A 1000-mL sample of raw wastewater collected at a municipal sewage treatment plant in Argentina was filtered through 0.45µm membranes, which were cultivated for 18 h at 37°C in MacConkey broth supplied with 10 µg/mL meropenem, to select for bacteria with reduced carbapenem susceptibility. The resulting suspension was then streaked on MacConkey agar and colonies showing different morphologies were selected and assessed by PCR, using primers targeting bla _IMP, bla _VIM, bla _NDM, bla _OXA-48 and bla _KPC (Poirel et al., 2011). Two bla _KPC positive strains were isolated and bla _KPC-2 genes were identified in both strains by PCR amplification with primers KPC-F (5´ - ATGTCACTGTATCGCCGTCT - 3´) and KPC-R (5´- TTTTCAGAGCCTTACTGCCC - 3´) (Saba Villarroel et al., 2017) followed by direct amplicon DNA sequencing. Carbapenemase production was checked by a modified Hodge test and a positive result for synergy between imipenem (30 μg) and phenyl boronic acid (300 μg) containing disks. Both isolates were identified to species level by MALDI-TOF/MS (matrix-assisted laser desorption/ionization – time of flight mass spectrometry) (Bruker Daltonics GmbH, Bremen, Germany), and designated WW14A (Klebsiella pneumoniae) and WW19C (E. asburiae).

Antibiotic susceptibility testing was performed by disk diffusion for 14 antimicrobial compounds: ampicillin (10 µg), amoxicillin/clavulanic acid (20/10 µg), cefazolin (30 µg), cefoxitin (30 µg), ceftriaxone (30 µg), ceftazidime (30 µg), aztreonam (30 µg), meropenem (10 µg), gentamicin (10 µg), amikacin (30 µg), ciprofloxacin (5 µg), trimethoprim/sulfamethoxazole (23.75/1.25 µg), chloramphenicol (30 µg) and tetracycline (30 µg). Colistin susceptibility was assessed by broth dilution, while minimal inhibitory concentration (MIC) of the following β-lactam antibiotics was carried out by agar dilution according to CLSI 2021 guidelines (CLSI, 2021): ampicillin, piperacillin, piperacillin/tazobactam, cephalothin, cephalexin, cefoxitin, ceftriaxone, ceftazidime, cefepime, imipenem and meropenem.

The transmissibility of plasmids was tested with the filter mating protocol using E. coli J53 (sodium azide resistant) as recipient. Mating was initiated by mixing 0.5 mL donor bacteria suspension with 0.5 mL recipient bacteria (16 h culture in lysogeny broth; LB) followed by immobilization onto a 0.22-µm nitrocellulose membrane filter. Control cultures were prepared identically with either donor or recipient bacteria alone. Filters were incubated overnight at 35°C on LB agar plates (no antibiotics). Biomass was removed with a sterile cotton swab and re-suspended in 1 mL sterile saline. Fifty µL of suspended bacteria, either donor, recipient, or the mating combination, were spread onto LB agar plates supplemented with 50 µg/mL sodium azide and 10 µg/mL imipenem and incubated at 37°C for 18 h.

For whole genome sequencing (WGS) analyses, strains WW14A and WW19C were streaked to single colonies on MacConkey agar plates containing 2 µg/ml imipenem and then grown for 18 h at 37°C in 3 mL of LB. Total genomic DNA was extracted using PureLink quick gel extraction kit (Life Technologies, CA). DNA was quantified using the Qubit dsDNA HS assay system (Life Technologies, CA). Genomic library was constructed using the Nextera DNA Flex library preparation kit (Illumina, San Diego, CA) and, subsequently, sequenced on an Illumina MiSeq platform to generate 2 × 250 base paired-end reads. Sequenced reads were de novo assembled using SPAdes (v3.11.1) (Bankevich et al., 2012). Automated annotation was performed using PROKKA (1.14.6) (Seemann, 2014) and RAST server (https://rast.nmpdr.org/). Species identification was confirmed from assemblies by ANI analysis (http://enve-omics.ce.gatech.edu/ani/) and using the free online resource Pathogenwatch (https://pathogen.watch/). STs were defined by submitting genome sequences to official MLST databases for Klebsiella pneumoniae (https://bigsdb.pasteur.fr/klebsiella/klebsiella.html) and ECC (https://pubmlst.org/organisms/enterobacter-cloacae).

Capsular type for WW14A was assessed using BIGSdb (https://pubmlst.org/software/bigsdb) and KAPTIVE (https://kaptive-web.erc.monash.edu/) databases. Plasmid types and incompatibility (Inc) groups were defined based on their replicon genes (rep) using the PlasmidFinder database available at the Center for Genomic Epidemiology (CGE) (http://www.genomicepidemiology.org/). Other databases available at the CGE were also used to seek antimicrobial resistance determinants (ResFinder), virulence genes (VirulenceFinder) and mobile genetic elements (MGE). Prophages sequences were identified and annotated using the web server PHASTER (https://phaster.ca/), while the detection of CRISPR signatures was carried out using the online tool CRISPRfinder (https://crisprcas.i2bc.paris-saclay.fr/) (Grissa et al., 2007; Couvin et al., 2018).

Pangenome analysis of WW14A and WW19C was performed for constructing whole genome single nucleotide polymorphism (SNP)-based within-genus phylogenetic trees, selecting isolates analyzed in other studies ( Supplementary Table S1 ), i.e., isolates analyzed by Chavda et al. (2016), Gomi et al. (2018) and Jousset et al. (2019) for WW19C; and isolates analyzed by Gomi et al. (2018) and all Klebsiella quasipneumoniae genomes present in NCBI Assembly database (June 5^th, 2020) for WW14A, aiming to place WW19C and WW14A in broader phylogenetic contexts. Briefly, annotated genomes were used in a pangenome analysis with Roary 3.13.0 to generate a core-gene alignment (using a blastp percentage identity of 95% and a core definition of 99%) (Page et al., 2015). This core-genome alignment was used to generate a SNP alignment with SNP-sites (v 2.5.1) (Page et al., 2016) which was later used to construct a maximum likelihood (ML) phylogenetic tree with RAxML (v 8.2.12) (Stamatakis, 2014) under the generalized time reversible model (GTR) and 100 bootstrap replicates.

Nanopore sequencing was performed according to the manufacturer’s instructions. Briefly, DNA libraries were prepared using a rapid sequencing kit (Oxford Nanopore Technologies), and the prepared library was subsequently loaded into a MinION flow cell (R9.4; Oxford Nanopore Technologies). Raw data (FAST5 files) were base-called, converted to FASTQ format and trimmed of barcode and adapter sequences using Guppy v3.0.3 (Ueno et al., 2003). Hybrid assemblies of both short and long reads were performed using Unicycler v0.4.8 (Wick et al., 2017) and annotated with Prokka (1.14.6) (Seemann, 2014) and RAST server (https://rast.nmpdr.org/). Manual annotation of IncP-6 plasmids was additionally performed using National Center for Biotechnology Information (NCBI) BLASTn (Altschul et al., 1990) and Uniprot databases (Consortium, 2019). The annotation of ISs was performed using ISFinder (https://isfinder.biotoul.fr/) (Siguier et al., 2006). Final hybrid assemblies were compared to publicly available IncP-6 plasmids using BRIG (Alikhan et al., 2011), and comparison of plasmids was facilitated using the Artemis Comparison Tool (Carver et al., 2005).

Strain WW14A was initially identified as Klebsiella pneumoniae by MALDI-TOF/MS. K. pneumoniae (phylogroup Kp1/KpI) is phylogenetically related to K. quasipneumoniae [subsp. quasipneumoniae (Kp2/KpIIA) and subsp. similipneumoniae (Kp4/KpIIB)], K. variicola (Kp3/KpIII) and two unnamed phylogroups (Kp5 and Kp6). Together, Kp1 to Kp6 make-up the K. pneumoniae complex. The phylogroups can be reliably identified based on genome sequencing but only recently Rodrigues et al. demonstrated the potential of MALDI-TOF/MS for precise identification of K. pneumoniae complex members. Incorporation of spectra of all K. pneumoniae complex members into reference MALDI-TOF/MS spectra databases, in which they are currently lacking, is desirable (Rodrigues et al., 2018).

Sequence analysis revealed that WW14A isolate includes a genome of 6.01 Mb with 57.3% GC content. It was confirmed as Klebsiella quasipneumoniae subsp. quasipneumoniae and assigned as ST2787. Moreover, a core genome phylogenetic analysis indicated that strain WW14A is closely related to a GES-5 producing Klebsiella quasipneumoniae subsp quasipenumoniae isolate from a Taiwanese hospital wastewater (Gomi et al., 2018) and is clearly distinct from strains isolated recently in Argentina (Faccone et al., 2020) and Brazil (Fuga et al., 2020; Furlan et al., 2020) ( Figure 1 ).

According to BIGSdb and KAPTIVE databases, the closest capsular type assigned for this strain is wzi88-KL70, considering there are 2 mismatches with the reference wzi cluster 88 and 78.7% identity with KL70 locus (9 from 20 genes present). The strain displayed a negative string test, so it may not be classified as hypermucoviscous. On the other hand, virulome analysis showed a hypervirulent profile, carrying genes encoding for aerobactins, enterobactins, hemolysins, pullulanase secretion, hyperadherence and biofilm formation. One intact phage-related sequence (Haemophilus phage SuMu [39.8 kb]) was identified, besides other ten incomplete (7.2 to 34.7 kb) or questionable phages from Pseudomonas, Salmonella, E. coli and Cronobacter. Also, two confirmed (178 and 459 bp) and four questionable (94 to 121 bp) CRISPRs signatures were identified. The presence of phages and CRISPRs in WW14A genome shows its adaptative signatures, i.e., a memory of past genetic interactions with bacteriophages and plasmids ( Table 1 ).

Resistome analysis showed antimicrobial resistance encoded by chromosomal mutations and by acquired resistance genes for fluoroquinolones, aminoglycosides, sulfonamides, trimethoprim and polymyxins. TEM-1A, OXA-9 and KPC-2 β-lactamases were present. Also, WW14A carried resistance genes for several metals (arsenic, cadmium, cobalt, magnesium, manganese, mercury, nickel, silver) and organic compounds like detergents and pesticides ( Table 2 ).

WW14A was resistant to all tested β-lactams, ciprofloxacin, amikacin, and trimethoprim-sulfamethoxazole but susceptible to gentamicin, tetracycline, chloramphenicol, and colistin. β-lactam MICs showed the strain was highly resistant to ceftazidime, ceftriaxone, imipenem, meropenem, piperacillin/tazobactam, ampicillin, piperacillin, cephalothin, cefalexin and cefoxitin ( Table 3 ).

Seven plasmids were detected, but only five plasmid incompatibility groups could be assigned by PlasmidFinder: pWW14A-IncFII/IncR (228495 bp), pWW14A-3 (untypable; 113444 bp), pWW14A-IncFIB/IncHI1B (96771 bp), pWW14A-KPC2 (IncP6; 40407 bp), pWW14A-6 (untypable; 8819 bp), pWW14A-7 (untypable; 3478 bp) and pWW14A-8 (untypable; 2954 bp). IncFII(K) contained the Tra conjugative system, while IncP-6 harbored the KPC-2 transposon.

WW14A is the first non-clinical KPC-2-producing isolate from Argentina which has been fully sequenced. Klebsiella quasipneumoniae was recently defined as a new species and was originally thought to be associated exclusively with environmental niches (Venkitapathi et al., 2021). However, despite there being relatively few reports to date, the true prevalence of this organism in clinical settings is likely underestimated as it is not generally distinguished from K. pneumoniae in routine testing of clinical laboratories (Mathers et al., 2019).

WW19C was identified as E. asburiae with a 5.2 Mb genome and 55.2% GC content. The phylogenomic analysis confirmed that WW19C was related to E. asburiae EN3600, a clinical isolate co-producing IMP-8, CTX-M-14, CTX-M-3, and QnrS1 from China ( Figure 2 ) (Yuan et al., 2019). MLST analysis revealed a new allele for fusA gene (allele 245), which associated with other alleles (dnaA [24], gyrB [43], leuS [52], pyrG [27], rplB [18], and rpoB [21]) generated the new ST1407.

Regarding its virulome, when compared to WW14A, WW19C shows the same categories of virulence, but with a broader variety of genes. Four intact phage-related sequences were identified: Enterobacter phage Tyrion (46.4 kb), Escherichia phages HK75 (34 kb) and 186 (26 kb), and Salmonella phage SEN34 (42.5 kb), the last being also found as an incomplete phage sequence in WW14A. In addition, WW19C also carried two incomplete (6.8 kb and 28.1kb) and two questionable phage sequences. Three questionable CRISPRs (96, 134 and 148 bp) were also identified ( Table 1 ).

Resistome analysis showed ARGs for fluoroquinolones, aminoglycosides, sulfonamides, trimethoprim, macrolides and polymyxins. Chromosomal β-lactamases TEM-1-like (A213T variant) and AmpC ACT-57, and plasmid-borne KPC-2 were present. Also, several metal resistance genes, efflux systems and organic compounds resistance genes could be identified, as detailed in Table 2 .

WW19C expressed a MDR phenotype, showing susceptibility to tetracycline, intermediate susceptibility to amikacin and resistance to all tested β-lactams, ciprofloxacin, gentamicin, trimethoprim-sulfamethoxazole, chloramphenicol and colistin. β-lactam MICs showed the strain was highly resistant to ceftazidime, ceftriaxone, imipenem, meropenem, piperacillin/tazobactam, ampicillin, piperacillin, cephalothin, cefalexin and cefoxitin ( Table 3 ).

Eight plasmids were detected but only four incompatibility groups could be assigned by PlasmidFinder: pWW19C-IncHI2 (332184 bp), pWW19C-IncFIB (139140 bp), pWW19C-KPC2 (IncP6; 34721 bp), pWW19C-5 (untypable; 4935 bp), pWW19C-6 (untypable; 4152 bp), pWW19C-7 (untypable; 3872 bp), pWW19C-8 (untypable; 3108 bp), pWW19C-9 (untypable; 1282 bp).

Members of the ECC are opportunistic pathogens responsible for a wide range of healthcare-associated infections and hospital outbreaks, especially in intensive care units (De La Cadena et al., 2018). The emergence of carbapenem-resistant ECC like WW19C is of great concern as diverse plasmids of incompatibility group IncP-6 (20 – 60 kb) were the most common carriers of bla _KPC among ECC clinical isolates from Colombia, suggesting that KPC dissemination is not just due to horizontal transmission of a single plasmid-borne bla _KPC, but that multiple rearrangements and transposition events can occur (Rojas Coy et al., 2019).

IncP-6 plasmids identified in this study, designated pWW14A-KPC2 and pWW19C-KPC2, could be completely closed, analyzed, and compared with other closed plasmids belonging to the same incompatibility group. They both share 72% of coverage, with 99.99% of sequence similarity. Figure 3 shows a comparison of IncP-6 plasmids described herein with other complete IncP-6 described in the literature (Naas et al., 2013; Dai et al., 2016; Wang et al., 2017; Yao et al., 2017).

The bla _KPC-2-containing plasmid pWW14A-KPC2 was 40,407 bp in size with an average GC content of 57.9%. It comprised 55 open reading frames (ORF) according to RAST annotation, of which 25 encoded proteins with known functions and 30 were hypothetical proteins. Further analysis showed that the structure was highly similar to those of several bla _KPC-2-containing plasmids of both environmental and clinical origin deposited in GenBank. Plasmids pKOX3-P5-KPC, from a clinical K. oxytoca in China (GenBank accession no. KY913901) (Wang et al., 2017), p121SC21-KPC2 from Spanish wastewater Citrobacter freundii 121SC21 (Genbank accession no. NZ_LT992437.1) (Yao et al., 2017), and p10265-KPC, first reported in China from a clinical P. aeruginosa 10265 (GenBank accession no. KU578314) (Dai et al., 2016) were highly similar over the entire region ( Figure 3 ). The bla _KPC-2 gene is the only determinant of antimicrobial resistance located in plasmid pWW14A-KPC2. This is the first report of a typical IncP-6 plasmid carrying bla _KPC-2 in a K. quasipneumoniae isolate.

On the other hand, pWW19C-KPC2 was 34,721 bp with an average GC content of 52.9%. It comprised 46 open reading frames (ORF) according to RAST annotation, of which 25 encoded proteins with known functions and 21 were hypothetical proteins. It showed the highest similarity with pECL189-1 from Enterobacter hormaechei ECL189 (GenBank accession no. CP047966.1, 72% coverage and 100% identity), a strain isolated from a Chinese hospital co-producing KPC-2, NDM-1, TEM-1 and SHV-66 ( Figure 3 ).

Both plasmids belong to the IncP-6 incompatibility group according to replicon-based schemes because they carry the replicase gene repA, which constitutes an IncP-6-type consecutive par-rep gene cluster, together with the partition genes parABC. The repA gene and the parABC locus of pWW14A-KPC2 and pWW19C-KPC2 are identical and show 100%, >96%, >98%, and >98% of nucleotide sequence identity with the IncP-6 plasmids p10265-KPC, Rms149, pRIO-5, and pCOL-1, respectively. RepA in Rms149 has shown to confer the plasmid’s replication ability in E. coli, P. aeruginosa, and P. putida and its parABC locus is known to promote plasmid mobilization in E. coli (Haines et al., 2005), while RepA in pRIO-5 allows replication in Serratia marcescens and Acinetobacter baumannii but not in P. aeruginosa (Bonnin et al., 2012). pWW19C-KPC2 also contains and extra plasmid replication initiation protein with 100% nucleotide identity with the replicon of pVPS18EC0801-5, a short non-typeable plasmid of 4910 bp present in foodborne E. coli strain 18GA07VL07-EC isolated from retail veal in the United States in 2018 (GenBank accession no. CP063722.1, unpublished). The presence of two replication initiation genes, IncP-6 replicon repA and the replication initiation protein found in pVPS18EC0801-5 suggests pWW19C-KPC2 may be a novel hybrid plasmid ( Supplementary Figure S1 ). The presence of extra replicon genes may expand pWW19C-KPC2 ability of maintenance and dissemination.

The IncP-6 backbone of pWW14A-KPC2 was compared with p10265-KPC and it contained the same plasmid maintenance genes, as follows: kfrA; a 5.6 kb MOBP family mobilization module composed of genes mobA (relaxase/primase fusion protein), mobB (oriT recognition-like protein), mobC (relaxosome protein), mobD and mobE (auxiliary proteins); the anti-oxidative system msrB-msrA-yfcG-corA-orfX gene cluster; and endonuclease paeR7IR. Within the backbone, pWW14A-KPC2 harbored two accessory modules: on one hand, a truncated Tn5045-associated mercury resistance operon disrupted by IS4321 (the latter being absent in p10265-KPC), and on the other, ISPa19. Linear comparison of the above-mentioned plasmids with pWW19C-KPC2 showed notable differences ( Figure 4 ):

None of the plasmids could be transferred to E. coli J53 via conjugation, despite repeated attempts. The examination of plasmid sequences confirmed the absence of conjugative transfer genes involved in plasmid transfer, such as the tra or trb operons, in accordance with the fact that typical IncP-6 plasmids are mobilizable rather than self-transmissible if a conjugative plasmid is also present (Botts et al., 2017). The mob gene cluster in pWW14A-KPC2 is functional for plasmid mobilization in E. coli (Yuan et al., 2019), being mobA, mobB, and mobC essential for functionality while mobD and mobE are non-essential but greatly enhance mobilization frequency (Dai et al., 2016). In the case of pWW19C-KPC2, which lacks this module, we inferred that this function is assumed by the relaxosome proteins coded in the unique accessory region related to the short non-typeable plasmid pVPS18EC0801-5 ( Figure 4 ).

While IncP-6 plasmids are naturally isolated from P. aeruginosa, their association with bla _KPC-2 have recently been reported in various species from both clinical and environmental sources, suggesting that they have a broad host range and the ability to persist in the environment for long periods (Naas et al., 2013; Dai et al., 2016; Wang et al., 2017; Yao et al., 2017). Relatively few IncP-6 plasmids had been fully sequenced up to 2017, including Rms149 (AJ877225), pCOL-1 (KC609323), and p10265-KPC (KU578314) and pPAEC79 (CP040685.1) from clinical strains of P. aeruginosa (Haines et al., 2005; Naas et al., 2013; Dai et al., 2016; Wang et al., 2021); pRIO-5 (JF785550) from a clinical strain of S. marcescens (Bonnet et al., 2000; Bonnin et al., 2012); pRSB105 (DQ839391), a plasmid detected in a sample from a WWTP in Germany (Schlüter et al., 2007); and pHH2-227 (JN581942), an uncultured IncW/IncP-6 hybrid plasmid that was exogenously captured from an arable soil (Heuer et al., 2009; Król et al., 2013; Botts et al., 2017). Interestingly, only 16 unique plasmids containing both the IncP-6 replicon and the bla _KPC-2 gene had been documented in the GenBank database until May 22^nd 2019 (Dong et al., 2020), while the number rises up to 25, according to plasmid database PLSDB by April 13^th 2021 (https://ccb-microbe.cs.uni-saarland.de/plsdb/) (Kukla et al., 2018; Galata et al., 2019).

It is of notice that two bla _GES-1/5 IncP-6 plasmids have been recently published, pKRA-GES-5 (MN436715.1) found in clinical isolates of K. pneumoniae ST45 from Poland, and pN260-3 (AP023450) in a clinical Enterobacter roggenkampii co-harboring bla _IMP-1 from Japan (Literacka et al., 2020; Umeda et al., 2021). Compared with IncP-6 archetype Rms149 and pWW14A-KPC, they had close to 100% of similarity with approximately 12.7-kb of the backbone containing partitioning, replication, and mobilization loci (Literacka et al., 2020), evidencing the remodeling ability of this platform.

The region harboring the bla _KPC-2 gene in pWW14A-KPC2 is ~11 kb long and contains a Tn3-based transposon disrupted by an ISApu-flanked element, where the core sequence is composed by ΔISKpn6/bla _KPC-2-Δbla _TEM-1-ISKpn27. This core module is connected with the gene cluster korC-orfX-klcA-orfX-repB. The resulting structure matched 99% to clinical isolates C. freundii M9169 and E. cloacae M11180 from Argentina (Gomez et al., 2011) and is associated with a truncated ISEc33 element. The resulting ΔISEc33-associated element is not bracketed by IRs and DRs, suggesting that its mobilization could be attributed to homologous recombination-based insertion of a foreign element Tn3-ISKpn27-Δbla _TEM-1-bla _KPC-2-ISKpn6-korC-orf-klcA-repB into a pre-existent intact ISEc33 element (making it truncated at 3´ end), rather than resulting from a transposition event of the whole ISEc33-associated element followed by the deletion of its adjacent extremities removing IR and DR sequences (Dai et al., 2016). This core platform was initially discovered in p10265-KPC. In the bla _KPC-2 gene cluster of p10265-KPC, the primary genetic structure, Tn3-ISKpn27-bla _KPC-2-ΔISKpn6-korC-orf-klcA-ΔrepB, may have undergone two evolutionary events: (i) insertion of a bla _TEM-1 gene between ISKpn27 and the Tn3 IRR (right inverted repeat) and (ii) disruption of the tnpA gene (transposase) from Tn3 by insertion of a composite transposon, ISApu1-orfX-ISApu2 (Dong et al., 2020). Interestingly, pWW19C-KPC lacked ΔISEc33 insertion sequence, indicating that the insertion of the bla _KPC-2 cluster occurred at a different position in an IncP-6 backbone and seems to have a different evolutionary history of genetic assembly and transposition ( Figure 4 ). The acquisition of the KPC-2 encoding region by IncP-6 replicons is in agreement with the results of previous studies, in which similar plasmids show remnants of multiple events, with intact or partial mobile elements dispersed throughout their sequences (Botts et al., 2017). Its dissemination could be an important contribution to the establishment of emerging clones as major nosocomial pathogens (Cejas et al., 2019).