sequence of tet(X)s-carrying plasmid, we combined the sequencing data from the genomic DNA and the plasmids, and predicted gaps were closed by PCR and Sanger sequencing using primers listed in Supplementary Table S3. Plasmid replicons, insertion sequences, antimicrobial resistance determinants and virulence factors were determined using online tools.¹ Easyfig was used to visualize the genetic comparisons.

In this study, the prevalence of tet(X)s-positive K. pneumoniae was 0.22% (2/921). The two tet(X4)-carrying strains, designated as TKPN_3 and TKPN_8, were collected from different swine nasal swabs (one was isolated from the pig farm; the other was isolated from a swine slaughter house). No other tet(X) variants were detected in K. pneumoniae strains.

The MLST analysis declared that K. pneumoniae TKPN_3 and TKPN_8 belong to ST35 and ST193, respectively (Table 1). To conduct the phylogenetic analysis of TKPN_3 and TKPN_8, the whole-genome sequencing data of 45 K. pneumoniae strains were download from NCBI database, including 40 tet(X4)-negative strains of ST35 and ST193, and five tet(X4)-positive strains belonging to ST727, ST43, ST1418, ST35, and ST327. A phylogenetic tree for 47 K. pneumoniae strains based on SNPs of core genomes was constructed. It showed that the K. pneumoniae strains were clustered into three distinct clusters, and tet(X4)-positive strains were dispersed on different branches, all of which were isolated from China between 2019 and 2021 (Figure 1). For source tracking bacterial pathogens, the genome sequences of TKPN_3 and TKPN_8 were submitted to BacWGSTdb and none close strains were found based on SNP (sequence-based) strategy. The core-genome-based MLST (cgMLST) analysis showed that 86 and 8 strains were, respectively, closed to TKPN_3 and TKPN_8, but had more than 340 different alleles (Supplementary Figure S1). In order to explore the evolution and transmission of the tet(X)s-positive K. pneumoniae, more sequencing data may be needed for phylogenetic analysis.

Klebsiella pneumoniae TKPN_3 and TKPN_8 were resistant to tigecycline, tetracycline, rifampin, chloramphenicol and ampicillin, but were susceptible to gentamicin, imipenem, amikacin, fosfomycin and colistin. TKPN_3 and TKPN_8 showed resistance to tigecycline with MIC of 256 and 32 mg/L, respectively. In addition, TKPN_3 was also resistant to ciprofloxacin, ceftazidime, cefotaxime and trimethoprim-sulfamethoxazole, while TKPN_8 was susceptible to them (Table 1).

Except for the tet(X4) gene, WGS data revealed that TKPN_3 harbored other 18 antibiotic-resistant genes (ARGs), including three quinolone resistance genes (oqxB, oqxA, qnrS1), two β-lactams resistance genes (bla_CTX-M-3, bla_SHV-33), one sulphonamide resistance gene (sul2), one trimethoprim resistance gene (dfrA27), one rifampicin resistance gene (arr-3), one fosfomycin resistance gene (fosA6), one macrolide resistance gene [mph(A)], one tetracycline resistance gene [tet(A)], five aminoglycoside resistance genes [aac(6′)-Ib-cr, aadA16, aph(3″)-Ib, aph(6)-Id, ant(3″)-Ia], one chloramphenicol resistance gene (floR) and one lincomycin resistance gene [lnu(G)]. In the isolate TKPN_8, multiple resistance genes were also identified, including tet(X4) along with other 11 ARGs including three quinolone resistance gene (oqxB, oqxA, qnrS1), one β-lactam resistance gene (bla_SHV-61), one sulphonamide resistance gene (sul2), one fosfomycin resistance gene (fosA6), one tetracycline resistance gene [tet(A)], two aminoglycoside resistance genes [aph(3″)-Ib, aph(6)-Id], chloramphenicol resistance gene (floR) and one lincomycin resistance gene [lnu(G); Table 1]. In general, these two tet(X4)-positive K. pneumoniae strains have similar antimicrobial resistance phenotype and genotype. Moreover, TKPN_3 showed higher level of tigecycline resistance compared with TKPN_8, and correspondingly, mutations involving regulators of multidrug efflux pump were also found in strain TKPN_3. Amino acid mutations were identified in RamR at T141I and in Lon at A299T.

In addition to resistance genes, a total of 78 different virulence factors were identified among these two tet(X4) positive isolates. These genes include various siderophores such as enterobactin (encoded by entABCEF, fepABCDG, fes, and ybdA genes), aerobactin (encoded by iucABCD and iutA genes) and salmochelin (encoded by iroE gene). Moreover, the gene clusters tssABCD and tssFGHIJKLM encoding type VI secretion system, which was proved to be beneficial to bacterial competition, cell invasion, type-1 fimbriae expression and in vivo colonization in K. pneumoniae (Hsieh et al., 2019), were detected in TKPN_3 and TKPN_8. In addition, the capsular serotypes of the TKPN_3 and TKPN_8 strains were identified as KL71 and KL63, respectively. Then the virulence of two strains was demonstrated by Galleria mellonella infection model. After 12 h post-infection, the survival rate of G. mellonella larvaes injected with TKPN_3 and TKPN_8 was 50%. Overall, no difference in survival was observed for G. mellonella infected by TKPN_3 and TKPN_8 compared with HvKP4 (TKPN_3 vs. HvKP4, p = 0.606; TKPN_8 vs. HvKP4, p = 0.326; Supplementary Figure S2).

To further determine the transferability of tet(X4) gene, conjugation assays were conducted. It showed that the tet(X4)-carrying plasmids form TKPN_3 and TKPN_8 could be successfully transferred from K. pneumoniae to E. coli J53. The two tet(X4)-harboring plasmids were denoted as pTKPN_3-186k-tetX4 and pTKPN_8-216k-tetX4. The conjugation frequencies of pTKPN_3-186k-tetX4 and pTKPN_8-216k-tetX4 were (1.0 ± 0.1) × 10⁻⁴ and (5.4 ± 0.7) × 10⁻⁵ cells per recipient, respectively. Compared with the recipient strain E. coli J53, the MICs of the transconjugants J53/pTKPN_3-186k-tetX4 and J53/pTKPN_8-216k-tetX4 for tigecycline and tetracycline increased by 128-fold, respectively.

The genetic features of tet(X4)-carrying plasmid pTKPN_3-186k-tetX4 of TKPN_3 was analyzed in depth. To obtain the complete sequence of this plasmid, the sequencing data of the genomic DNA and the plasmids DNA were combined, and predicted gaps were closed by PCR and Sanger sequencing. pTKPN_3-186k-tetX4 was a 186,211 bp plasmid and encoded 218 predicted open reading frames (ORFs). The backbone of the tet(X4)-carrying plasmid pTKPN_3-186k-tetX4 showed a typical mosaic structure with multiple replicon types including IncHI1B(R27), IncHI1A and IncFIA (HI1). The genes involved in conjugation were identified including traC, traD, traG, traI, traJ, traU, and mobI. The tet(X4) gene and another five antimicrobial resistance genes bla_TEM-1B, ant(3″)-Ia, qnrS1, floR, and lnu(G) were found in pTKPN_3-186k-tetX4. The mobile element IS1R (IS1 family, 768 bp) was located upstream of tet(X4) and delta ISCR2 (IS91 family, 977 bp) was found in downstream of tet(X4), which may contribute to the transmission of the tet(X4) gene (Figure 2). Additionally, two complete insertion sequence IS26 and two truncated transposons Tn2 were identified downstream of tet(X4) in pTKPN_3-186k-tetX4. The pTKPN_3-186k-tetX4 shared a similar plasmid backbone (99% coverage and 100% identity) against p1919D62-1 (GenBank accession number: CP046007.1), which isolated from a swine E.coli strain 1919D62 in China. The other two plasmids with high similarity with our plasmid were pSZ6R-tetX4 (GenBank accession number: MW940627.1) and pAB4-4-tetX4 (GenBank accession number: MW940615.1). The two plasmids pSZ6R-tetX4 and pAB4-4-tetX4 were isolated from Citrobacter sp. and Klebsiella sp. strains, respectively, (Figure 2).

The complete sequence of another tet(X4)-positive plasmid pTKPN_8-216k-tetX4 was obtained by long-read sequencing. pTKPN_8-216k-tetX4 was 216,814 bp in size and also belongs to an IncHI1B(R27)/IncHI1A/IncFIA(HI1) type hybrid plasmid. pTKPN_8-216k-tetX4 exhibited a high similarity (74% coverage, 100% identity) with pTKPN_3-186k-tetX4, in particular, it had an additional ~30 kp fragment. This fragment was similar to pCY814036-iucA (GenBank accession number: CP093152.1; Figure 2). pCY814036-iucA was a multi-drug resistant and hypervirulent hybrid plasmid and isolated from a K. pneumoniae strain in China. pTKPN_8-216k-tetX4 also carried the conjugation-related genes traC, traD, traG, traI, traJ, traU, and mobI. In addition to tet(X4) gene, aph(3″)-Ib, aph(6)-Id, floR, sul2 and tet(A) were identified in pTKPN_8-216k-tetX4. It was worth noting that similar to pTKPN_3-186k-tetX4, tet(X4) gene was flanked by mobile elements IS1R and ISCR2 in pTKPN_8-216k-tetX4 (Figure 2).