Emergence of KPC-2-Producing Raoultella ornithinolytica Isolated from a Hospital Wastewater Treatment Plant

LETTER

Bacterial carbapenem resistance is a threat to the public health worldwide (1). Klebsiella pneumoniae carbapenemase (KPC) is the most common carbapenemase and can be obtained from carbapenem-resistant Enterobacteriaceae (CRE) that are distributed globally (2, 3). Although bla_KPC genes are usually recovered from samples from patients and healthy humans, transmission of bla_KPC genes between environmental isolates remains largely unknown (4). In this study, we isolated Raoultella ornithinolytica SRo1810 carrying bla_KPC-2 from a wastewater treatment plant (WWTP) located at the First Affiliated Hospital of Zhejiang University (FAHZU). Comprehensive analysis of R. ornithinolytica SRo1810 was performed to understand the resistant profiles and molecular characteristics of this isolate.

R. ornithinolytica SRo1810 was isolated from the WWTP in July 2017 by using selective medium as described previously (5). The carbapenemase gene was identified using PCR and DNA sequencing. The plasmid was characterized using S1-pulsed-field gel electrophoresis (S1-PFGE), and the location of bla_KPC-2 was identified with Southern hybridization. A plasmid conjugation experiment was carried out using Escherichia coli J53 as the recipient strain (5). MICs were established using a broth dilution method and were interpreted according to the EUCAST guidelines (http://www.eucast.org/). Genomic DNA was extracted using a DNA kit (Omega Bio-tek, Norcross, GA, USA). The DNA was subsequently sequenced using Illumina-HiSeq 4000-PE150 (Illumina, San Diego, CA, USA) and PacBio RS II platforms (Pacific Biosciences, Menlo Park, CA, USA). After sequencing, the complete genomic sequence of R. ornithinolytica SRo1810 was generated using Unicycler (6) by combining the sequencing results. Bioinformatic analysis was conducted as described previously (7–9).

As shown in Table S1 in the supplemental material, SRo1810 was resistant to all β-lactams tested but was susceptible to aminoglycosides, fluoroquinolones, tigecycline, and colistin (Table S1). Another β-lactamase gene, bla_TEM-1B, was further identified using the ResFinder database (https://cge.cbs.dtu.dk/services/ResFinder/). Interestingly, a variant of fosA7, which encodes fosfomycin resistance via a glutathione transferase, was detected in the genome via BLAST search analysis (Fig. S1A and B). Antimicrobial susceptibility tests confirmed that SRo1810 was resistant to fosfomycin. bla_TEM-1B and bla_KPC-2 were located in the plasmid, while fosA7 was located on the chromosome. In addition, we investigated the genetic environment surrounding fosA7, and the results suggest that mobile genetic elements may promote the transmission of the fosA7 gene. The insertion sequences ISL3 and IS3 are located upstream of the fosA7 gene, and IS110 and IS6 are located downstream of the fosA7 gene (Fig. S1C). Southern blot analysis demonstrated that SRo1810 carried at least five plasmids with sizes of approximately 150 kb, 120 kb, 80 kb, 40 kb, and 30 kb. KPC-2 was located on an ∼42-kb plasmid (designated pSRo1810-KPC; Fig. 1B and C). Whole-genome sequencing results indicate that plasmid pSRo1810-KPC is an IncN-type plasmid with a length of 42,848 bp and contains 80 protein-coding genes with a GC content of 49.7% (Table S2). The PFGE profiles of SRo1810 were distinct from the PFGE profiles of Ro24724 (Fig. 1A). In addition, comparative genomics analysis indicates that there are no isolates similar to this strain in the database (Fig. 1F).

• Open in new tab
• Download powerpoint

FIG 1

Comparison of genomic and plasmid profiles of R. ornithinolytica from clinical and sewage isolates. (A) PFGE patterns generated by XbaI digested total DNA of two isolates. (B) Plasmid profiles of two KPC-producing R. ornithinolytica isolates. (C) Southern blot hybridization of S1-nuclease digested DNA using a bla_KPC-2 specific probe. Lane1, sewage isolate (SRo1810); lane2, clinical isolate (Ro24724); M, XbaI digested total DNA of Salmonella enterica serotype Braenderup H9812 as a size marker. (D) Comparative analysis of genetic structures of bla_KPC-2 in pSRo1810-KPC with other plasmids. Arrows represent open reading frames (ORFs) and their direction of transcription. The shaded areas indicate regions of synteny between adjacent schematics; the matching percentage nucleotide sequence identity for each such region is indicated in the scale bar. Gene labels identify the locations of genes on pHS062105-3. The GenBank accession numbers of pHS062105-3 and pKPC-ECN49 are KF623109 and KP726894, respectively. (E) Sequence comparison of scaffolds (portions of genome sequences reconstructed from whole-genome sequencing [WGS] data) identified in pSRo1810-KPC with 2 bla_KPC-2-bearing plasmids, pHS062105-3 and pKPC-ECN49, in BRIG (http://brig.sourceforge.net/). Arrows indicate positions and direction of transcription of genes. Reference plasmid pHS062105-3 is indicated in pink in the inner circle. The color intensity in each ring represents the BLAST match identity. (F) Comparative genomics analysis of R. ornithinolytica isolates based on single-nucleotide polymorphism (SNP) analysis shown as a maximum likelihood-based phylogenetic tree built from core genome SNPs of 45 R. ornithinolytica isolates mapped to the reference genome of R. ornithinolytica strain 10-5426.

By comparison, we found that pSRo1810-KPC is an IncN3-type plasmid and matches the 42.85-kb plasmid pHS062105-3 (99.5% query coverage, 99.9% identity, GenBank accession number KF623109) from clinical KPC-2-producing K. pneumoniae in China. These two plasmids were found to share a similar backbone with a 41.3-kb plasmid, pKPC-ECN49, which was recovered from clinical Enterobacter cloacae in China (Fig. 1E). Compared to pHS062105-3, the Tn3-tnpA-tnpR element downstream of bla_KPC-2 was missing from pKPC-ECN49. These three plasmids were classified as IncN3-type plasmids and were found to share a similar 14-kb P-type type IV secretion system, which usually encodes a rigid pilus and is found on virulence plasmids (10); they also possess an unknown maintenance mechanism. All these plasmids lacked a tra module which encodes a primary pilus for conjugation, which explains the failure to transfer pSRo1810-KPC by conjugation.

After annotating the contig with the carbapenem resistance gene, we compared SRo1810 with three similar gene sequences for genetic environmental analysis. The bla_KPC-2 gene was found to be located on an 11-kb Tn3 transposon element in pSRo1810-KPC with the following linear structure: Tn3-ISKpn8-bla_KPC-2-ΔISKpn6-korC-klcA-unkown ORF-ΔrepB-Tn1721 (Fig. 1D). It is worth noting that SRo1810 shares the same genetic environment with pHS062105-3 from a clinical K. pneumoniae isolate. The Tn1721-based transposon is one of the most common elements carrying bla_KPC-2 and is widespread among Enterobacteriaceae tested at different locations in China (11). The backbone of pSRo1810-KPC was distinct from that of pRo24724. pRo24724 is approximately 450 kb long and belongs to the IncU incompatibility group (12). In addition, PFGE results indicate that SRo1810 is not directly related to the clinical strain Ro24724. There are few studies of R. ornithinolytica carrying KPC-2. However, an R. ornithinolytica sample isolated from wastewater from a hospital in Spain in 2013 carried an IncP-6-type plasmid with KPC-2; bla_KPC-2 was also located on the Tn3 transposon.

In summary, we report for the first time a KPC-2-producing R. ornithinolytica strain, SRo1810, isolated from a WWTP. To our knowledge, this is the first study to obtain an environmental R. ornithinolytica isolate carrying the carbapenemase gene. The clinical and sewage isolates of R. ornithinolytica obtained from the hospital may have distinct origins, but they share the same bla_KPC-2 gene and carbapenem resistance mechanism. This work may have important implications regarding the transmission of the bla_KPC-2 gene in hospitals. We emphasize the importance of improved multisectoral surveillance for carbapenemase-producing isolates in WWTP effluent, which may contribute to the spread of antimicrobial resistance genes in clinical settings and communities.