Comprehensive Genome Analysis of a Human-Derived β-Lactam-Resistant Klebsiella variicola Isolate from China

Abstract

Introduction

Klebsiella variicola is an emerging bacterial pathogen commonly associated with opportunistic human infections.

Methods

In this study, we characterized and analyzed the core genome of a β-lactam-resistant K. variicola subsp. variicola isolate from a human in Kunming, Yunnan, China.

Results

The human-derived β-lactam-resistant K. variicola genome was assembled into 37 contigs, with a total genome size of 5,667,700 bp. K. variicola comprises multiple antimicrobial resistance (AMR) genes, including blaLEN-22, fosA, OqxA, and OqxB, as well as the virulence-associated factors fimH, nlpI, and iutA. The β-lactam-resistant K. variicola genome comprised K and O (KL30 and O3/O3a) serotypes. We also identified 29 SNPs from three different snpEff categories: 18 low, five moderate, and six modifiers. Subsequently, we identified five assembled replicons, including three plasmids. Plasmid pKp5-1, plasmid p15WZ-82_res, and plasmid pKP91, and two phage regions: phage region-1, which resembled Entero-phage-HK446, and phage region-2, which was questionable with phage-Erwini-vB-EhrS-59. Furthermore, secondary metabolites, such as redox-cofactor, butyrolactone, azole-containing-RiPP, terpene-precursors (two distinct clusters), NRP-metallophores, and RiPP-like gene clusters were also identified in the K. variicola genome.

Discussion

Identifying key antibiotic resistance determinants, virulence factors, capsule serotypes, secondary metabolites, plasmids, and phage replicons emphasizes the zoonotic potential of this pathogen. Given the potential zoonotic implications, a multidisciplinary approach should be used to prevent the spread of β-lactam-resistant K. variicola.

Introduction

The Klebsiella pneumoniae complex belongs to the Enterobacteriaceae family and comprises seven related species, including Klebsiella variicola.¹ K. variicola is an emerging bacterial pathogen that can colonize humans, animals, and birds.^1,2 K. variicola was first identified as a new Klebsiella species in plants in 2004,³ followed by Klebsiella quasipneumoniae and Klebsiella similipneumoniae in 2014.⁴ The clinical significance of K. variicola is obscured by imprecise diagnoses that understate its prevalence, resulting in knowledge gaps and clinical implications for the healthcare system. According to Martínez-Romero et al (2018), some K. variicola and K. quasipneumoniae isolates previously thought to be K. pneumoniae were correctly identified using rpoB phylogenies and average nucleotide identity (ANI) methods.⁵ Advanced diagnostic tools, such as MALDI coupled to time-of-flight mass spectrometry can correctly identify bacterial pathogens.⁶ In addition, whole genome sequencing (WGS) and multilocus sequence typing (MLST) based on seven gene alleles (gapA, infB, mdh, pgi, phoE, rpoB, and tonB) analysis can identify emerging bacterial pathogens.⁶

The current global antibiotic burden may increase gene mutations, recombination, lateral gene transfer, and phage populations in specific environments.⁷ Improper waste discharge from human communities, hospitals, and poultry farms into the open environment contributes to the spread of AMR (antimicrobial resistance) pathogens.^7,8 A study tested 55 positive K. pneumoniae isolates from 120 hospitalized patient samples, and most were resistant to cefazolin, ceftriaxone, ampicillin, ceftazidime, and cefepime. Some isolates tested positive for extended-spectrum β-lactamases, with the overall rates of multidrug-resistant (MDR) and extensively drug-resistant isolates being 57.2% and 10.9%, respectively.⁹ Consequently, the presence of resistant pathogens and antibiotic residues in surface and underground water promotes the growth of resistomes.^7,10 K. variicola, similar to other Klebsiella, causes hospital- and community-acquired bloodstream, respiratory tract, and urinary tract infections, as well as surgical wounds in humans and animals.⁸

Several human-derived K. variicola isolates are known to carry high-risk AMRs and virulence factors (VFs). A study in the United States found five K. variicola clones associated with infections in trauma patients. Compared to K. pneumoniae, K. variicola is more likely to be involved in bloodstream infections and less likely to be MDR phenotypes. None of the K. variicola isolates were found to have hyper-microviscosity.¹¹ De-Campos et al (2021) characterized five K. variicola strains previously identified as K. pneumoniae. A closer genomic look revealed that three of the five strains shared a close genomic cluster with K. variicola, with the remaining two being K. pneumoniae and K. quasipneumoniae. All five strains were MDR, with three exhibiting virulence phenotypes, ie, the ability to invade epithelial cells and survive in human blood and serum.¹² These findings revealed the emergence of new K. variicola clones with pathogenic potential for various hosts, raising serious concerns about human health. K. variicola, isolated from the blood and cerebrospinal fluid of patients with post-surgical meningitis in Japan,¹³ may aid physicians in identifying pathogens and managing Klebsiella phylogroup infections effectively. MDR K. variicola str. UFMG-H9 and UFMG-H10, isolated from the urine of healthy zebuine cattle (Gyr heifers) at the Agricultural Institute of Minas Gerais in Brazil, had a higher genome similarity to isolates from the human urinary tract than those from bovine feces/milk samples.¹⁴ Given K. variicola’s presence in human feces, the observed MDR phenomenon raises concerns about its potential as a human pathogen. Some Klebsiella species develop drug-resistance phenomena, making treating infections more difficult.⁸ In this context, antibiotic-resistant Klebsiella infections have responded to bacteriophages (simply phages) that target specific bacteria;¹⁵ however, this is not always the case, as phage-mediated transduction can spread antibiotic-resistance determinants in various niches.¹⁶

Phages are abundant and diverse predators that influence the structure of the microbial community in the ecosystem and have tremendous therapeutic potential for bacterial infections. Phages can have two lifestyles: virulent, which means that they always kill their host through direct replication, and temperate, which means that they can integrate into the host genome as prophages to form lysogens.¹⁷ WGS analysis has helped phage therapy against emerging bacterial pathogens by identifying species-specific phage receptors, AMRs, and VF determinants, which are critical for developing effective phage-based treatments.^18,19 Recent genomic analyses of K. variicola have revealed distinct outer membrane structures and capsule serotypes that influence phage susceptibility, which can guide the selection and engineering of lytic phages for this species.¹⁹ In addition, phage detection within the K. variicola genome sheds light on phage-host interactions and the potential resistome. Our findings may be essential for optimizing phage combinations, preventing the evolution of resistance, and ensuring the efficacy of phage therapy as an alternative to antibiotics in the treatment of Klebsiella infections. Similar to phages, plasmids are well-known mobile genetic elements (MGEs) that allow for the horizontal transfer of resistance genes that encode β-lactams and VFs.¹⁵ Molecular techniques, such as PCR-based replicon typing and conjugation assays, have long been used to identify plasmids; however, WGS has significantly improved plasmid detection by enabling the precise characterization of plasmid gene content and evolutionary relationships. K. variicola harbors a variety of plasmids, some of which are shared with K. pneumoniae, indicating frequent intraspecies genetic exchange.¹⁹ Plasmid-associated genes for siderophore production, capsule biosynthesis, and other VFs have been shown to improve bacterial fitness in ecosystems.^15,19

In this study, we characterized and comprehensively analyzed the core genome of a β-lactam-resistant K. variicola isolate from a diarrhea patient in China. We performed taxonomic annotation, phylogenetic analysis, MLST, and capsule serotyping of the β-lactam-resistant K. variicola isolate. We also revealed several AMR genes, VFs, and MGEs, including plasmids and phage types, in the β-lactam-resistant K. variicola. Our findings emphasize the importance of human-derived β-lactam-resistant K. variicola genomic surveillance in tracking the emerging drug resistance threats.

Materials and Methods

Isolation and Identification of Klebsiella Species

Clinical human fecal samples (suffering from diarrhea) were collected aseptically in 5 mL sterile tubes in Kunming, Yunnan, China, on April 22, 2025. Bacteria were grown and isolated on EMB agar using the serial dilution method following overnight incubation in a 37°C shaker. Three suspected Klebsiella isolates from the MacConkey agar (Hope Bio-Technology Co., Qingdao, China) plates were transferred to 50 mL sterile tubes containing BHI broth and incubated at 37°C for 24 h. After incubation, the culture tubes were centrifuged at 10,000 g for 10 min. The collected bacterial pellet was resuspended in sterile PBS (1.5 mL) and centrifuged for purification. A QIAamp DNA mini kit (Qiagen, Germany) was used to extract bacterial DNA, according to the manufacturer’s instructions. High-quality extracted DNA was tested for purity and concentration. The Nanodrop-2000 detected DNA purity, and a Qubit-4 fluorometer (Thermo Fisher Scientific, USA) was used. Finally, the extracted DNA was confirmed using a 1% agarose gel and stored at –80°C until further analysis.

This study complies with the Declaration of Helsinki and was approved by the ethics committee of Yunnan International Travel Healthcare Center, Kunming, Yunnan, China (protocol no. 2023001). Informed consent was obtained from the study participants before study commencement.

The extracted DNA was used to amplify the 16S rRNA gene using specific primers 27F and 1492R for species identification. The PCR assay was performed in a 25 μL reaction mixture containing 12.5 μL of 2× Taq Green PCR master mix (Thermo Fisher Scientific, USA), 0.5 μL of each forward and reverse primer, two μL of the DNA template, and 9.5 μL of nuclease-free water. The PCR conditions were set as an initial denaturation step at 95°C for 3 m, followed by 32 cycles of denaturation at 94°C for 20s, annealing at 50°C for 20s, and extension at 72°C for 1.2 m, and a final extension at 72°C for 10 m. The amplified PCR product was cleaned with ExoSAP-IT before sequencing using the BigDye terminator (v. 3.1 kit: Thermo Fisher Scientific, USA) and the SeqStudio Genetic Analyzer. Chromas (v. 2.6.6) was used to inspect raw chromatograms and assess sequence quality. Low-quality and primer sequences were trimmed to produce high-confidence reads. BioEdit (v. 7.2.6) was used to assemble forward and reverse sequences into consensus sequences, which were then compared to the NCBI nucleotide (BLASTn) database for closest taxonomy matches based on sequence identity and E-values.

Whole Genome Sequencing and Analysis

DNA Library Preparation, Draft Genome Sequencing, and Assembly

Following high-quality bacterial DNA isolation using the QIAamp DNA mini kit (Qiagen, Germany), a DNA library was prepared using the Rapid Barcoding kit (RBK004), following the manufacturer’s instructions. Libraries containing a Klebsiella isolate barcode were sequenced using GridION (flow cells, active pore number ≥800). The fast5 format was converted to fastq files using Guppy’s high-accuracy mode and the MinKNOW software. The Klebsiella isolate was de-barcoded using the same software used for base calling. The quality of the trimmed data was checked using MultiQC (v. 1.25). FiltLong (v. 0.2.0) was used for quality filtering. The filtered reads were assembled using the Unicycler assembler (v. 0.4.8). We then used Racon (v. 1.4.12) in conjunction with Pilon software with a minimum contig length of 300 bp and contig coverage of 5 bp to improve the draft genome assembly and call sequence variants of various sizes, such as large insertions and deletions, and to fill gaps. The CheckM (v. 2) software was used to assess the quality of the assembled genome. The QUAST quality assessment tool (v. 2.3) was used to evaluate the draft genome assembly. The assembled genome was annotated with the NCBI prokaryotic genome pipeline and the rapid annotation subsystem technology (RAST) tool, while VIGOR4 annotated phage sequences in the bacterial genome.

Taxonomic Annotation and Phylogenetic Analysis

The human-derived β-lactam-resistant K. variicola genome was annotated and compared with other Klebsiella to understand potential genes for functional categorization (systems and subsystems), phylogenetic analysis, and MLST typing, which was determined using a PubMLST database that uses seven housekeeping gene alleles (gapA, infB, mdh, pgi, phoE, rpoB, and tonB). The assembled K. variicola genome was annotated using Pharokka (v. 1.2.1). The MLST profile was used for in silico taxonomic classification. Pairwise BLASTp and Markov cluster algorithms were used to find similarities between the coding sequences (CDSs). The Krona and Sankey tools generate Krona and Sankey plots, space-filling displays with coloring, and interactive polar coordinates, allowing complex taxonomic hierarchies. Single-nucleotide polymorphisms (SNPs) with distinct snpEff impact categories, such as low, moderate, and modifiers, were revealed in the K. variicola genome using the PATRIC database. The PATRIC database contains reference and representative Klebsiella genomes that were used to build phylogenetic trees using fast bootstrapping to support the tree values. The MinHash tool identified the genome most closely, and a phylogenetic tree was constructed using PATRIC global protein families from related Klebsiella genomes. The MUSCLE program aligned multiple-protein sequences, and the nucleotides of each sequence were mapped to the protein alignment. Amino acid and nucleotide alignments were combined to create a data matrix, which was then analyzed as a phylogenetic tree using the RA×ML program.

Prediction of AMR Genes, VFs, Capsule Serotypes, and Secondary Metabolites

The ABRicate and AMRFinderPlus tools evaluated AMR determinants and VFs in the human-derived β-lactam-resistant K. variicola. In addition, ResFinder (with strict parameters) and the Comprehensive Antibiotic Resistance Database (CARD; v. 4.0.0: with strict/perfect criteria, ie, cut_off: strict/perfect) were used to identify the AMRs. Efflux pump-coding genes and VFs were annotated using the ABRicate pipeline with minimum percentage identity (70) and coverage thresholds (50). Kaptive (v. 2.0) was used to determine capsule serotypes in the K. variicola genome. Minimum inhibition concentration (MIC) assay was performed to determine the resistance potential of the K. variicola isolate. In addition, the K. variicola genome sequence was examined for the presence of secondary metabolite biosynthetic gene clusters using the antiSMASH server (v. 7.0).

Prediction of Plasmid Types and Phage Replicons

PlaScope (v. 0.9) was used to identify and locate the plasmid replicons. Circular plasmid maps were created using the Proksee server. The quality and completeness of the phage genomes were evaluated using QUAST and CheckV tools (with default settings). Pharokka was used to complete the draft genome annotation, which was screened for the presence of tRNAs using the tRNAscan-SE (v. 2.0.11). Circular plots were created using the Pharokka software. Proteome-based phylogeny helped us to understand the relationships between the human-derived β-lactam-resistant K. variicola phages and previously characterized phages using NCBI BLASTn. The VICTOR tool was used to study the evolutionary relationships among the Enterobacteriaceae phages. We used a viral intergenomic distance calculator to compare the ANI of our phages with that of other phage genomes. Genomes with >95% ANI were identified to belong to the same phage species. The Clinker pipeline was used to compare gene content and synteny between phage genomes based on a 50% similarity threshold. The phage TB server determined phage interactions with the host bacteria.

Results

Klebsiella Species Identification

Two suspected Klebsiella colonies were isolated from EMB agar plates. A 16S rRNA gene PCR assay confirmed the presence of the K. variicola isolates. The 16S rRNA gene query for K. variicola showed the highest similarity (100%) with K. variicola str. F2R9 (NR_025635.1), 99.71% with K. pneumoniae str. JCM1662 (NR_112009.1), 99.63% with K. pneumoniae str. DSM 30104 (NR_117686.1), and K. pneumoniae str. ATCC 13883 (NR_119278.1) in the GenBank database. After WGS quality filtering, the total genomic length of the human-derived β-lactam-resistant K. variicola was 5,667,700 bp, with an average depth of 261.6.

In silico Taxonomic Classification and Phylogenetic Analysis

Krona and Sankey plots visualized bacterial taxonomy at the species level. Kraken’s taxonomic classification system, which employs exact k-mers, was matched to its lowest ancestor. The Kraken’s classified human-derived β-lactam-resistant K. variicola to species K. variicola, K. pneumoniae, K. michiganensis, and other closely related species (Supplementary Figure S1). In addition, the human-derived β-lactam-resistant K. variicola subspecies variicola sequence reads were assembled into 37 contigs with an average GC content of 57.23%, an N50 of 516,484 bp, 5529 CDSs, 80 tRNAs, three rRNA genes, three plasmids, and two phage regions (Table 1).

Table 1.

Genome Assembly and Annotations of a Human-Derived β-Lactam-Resistant K. Variicola

Genome Annotation	Number
Contigs	37
GC content	57.23%
Contig L50	4
Genome length	5,667,700 bp
Contig N50	516,484
CDS	5519
tRNA	80
rRNA	3
Hypothetical protein	752
Proteins with functional assignments	4767
Proteins with EC number assignments	1482
Proteins with GO assignments	1227
Proteins with pathway assignments	1095
Proteins genus-specific family assignments	5343
Proteins with cross-genus family assignments	5373

PATRIC genome annotation revealed that a human-derived β-lactam-resistant K. variicola isolate exhibited high similarity to reference genomes. Comparative genomic profiling revealed the closest matches to K. variicola str. LEMB11 (GenBank: CP045783), K. pneumoniae str. HS11286 (CP003223), K. quasivariicola str. 08A119 (CP084768), K. africana str. 200023 (CP084874), and K. quasipneumoniae str. KqPF26 (CP065838). In addition, our findings suggest a close phylogenetic relationship between the human-derived β-lactam-resistant K. variicola isolate and other members of the K. pneumoniae complex. Notably, our human-derived β-lactam-resistant K. variicola isolate clustered in the K. variicola str. LEMB11 clade, which was previously isolated from Garam Masala in the United Kingdom. Phylogenetic analysis of similar Klebsiella genomes suggested shared evolutionary features and potential functional overlap between specific taxa (Figure 1). Furthermore, ribosomal MLST analysis identified the assembled genome “rST 200763” with 100% similarity to K. variicola. According to the core genome MLST based on seven gene alleles (gapA, infB, mdh, pgi, phoE, rpoB, and tonB), the identified K. variicola subspecies variicola was grouped into sequence type 146 (ST146).

Figure 1.

Phylogenetic analysis based on WGS of K. variicola genomes and its closely related species. Our human-derived β-lactam-resistant K. variicola isolate from China (as shown with black bold texts) was clustered into the K. variicola LEMB11 clade, previously isolated in the United Kingdom. Phylogenetic construction determined the position of our K. variicola isolate among other reference genomes obtained from the PATRIC global protein families. Protein sequences of PGFams were aligned with the MUSCLE tool, and nucleotide sequences were overlaid onto the protein alignments for codon structure maintenance. Concatenated amino and nucleotide alignments were utilized to create a data matrix that was subsequently used to input phylogenetic inference into RaxML.

K. variicola Genomic Features and SNP Annotations

Comprehensive genome analysis was performed using the RAST tool, which assigned a unique genome identifier (244366.1524) to the human-derived β-lactam-resistant K. variicola. This pathogen contains 752 hypothetical proteins and 4767 functionally assigned proteins. Functional assignments included 1482 proteins with Enzyme Commission (EC) numbers, 1227 proteins with Gene Ontology (GO), and 1095 proteins assigned to pathways. PATRIC identified two types of protein families, ie, 5345 proteins belonging to the genus-specific families and 5376 belonging to the cross-genus families (Table 1). A circular K. variicola genome structure was used to assess the genetic diversity of core structure and functional genes (Figure 2a). These include, from the outer to the inner rings, contigs, CDS on both strands, RNA genes, CDS with homology to known AMR genes, VFs, and GC content. The CDS colors on the forward and reverse strands represent systems and subsystems (a collection of proteins) collaborating to perform specific biological, molecular, and cellular processes in the human-derived β-lactam-resistant K. variicola (Figure 2b).

Figure 2.

Beta-lactam-resistant K. variicola circular genomic graph. (a) The K. variicola circular graph was created to evaluate the diversity of core and accessory genes. (b) K. variicola genome structure and functional annotation to various subsystems.

The human-derived β-lactam-resistant K. variicola genome analysis revealed 29 SNPs with three distinct snpEff impact categories (18 low, five moderates, and six modifiers). Moderate missense mutations are known to alter protein sequences. Low-category snpEff impact synonymous mutations do not affect protein function, whereas modifier-category mutations occur in non-coding intron regions. The impact is determined by the type of mutation (whether the SNP causes a non-synonymous mutation or is located in the coding regions, introns, or promoter regions). Such classification allows us to understand how damaging each SNP may be associated with gene function. We identified 16 synonymous mutations in the low-snpEff impact category. A non-synonymous SNP occurs in the coding region (exon), altering the amino acid sequence and, as a result, gene function. All five non-synonymous mutations found in this study were in the moderate category, with only one insertion mutation in the modifier category (Supplementary Table S1).

GO Enrichment Analysis

GO enrichment analysis classified the human-derived β-lactam-resistant K. variicola genome into biological, cellular, and molecular processes (Figure 3). According to the biological process, the majority of the genes were annotated to cellular process (n = 1066), metabolic process (n = 906), biosynthesis process, gene expression, and translation (Figure 3a). In addition, the GO molecular function category was enriched with catalytic activity (n = 805), binding (n = 775), small molecule binding (n = 294), transferase activity (n = 270), and enzyme regulator activity (n = 21) (Figure 3b). Moreover, the GO cellular process was enriched with anatomical entity (n = 1301), cytoplasm (n = 746), intracellular (n = 825) activities, intracellular non-membrane-bounded organelle (n = 68), ribonucleoprotein complex (n = 15), and proton-transporting ATP synthase complex (n = 8) (Figure 3c). The detailed values of genes involved in GO enrichment processes are shown in the Supplementary Table S2.

Figure 3.

GO enrichment analysis classified all the genes into biological process, molecular function, and cellular process. (a) In the biological process, the majority of the genes were enriched in cellular, metabolic, biosynthetic, and gene expression. (b) The GO cellular category was enriched with anatomical entities, cytoplasm, and intracellular activities. (c). The GO molecular category was enriched with catalytic activity, binding, and small molecule binding.

Phenotypes, Acquired AMRs, Mutations Mediating AMR Genes, and VFs

The human-derived K. variicola phenotype comprised multiple AMR genes, including OqxB resistance to fluoroquinolone, ciprofloxacin, nalidixic acid, rimethoprim, and chloramphenicol. fosA was resistant to fosfomycin, whereas β-lactam-LEN22 (blaLEN22) was resistant to amoxicillin, ampicillin, iperacillin, ticarcillin, and cephalothin (Table 2). In addition, several acquired AMR genes were also revealed in the human-derived β-lactam-resistant K. variicola isolate. A chromosomal gene blaLEN22 (accession no. AM850912) was resistant to amoxicillin, ampicillin, piperacillin, ticarcillin, and cephalothin. FosA (ACZD01000244) was resistant to fosfomycin, whereas OqxB and OqxA (EU370913) were resistant to ciprofloxacin, nalidixic acid, trimethoprim, and chloramphenicol. We also revealed several chromosomal mutations mediating AMR genes, such as acrR (accession no. AJ318073.1), ompK35 (accession no. AJ011501.1), ompK36 (Z33506.1), ompK37 (AJ011502.1), ramR (KY465996.1), gyrB, and rpsL (Supplementary Table S3). In addition, CARD annotation (with strict cutoff parameters) identified multiple MDR efflux pump genes (KpnF, KpnE, and qacG) and major facilitator superfamily (MFS) efflux pump genes (leuO, emrR, KpnG, and KpnH). We also identified a general bacterial porin gene, OmpK37, with reduced permeability to β-lactam in human-derived K. variicola (Supplementary Table S4). Moreover, the presence of mobile-virulence-associated genes fimH (type-1 fimbriae), nlpI (NlpI lipoprotein precursor), and iutA (ferric aerobactin receptor) could play key roles in bacterial pathogenesis (Table 2). Moreover, the MIC susceptibility profile of the K. variicola isolate revealed an MDR phenotype, which was resistant to a variety of antibiotics, including β-lactam antibiotics, aminoglycosides, ciprofloxacin, enrofloxacin, doxycycline, tetracycline, chloramphenicol, fosfomycin, and linezolid. The isolate was resistant to imipenem but susceptible to meropenem. Amikacin, colistin, azithromycin, florfenicol, and sulfadiazine sensitivity remained unchanged. Overall, the findings suggest that the isolate has broad-spectrum resistance (Supplementary Table S5). Further, the β-lactam-resistant K. variicola genome was predicted to have human pathogenic capacity due to the presence of virulence-associated proteins, bacteriocin, invasin, type-VI secretion protein VasK, hydrolase, fimbrial adhesin, quinoprotein glucose dehydrogenase, ferric hydroxamate uptake, HTH-type transcriptional regulator MalT, fimbrial outer membrane usher protein, phosphoglycerol transferase, LPS-assembly protein LptD, translocation and assembly subunit TamB, ferrienterobactin receptor, ß-mannosidase B, glucose-1-phosphatase, outer membrane protein-W, and periplasmic alpha-amylase (Supplementary Table S6).

Table 2.

Core Genomic Characteristics of a Human-Derived β-Lactam-Resistant K. Variicola

MGE_resistance	Phenotype	Accession no.	Position in Contig	Coverage (%)	Identity (%)
blaLEN22	Piperacillin, amoxicillin, cephalothin, ampicillin, ticarcillin	AM850912	22,246-23,106	100.0	99.9
fosA	Fosfomycin	ACZD01000244	24,673-25,092	100.0	94.0
OqxB	Trimethoprim, chloramphenicol, cetylpyridinium chloride, benzylkonium chloride, ciprofloxacin, nalidixic acid	EU370913	422,497-425,649	100.0	97.7
OqxA	Trimethoprim, chloramphenicol, cetylpyridinium chloride, benzylkonium chloride, ciprofloxacin, nalidixic acid	EU370913	425,673-426,848	100.0	96.3
MGE_virulence	Protein function	Accession no.	Position in contig	Coverage (%)	Identity (%)
fimH	Type 1 fimbriae	NA	66540-67,028	100.0	95.1
mrkA	Type 3 fimbrial major subunit	ABW83989	84,741-85,355	100.0	92.4
iutA	Ferric aerobactin receptor	FLWH01000001	270,692-272,881	100.0	94.6
iutA	Lipoprotein NlpI precursor	CP000243	63,206-64,090	99.9	90.2
Plasmid/replicon types	Accession no.	Position in contig	Coverage (%)	Identity (%)	Mobility
K. variicola plasmid pKp5-1/FIA	KF954760	21,700-21,994	100.0	97.3	Non-mobilizable
K. variicola plasmid p15WZ-82_res/IncFII	CP000966	39,407-39,636	100.0	100.0	Non-mobilizable
K. variicola plasmid pKP91/IncFIB	JN233704	4006-4561	99.3	96.1	Non-mobilizable
Surface polysaccharide typing
K locus	Predicted capsule type	Confidence	O locus	Capsule type	Confidence
KL30	K30	Typeable	O3/O3a	O3/O3a	Typeable

K. variicola Serotypes and Metabolic Traits

The human-derived β-lactam-resistant K. variicola also contained capsule serotypes, the K locus (KL30) based on the wzi gene, and the O locus (O3/O3a). Seven biosynthetic gene clusters (BGCs) involved in the biosynthesis of secondary metabolites were also identified in the K. variicola genome (Figure 4). BGCs associated with the production of redox cofactors (Figure 4a), butyrolactone (Figure 4b), azole-containing RiPP (ribosomally synthesized and post-translationally modified peptides) (Figure 4c), two distinct terpene precursors (Figure 4d and e), NRP-metallophores (non-ribosomal peptide siderophores) (Figure 4f), and RiPP-like compounds (Figure 4g) were identified in the bacterial genome. The presence of these BGCs may suggest K. variicola’s potential metabolic capacity for producing bioactive compounds, which may help bacterial adaptation and survival in host tissues and extreme environments.

Figure 4.

Prediction of BGCs in the K. variicola genome. (a) BGCs associated with the redox cofactor, (b) butyrolactone, (c) azole-containing RiPP, (d and e) two terpene precursors, (f) NRP-metallophores, and (g) RiPP-like compounds were annotated in the human-derived β-lactam-resistant K. variicola assembled genome.

Annotation of Plasmid and Phage Replicons

The human-derived K. variicola genome comprised multiple AMR genes, VFs, and other crucial MGEs, some of which are shown in the circular graph (Figure 5a). We identified three non-mobilizable plasmids in the β-lactam-resistant K. variicola that were compatible with other K. variicola strains. Among these plasmids, K. variicola plasmid p15WZ-82_res/FIA, which was a non-mobilizable plasmid, showed 97.3% nucleotide identity with the previously sequenced plasmid (accession no. KF954760). In addition, K. variicola plasmid p15WZ-82_res/IncFII showed 100% nucleotide identity and coverage with the previously known plasmid (CP000966). K. variicola plasmid pKP91/IncFIB showed 96.1% nucleotide identity with the known plasmid (JN233704). Notably, several AMR, acquired, and other genetic elements with potential transfer capability were revealed on the plasmids. MGEs involved in integration/excision, replication/recombination, stability, transfer, and defense were shown on the K. variicola plasmid pKp5-1 (Figure 5b). K. variicola plasmid p15WZ-82_res carried AMR genes tral, traU, and traN, along with other MGEs (Figure 5c). K. variicola plasmid pKP91 carried a few crucial MGEs, such as relE and tnpA, compared to the other plasmids (Figure 5d). Moreover, the human-derived β-lactam-resistant K. variicola genome was intact, with two phage regions, phage region-1 and phage region-2. Phage region-1, comprising 62 phage genes, 42,050 bp in length, and a GC content of 48.78%, resembles Entero-phage-HK446 (accession no. NC_019714.1). Phage region-2 of 20,961 bp with a GC content of 51.79% was questionable with phage-Erwini-vB-EhrS-59 (NC_048198). The number and position of phage capsids, integrases, tails, hypothetical proteins, transposases, portals, terminases, and proteases are shown in Figure 6.

Figure 5.

Beta-lactam-resistant K. variicola genomic graph and plasmids with various genes. (a) Several important genes are shown on the bacterial circular genome. Three plasmids were identified in the β-lactam-resistant K. variicola. (b) K. variicola plasmid pKp5-1 with IncFIA replicon, (c) K. variicola plasmid p15WZ-82_res with IncFII replicon, and (d) K. variicola plasmid pKP91 with IncFIB(K) replicon were compatible with the K. variicola species.

Figure 6.

Identification of two intact phage regions with the K. variicola genome. The K. variicola genome was intact with two phage regions: phage region-1, which resembled Entero-phage-HK446, and phage region-2, which was completely questionable with phage-Erwini-vB-EhrS-59.

Discussion

The global emergence of a β-lactam-resistant K. variicola in humans has raised concerns about its potential role in zoonotic transmission. In this study, we characterized and comprehensively analyzed the genome of a ß-lactam-resistant K. variicola str. carrying blaLEN-22 isolated from human feces in Kunming. blaLEN-22 has shown resistance to several drugs, including ampicillin, amoxicillin, cephalothin, piperacillin, and ticarcillin, demonstrating K. variicola’s adaptability to antibiotic pressure. Our findings focused on AMRs, VFs, capsule serotypes, and plasmid and phage replicons in K. variicola, thereby emphasizing their epidemiological relevance. In addition, this study contributes to the importance of surveillance and genomic features in preventing drug-resistant microbes in humans and animals. K. variicola genomic surveillance revealed blaLEN-22 is a primary β-lactam determinant that degrades β-lactam antibiotics and makes them ineffective.²⁰ Our findings are consistent with a previous study in which several ß-lactamase genes, including blaLEN-2, blaLEN-16, blaLEN-17, blaLEN-19, blaLEN-21, blaLEN-22, and blaLEN-24, were found in K. variicola isolated from humans and animals,¹⁵ demonstrating their widespread presence in various niches.

In addition to blaLEN-22, the human-derived K. variicola str. contained other AMR genes acrA, acrAB, emrR, kpnG, kpnH, rsmA, and oqxA, which confer resistance to ciprofloxacin, tetracycline, gentamicin, fosfomycin, amoxicillin, chloramphenicol, and fluoroquinolone, indicating an MDR phenotype. Furthermore, efflux pump genes and the porin modification system suggest a multifactorial resistance mechanism that enhances the ability of the bacterium to resist broad-spectrum antimicrobial therapies.²¹ Our findings are consistent with previous reports that K. variicola exhibits high levels of drug resistance in humans.²² Klebsiella species, particularly K. variicola, harbor VFs, including iroE (siderophore esterase-iroE), clpV (type VI secretion system tssH), icmF (type VI secretion protein), iutA (ferric aerobactin receptor), acrB (acriflavine resistance protein-B), acrB (acriflavine resistance protein-B), KPK_RS07970 (capsule assembly protein Wzi), wbdA (LPS immune modulation), hlyA (hemolysin-A exotoxin), and gspD (T2SS secretin variant PulD). These VFs allow bacteria to colonize and persist in various hosts, resulting in chronic infections. K. variicola biofilm formation capability is particularly concerning because biofilms act as AMR reservoirs and facilitate horizontal gene transfer.²³ The interaction between resistance and VFs demonstrates that β-lactam-resistant K. variicola can adapt to different niches, posing infection prevention and treatment challenges.

In this study, the β-lactam-resistant K. variicola was closely related to isolates from human and animal infections, indicating potential human-animal transmission dynamics.²³ Our findings are consistent with those of previous studies showing zoonotic potential of K. variicola as a reservoir for various AMRs and VFs.^8,15 In addition, detection of capsular K and O loci in K. variicola sheds light on its pathogenic potential. The K KL30 and O O3/O3a loci predicted the conserved capsular serotypes for this species. KL30 has been previously identified in clinical ST29 isolates and may be associated with increased virulence and immune evasion.²⁴ Similarly, the presence of the O3b serotype in K. variicola may influence serum resistance and immune system interactions, allowing the pathogen to survive and cause infections.²⁴ The presence of these serotypes in the K. variicola genome demonstrates their importance in evolutionary conservation and potential pathogenicity.

Identifying biosynthetic gene clusters in the β-lactam-resistant K. Variicola emphasizes its ability to produce a diverse range of secondary metabolites.²⁵ Interestingly, we discovered gene clusters for the biosynthesis of redox cofactors, butyrolactones, azole-containing RiPPs, two terpene precursor clusters, NRP metallophores, and RiPP-like molecules. The presence of two distinct terpene precursor clusters demonstrates the metabolic versatility of the pathogen. Redox cofactors play central roles in regulating intracellular redox balance and can influence reactive oxygen species production, further enhancing microbial survival in extreme environments.²⁶ Butyrolactones are quorum-sensing signaling molecules that regulate gene expression according to the cell density. Azole-containing molecules, such as RiPPs and RiPP-like compounds, have shown antibacterial activity and an intricate architecture, suggesting that they may be involved in habitat colonization and species competition. Terpenes produced from terpene precursor clusters exhibit a wide range of biological activities, including antimicrobial and immunomodulatory functions. NRP metallophores are selective siderophores that aid metal ion acquisition under limiting conditions. The presence of these secondary metabolites suggests that K. variicola may not only exist in different environmental niches but also actively engage in microbial interactions with hosts.²⁷ The genomic potential of this organism may be the key to its involvement in beneficial and harmful interactions with hosts. Additional metabolomic profiling is required to determine the precise function of these biosynthetic gene clusters and assess their biomedical relevance.^26,27

The presence of MGEs, including plasmids and phage replicons, increases the likelihood of horizontal gene transfer, which can accelerate the spread of resistance in bacterial populations.¹⁵ These findings emphasize the importance of a One-Health approach to monitor and control the spread of MDR pathogens across different ecological niches. Three plasmids (IncFIA (pBK30683), pBK30683, and IncFII) in a human-derived β-lactam-resistant K. variicola have raised serious concerns because of their essential roles in AMRs and VF transmission. IncF plasmids, particularly those from the FIA and FII replicon families, are among the most frequently associated with MDR in Enterobacterales, particularly in Klebsiella species.²⁸ These plasmids contain β-lactamase determinants with a high transfer efficiency through conjugation, making them important contributors to global antibiotic resistance. Plasmid pBK30683, which belongs to the IncFIA and IncFII types, is likely a hybrid replicon capable of hosting AMR.²⁹ IncFIA and IncFII within a single plasmid structure can increase stability and adaptability, allowing AMRs to persist in Klebsiella hosts even without selective pressure.

IncFII plasmids contain blaCTX-M, blaNDM, and mcr1, which confer resistance to broad-spectrum β-lactamases, carbapenemases, and colistin, respectively.²⁹ Their ability to incorporate resistance islands from other MGEs accelerates the spread of MDR phenotypes.³⁰ Given the increasing prevalence of IncF-type plasmids in clinical and environmental niches, healthcare centers should investigate their stability, transmission, and recombination potentials. Molecular epidemiology and WGS studies can be used to monitor the spread of these plasmids and to detect emerging resistance patterns. Strategies, such as plasmid curing, phage therapy, and novel antimicrobial agents, should be investigated to address the threat of resistance-encoding determinants in emerging pathogens.^29,30 In addition to plasmids, the discovery of two phage replicons, Entero-phage-HK446 and phage-Erwini-vB-EhrS-59, which contain 62 phage genes, sheds light on the bacterial ecosystem diversity and functional potential. Entero-phage-HK446 and phage-Erwini-vB-EhrS-59 belong to the enterobacteria and Erwinia phage groups, respectively, and are known to infect Enterobacterales.³¹ Genetic analysis of these phages revealed a wide range of genes involved in phage functions such as replication, virion assembly, and host cell lysis. Phage lysin, holins, and tail fiber protein genes are especially important for host recognition and infection.^31,32 The detection of such phages in emerging pathogens may highlight their therapeutic potential, particularly in phage therapy against AMR bacterial infections. Furthermore, phage genes may underlie the complexities of phage-bacteria interactions, implying that these phages could play essential roles in the microbial ecosystem, promoting horizontal gene transfer and bacterial evolution.¹⁶ Future research should focus on functional validation of these genes and the application of phages in biocontrol strategies, particularly in human and animal settings.

Beta-lactam-resistant K. variicola isolated from human feces emphasizes the importance of enhanced antimicrobial stewardship and surveillance programs. Local clinics and hospitals are essential to monitor emerging pathogens in the human population. The high levels of AMR genes in K. variicola may emphasize routine antibiotic use, which may contribute to the evolution of resistance phenotypes. Stricter preventive measures (such as prudent antibiotic use) are required to slow the spread of resistant pathogens.³³ Furthermore, genomic surveillance should track emerging bacterial pathogens, their resistance, and their virulence determinants.

Conclusion

In conclusion, our findings offer genomic insights into the human-derived β-lactam-resistant K. variicola isolated from humans. Identifying key AMR genes, virulence-associated factors, capsule typing, secondary metabolite genes, plasmids, and phage determinants highlights the importance of the ongoing surveillance of this emerging bacterial pathogen. Given the potential zoonotic consequences, a One-Health approach should be used to prevent β-lactam-resistant K. variicola infections in humans.

Funding Statement

This study was supported by the National Key R&D Program of China (2022YFC2302700 & 2023YFC2307400).

Data Sharing Statement

The whole-genome nucleotide sequence data of this isolate was deposited in the GenBank database under bioproject number PRJNA1255725.

Disclosure

The authors declare no competing interests in this work.