Detection of mcr-1 Gene in Undefined Vibrio Species Isolated from Clams

The increase of antimicrobial resistant strains is leading to an emerging threat to public health. Pathogenic Vibrio are responsible for human and animal illness. The Enterobacteriaceae family includes microorganisms that affect humans, causing several infections. One of the main causes of human infection is related to the ingestion of undercooked seafood. Due to their filter-feeding habit, marine invertebrates, such as clams, are known to be a natural reservoir of specific microbial communities. In the present study, Vibrionaceae and coliforms microorganisms were isolated from clams. A microbial susceptibility test was performed using the disk diffusion method. From 43 presumptive Vibrio spp. and 17 coliforms, three Vibrio spp. with MICs to colistin >512 mg L−1 were found. From the 23 antimicrobial resistance genes investigated, only the three isolates that showed phenotypic resistance to colistin contained the mcr-1 gene. Genotypic analysis for virulence genes in EB07V indicated chiA gene presence. The results from the plasmid cure and transformation showed that the resistance is chromosomally mediated. Biochemical analysis and MLSA, on the basis of four protein-coding gene sequences (recA, rpoB, groEL and dnaJ), grouped the isolates into the genus Vibrio but distinguished them as different from any known Vibrio spp.


Introduction
The genus Vibrio contains more than 130 species [1,2] and includes both non-pathogenic and pathogenic species, being highly genetically diverse [3,4]. Vibrio spp. are Gram-negative motile bacteria, facultative anaerobic, and non-spore-forming. With cells shaped as curved or straight rods, and typically found in freshwater, estuarine and marine environments [1,5]. Pathogenic species are associated with several human diseases caused by the natural microbiota of aquatic environments [6] and are mostly linked to the ingestion of undercooked seafood [7]. Vibriosis is one of the most prevalent bacterial diseases associated with marine fish and shellfish [8]. Non-pathogenic Vibrio, on the other hand, are not associated with human illnesses, and are, therefore, not routinely identified [9]. These so-called "marine species" or "marine vibrios", have nonetheless been frequently isolated, thus receiving more attention in recent years since they have an important role in specific ecological niches and belong to autochthonous marine microbial communities [4]. These niches are often fish farms, where antibiotics are frequently used, and Vibrio can become reservoirs of antimicrobial resistance genes [10,11].
Enterobacteriaceae possess a bacilli shape, are non-spore-forming, have variable motility, and are generally found inside other organisms. Usually, these microorganisms have the ability to mobilise their antimicrobial genetic determinants in plasmids via conjugation, which means that they can be resistant to almost every existing antibiotic [12]. Hospitals throughout the world have been reporting clinical cases of people that contract severe bacterial infections due to the presence of resistant lineages of major infectious agents such as Escherichia coli [13] and Klebsiella pneumoniae [14].
The abusive and erroneous use of antibiotics by human and veterinary medicine [15,16] promotes a selective pressure for antimicrobial resistant bacteria. The increase of resistant strains is leading to an emerging threat to public health [17,18]. It is estimated that 10 million lives could be lost due to antimicrobial resistant bacteria annually by 2050, if no action is taken [19].
For many years , polymyxin antibiotics were commonly employed in clinical use, until they were gradually withdrawn because of toxicity issues. However, nowadays, the available antibiotics are not enough to cope with the emergence of multidrugresistance bacteria, resulting in the re-introduction of polymyxins as the last line of defence for clinical treatment [20]. Among polymyxins, the two most used are polymyxin E (colistin) and polymyxin B [21]. Until recently, all the findings about polymyxin-resistant mechanisms were chromosomally mediated [22][23][24]. However, the plasmid-mediated colistin resistance gene mcr has already been reported. The first report of mcr-1 was in E. coli and K. pneumoniae [25] in China. Since then, the mcr-1 gene has emerged in strains collected in many parts of the world (e.g., Europe- [26][27][28][29]; Asia- [30,31]; North America- [32]; Africa- [33]). A very large study of bacterial genomes concluded that almost all mobile colistin resistance genes (from mcr-1 to mcr-9) originated from environmental bacteria, mainly from water sources [34]. The origin of mcr-1 is believed to be from Moraxella species [34,35]. As far as the authors know, up to this date, the presence of mcr-1 has been mostly reported in the Enterobacteriaceae family and only once in a non-Enterobacteriaceae, Vibrio parahaemolyticus [36].
Bivalve molluscs such as mussels, clams, or oysters are commercially valuable as a food delicacy. It means that these organisms are regularly monitored for the presence of Vibrio spp. and enteric bacteria as indicative of pollution [37]. Furthermore, bivalves are filter-feeders and can filtrate between 20 and 100 L of water per day (species-specific), retaining several microorganisms [38,39].
In this report, environmental bacteria were isolated from two clam species, Ruditapes decussatus, and Ruditapes philippinarum, after selective isolation for Vibrio microorganisms and coliform bacteria. The aim of this study was to evaluate the profile of antimicrobial resistance to three important groups of antimicrobials (polymyxins, carbapenems, and quinolones) in the microbial isolates, and the evaluation of the presence of virulence genes. Three Vibrio spp. were found resistant to colistin, with the presence of mcr-1, and one of them also presented the virulent chiA gene. The phenotypic and phylogenetic analysis strongly support that the three isolates belong to the genus Vibrio and represent undescribed species.
Enumeration of Vibrionaceae and coliforms was performed by serial dilutions, of the molluscs edible part and inter-valve liquid, in saline solution (0.85% w/v) and plating in thiosulfate citrate bile salts sucrose agar (TCBS; VWR, Radnor, PA, USA) and Chromocult ® Coliform agar (Merck, Darmstadt, Germany), respectively. Plates were incubated for 24 h at 28 • C for Vibrio spp. and at 37 • C for coliforms. Colony-Forming Units (CFUs) with distinct morphology on TCBS agar were inoculated onto Tryptic Soy Agar (TSA; Biolife, Milano, Italy) supplemented with NaCl (1.5% w/v), and incubated for 24 h at 28 • C. CFU that had distinct morphology on Chromocult ® were inoculated onto Nutrient Agar (NA; VWR, Radnor, PA, USA), and incubated for 24 h at 37 • C. The isolates were conserved at −80 • C in cryovials (VWR, Radnor, PA, USA).

Antimicrobial Susceptibility Test
Antimicrobial susceptibility tests were performed by the disk diffusion method, according to the EUCAST guidelines [40] against the antimicrobial: colistin (10 µg), cefotaxime (30 µg), imipenem (10 µg), ciprofloxacin (5 µg) and meropenem (10 µg). The disks were from Thermo Fisher Scientific. Briefly, a microbial suspension in 0.85% saline solution was adjusted to the 0.5 McFarland turbidity standard on a densitometer (VWR, Radnor, PA, USA). For coliforms, the colonies were grown in Mueller-Hinton agar (MHA) plates (Thermo Fisher Scientific, Waltham, MA, USA), at 37 • C for 24 h, and for Vibrionaceae in MHA with 1.5% (w/v) NaCl at 28 • C for 24 h. The antimicrobial inhibition halos were then measured after this period.
The minimum inhibition concentration (MIC) was determined by the microdilution method, according to the EUCAST guidelines [41], with the addition of 2% (w/v) NaCl. The plates were incubated at 28 • C for 24 h and the MIC was recorded as the lowest concentration of the colistin that completely inhibited the growth.

Biochemical Characterisation
The Gram staining procedure was performed according to Coico [42] together with the KOH test [43]. Oxidase reaction was made using Cytochrome Oxidase test (bioMérieux, Marcy l'Etoile, France). The biochemical analysis of the three isolates was conducted using API 20NE (bioMérieux, Marcy l'Etoile, France) at 28 • C. The inocula were prepared in saline solution (2% w/v), adjusted to the 0.5 McFarland turbidity standard. The galleries were filled according to the manufacturer's instructions. Reads were carried out at 24 h and 48 h. The tests were performed in triplicate.

Lipid Profile
Each isolate was grown in TSA (BD, Sparks, MD, USA) at 30 • C for 24 ± 1 h. The fatty acids of the isolates were extracted and methylated to fatty acid methyl esters (FAMEs) using the Sherlock Instant FAME TM method (MIDI, Inc., Newark, DE, USA). FAMEs were analysed by gas chromatography on a gas chromatograph (GC Agilent Technologies 6890N, Santa Clara, CA, USA), with a flame detector and a 7683 B series injector, using a 25 m Agilent J&W Ultra 2 capillary column. The FAMEs profile of each isolate was determined by the Sherlock ® software package, v. 6.2 (MIDI, Inc.).

DNA Extraction
The DNA from the bacterial isolates was extracted by a simple boiling method. Briefly, the isolates were grown overnight in Tryptic Soy Broth (TSB; Biolife, Milano, Italy) with a final concentration of 2% (w/v) NaCl, at 28 • C for Vibrionaceae, and at 37 • C for coliforms. Then, 1 mL of each bacteria suspension was centrifuged (10,000× g, 1 min) and the supernatant was removed. The pellet was resuspended in 100 µL of sterile Milli Q water and heated at 100 • C for 10 min, followed by centrifugation at 10,000× g for 5 min. The supernatant was collected and kept at −20 • C until further use.

Phylogenetic Analysis
All phylogenetic analyses based on partial sequences of 16S rRNA, recA, rpoB, groEl and dnaJ were performed on Molecular Evolutionary Genetics Analysis (MEGA-X) software [44]. Multilocus Sequence Alignments (MLSA) with partial concatenated sequences of recA, rpoB, groEl and dnaJ from the isolates were aligned by ClustalW with full-length reference genes obtained from closely related species belonging to genus Vibrio and other genera (Photobacterium and Grimontia), available on GeneBank from NCBI (accession numbers are available in Table S2). Photobacterium damselae 9046-81 and Grimontia hollisae FDAARGOS 111 were used as the outgroup. A phylogenetic tree was constructed with the Maximum Likelihood (ML) method [45] and Kimura 2-parameter model. Initial trees for the heuristic search were obtained by applying the Neighbor-Joining method to a matrix of pairwise distances estimated using the Maximum Composite Likelihood (MCL) approach. A discrete Gamma distribution was used to model evolutionary rate differences among sites (5 categories; +G, parameter = 0.8508). The topologies of the phylogenetic tree were determined using bootstrap analyses based on 1000 replicates.

Plasmid Extraction and Bacteria Transformation
The Vibrio spp. isolates were grown overnight in TSB supplemented with NaCl (1.5% w/v), at 28 • C. GeneJET Kit (Thermo Fisher Scientific, Waltham, MA, USA) was used for plasmid extraction, following the manufacturer's instructions. Transformation by heat shock using TOP10 chemically competent E. coli (Thermo Fisher Scientific, Invitrogene, Waltham, MA, EUA) was performed. Briefly, the extracted plasmid was added to TOP10 competent E. coli and left on ice for 20 min. Then, the mixture was heated for 50 s at 42 • C and placed on ice for 2 min. 500 µL of Luria-Bertani (LB; VWR, Radnor, PA, USA) medium (at room temperature) was added and the cells were incubated at 37 • C for 1 h. The tubes were centrifuged (10,000× g, 1 min) and the supernatant was discarded. The pellet was resuspended and inoculated on LB agar supplemented with colistin (PanReac AppliChem, Barcelona, Spain; 2 mg L −1 ). The plates were incubated overnight at 37 • C.

Plasmid Curing
Bacteria plasmid curing involves the loss of plasmids when the growth occurs in the presence of a plasmid curing agent, like ethidium bromide, acridine orange, or sodium dodecyl sulphate in sublethal concentrations [46]. If bacteria lose antimicrobial resistance after curing, it is reasonable to believe that resistance genes are located in the lost plasmid.
The Vibrio spp. isolates were grown overnight at 28 • C, in MHA (Thermo Fisher Scientific, Waltham, MA, EUA) supplemented with 2% NaCl (w/v) and 2 mg L −1 colistin. The inoculum was prepared in a 0.85% (w/v) saline suspension and adjusted to the 0.5 McFarland turbidity standard. Then, sequential dilutions up to the 10 −3 dilution were made, and ethidium bromide was added in the concentration of 0.025, 0.05, and 0.1 g L −1 . The microorganisms were incubated at 28 • C (200 rpm) until their growth reached the end of the exponential phase. After that, the isolates were grown on both MHA with and without colistin (2 mg L −1 ), at 28 • C for 24 h. Curing efficiency was determined by comparing the number of growing colonies.

Enumeration of Vibrionaceae and Coliforms Isolates
The CFU per g of molluscs, of Vibrio spp. and coliforms were determined for each clam species isolated at different locations. The counts are available in Table 1.
The CFU with distinct morphology on TCBS and on ChromoCult ® Coliform agar were isolated, totalising 38 presumptive Vibrio spp. and 17 coliforms. Among the 17 coliforms isolated, nine were E. coli (blue/violet colonies) and three were Citrobacter freundii (red colonies). Food safety criteria regarding microorganisms are available at the European Commission regulation [47]. There is increased concern regarding food safety in shellfish products. For live bivalve molluscs, E. coli has been used as an indicator of hygiene quality. The method used to enumerate E. coli is the most probable number (MPN). The range between 230 and 700 MPN per 100 g of bivalve flesh and intra-valvular liquid, is acceptable for direct human consumption according to ISO TS 16649-3:2015 (https: //www.iso.org/standard/56824.html, accessed on 12 December 2021). A standard bacterial analysis does not reveal the presence of enteroviruses or microorganisms of the genus Vibrio. Thus, there is a need for considering new sanitary quality indicators [48,49].

Antimicrobial Resistance of the Isolates
All 60 isolates (both Vibrio spp. and coliforms) were tested for susceptibility/resistance to colistin, cefotaxime, imipenem, ciprofloxacin, and meropenem (Table S3). Almost all isolates showed some degree of sensitivity since a halo was formed. However, three Vibrio spp. isolates, EB07V, NJ21V, and NJ22V, showed phenotypic resistance to colistin since an inhibitory halo was not formed (Table S3). The EUCAST does not have Epidemiological Cut off Values (ECOFF) or clinical breakpoint for Vibrio spp. against colistin. Still, considering the ECOFF for Escherichia coli [50], the isolates (EB07V, NJ21V and NJ22V) had a MIC of >512 mg L −1 and can be considered resistant to colistin. EB07V was isolated from R. decussatus collected at Óbidos Lagoon, and NJ21V plus NJ22V from R. philippinarum collected at Ria Formosa.
The presence of antimicrobial resistance genes in all 60 isolates was determined and from the 23 antimicrobial resistance genes investigated, only the three isolates that showed phenotypic resistance to colistin presented the mcr-1 gene (Tables S4-S6). The mcr-1 gene was confirmed by sequencing the three DNA fragments, and compared to those publicly available (BLAST, https://blast.ncbi.nlm.nih.gov/Blast.cgi, accessed on 21 January 2022). The three mcr-1 sequences shared 100% of identity with plasmid-mediated mcr-1 gene sequences published (E. coli, Salmonella enterica and Klebsiella pneumoniae).
Considering these results, the presence of virulence genes was assessed on the three isolates. Out of the 19 virulence genes tested, only the chiA gene (chitinase A) was found in EB07V. Chitin is the main structural polysaccharide of crustaceans [51], being a valuable carbon and nitrogen source in marine environments [52]. To use chitin, marine bacteria such as V. vulnificus [53], V. harveyi [54], or V. anguillarum, and V. parahaemolyticus [55], require a group of enzymes called chitinases and chitin-binding proteins. Chitinases are able to convert chitin into useful biomolecules [51]. The relationship of V. cholerae with live copepods (which possess chitin) has already been observed, resulting in an increase in bacteria survivability and culturability [56]. The associations between Vibrio spp. with copepods suggest that copepods may be an environmental reservoir for the pathogens and a source of their dissemination [57]. Studies regarding the consumption of copepod nauplii by clams in freshwater, marine, and estuarine environments have been published [58,59]. This grazing can be termed "incidental predation". However, bivalves can devastate zooplankton populations [60]. Thus, the isolates from this study could have been transmitted through this interaction between copepods and clams.
Since the three Vibrio spp. isolates presented the mcr-1 gene, it was important to ascertain whether the resistance was plasmidic or genomic. Thus, an attempt to transform competent E. coli cells and cure the isolates was conducted. In the case of the transformation, no transformed cells were obtained. When attempting to cure the isolates, no differences were observed when the number of growing colonies in medium MHA with and without colistin were compared ( Figure S1). These results indicate that the observed resistance to colistin in all three isolates is chromosomally mediated. Lei et al. [36] found the mcr-1 gene in V. parahaemolyticus isolates obtained from shrimp and tested its mobility by conjugation. Their results show that the mcr-1 gene is plasmid-mediated, and it is allocated in IncX4 plasmid. Indeed, IncX4 plasmids were previously found to be associated with the dissemination of mcr-1 in enterobacteria and probably were successfully mobilised to Vibrio spp. [36,61]. Through in silico analysis in the database (www.ncbi.nlm.nih.gov, accessed on 12 December 2021), it was possible to verify the presence of mcr-1 in the genome of two V. parahaemolyticus (accession number: NNEB01000260.1 and NNHN01000221.1) isolated in China in 2017 and 2019.
Until 2015, colistin resistance was exclusively related to chromosomic mutations in genes involved in lipid A decoration, i.e., pmrA/B, phoP/Q, ccrA/B, lpxACD, or mgrB genes that resulted in a modification of bacterial membranes by adding sugar to the lipid A moiety [62,63]. The colistin resistance processes can also involve genes encoding phosphoethanolamine transferase (PET) and/or glycosyltransferase proteins that are essential for membrane phospholipid biosynthesis [64][65][66]. It was in 2015 that a mobile colistin resistance mechanism was found for the first time: the mcr-1 variant encoding for a PET [25]. Khedher et al. [34] showed that PET genes are ubiquitous in bacteria since they were detected in 1047 species, which could not be explained by the abusive overuse of antibiotics for human and veterinary medicine, or agricultural purposes. These findings highlighted a question regarding the role of these enzymes and suggested a defense system in the biosphere against bacteriophages in aquatic environments, but also against cationic antimicrobial peptides secreted by vertebrates (e.g., humans) and invertebrates. The authors also found that bacteria that are intrinsically resistant to colistin, i.e., Serratia spp. and Proteus spp., also contain MCR variants. V. vulnificus and V. cholerae have chromosomally mediated colistin resistance [67]. In the case of the biotype EI Tor of Vibrio cholerae O1, for example, an eptA gene is present in addition to the almEFG operon [68][69][70]. Apparently, it is the simultaneous expression of the eptA gene [70] and almEFG operon [71,72] that trigger the appropriate lipopolysaccharide-lipid A modification, promoting polymyxin resistance [73]. In the case of V. vulnificus, it has been reported that the potassium uptake protein TrkA, was responsible for the resistance to polymyxin B [74]. Taking into account that the three isolates did not pass their colistin resistance gene through horizontal transfer models, the results clearly indicate that the isolates contain the mcr-1 variant in their chromosomes and not in a plasmid.

Biochemical Analysis of the Isolates
The three isolates, EB07V, NJ21V, and NJ22V were positive in the KOH test which indicates the isolates are Gram-negative. Vibrio spp. isolates were biochemically characterised, using API 20NE strips ( Table 2).
All three isolates were positive for nitrate reduction and cytochrome oxidase activity, like most members of the genus Vibrio, with the exception of V. metschnikovii and V. gazogenes [75]. In addition, the isolates were positive for tryptophane production, esculin, gelatin hydrolysis, and β-galactosidase activity. Conversely, arginine dihydrolase and urease were negative in all isolates. The isolates were able to assimilate glucose, mannitol, N-acetyl-glucosamine, maltose, potassium gluconate, and malate. However, none of the isolates could assimilate arabinose, and capric and phenylacetic acids. Mannose was only assimilated by EB07V and NJ21V. EB07V and NJ22V were not capable of assimilating adipic acid and, in the case of NJ21V, the results were not conclusive. Only EB07V was able to ferment glucose and assimilate trisodium citrate. It is known that Vibrio strains have the ability to ferment glucose anaerobically [9,76]. However, NJ21V and NJ22V were unable of doing so. Nevertheless, NJ21V and NJ22V were able to grow on TCBS medium, which has an alkaline pH of 8.6, bile salt, and NaCl which suppress the growth of most interfering bacteria. Thus, due to the selectivity of the medium and subsequent genetic analysis, the isolates NJ21V and NJ22V were considered as Vibrio spp.

Phylogenetic Analysis
After comparing the 16S rRNA sequences of the three isolates to those publicly available (BLAST, https://blast.ncbi.nlm.nih.gov/Blast.cgi, accessed on 21 December 2021), the isolates were identified as belonging to the Vibrio genus, but it was not possible to identify them to the species level. The isolate EB7V shared 99.47% of similarity with V. hangzhouensis and V. mediterranei; NJ21V shared 98.68% of similarity with several Vibrio spp. (e.g., V. xuii, V. parahaemolyticus, V. galatheae); and NJ22V shared 98.71% of similarity with some Vibrio spp. (e.g., V. xuii, V. parahaemolyticus, V. scophthalmi). The genus Vibrio contains a large number of closely related species sharing a high level of sequence similarity with just 0.1 or 0.2% difference in the nucleotide sequence of 16S rRNA [84]. Based on that, the use of other genes (e.g., rpoA, recA, pyrH, and atpA) to discriminate closely related species has been proposed [85,86]. Four other house-keeping genes were chosen, recA, rpoB, groEL, and dnaJ that were sequenced and aligned. The gene groEL, for example, was reported to be extremely conserved in nature [87] and was already used to identify Vibrio isolates [88].
The phylogenetic tree for concatenated sequences for the house-keeping genes was constructed using the ML method ( Figure 2). The MLSA is based on partial sequences of the four house-keeping genes for 32 Vibrio spp., Photobacterium damselae 9046-81 and Grimontia hollisae FDAARGOS 111 (used as outgroup) and the three isolates. MLSA has been proposed as a useful technique for the identification of Vibrio isolates and for studying the phylogeny in this genus [85,89]. Moreover, this technique has been utilised for the identification of some new Vibrio species, such as V. gigantis and V. crassostreae [90,91] and, more recently for V. barjaei and V. nitrifigilis [92,93].
However, the three isolates clustered in the genus Vibrio but do not cluster with any Vibrio clade, which suggests that the three isolates may represent new undescribed species of the genus Vibrio. In the case of NJ21V and NJ22V, it appears that they belong to the same species, although being different strains. Further studies using these isolates are needed to prove this statement. Phylogenetic construction based on the concatenation of partial sequences of four housekeeping genes (recA, rpoB, groEL, and dnaJ) by the ML method algorithm and Kimura 2-parameter model [94]. The tree with the highest log likelihood (−140419,60) is shown. The percentage of trees in which the associated taxa clustered together is shown next to the branches. Initial trees for the heuristic search were obtained by applying the Neighbor-Joining method to a matrix of pairwise distances estimated using the Maximum Composite Likelihood (MCL) approach. A discrete Gamma distribution was used to model evolutionary rate differences among sites (5 categories; +G, parameter = 0.8284)). The tree was drawn to scale, with branch lengths measured in the number of substitutions per site, 0.20. There were a total of 8130 positions in the final dataset. Photobacterium damselae 9046-81 and Grimontia hollisae FDAARGOS 111 were used as the outgroup. Table 2. Biochemical characterisation of the three isolates that show colistin resistance. A "+" indicates a positive reaction, whereas a "−" denotes a negative reaction. NC-not conclusive.

Conclusions
In this study, the presence of the mobilisable colistin resistance gene mcr-1 in three Vibrio spp. (EB07V, NJ21V, and NJ22V), isolated in microbial communities from clams, highlights a potential threat to public health. Clams are an appreciated food delicacy and colistin is one of the last line of defence antibiotics used for clinical treatment [20]. One of the three mcr-1 positives isolates, EB07V, also carried the gene chiA, responsible for the invasion of tissues of chitin-containing organisms, which may increase its pathogenic potential. The location of the mcr-1 resistance gene appears to be chromosomal which indicates a low potential for mobilisation. Biochemical analysis and MLSA on the basis of four protein-coding gene sequences (recA, rpoB, groEL and dnaJ) grouped the isolates into the genus Vibrio but distinguished them as different from other species of Vibrio, suggesting we might be in the presence of new species. To corroborate this statement further investigation is required. However, the fact that the microbial communities in the environment possess the colistin mobilise resistant determinant mcr-1, highlights the importance of this study.
-indicates non detected gene and + indicates detected gene. Table S5: Antimicrobial resistances genes search on all the isolates. The second class was β-lactams (bla IMP , bla VIM , bla KPC , bla NDM , bla OXA , bla CTX-M , bla SHV and bla TEM ). -indicates non detected gene and + indicates detected gene. Table S6: Antimicrobial resistances genes search on all the isolates. The third antimicrobial class searched was quinolone (qnrA, qnrB, qnrC, qnrD, qnrS and qepA). -indicates non detected gene and + indicates detected gene. Figure S1: Plasmid curing of the isolates EB07V, NJ21V and NJ22V. The number of CFU was determined in MHA plates with and without colistin. References  have been cited in the Supplementary materials.

Conflicts of Interest:
The authors declare no conflict of interest.