Introduction

VEB-1 (Vietnamese extended-spectrum β-lactamase) was initially described in 1999 in an Escherichia coli isolate obtained from a Vietnamese patient hospitalised in Paris [1]. The enzyme was found to confer resistance to broad-spectrum cephalosporins and be plasmid-encoded and located within a class 1 integron [1]. Since then, this Ambler class A enzyme along with its variants have been predominantly described in Enterobacterales, Pseudomonas spp. and Acinetobacter spp., and have been reported globally [2, 3]. To date, 31 VEB variants have been identified (http://bldb.eu/).

Avibactam, a non-β-lactam-based diazabicyclooctane (DBO) molecule, used in combination with the third-generation cephalosporin ceftazidime, has been shown to have potent inhibitory activity against Ambler classes A and C and to some class D β-lactamases [4]. Hence, the ceftazidime-avibactam (CAZ-AVI) combo is being used as last-resort option for the treatment of serious infections caused by carbapenem-resistant Enterobacterales, especially those producing KPC-type carbapenemases against which it usually exhibits excellent activity [5]. However, in recent years, resistance to CAZ-AVI has emerged among KPC-producing Klebsiella pneumoniae (KPC-Kp), usually mediated via the production of KPC variants and at the cost of carbapenem resistance that is subsequently almost fully lost [6]. VEB variants are known to be effectively inhibited by avibactam [7]. However, a novel VEB variant, namely VEB-25, which was initially reported in 2019 from two KPC-2-producing K. pneumoniae isolates obtained from patients hospitalized in two Greek hospitals, has been reported to be a source of CAZ-AVI resistance [8]. This study identified these two isolates as belonging to Sequence Types ST147 and ST258, harbouring the blaVEB-25 gene on ca. 150 kb InC plasmid and a ca. 170 kb IncC/IncR plasmid, respectively. A subsequent study reported an outbreak within two ICUs in a single hospital in Athens, which affected seven patients within a 5-week period in 2019, and the strain responsible was identified as an ST147 K. pneumoniae isolate co-producing KPC-2 [9]. To our knowledge, this variant has only been reported in Greece so far.

We report here the phenotypic and genotypic characterisation of a CAZ-AVI-resistant K. pneumoniae isolate co-producing KPC-2 and VEB-25, recovered from a patient in Switzerland.

Materials and methods

Bacterial isolates

A carbapenem-resistant K. pneumoniae isolate, N3418, was sent to the Swiss National Reference Centre for Emerging Antibiotic Resistance (NARA) for further investigation. Species identification was confirmed using the EnteroPluri-Test (Liofilchem https://www.liofilchem.com) and UriSelect 4 agar (Bio-Rad, Cressier, Switerland). The isolate was subjected to the detection of carbapenemase activity by the Rapidec Carba NP test (bioMérieux) and then to NG-Test CARBA 5 test (NG Biotech), according to the manufacturer’s instructions. KPC alleles were identified by PCR amplification and Sanger sequencing.

Antimicrobial susceptibility testing

MICs of antimicrobial agents were determined by broth microdilution in cation-adjusted Mueller–Hinton broth (BioRad) for all antibiotics. The preparation of the different BL/BLIs, namely, CAZ-AVI, CAZ-relebactam (CAZ-REL), CAZ-tazobactam (CAZ-TAZ), CAZ-clavulanic acid (CAZ-CLA), CAZ-vaborbactam (CAZ-VAB), aztreonam-AVI (AZT-AVI), imipenem-REL (IPM-REL), meropenem-VAB (MEM-VAB), piperacillin-TAZ (PIP-TAZ), were performed according to the CLSI guidelines [10], with a fixed concentration of the inhibitors at 4 mg/L for AVI, REL, TAZ and CLA, and 8 mg/L for VAB.

Sequence analyses

The blaVEB-1 and blaVEB-25 alleles were amplified using primers VEB-Fw (5′-GATGATGAGCTCGCAGTCGCCCTAAAACAAAG-3) and VEB-Rev (5′-GATGATGGATCCCAAATTGCACTTCAACCCGC-3′). Sequencing of the amplicons and recombinant plasmids was performed by Microsynth AG (Balgach, Switzerland).

Cloning experiments

The blaVEB alleles were amplified using the above primers, and cloned into pCR-Blunt II-TOPO (Invitrogen, ThermoFisher), before transformation into Escherichia coli Top10. Transformants were selected on plates supplemented with kanamycin (50 mg/L). Successful transformants were confirmed by amplification and sequencing of the blaVEB alleles.

Conjugation experiments and plasmid transformation

Mating-out assays of the blaVEB-25-harbouring plasmid was attempted using a sodium azide-resistant E. coli J53 as recipient, as previously described [11], with selection on sodium azide at 100 mg/L and ceftazidime-avibactam at 8/4 mg/L. Following failure of conjugation attempts, plasmid pVEB-25_IncC was transformed into E. coli Top10 by electroporation and selected on LB agar plates containing ceftazidime-avibactam at 8/4 mg/L. Transformants were confirmed by the amplification and subsequent sequencing of the blaVEB-25 and IncC repA gene.

Whole genome sequencing (WGS)

WGS was performed using both short and long-read sequencing technologies as previously described, using either a MiSeq (Illumina) or MinION Mk1C (Oxford Nanopore Technologies, Oxford, UK) platforms [11].

Assemblies were performed using the Shovill pipeline (https://github.com/tseemann/shovill) and contigs were annotated using Prokka [12]. STs, the presence of resistance genes and plasmid replicon types were determined using MLST 2.0, ResFinder 4.1 [13] and PlasmidFinder 2.1 [14], on the Center for Genomic Epidemiology platform (https://cge.cbs.dtu.dk/services/). Long-read sequencing reads were trimmed and corrected using Canu [15]. Hybrid assembles, using both short and long-read data, were performed using UniCycler [16].

IC 50 measurements

IC50 measurements were performed as previously described [17]. Briefly, β-lactamases from crude extracts were prepared and used for the measurement of the specific activity in 100 mM sodium phosphate (pH 7.0). Measurements were performed in a Genesys 10S UV/VIS spectrophotometer (Thermo Scientific) spectrophotometer using a wavelength of 262 nm for cephalothin. The 50% inhibitory concentration (IC50) for KPC variants was determined as the concentration of clavulanic acid (CLA), tazobactam (TAZ), REL, AVI, or VAB, that reduced the hydrolysis rate of 100 μM cephalothin (CEF) by 50%. Extracts were preincubated with inhibitor for 3 min prior to the addition of CEF. All measurements were performed in triplicate.

Results and discussion

Phenotypic and genotypic profiling of N3418

K. pneumoniae isolate N3418 was obtained from the urine of an 84-year-old male who had previously been hospitalised in Greece but had not previously received CAZ-AVI treatment. Susceptibility testing showed that N3418 was resistant to all beta-lactams tested, including the carbapenems and cefiderocol (FDC), in addition to the BL/BLI combination CAZ-AVI. The isolate remained susceptible to IPM-REL and MEM-VAB, but exhibited resistance to colistin (COL; MIC of 16 mg/L) (Table 1). N3418 was also resistant to ATM-AVI (using a provisional breakpoint of 4 mg/L), and high MICs to unconventional BL/BLI combinations CAZ-REL (128 mg/L, CAZ-CLA (256 mg/L), and CAZ-VAB (32 mg/L). PCR and Sanger sequencing identified the KPC-2 variant that could therefore not explain the resistance to CAZ-AVI. To ascertain the mechanism of CAZ-AVI resistance, the isolate was subjected to WGS which subsequently led to the identification of blaVEB-25. WGS identified that this isolate belonged to ST323 and harboured the blaKPC-2, blaVEB-25, and blaOXA-10 beta-lactamase genes (Table 2). ST323 has been identified as one of the most dominant lineages associated with ESBL phenotype in Australia, but is not related to the ST147 and ST258 previously identified as VEB-25 producers in Greece [17]. Plasmid replicon type analyses revealed a number of replicon types present, comprising Col440II, ColRNAI, IncC, IncFIB(pKPHS1), IncFIB(pQIL), IncFII(K), IncFII(pKP91), and IncR. Using a combination of long and short read sequencing, we could identify that the blaKPC-2 gene was encoded on a ~ 100 kb FIB(pQIL) plasmid and blaVEB-25 was encoded on a ~ 114 kb IncC plasmid. A deletion of the mgrB gene (involved in lipopolysaccharide synthesis) and disruption into the fiu gene were also found, likely contributing to COL and FDC resistance, respectively [19, 20].

Table 1 Susceptibility testing resulting of N3418 and recombinant E. coli strains producing VEB-1 and VEB-25
Full size table
Table 2 Genotypic characteristics of N3418 and pVEB-25_IncC
Full size table

Phenotypes of the recombinant E. coli strains producing VEB-25 and VEB-1, and relative enzyme kinetic measurements

Susceptibility testing of recombinant E. coli strains producing VEB-1 and VEB-25 showed that, relative to VEB-1, the VEB-25 recombinant strain exhibited increased MICs for AMC, PTZ, CAZ-AVI, CAZ-REL, CAZ-TAZ, CAZ-CLA, and ATM-AVI (Table 1). Fold changes for each BLBI ranged from 2- to 128-fold, with the greatest change being observed between the BL/BLI combination encompassing the beta-lactams (CAZ and ATM) and the DBOs, AVI and REL.

IC50 assays showed that the inhibitory activity of AVI, REL, clavulanic acid and tazobactam were considerably decreased against VEB-25 compared to VEB-1, correlating with the susceptibility testing results (Table 3). The Lys234 residue is conserved in most class A beta-lactamases [21], and has been indicated as critical for the inhibitory activity of avibactam toward KPC enzymes. Accordingly, this finding correlates with what has been observed from in vitro mutation studies in KPC enzymes [22]. Noteworthy, in silico Genbank analysis identified the VEB-20 variant exhibiting the same Lys234Arg substitution as VEB-25, encoded on the chromosome of a Pseudomonas aeruginosa recovered from France (GenBank accession number: NG_063894). However, the phenotype conferred by this enzyme remains unknown.

Table 3 Inhibitory concentrations of beta-lactamase inhibitors against VEB-1 and VEB-25
Full size table

Plasmid pVEB-25_IncC encoding VEB-25

The natural blaVEB-25-harbouring plasmid pVEB-25_IncC was transformed into E. coli Top10. Susceptibility testing showed that pVEB-25_IncC conferred resistance to a range of beta-lactams including CAZ, ATM, AMX, PIP and the BL/BLIs CAZ-AVI, PTY and AMC (Table 1). Plasmid pVEB-25_IncC was found to be an IncC replicon type and 114,198 bp. In addition to blaVEB-25, the plasmid also encoded 15 further resistance genes encoding resistance to multiple antibiotic classes; blaOXA-10 (beta-lactams), tetA (tetracycline efflux pump), qnrS1 (quinolones), cmlA5 (phenicols), arr-2 (rifiampicin), dfrA23 (trimethoprim), sul1 and 2 copies of sul2 (sulphonamides), and six aminoglycoside resistance genes (aph(3″)-Ia, ant(2″)-Ia, ant(3″)-Ia, aph(6)-Id and 2 copies of aph(3″)-Ib) (Fig. 1). The ~ 9 kb multidrug resistance region harbouring blaVEB-25 comprised; IS10A, blaVEB-25, aadB, arr-2, cmlA5, blaOXA-10, aadA1, qacE∆1, ∆sul1, IS26. This is partially similar to the ~ 14 kb multidrug resistance region previously described by Voulgari et al., however lacking a ~ 5 kb region that includes rmtB1 and blaTEM-1 [8]. However, pVEB-25_IncC also harboured a mercury resistance operon (merACPTR) and citrate-dependent iron (III) uptake system (fecIRABCDE), in addition to a the HigAB type II toxin-antitoxin system that is typical of IncC plasmids [23]. Interestingly, the fecIRABCDE gene cluster has recently been linked to reduced susceptibility to FDC and its presence, alongside the fiu gene disruption, may explain the FDC resistance levels observed here [24]. Conjugation attempts were unsuccessful; however, pVEB-25_IncC could be transferred to E. coli Top10 by electroporation. Analysis of the plasmid sequence using Mob-Typer [25] predicted that the plasmid was mobilizable but not conjugative, and further analysis of the plasmid content revealed that it contained only a partial transfer (tra) region, comprising traF, traG and traH.

Fig. 1
figure 1

Plasmid pVEB-25_IncC

Full size image

Conclusions

We characterized here a VEB variant responsible for CAZ-AVI resistance that has previously only been rarely reported in Greece. The amino acid change, Lys234Arg, that differentiates VEB-1 from VEB-25, clearly plays an important role in the activity of beta-lactamase inhibitors — most notably the DBO compounds. Overall, VEB-mediated resistance to CZ-AVI in K. pneumoniae has been reported rarely but should be considered when encountering resistant isolates.