A carbapenem-resistant C. youngae strain was isolated from the peritoneal effusion at a tertiary hospital in Shaanxi Province, Northwest China. Strain identification was conducted using Illumina MiSeq technology, which is capable of generating millions of short (100–300 bp), low-error (0.1%) paired-end reads. Minimum inhibitory concentrations (MICs) were determined using the Clinical and Laboratory Standards Institute (CLSI) broth microdilution method. With the exception of tigecycline, colistin, and cefoperazone–sulbactam, all drug breakpoints were interpreted according to CLSI M100-33 guidelines (Clinical and Laboratory Standards Institute, 2023). Tigecycline MICs were interpreted using the US Food and Drug Administration (FDA) Enterobacterales breakpoint (FDA, 2021), and the colistin MICs were interpreted using the European Committee on Antimicrobial Susceptibility Testing (EUCAST) standards. Quality control strains, E. coli ATCC 25922 and Pseudomonas aeruginosa ATCC 27853, were used for antimicrobial susceptibility testing. Carbapenemase phenotypes were detected using the imipenem-EDTA double-disk synergy test and the NG-Test Carba-5, while carbapenemase genes (KPC, NDM, OXA, IMP, and VIM) were confirmed using the PCR.

Conjugation and transformation experiments were carried out to investigate plasmid transfer. Briefly, C. youngae, a bla_KPC-2-positive isolate, was used as a donor strain, while E. coli J53 (azide-resistant) and E. coli DH5α were used as recipient strains. The conjugation test was carried out on Mueller–Hinton (MH) agar containing azide (100 μg/mL) and ampicillin (50 μg/mL), while the transformation test was performed on MH agar with ampicillin (50 μg/mL). The presence of the bla_KPC-2 gene and other resistance genes essential for conjugation was confirmed using antimicrobial susceptibility testing, PCR, and DNA sequencing. Identification of each conjugate colony was performed using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) with MALDI Biotyper software (bioMerieux, located in Marcy-l’Etoile, France) and further confirmed by next-generation sequencing (NGS). Plasmids carrying the bla_KPC-2 gene from the conjugants were extracted using the Qiagen Midi Kit (Qiagen, Hilden, Germany) and sequenced on the Illumina NovaSeq (Illumina, San Diego, CA, United States) short-read sequencer (150-bp paired-end reads). Sequencing reads were trimmed using a sickle (GitHub) and assembled de novo using SPAdes 3.12.0. Baseline calibration and assembly evaluation were done using Pilon 1.18. Open reading frame prediction and annotation were performed using RAST version 2.02 and BLAST at NCBI. Plasmid replicon types were determined using a PCR-based plasmid replicon typing method (Carattoli et al., 2005). Plasmid comparisons were performed using BRIG (Alikhan et al., 2011).

The bla_CMY-190 gene and its upstream promoter region were amplified from C. youngae-YS01. The pHSG398 vector was digested with EcoRI and KpnI using the ClonExpress RII One-Step Cloning Kit (Vazyme Biotech Co., Ltd., Nanjing, China). The linearised vector, purified PCR product, buffer, and enhanced recombinase were mixed and incubated for 20 min to obtain the recombinant vector. This recombinant vector was then introduced into E. coli DH5α by chemical transformation, followed by the determination of minimum inhibitory concentrations (MICs). In the presence of β-lactamase inducer (cefoxitin 10 μg/mL or cefotaxime 8 μg/mL), E. coli DH5α-PHSG398 and the positive control S. marcescens were incubated at 37°C for 2 h. The β-lactamase activity was then determined by a spectrophotometric method using the Amplite Beta-Lactamase Activity Assay Kit (AAT Bioquest Inc., America). A chromogenic β-lactam substrate (cephalosporin) was used, which changes color from yellow to red upon hydrolysis by β-lactamase. The assay was performed using an absorbance microplate reader, measuring the OD ratio at wavelengths of 490 nm to 380 nm.

Genomic DNA was extracted from overnight cultures of single bacterial colonies using a genome extraction kit according to the manufacturer’s instructions (Vazyme, China). DNA was sequenced using Illumina short-read sequencing (150 bp paired-end reads) (Illumina, San Diego, CA, United States). De novo sequence assembly was performed using SPAdes 3.12.0. For isolates suspected of carrying plasmid-encoded carbapenemases, long-read sequencing was conducted using the MinION instrument (Oxford Nanopore Technologies, Oxford, United Kingdom). Long-read sequencing libraries were prepared and multiplexed using a rapid multiplex barcoding kit (Oxford Nanopore Technologies catalog number SQK-RBK004) and sequenced on R9.4.1 flow cells. Base calling of raw reads was performed using Guppy v2.3.1 (Oxford Nanopore Technologies, Oxford, United Kingdom), and hybrid assembly, incorporating both Illumina short reads and Oxford Nanopore long reads, was performed using Unicycler v0.4.8-beta (Wick et al., 2017). Antimicrobial resistance gene analysis and draft genome annotation were performed using BacWGSTdb. Multilocus sequence typing (MLST) was performed using the mlst tool.¹ Genomic comparison of closely related isolates was performed using Proksee.² Antimicrobial resistance genes and plasmids were identified using the BLASTn analysis of assembled contigs against the ResFinder and PlasmidFinder databases, using a cutoff of 80% sequence identity and 80% sequence coverage. A maximum likelihood phylogenetic tree was reconstructed using MEGA11.

Citrobacter youngae was recovered from the peritoneal effusion of a 62-year-old female patient hospitalized for acute liver failure. During her hospital stay, the patient developed a fever, marked abdominal distension, weakness, malaise, and chest tightness on exertion. Initial treatment included intravenous administration of meropenem (1 g Q8H) for 4 days. The patient had several comorbidities, including hepatic malignancy, decompensated cirrhosis secondary to hepatitis B, ascites, acute renal failure, uremia, and portal vein thrombosis. Subsequently, a KPC-2 enzyme-producing C. youngae was isolated from the ascitic fluid, and the intravenous therapy was changed to ceftazidime–avibactam (2.5 g Q8H) for 7 days. Following this treatment, the patient’s temperature normalized, and the infection was effectively controlled. She was discharged 21 days after admission. The antimicrobial susceptibility profile of the C. youngae isolate is detailed in Table 1. The isolate was susceptible to ceftazidime–avibactam (MIC = 1 μg/mL), amikacin (MIC = 2 μg/mL), tigecycline (MIC = 0.125 μg/mL), and colistin (MIC≤0.25 μg/mL). In contrast, it was resistant to imipenem (MIC = 8 μg/mL), meropenem (MIC = 16 μg/mL), cefepime (MIC≥32 μg/mL), cefoperazone–sulbactam (MIC = 64 μg/mL), aztreonam (MIC≥128 μg/mL), and ciprofloxacin (MIC = 8 μg/mL). The results of ATCC strains in this study were within the quality control range.

The study also identified the presence of the plasmid-encoded bla_KPC-2 gene in C. youngae, marking the first detection of this gene within this bacterial species. This finding raises concerns about the potential for horizontal gene transfer of resistance genes. PCR-based sequencing confirmed the presence of bla_KPC-2 in C. youngae. In addition, the plasmid carrying bla_KPC-2 was successfully transferred from the C. youngae strain to E. coli J53 and E. coli DH5α. Both conjugants, C. youngae-YS01-E. coli J53 and C. youngae-YS01-E. coli DH5α, exhibited resistance to piperacillin–tazobactam, cefoperazone–sulbactam, and aztreonam, with at least a 4-fold increase in MICs for imipenem and a 60-fold increase for meropenem. Notably, C. youngae-YS01-E. coli DH5α showed higher MIC values for carbapenem antibiotics than C. youngae-YS01-E. coli J53 (Table 1). The conjugate C. youngae-YS01-E. coli J53 was resistant to piperacillin–tazobactam and aztreonam but sensitive or with intermediate susceptibility to imipenem and meropenem. In contrast, the conjugate C. youngae-YS01-E. coli DH5α was resistant to piperacillin–tazobactam, cefoperazone–sulbactam, and aztreonam but showed intermediate susceptibility or resistance to imipenem and meropenem. The MICs of meropenem and imipenem in both conjugates and the transformants increased by at least 4- and 60-fold, and 16- and 120-fold, respectively, compared to the recipient strains E. coli J53 and E. coli DH5α (Table 1).

We conducted phenotypic and molecular characterization as well as a functional study of a new CFE subtype designated bla_CMY-190, which is encoded by the chromosomal ampC gene and produced by clinical isolates of C. youngae. Sequencing of C. youngae-YS01 revealed an open reading frame (ORF) of 1,143 bp in length, encoding a putative protein of 381 amino acids. This protein showed a high identity (88.05%) with the plasmid-encoded enzyme CFE-1 (accession number NG_048757) when compared with the amino acid sequence of AmpC β-lactamase. The sequence of the cloned DNA fragment and the associated β-lactam drug resistance pattern identified a new CFE-1-like gene distinct from the E. coli plasmid genes for the transcriptional regulator AmpR and AmpC β-lactamase. The deduced amino acid sequence confirmed that the new gene encoding the β-lactamase CMY-190 (Sequence Information: OR896917) is a variant of CFE-1, characterized by 26 amino acid substitutions. In particular, bla_CMY-190 confers resistance to third-generation cephalosporins, including ceftriaxone, cefotaxime, and ceftazidime. Compared to the recipient E. coli DH5α-PHSG398, the MICs of cefotaxime, ceftriaxone, and ceftazidime increased 4-fold, while the MICs of cefepime and cefoxitin increased 8- and 32-fold, respectively (Table 2). To assess the inducibility of bla_CMY-190 expression, we measured β-lactamase activity in E. coli DH5α strains carrying plasmid pHSG398 using the spectrophotometric method with the chromogenic β-lactam (cephalosporin) nitro-cephalosporin as a substrate. After exposure to a β-lactam inducer (cefoxitin 10 μg/mL or cefotaxime 8 μg/mL), the β-lactamase activity in the bacterial cells was evaluated. The results showed limited induction of β-lactamase activity after exposure to cefotaxime, whereas cefoxitin did not induce further β-lactamase activity. In a parallel experiment, S. marcescens was used as a positive control. When exposed to cefotaxime (8 μg/mL), β-lactamase activity increased 4-fold. In contrast, E. coli DH5α transformants lacking the AmpR plasmid exhibited lower β-lactamase activity and MICs than those carrying AmpR, indicating that the expression of the bla_CMY-190 gene requires a functional AmpR regulator.

Whole genome sequencing (WGS) has identified several resistance genes, including the β-lactamase genes bla_CMY-190, bla_KPC-2, and bla_CTX-M-3; the aminoglycoside resistance gene aac (3)-II; and the fluoroquinolone resistance gene qnrS1. Together, these genes confer resistance to carbapenems, aminoglycosides, and quinolones. Restriction maps and nucleotide sequences were generated to further characterize these elements. Analysis of the bla_CMY-190 sequence revealed the presence of the genes blc (encoding an outer membrane lipoprotein), RcsA (encoding a transcriptional regulatory protein), and epmB (L-lysine 2, 3-aminomutase) downstream of bla_CMY-190. In addition, the ampR gene and the frdABCD operon of C. youngae were identified upstream of the ampC gene. The region surrounding the ampR–ampC genes in C. youngae contained both blc and frdABCD, suggesting a likely common chromosomal location. This genetic element appears to be associated with all known bla_CMY-190-carrying structures, whether chromosomal or plasmid-based (Figure 1). The bla_KPC-2 gene was found to be flanked by mobile elements related to Tn1721 and IS26, consistent with the predominant genetic structures carrying bla_KPC-2 in the domestic environment (Figure 2). This genetic configuration has recently been identified in several Citrobacter species (Zhu et al., 2020; Qiao et al., 2023; Hu et al., 2019). Sequencing results from C. youngae revealed that the CFE gene is located on chromosome 4878723bbp within the ST195 sequence type. Other chromosomal resistance genes include bla_CTX-M-3 and QnrB12. Comparative BLAST analysis indicated high similarity between this C. youngae chromosome sequence and a previously isolated strain from Korea (GenBank accession no. CP021963). A phylogenetic tree illustrating the related bla_CMY variants is shown in Figure 3. The bla_KPC gene was identified on a 64,927 bp IncN plasmid, which, based on BLAST analysis, shows high similarity to the plasmid sequence of an E. coli strain isolated from a hospital in Shanghai, China (79% coverage, 99.92% identity, GenBank accession no. CP028486) (Figure 4).