Abstract

Introduction:

Klebsiella pneumoniae infections are a significant healthcare concern due to advancing antimicrobial resistance rates and virulence genes’ detection. Specifically, ceftazidime/avibactam resistance is increasingly diffused among K. pneumoniae clinical isolates, limiting therapeutic choices and complicating patients’ outcomes. Some authors documented the simultaneous presence of ceftazidime/avibactam resistance markers and specific virulence genes. According to these premises, we characterized ceftazidime/avibactam-resistant K. pneumoniae strains, exploring their diffusion at the Mediterranean Institute for Transplants and Highly Specialized Therapies (ISMETT) of Palermo, Italy.

Materials and methods:

A total of 14 ceftazidime/avibactam-resistant strains were retrospectively analyzed using whole-genome sequencing and Sanger validation.

Results:

Most strains belonged to ST101 and carried blaKPC-31, blaKPC-3, or blaKPC-34 genes, often with resistance-associated mutations. The analysed strains revealed multidrug resistance, frequent outer membrane protein alterations, and a significant spectrum of virulence genes. Additionally, plasmid analysis revealed sequence type–specific replicons potentially involved in gene dissemination.

Discussion:

These findings highlight a convergence of resistance and virulence traits in high-risk K. pneumoniae clones, enhancing the importance of MDR pathogens’ genomic surveillance in critical healthcare settings.

1 Introduction

Antimicrobial resistance reached a concerning worldwide incidence during the last decades, specifically referring to Klebsiella pneumoniae as one of the most diffused multidrug-resistant bacteria. This Gram-negative pathogen often reports resistance to the beta-lactam group, mainly depending on carbapenemase production. Furthermore, decreased cell permeability, loss of outer membrane proteins (Omps), and efflux pumps may contribute to beta-lactam resistance episodes (Karampatakis et al., 2023; Carattoli et al., 2021).

On the premise of K. pneumoniae’s remarkable colonization and dissemination within the human host, carbapenemase-producing strains became a major threat within healthcare settings, complicating therapeutic choices and critical patient management (Karampatakis et al., 2023; Carattoli et al., 2021). Literature data documented eight major multi-drug resistant (MDR) K. pneumoniae clonal groups (CG), whose denominations are CG15, CG20, CG29, CG37, CG147, CG101, CG258, and CG307. Remarkably, CG307, CG101, CG258, and CG147 emerged as common healthcare-associated infection etiological agents (Carattoli et al., 2021).

Novel beta-lactam/beta-lactamase inhibitor combinations (BLICs) contrasted the MDR for K. pneumoniae infections within critical hospital settings during the last decade. Ceftazidime/avibactam (CZA) combines a broad-spectrum third-generation cephalosporin (ceftazidime) and a beta-lactamase inhibitor (avibactam). This combination showed promising effectiveness due to the avibactam active action against Ambler classes A, C, and D beta-lactamases, along with numerous extended-spectrum beta-lactamases (ESBL), K. pneumoniae carbapenemases (KPC), oxacillinases (OXA-48), and cephalosporinases (AmpC) (van Duin and Bonomo, 2016; Stachyra et al., 2010; Aktaş et al., 2012). Several data demonstrated ceftazidime/avibactam reduced susceptibility or resistance among K. pneumoniae isolates, mainly due to metallo-beta-lactamase (MBL) production. Additionally, beta-lactamase variants, membrane permeability alterations, penicillin-binding proteins (PBPs), and efflux pumps contribute to the same phenomenon (Campogiani et al., 2023). Regarding KPC-producing K. pneumoniae, previously published studies reported Ω-loop mutations (insertion, substitutions, or deletion) in the blaKPC gene, such as the D179Y deletion (Alsenani et al., 2022). The ceftazidime/avibactam-resistant K. pneumoniae isolates further reduced therapeutic alternatives in the case of severe infections. Moreover, these MDR strains showed a consistent dissemination capability, inspiring hypotheses about the simultaneous virulence factors (Xiong et al., 2022; Antinori et al., 2020; Livermore et al., 2015; Winkler et al., 2015; Cuzon et al., 2011). These considerations required specific investigation and infection control measures to contain these clones’ diffusion within critical healthcare settings.

According to these assumptions, we propose an experimental analysis of ceftazidime/avibactam-resistant K. pneumoniae strains belonging to immunocompromised patients. The study aimed to investigate ceftazidime/avibactam resistance molecular mechanisms and eventual K. pneumoniae virulence factors through next-generation sequencing technologies.

2 Materials and methods2.1 Strains sampling

A retrospective study included ceftazidime/avibactam-resistant K. pneumoniae strains from the Mediterranean Institute for Transplantation and Highly Specialized Therapies (ISMETT) in Palermo, Italy. This Institute includes high-specialization transplant units for abdominal and cardiothoracic surgeries, along with their corresponding intensive care units. We analyzed single isolates from fourteen different patients recovered during a 1-year and 10-month period (February 2021–December 2022). The analysis specifically regarded ceftazidime/avibactam-resistant K. pneumoniae strains isolated from infection episodes, excluding colonizing isolates. Specifically, deep respiratory, blood, and urine K. pneumoniae isolates were included as infection specimens. Otherwise, K. pneumoniae emerging from surveillance samples (e.g., rectal or oropharyngeal swabs) were excluded, only indicating a bacterial colonization.

Identification and antimicrobial susceptibility had previously been performed by the Vitek 2 system (bioMérieux, Marcy l’Etoile, France) at the above-mentioned hospital. These strains were forwarded to the Molecular Microbiology and Antimicrobial Resistance (MMAR) Laboratory at the University of Catania (Catania, Italy), aiming to investigate K. pneumoniae resistome and virulome patterns. Clinical data and microbiological reports from all the involved patients were collected into a datasheet according to the Ethical Committee’s favourable opinion.

2.2 Phenotypic assays and antimicrobial susceptibility testing

The phenotypic assay known as the “string test” was performed on the included isolates. Despite the subsequent availability of whole-genome sequencing, the string test was used as an initial phenotypic, low-cost, and immediate screening tool to assess hypermucoviscosity, enabling the integration of genomic data with functional phenotypic information. The assay tested positive in the case of a thread-like string longer than 5 millimeters for a stretched K. pneumoniae colony through a sterile loop from the agar plate. The authors aimed to register this result for a subsequent comparison to the virulome patterns details. The EUCAST broth microdilution (https://www.eucast.org/fileadmin/src/media/PDFs/EUCAST_files/MIC_testing/Reading_guide_BMD_v_5.0_2024.pdf) confirmed all the K. pneumoniae strains’ susceptibility profiles, reporting definitive minimum inhibitory concentrations (MICs) values for ceftazidime/avibactam (CZA), amoxicillin/clavulanate (AMC), piperacillin/tazobactam (TZP), cefepime (FEP), ceftazidime (CAZ), meropenem (MEM), imipenem (IMI), aztreonam (AZT), meropenem/vaborbactam (MEV), ciprofloxacin (CIP), amikacin (AK), gentamycin (CN), and trimethoprim/sulfamethoxazole (SXT). All the gathered MIC values followed the EUCAST guidelines interpretation criteria (EUCAST, 2025).

2.3 Whole-genome sequencing

All the collected strains underwent a DNA extraction procedure through QIAGEN QIAamp® DNA Mini Kit (Ref. 51,304, QIAGEN, 40724 Hilden, Germany), according to the manufacturer’s instructions. Subsequent DNA quantification used the Eppendorf BioPhotometer® D30 and the fluorimeter Qubit dsDNA BR Assay Kit. These tools allowed a purity and quantity evaluation of the initial sample (Ref. 32,850, Invitrogen, 92,008 Carlsbad, CA, USA) (Bongiorno et al., 2022). The Illumina MiSeq platform managed the next-generation sequencing (NGS) analysis using 100 ng of each extracted sample.

This procedure followed the manufacturer’s instructions provided by the Watchmaker’s DNA Library Prep kit with Fragmentation–Watchmaker Genomics® (Ref. 7 K0013-024, 5,744 Central Avenue, Suite 100, Boulder, CO 80301, USA). Twist Universal Adapter System provided indexes (16 Indexes, 16 Samples) (Ref. 101,307, Twist Bioscience, HQ 681 Gateway Blvd, South San Francisco, CA 94080FAQ). The fluorometric Qubit dsDNA HS Assay Kit (Ref. Q32851, Invitrogen, Carlsbad, CA 92008, USA) quantified libraries, and the Agilent® High Sensitivity DNA Kit (Ref. 5,067–4,626) evaluated their quality. Illumina® provided the “Denature and Dilute Libraries Guide” to denature and dilute libraries, choosing 8.5 pM as the loading concentration. Finally, sequencing analysis utilized MiSeq Reagent Kits v3 (Ref.15043895, Illumina, Inc., 92,122, San Diego, CA, USA). The Local Run Manager v3 software created the Sample Sheet according to the corresponding software guide by Illumina (Calvo et al., 2024; Nicitra et al., 2025).

2.4 Sanger sequencing

In order to confirm mutations potentially involved in resistance to ceftazidime/avibactam and to achieve optimal coverage and very high base-level accuracy in the specific region assessed for the presence of a point mutation, all KPC-3 strains underwent Sanger sequencing of this gene. Primers were designed to cover the gene region responsible for resistance (primer sequences available in Supplementary Table 2). Polymerase chain reaction (PCR) was performed on strains harboring the blaKPC-3 gene (isolates 1, 4, 11, 13, and 14) using the QIAGEN Multiplex PCR Kit (Ref. 206,143, QIAGEN, 40724 Hilden, Germany). PCR products were then purified using the ExoSAP-IT™ PCR Product Cleanup Reagent (Ref. 78,205, Applied Biosystems™, Thermo Fisher Scientific Inc., 168 Third Avenue, Waltham, MA 02451, USA) before sequencing and labelling with the BigDye™ Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems™, Thermo Fisher Scientific Inc., 168 Third Avenue, Waltham, MA 02451, USA). Further purification was carried out using the BigDye XTerminator™ Purification Kit (Applied Biosystems™, Thermo Fisher Scientific Inc.). Prior to sequencing, all purified products were denatured at 93 °C for 3 min using 10 μL of formamide. The final products were then sequenced using the Applied Biosystems™ SeqStudio™ Genetic Analyzer (Applied Biosystems™, Thermo Fisher Scientific Inc.). Sequences were subsequently analyzed and compared with the KPC-3 wild-type gene (ARO:3002313) using FinchTV (Chromatogram Viewer v1.5).

2.5 Data analysis

The QIAGEN CLC Genomics Workbench software analyzed the obtained data following the User Manual for the CLC Microbial Genomics Module v22.0 released on 4 January 2022 (QIAGEN, Aarhus, 8,000 Denmark). This software used the CARD database (https://card.mcmaster.ca/) to define resistance, virulence, and sequence typing (MLST) genes. Fourteen paired-end bacterial raw reads were trimmed with TrimGalore (v0.5.0) (Martin, 2011; Krueger, 2025) to remove the adapter sequence. Furthermore, the bacterial genome was de novo assembled using Unicycler (v0.4.8) (Wick et al., 2017) with the Illumina-only assembly modality. Virulence factors, resistance genes, and capsule loci are derived from Kleborate (v2.2.0) and the Kaptive command (Lam et al., 2021; Wyres et al., 2016). The software Prokka (v1.13) defined bacterial annotation. Moreover, Unicycler output assemblies were aligned with several protein sequences to identify punctual mutations (Seemann, 2014). Heatmaps emerged from the R package Complex Heatmap software (Danecek et al., 2021). Plasmid analysis used the software Bactopia (https://pubmed.ncbi.nlm.nih.gov/32753501/) and its “bactopia tools” PlasmidFinder (Li, 2011).

The software Tormes confirmed virulence and resistance genes (Sievers et al., 2011). The samples were screened for the magA gene using AC: AY762939. ClustalO (Price et al., 2010) allowed phylogenetic tree creation, reporting alignment of sequences, and FastTree (Paradis and Schliep, 2019). Additionally, the tree illustration was developed from the R packages showing ape (Yu et al., 2017), ggtree (RColorBrewer, 2025), RColorBrewer (ggplot2, 2025), ggplot2, and ggn new scale (Campitelli, 2019). The reports of the quality score for the sample are available in Supplementary Table 3.

3 Results3.1 Ceftazidime/avibactam exposure

The 14 analyzed strains belonged to immunocompromised patients from an Italian Transplant Unit, during 1 year (2021–2022). Seven patients (50%) recorded previous CZA exposure, while six isolates (42.8%) reported carbapenem regimens. Microbiological investigations isolated CZA-resistant K. pneumoniae within 60 days after CAZ or carbapenem therapeutic approaches. Four patients (28.6%) revealed a KPC-producing K. pneumoniae colonization before the infection episode. Among these cases, strains 4, 5, 10, and 12 emerged from patients documenting previous stays in other hospital settings before the ISMETT admission. Regarding strains 5 and 12, the patient’s clinical history is registered as having Nigerian and Greek origins. The clinical reports recorded death for five patients (35.7%) due to CZA-resistant K. pneumoniae severe infection. Patients’ clinical features and microbiological details are reported in table 1 and Supplementary Table 1.

3.2 Sample features

The study analyzed 14 clinical isolates demonstrating CZA resistance and infection aetiology. The susceptibility profile confirmed resistance to amoxicillin/clavulanate, piperacillin/tazobactam, cefepime, ceftazidime, ciprofloxacin, and gentamycin for all the strains (100%). Otherwise, 13 isolates (92.8%) reported aztreonam and amikacin resistance. Ten K. pneumoniae (71.4%) showed trimethoprim/sulfamethoxazole resistance. Finally, resistance to meropenem, imipenem, meropenem/vaborbactam, and colistin emerged in two isolates (14.2%).

The NGS analysis revealed three different sequence typings. Specifically, this procedure confirmed ST101 for Italian patients (85.7%), while ST39 and ST307 (cases 5 and 12) emerged for 1 strain each (7.1%), respectively from Nigeria and Greece. All the isolates revealed a similar genotypic profile, reporting LPS O-antigen D-galactans encoded by the O1/O2v1 serotype in ST101 or O1/O2v2 serotype in ST39 and ST307. These serotypes differed by rearrangements in the rfb region. The analysis registered different capsular polysaccharides (k-locus) as K17, K23, and K102, respectively related to the wzi gene sequences 137, 172, and 137. Supplementary Table 1 summarizes the year, source, susceptibility profile, and genomic sample features.

3.3 Resistome analysis

We compared phenotypic susceptibility profiles to genotypic details derived from the sequencing analysis. Figure 1 includes antibiotic susceptibility (S) or resistance (R), detected beta-lactam resistance genes, and the eventual outer membrane proteins (Omps) gene mutations. As regards the blaKPC gene, several variants emerged: five strains (35.7%) documented blaKPC-3, seven strains (50%) showed blaKPC-31, which is directly involved in CZA resistance episodes; two isolated 14.3% revealed blaKPC-34 with D179Y and H281Y associated mutations.

Resistome details. The image reports antibiotic susceptibility (S) or resistance (R), detected beta-lactams’ resistance genes, and omp genes mutations for all the collected strains. OmpK36 dark green indicates the following mutations: A193G-Y201F-Y210W-N221H-L225T-G226D-206 D227E-K231V-L232P-T258S-E308D-H349R; case 5 mutations regard the following positions: A183T-T192G-Y201F-N221H-L225N-N229K-207$2301-S271F-L3071-1315 L-D344E-S346DH349R-N350R; case 12 instead is mutated only in two positions (V178P-H349R). OmpK37 in case 5 and case 12 are mutated in: N26169M–1,127 M-N229G-Q234H-E243D- D274G-V2761; the other cases are mutated in: N26-169 M-1127 M. Finally, LamB in case 12 is mutated in K49Q-L75V-N94Q-D97N-S103N-A109V–1,111 V; case 7, case 9, and case 11 possess a stop in position Q112,” other samples do not show mutations. S, susceptible; R, resistant, CZA, ceftazidime/avibactam; MEM, meropenem; MEM/VAB, meropenem/vaborbactam; MLST, multi-locus sequence typing: MIC, minimum inhibitory concentration.

Strains harboring blaKPC-3 recorded a CZA MIC range of 12–24 mg/L, while the highest values (MIC range: 64- > 256 mg/L) were developed from blaBPC-31 and blaKPC-34 isolates. Regarding blaKPC-3, 60% of the detected genes presented a two-amino acid deletion (Glu and Leu) at 167–168 positions. Isolates 1, 4, 11, 13, and 14 documented this condition, further analyzed and confirmed through Sanger sequencing (Figure 2). Furthermore, all strains showed blaSHV genes as blaSHV-212 (85.7%), blaSHV-28 (case 5), and blaSHV-187 (case 12). Three strains revealed blaOXA-1 and blaTEM-181, while one showed blaOXA-9 and blaTEM-181, and another showed blaOXA-1/9 and blaTEM-181.

Partial alignments of nucleotide and corresponding amino acid sequences of the blaKPC-3 gene from clinical Klebsiella pneumoniae isolates, illustrating a six-nucleotide deletion within the Q-loop region. The affected residues may contribute to reduced susceptibility to ceftazidime-avibactam.

On the other hand, one strain reported only blaTEM-181 without any blaOXA genes. Six isolates (42.8%) carried blaCTX-M-15. As regards membrane permeability alterations, all the strains outlined a wild-type ompK35. Moreover, all ST101 isolates shared the same mutation set in the ompK36 and ompK37 genes. All ST101 strains share the same mutation set in the ompK36 and ompK37 genes. An identical mutation pattern in the ompK36 and ompK37 genes originated from ST307 and ST39. Isolates 7, 9, and 11 revealed a synonymous mutation in lamB, while case 12 showed a significant lamB mutation pattern. A core resistance gene group appeared in all the collected isolates, reporting oqxA-B for efflux pump against quinolones, while five strains (35.7%) qnrB17 encoding peptides for alternative fluoroquinolones targets. Moreover, a group of analyzed strains reported markers related to aminoglycoside resistance (aph, AAC, aad, ANT, armA), sulfonamide resistance (sul1, sul2), trimethoprim resistance (dfrA12 and dfrA14), macrolides resistance (mphA and mphE), and phenicoles resistance (catB3 and catL).

3.4 Virulome analysis

Virulence gene detection was derived from a virulence factor database (VFD), revealing an extended panel for all the analyzed isolates. Figure 2 summarizes the recorded virulence markers. All strains showed salmochelin siderophore iroE genes, capsule-production-related gene galF, rcsA-B system, complete gene sets for type I (fimA-B-C-D-E-F-G-H-I-K) and type III (mrkA-B-C-D-F-H-I-J) fimbriae, and type VI secretion system genes (arcAB, tssF-G, sciN/tssJ, clpV/tssH, dotU/tssL, hcp/tssD, vasE/tssK, and vipA/tssB). Six isolates (42.8%) documented aerobactin genes (iucA-B-C-D and iutA) and the capsule-production-related gene rpmA2 (isolates 2, 7, 9, 11, 13, and 14). These isolates reported a negative string test result. Regarding enterobactin genes, all collected strains carried entB-C-D-E-F, fepB-C-D-G, fes, and ybdA loci. Furthermore, three isolates (21.4%) reported entA, and two isolates (14.3%) reported fepA. Focusing on yersiniabactin genes, 13 strains (92.8%) presented ybtP-Q-E-T-U-A genes, while 11 strains (78.6%) reported ybtS and ybtX genes. Lipopolysaccharide (LPS) synthesis genes were variably distributed among the analyzed strains. First, all the strains reported the wbbN and wzm genes, while most isolates (92.8%) had the glf gene. Twelve strains (85.7%) showed the wbbM gene, while four isolates (28.6%) revealed the kfoC gene and three (21.4%) the wbbO gene. For the VI secretion system, 13 isolates (92.8%) reported vipB/tssC, the same 2 strains (21.4%) (isolates 5 and 12) carried impA/tssA and KPHS_231 genes, and only 1 strain carried vgrG/tssI (7.14%) and tli1 genes (case 005). Figure 3 summarizes the virulome analysis results.

Virulome details for all the analysed Klebsiella pneumoniae isolates. Distribution of virulence genes across 14 different Cp-kp cases, with the genes grouped into distinct virulence categories, as indicated at the top. Each coloured square represents the presence of a specific virulence gene in a case. The categories include: Aerobactin (blue), Enterobactin (green), Capsule production (pink), Salmochelin (light blue), LPS (reddish purple), rcsAB (yellow), Fimbriae (type I–III) (purple), T6SS (orange), Yersiniabactin (light orange). The cases are listed on the right, with a colourful square indicating the presence of the corresponding virulence gene.

3.5 Plasmids and phylogenetic correlation

The analysis identified multiple plasmids related to various antimicrobial resistance genes. Specifically, Col(pHAD28), Col156, Col440II, ColRNAI, incFIA(HI1), IncHI1B (pNDM-MAR), IncFIB(pQil), IncFIB (K), IncFIB(pNDM-Mar), IncR, IncFII (K), IncFIC(FII), and IncI1-I(Gamma) emerged from the protocol.

All ST101 strains carried Col(pHAD28), IncR, IncFIA(HI1), Col156, and ColRNAI; the IncHI1B(pNDM-MAR) plasmid replicons appeared in isolates 2, 7, 9, 11, 13, and 14. The ST307 showed IncFIB(K) and IncFIB (pQil) plasmid replicons. Finally, the ST39 included IncFIB (pQil), ColRNAI, IncFIC(FII), and IncI1-I(Gamma) plasmid replicons. The distribution of the different plasmid replicons. The phylogenetic tree (Figure 4) explains the relationships between different CZA-resistant isolates depending on their genetic distances. The tree branches indicate how closely related these strains are, highlighting specific features such as sequence type (ST), the presence of diverse blaKPC genes, and plasmid replicons.

The phylogenetic tree displays the genetic relatedness among the 14 clinical isolates based on whole-genome sequencing data, revealing two major clusters. Sequence types (STs) are indicated by colored circles: pink for ST101, red for ST307, and light blue for ST39. Different blaKPC alleles are represented by distinct shapes: circle for blaKPC-3, triangle for blaKPC-31, and square for blaKPC-34. The case identifiers are followed by the isolation date (month.year). On the right side, colored bars indicate the presence of plasmid replicons in each isolate. Each color corresponds to a specific plasmid type: yellow: Col440II, electric blue: CoIRNAI, dark green: Col(pHAD28), red: Col156, light green: IncFIB(pQil), purple: IncFIB(K), pink: IncFII(K), dark blue: IncHIIB(pNDM-MAR), light purple: IncI1-I(Gamma), magenta: IncR, light orange: InSIA(HII), light grey: IncFIB(PNDM-MAR) replicons. This figure highlights the genomic clustering of isolates and the distribution of plasmid replicons, underscoring the sequence type-specific association of mobile genetic elements potentially involved in antimicrobial resistance dissemination.

According to the performed analysis, the strains could be divided into two major clusters, depending on their genetic relatedness. On the left, the image shows mostly ST101 isolates, along with sub-clusters indicating distinct groups within the same sequence types. On the right, the figure reports different sequence types (ST307 and ST39), highlighting their genetic divergence from the ST101 group. Plasmids were reported in the phylogenetic tree to support the difference between the two ST101 clusters.

4 Discussion

K. pneumoniae infections represent a significant healthcare challenge due to frequent resistance and virulence episodes. Published studies unequivocally demonstrated that clinical isolates frequently harbour both resistance and virulence genes (Muraya et al., 2022; Dong et al., 2022). These genes often correlate with specific STs, which have been previously documented among Southern Italian regions. Specifically, our study revealed a significant prevalence of ST101 according to other scientific data from the same geographical area (Loconsole et al., 2020; Mammina et al., 2012). Ceftazidime/avibactam resistance presents severe complications, drastically limiting therapeutic options (Alsenani et al., 2022). Our investigations explicitly identified distinct virulome characteristics in ceftazidime/avibactam-resistant K. pneumoniae strains, indicating a clear potential for hypervirulent and hypermucoviscous phenotypes (Dong et al., 2022).

We are undertaking a comprehensive analysis of these resistant strains to correlate the resistome and virulome information effectively. By using whole-genome sequencing (WGS), we reported interesting insights into our isolate collection, along with critical information regarding mobile genetic elements. Our results significantly enhanced our understanding of resistance mechanisms and the phenotypic traits of the strains. Notably, data from associated patients reveal that 50% had prior exposure to ceftazidime/avibactam, reinforcing previous findings (Aktaş et al., 2012). Otherwise, the remaining patients reported no exposure, indicating multiple pathways leading to ceftazidime/avibactam resistance. Variants of the KPC gene, particularly KPC-31, are definitively linked to ceftazidime/avibactam resistance (Guo et al., 2024), and our isolates containing KPC-31 displayed susceptibility to meropenem, consistent with findings in other ceftazidime/avibactam-resistant strains (Antonelli et al., 2019; Oliva et al., 2023). Antinori et al. (2020) described a KPC variant carrying D179Y deletion, demonstrating its ceftazidime/avibactam resistance. Our analysis registered a K. pneumoniae strain with the same characteristics. Moreover, KPC-31 variants occasionally evade phenotypic detection methods (Antonelli et al., 2019), highlighting the importance of integrating molecular technologies into microbiological surveillance and diagnostic workflows (Muresu et al., 2022).

Additionally, we recorded a strain with a two-amino acid deletion in the KPC-3 enzyme Ω-loop (positions 167–168). The above-mentioned deletion determines a mutant variant (Winkler et al., 2015) related to ceftazidime/avibactam resistance. This information leads us to hypothesize the introduction of specific phenotypic detection methods, such as chromogenic media, which previously allowed CZA resistance identification, along with targeted PCR for Ω-loop variants (Bianco et al., 2022). Livermore et al. reported ceftazidime/avibactam vulnerabilities in selecting mutations leading to resistance (Livermore et al., 2015). Furthermore, Winkler et al. documented the activity of CZA against a set of isogenic Escherichia coli strains carrying mutations within the U-loop of the KPC enzyme, demonstrating that substitution of amino acid at positions 165–179 enhanced ceftazidime affinity and prevented avibactam binding (Winkler et al., 2015).

During the experimental evaluations, we demonstrated some differences in KPC variants within the Italian country. Literature data reported KPC-31 also in other extended hospital settings located in Naples, Sassari, and Rome, along with KPC-66 to KPC-70, KPC-39, and KPC-49, which have not been described in Sicily (Carattoli et al., 2021; Muresu et al., 2022).

Our study evaluated membrane permeability mutations, correlating them to ceftazidime/avibactam phenotypic susceptibility profiles. Despite the presence of ompK36 and ompK37 mutated genes in all isolates, only some isolates registered high CZA MIC values due to a significant number of ompK gene mutations. Specifically, case 8 (ST101) reported numerous ompK36 mutations along with a CZA MIC value higher than 256 mg/L. Otherwise, case 12 (ST39) revealed a mutated lamB protein gene along with two ompK36 mutations, showing a ceftazidime/avibactam MIC value of 48 mg/L.

Certainly, plasmids play a fundamental role in KPC gene dissemination, along with spontaneous recombination and mutation episodes. For instance, the Tn4401 transpositions contribute to the blaKPC gene copy number increase, mobilizing them onto higher-copy-number plasmids (Cuzon et al., 2011). Our study demonstrated sequence-typing-related plasmid distribution. Among our ST101 strains, we identified the IncFIB(pNDM-Mar) and ColRNAI plasmid replicons, which are well-known in K. pneumoniae for carrying carbapenem-resistance genes and virulence markers such as aerobactin or rmpA2 (Zhao et al., 2022; Henson et al., 2017; Maugeri et al., 2025). Additionally, the ColRNAI plasmid replicon was also found in ST39, demonstrating similar characteristics. We further documented the IncHI1B (pNDM-MAR) replicon, potentially associated with OXA and CTX-M resistance genes, as well as aerobactin virulence genes. Other plasmid replicons identified, including Col156 and Col440II, currently lack sufficient empirical data regarding their resistance or virulence genes.

Unique to this study, we observed the IncFIA (HI1) and ColpHAD28 plasmid replicons, suggesting strong correlations with conjugation and resistance markers (Sherburne et al., 2000). Nonetheless, connections between these replicons and known resistance or virulence elements could not be established at this stage. The IncR plasmid replicon documented in our ST101 isolates may associate with markers for aerobactin and yersiniabactin, as well as SHV resistance genes (Loconsole et al., 2024). The IncFIB(pQil) derived from ST39 and ST307, highlighting these strains’ capability to express ceftazidime/avibactam high-level resistance due to the blaKPC-31 gene plasmidic localization (Loconsole et al., 2024). All our analyzed strains reported the IncFIIK plasmid replicon, which can be related to extended beta-lactamases (ESBL) production (Arcari et al., 2023). As a confirmation, all our strains revealed at least one ESBL resistance gene. All the investigations on the plasmid replicons had only an epidemiological impact in collecting preliminary information about the Inc. groups. A study limitation is the absence of a complete plasmid sequencing analysis; thus, future third-generation sequencing methodologies may be essential to define long-read analysis of plasmids. Our virulome analysis detected the rpmA2 gene for several ST101 strains.

Remarkably, this gene correlates with capsule hyperproduction and strain hypermucoviscosity (Arena et al., 2022; Russo and Marr, 2019; Gali et al., 2023) and was not documented among K. pneumoniae isolates in Northern Italy. Those isolates frequently showed yersiniabactin genes, while aerobactin genes were rarely associated with clinical isolates (Thorpe et al., 2022). Our ST101 isolates also reported aerobactin and yersiniabactin genes, accounting for a Kleborate virulence score of 4. Although the negative string test result, this score may define a hypervirulent strain according to previously published data and Kleborate assessments (Arcari et al., 2023; Arena et al., 2022; Russo and Marr, 2019; Bi et al., 2018; Russo et al., 2024; Choby et al., 2020). The resistome, virulome, and plasmid analyses revealed K. pneumoniae strains simultaneously harbouring resistance mechanisms, virulence genes, and markers diffusion through mobile elements. As a demonstration, all the investigated STs concurrently carried beta-lactam resistance markers and at least one siderophore gene category. A consistent percentage of similar markers could be carried by a plasmid replicon, whose content may be further investigated through ultimate generation sequencing analysis.

Unfortunately, the absence of long-read sequencing methodologies was one of the study limitations, suggesting the importance of searching for external collaboration in completing experimental analysis. Furthermore, the study included a few isolates, highlighting the need to extend experimental evaluations to other patient categories.

Unfortunately, the use of short-read sequencing methodologies prevented the complete reconstruction of plasmid structures and the unambiguous assignment of virulence and resistance genes to specific plasmid backbones. This limitation underscores the importance of establishing further corroborations to implement long-read or hybrid sequencing approaches in future studies, which would allow a more accurate resolution of plasmid architecture and gene linkages. Furthermore, the study included a few isolates, highlighting the need to extend experimental evaluations to other patient categories.

A future perspective may consist of multicenter evaluations of fragile patients carrying carbapenem-resistant K. pneumoniae from surveillance or significant clinical samples. This perspective could emphasize the essential role of molecular surveillance protocols in characterizing and identifying particular K. pneumoniae clones in critical hospital settings.

StatementsData availability statement

The original contributions presented in the study are publicly available. This data can be found here: NCBI BioProject, accession PRJNA1192262.

Ethics statement

The studies involving humans were approved by the Ethics Committee of ISMETT (IRRB/14/23 with approval of 5/6/23). The studies were conducted in accordance with the local legislation and institutional requirements. The human samples used in this study were acquired from a by-product of routine care or industry. Written informed consent for participation was not required from the participants or the participants’ legal guardians/next of kin in accordance with the national legislation and institutional requirements.

Author contributions

EN: Conceptualization, Investigation, Methodology, Visualization, Writing – original draft, Writing – review & editing. MC: Data curation, Writing – original draft, Writing – review & editing. IMPF: Investigation, Methodology, Writing – review & editing. NM: Formal analysis, Software, Writing – original draft. GP: Formal analysis, Software, Writing – original draft. DL: Writing – original draft. AC: Writing – original draft. FM: Writing – original draft. AM: Supervision, Visualization, Writing – review & editing. SS: Supervision, Visualization, Writing – original draft. DB: Conceptualization, Data curation, Formal analysis, Supervision, Visualization, Writing – original draft, Writing – review & editing.

Funding

The author(s) declared that financial support was received for this work and/or its publication. This research was partially supported by the EU funding with the MUR PNRR Extended Partnership Initiative of Emerging Infection Disease Project n. PE 00000007 INF-ACT.

Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declared that Generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

References

AktaşZ.KayacanC.OnculO. (2012). In vitro activity of avibactam (NXL104) in combination with β-lactams against gram-negative bacteria, including OXA-48 β-lactamase-producing Klebsiella pneumoniae. Int. J. Antimicrob. Agents39, 86–89. doi: 10.1016/j.ijantimicag.2011.09.012,

AlsenaniT. A.VivianiS. L.KumarV.TaracilaM. A.BethelC. R.BarnesM. D.et al. (2022). Structural characterization of the D179N and D179Y variants of KPC-2 β-lactamase: Ω-loop destabilization as a mechanism of resistance to ceftazidime-avibactam. Antimicrob. Agents Chemother.66:e0241421. doi: 10.1128/aac.02414-21,

AntinoriE.UnaliI.BertoncelliA.MazzariolA. (2020). Klebsiella pneumoniae carbapenemase (KPC) producer resistant to ceftazidime-avibactam due to a deletion in the blaKPC3 gene. Clin. Microbiol. Infect.26, 946.e1–946.e3. doi: 10.1016/j.cmi.2020.02.007,

AntonelliA.Giani