Whole-Genome Sequencing Investigation of a Large Nosocomial Outbreak Caused by ST131 H30Rx KPC-Producing Escherichia coli in Italy

KPC-producing Escherichia coli (KPC-Ec) remains uncommon, being mainly reported as the cause of sporadic episodes of infection rather than outbreak events. Here we retrospectively describe the dynamics of a large hospital outbreak sustained by KPC-Ec, involving 106 patients and 25 hospital wards, during a six-month period. Twenty-nine representative KPC-Ec isolates (8/29 from rectal swabs; 21/29 from other clinical specimens) have been investigated by Whole-Genome Sequencing (WGS). Outbreak isolates showed a multidrug-resistant profile and harbored several resistance determinants, including blaCTX-M-27, aadA5, dfrA17, sulI, gyrA1AB and parC1aAB. Phylogenomic analysis identified the ST131 cluster 1 (23/29 isolates), H30Rx clade C, as responsible for the epidemic event. A further two KPC-Ec ST131 clusters were identified: cluster 2 (n = 2/29) and cluster 3 (n = 1/29). The remaining KPC-Ec resulted in ST978 (n = 2/29) and ST1193 (n = 1/29), and were blaKPC-3 associated. The KPC-Ec ST131 cluster 1, originated in a previous KPC-Kp endemic context probably by plasmid transfer, and showed a clonal dissemination strategy. Transmission of the blaKPC gene to the globally disseminated high-risk ST131 clone represents a serious cause of concern. Application of WGS in outbreak investigations could be useful to better understand the evolution of epidemic events in order to address infection control and contrast interventions, especially when high-risk epidemic clones are involved.


Introduction
The ongoing rise of carbapenemase-producing Enterobacterales (CPE) represents an important threat to public health worldwide, in both healthcare and community settings. The pervasive dissemination of CPE substantially impacts on patient safety since few therapeutic alternatives remain. Overall, the worldwide predominant carbapenemase is the Ambler class A Klebsiella pneumoniae carbapenemase (KPC) enzyme, encoded by alleles of the bla KPC gene, with KPC-producing K. pneumoniae (KPC-Kp) being the most common among KPC-producing CPE. Although less prevalent than in K. pneumoniae, KPC production in other species belonging to Enterobacterales is increasingly reported [1].

Epidemiological Context
Epidemiological surveillance data showed that the epidemic event followed a previous KPC-Kp outbreak, starting in October 2015 and partially overlapping the KPC-Ec outbreak (Supplementary Figure S1). The prevalence of KPC-Ec increased from 0% in 2014 and 2015 to 2.8% in 2016. This trend seems to be related to the increase of KPC-Kp prevalence, from 12.4% in 2014, to 17.3% in 2015 and to 23.1% in 2016 (Supplementary Figure S1). Notably, 50 (47.2%) out of 106 patients' results were positive to both KPC-Ec and KPC-Kp during previous hospitalizations. In particular, 20 out of 50 patients (40%) were previously colonized by KPC-Kp, while in 23 cases (46%) the isolation of both species was concomitant, indicating the co-presence at the intestinal level. Only in seven patients (14%), however, was KPC-Ec isolated previously with respect to KPC-Kp.
The virulence genes sat (secret autotransporter toxin), iss (increased serum survival) and gad (glutamate decarboxylase) were found in all ST131 strains, while the iha (adherence protein) determinant was found in all isolates but one ST131. The senB (enterotoxin) gene was present in the majority of ST131 isolates, while the cnf1 (cytotoxic necrotizing factor type 1) was only identified in the isolate sk35y35t, which was the first KPC-Ec that emerged in the outbreak period. A different pattern of virulence genes was detected in the two ST978 strains, consisting of vat (vacuolating autotransporter toxin), pic (serine protease) and gad determinants. Lastly, the ST1193 isolate harbored iha, sat and vat as virulence genes ( Table 3).
The comparison of the 29 KPC-Ec isolates with other E. coli genomes deposited in the PATRIC database is shown in Figure 1. Notably, most of them showed a major similarity with KPC-Ec strains collected in the UK.    of which belonged to ST131 ("ST131 cluster 1", "ST131 cluster 2" and "ST131 cluster 3"), one to ST978 and another to ST1193 (Figure 1). The ST131 cluster 1, characteristic of the majority of isolates (n = 26), was found in all wards with the exception of the Intensive Care and the Infectious Diseases Units (Figure 2), thus representing the outbreak epidemic clone.
Notably, sporadic isolates belonging to the ST131 cluster 1 were observed till June 2017. Based on phylogenetic data, the first isolate belonging to the ST131 cluster 1 (i.e., the outbreak index strain, coded sk36y36t) was isolated in the Nephrology Ward. The patient had been previously hospitalized for long periods in the past years in both Nephrology and ICU. Of note, the index patient was previously colonized by KPC-Kp (strain code, sk138y138t) ( Table 2). The ST131 cluster 2 included two isolates, collected in Surgery and ICU in June and September 2016, respectively. The ST131 cluster 3 was represented by a single isolate, obtained in February 2016 from the Infectious Diseases Ward. ST978 strains were recovered in March 2016 from the blood culture and rectal swab of the same patient (Table 2 and Figure 2). Lastly, the ST1193 strain was isolated in November 2016 in the Infectious Diseases Ward. Figure 2 shows the ward distribution of the KPC-Ec strains belonging to different clusters and STs.  The phylogenetic reconstruction showed the presence of five distinct clusters, three of which belonged to ST131 ("ST131 cluster 1", "ST131 cluster 2" and "ST131 cluster 3"), one to ST978 and another to ST1193 (Figure 1). The ST131 cluster 1, characteristic of the majority of isolates (n = 26), was found in all wards with the exception of the Intensive Care and the Infectious Diseases Units (Figure 2), thus representing the outbreak epidemic clone.
Notably, sporadic isolates belonging to the ST131 cluster 1 were observed till June 2017. Based on phylogenetic data, the first isolate belonging to the ST131 cluster 1 (i.e., the outbreak index strain, coded sk36y36t) was isolated in the Nephrology Ward. The patient had been previously hospitalized for long periods in the past years in both Nephrology and ICU. Of note, the index patient was previously colonized by KPC-Kp (strain code, sk138y138t) ( Table 2). The ST131 cluster 2 included two isolates, collected in Surgery and ICU in June and September 2016, respectively. The ST131 cluster 3 was represented by a single isolate, obtained in February 2016 from the Infectious Diseases Ward. ST978 strains were recovered in March 2016 from the blood culture and rectal swab of the same patient (Table 2 and Figure 2). Lastly, the ST1193 strain was isolated in November 2016 in the Infectious Diseases Ward. Figure 2 shows the ward distribution of the KPC-Ec strains belonging to different clusters and STs.
Recent studies have focused on deciphering the genomic evolution and diversity within the ST131 lineage [4,20,21]. Since three different ST131 clusters were identified, we compared their genomes with those published by Petty and colleagues [20]. As shown in Figure 3, all the studied strains could be assigned to the ST131 clade C, characterized by the fimbrial variant fimH30Rx, and the gyrA1AB and parC1aAB alleles, associated with fluoroquinolone resistance. Phylogeny highlighted that the only difference was the presence of the bla CTX-M-27 gene variant instead of bla CTX-M-15 for ST131 clusters 1 and 3. Recent studies have focused on deciphering the genomic evolution and diversity within the ST131 lineage [4,20,21]. Since three different ST131 clusters were identified, we compared their genomes with those published by Petty and colleagues [20]. As shown in Figure 3, all the studied strains could be assigned to the ST131 clade C, characterized by the fimbrial variant fimH30Rx, and the gyrA1AB and parC1aAB alleles, associated with fluoroquinolone resistance. Phylogeny highlighted that the only difference was the presence of the blaCTX-M-27 gene variant instead of blaCTX-M-15 for ST131 clusters 1 and 3.   [20]. The branches corresponding to the strains of this study are highlighted in red. Metadata have been represented as follows: the first column indicates the origin of the strains; the second one the ST131 clades as called by Petty et al. [20]. The subsequent four columns report the allele variants of the fimH, parC, gyrA and blaCTX-M genes.  [20]. The branches corresponding to the strains of this study are highlighted in red. Metadata have been represented as follows: the first column indicates the origin of the strains; the second one the ST131 clades as called by Petty et al. [20]. The subsequent four columns report the allele variants of the fimH, parC, gyrA and bla CTX-M genes.
The six KPC-Kp strains, isolated during the early stages of the outbreak, were investigated for ST and plasmid incompatibility group arrangement. KPC-Kp strains belonged to ST35 (n = 2, from Nephrology and Urology), ST17 (n = 2, from the Rehabilitation Unit and Internal Medicine), ST3033 (n = 1, from Cardiology) and ST2279 (n = 1, from Infectious Diseases).
The positive colonization date (and co-isolation date, when it occurred) and related clinical samples of KPC-Ec and KPC-Kp isolates involved in WGS analyses are shown in Supplemental Tables S1 and S2.
Regarding plasmids, IncFIBpQil was found in all the ST131 and ST1193 KPC-Ec strains, as well as in all the KPC-Kp strains. IncFII was found in all the ST131 cluster 1 and cluster 2 KPC-Ec strains; Col156 was found in all the ST131 cluster 1 isolates but one. IncX3 was detected in both the ST978 isolates. Lastly, all the KPC-Kp strains harbored IncFIB(K)_Kpn3. The KPC-Ec strains belonging to the epidemic ST131 of both cluster 1 and cluster 3 produced the KPC-2 enzyme, as well as the KPC-Kp isolates. The bla KPC-3 allele was instead found in the ST131 cluster 2, ST1193 and ST978 strains. Analysis of the bla KPC surrounding genetic environment showed that all clusters, except ST978 and ST1193, were characterized by a conserved bla KPC scaffold (Figure 4). ST978 and ST1193 showed two different scaffolds. All the clusters harbored the Tn4401a, a structural variant of the Tn4401 transposon. The six KPC-Kp strains, isolated during the early stages of the outbreak, were investigated for ST and plasmid incompatibility group arrangement. KPC-Kp strains belonged to ST35 (n = 2, from Nephrology and Urology), ST17 (n = 2, from the Rehabilitation Unit and Internal Medicine), ST3033 (n = 1, from Cardiology) and ST2279 (n = 1, from Infectious Diseases).
The positive colonization date (and co-isolation date, when it occurred) and related clinical samples of KPC-Ec and KPC-Kp isolates involved in WGS analyses are shown in Supplemental Tables S1 and S2.
Regarding plasmids, IncFIBpQil was found in all the ST131 and ST1193 KPC-Ec strains, as well as in all the KPC-Kp strains. IncFII was found in all the ST131 cluster 1 and cluster 2 KPC-Ec strains; Col156 was found in all the ST131 cluster 1 isolates but one. IncX3 was detected in both the ST978 isolates. Lastly, all the KPC-Kp strains harbored IncFIB(K)_Kpn3. The KPC-Ec strains belonging to the epidemic ST131 of both cluster 1 and cluster 3 produced the KPC-2 enzyme, as well as the KPC-Kp isolates. The blaKPC-3 allele was instead found in the ST131 cluster 2, ST1193 and ST978 strains. Analysis of the blaKPC surrounding genetic environment showed that all clusters, except ST978 and ST1193, were characterized by a conserved blaKPC scaffold (Figure 4). ST978 and ST1193 showed two different scaffolds. All the clusters harbored the Tn4401a, a structural variant of the Tn4401 transposon. Given the presence of the IncFIBpQil, a well-known KPC-harboring plasmid, the detection of the KPC-2 enzyme, and evaluating the extreme similarity of the blaKPC genomic environment in KPC-Kp and KPC-Ec isolates belonging to the ST131 cluster 1; a plasmidmediated transmission of the KPC-2 determinant from KPC-Kp to KPC-Ec isolates could reasonably have happened.
Although different temperature conditions were tested, conjugation assay results were negative, thus indicating that the resistance determinants were not located on a conjugative plasmid.

Discussion
Our study describes a large intrahospital outbreak caused by KPC-Ec, involving a total of 106 inpatients in 25 wards. To the best of our knowledge, this is the largest outbreak caused by KPC-Ec reported worldwide. Genomic analyses allowed us to ascertain that five different KPC-Ec clusters were involved, three of which belonged to the highrisk ST131 clone. Notably, only the "cluster 1" was responsible for the epidemic event and was associated to the H30Rx clade C sub-clone. The ST131 cluster 1 strains co-harbored blaCTX-M-27 and blaKPC determinants and were the only ones provided with the senB virulence gene, coding for the secreted enterotoxin TieB.
Although its relevant ability to spread among patients in different wards, ST131 KPC-Ec showed a low propensity to cause infections. In fact, most samples were collected Given the presence of the IncFIBpQil, a well-known KPC-harboring plasmid, the detection of the KPC-2 enzyme, and evaluating the extreme similarity of the blaKPC genomic environment in KPC-Kp and KPC-Ec isolates belonging to the ST131 cluster 1; a plasmid-mediated transmission of the KPC-2 determinant from KPC-Kp to KPC-Ec isolates could reasonably have happened.
Although different temperature conditions were tested, conjugation assay results were negative, thus indicating that the resistance determinants were not located on a conjugative plasmid.

Discussion
Our study describes a large intrahospital outbreak caused by KPC-Ec, involving a total of 106 inpatients in 25 wards. To the best of our knowledge, this is the largest outbreak caused by KPC-Ec reported worldwide. Genomic analyses allowed us to ascertain that five different KPC-Ec clusters were involved, three of which belonged to the high-risk ST131 clone. Notably, only the "cluster 1" was responsible for the epidemic event and was associated to the H30Rx clade C sub-clone. The ST131 cluster 1 strains co-harbored bla CTX-M-27 and bla KPC determinants and were the only ones provided with the senB virulence gene, coding for the secreted enterotoxin TieB.
Although its relevant ability to spread among patients in different wards, ST131 KPC-Ec showed a low propensity to cause infections. In fact, most samples were collected from rectal swabs, whereas only 20 isolates (16.2%) were obtained from other biological specimens. Of note, only two isolates (1.6%) were from blood cultures. Overall, no deaths were attributable to infection caused by the ST131 KPC-Ec cluster 1 outbreak clone. Conversely, the treatment of infections caused by ST131 KPC-Ec was a challenge. In fact, isolates were mostly nonsusceptible to beta-lactams, including carbapenems and ceftolozane/tazobactam, due to the presence of the blaKPC and blaCTX-M determinants. KPC-Ec were also mostly resistant to ciprofloxacin and trimethoprim-sulphamethoxazole, another typical feature related to the pandemic ST131 lineage. Therapeutic options were aminoglycosides, tigecycline, colistin and the ceftazidime/avibactam combination.
The dissemination of the bla KPC determinant in E. coli, largely due to horizontal transfer of plasmids or other mobile elements into diverse genetic backgrounds, has been previously described [19]. Since the acquisition of bla KPC by E. coli is a very uncommon event, no data are to date available in the literature about a different prevalence between bla KPC-2 and bla KPC-3 variants. Nonetheless, in our experience, the bla KPC-2 gene seems to be the most represented in Italy, mainly associated with ST131 (unpublished data from a multicentric clinical study). The bla KPC genetic background was conserved among both the KPC-Ec ST131 cluster 1 and the KPC-Kp strains of the same period, suggesting that transmission events of plasmid/mobile elements occurred. More importantly, the index patient was colonized during the same time period by both KPC-Kp (strain code sk138y138t) and KPC-Ec (strain code sk36y36t) strains (Table 2), the last one reasonably representing the origin of the outbreak. On the other hand, conjugation assays showed that the bla KPC gene was not located on a conjugative plasmid, highlighting that the outbreak was caused by the spread of a dominant clone that had acquired the plasmid, rather than by the dissemination of a resistance plasmid to unrelated strains. Transmissions were drastically interrupted in May 2016, thanks to the adoption of a strict cohorting of both colonized and infected patients, that was assisted by dedicated healthcare staff. During the outbreak period, infection prevention and control measures were implemented. Training courses for the staff based on infection control, contact precautions and hand washing campaigns were promoted. Only sporadic cases related to KPC-Ec (ST131 cluster 1) were observed till June 2017, representing the tail of the outbreak. After this period, no other cases related to KPC-Ec were observed.
Taking together epidemiological and WGS data, we can speculate that the KPC-Ec outbreak clone developed in the previous context of the KPC-Kp outbreak. Furthermore, the outbreak appeared to be caused by the diffusion of a dominant clone (that probably acquired a bla KPC -harboring plasmid), and not by the dissemination of a resistance plasmid among the strains of the different clusters.
Rapid application of WGS in outbreak investigations could be useful to better understand the dynamics of epidemic events in order to address infection control and contrast interventions.

Epidemiological Context and Characterization of Bacterial Isolates
We retrospectively studied an outbreak caused by KPC-Ec that occurred at the Hospital of Lecco (Northern Italy, close to Milan) across a six-month period (February to July 2016). The hospital accounts for about one thousand beds, and has a catchment area of about 340,000 inhabitants. During the outbreak period, as a part of the surveillance activity of the hospital team for infection control, 123 KPC-Ec nonrepetitive isolates were collected from 106 patients. A total of 103 isolates were from colonization surveillance rectal swabs. Rectal swabs have been used to screen intestinal colonization by KPC-Ec, as indicated by international guidelines (Centers for Diseases Control and Prevention, CDC, 2015; https://www.cdc.gov/hai/pdfs/cre/cre-guidance-508.pdf, accessed on 14 June 2021). Twenty strains from other sites were isolated from symptomatic patients with suspected infection. These isolates were from urine (n = 11), blood (n = 2), purulent exudate (n = 2), respiratory secretions (n = 2), drainage fluid (n = 1), peritoneal fluid (n = 1) and a surgical wound swab (n = 1). Strains from the same patients were included only when isolated from different sites. A further four KPC-Ec isolates were sporadically collected after the epidemic event and until June 2017 and were also investigated to verify their clonal relationship to outbreak strains. To better clarify the origin of the bla KPC gene in the KPC-Ec strains, six KPC-Kp isolates were collected from patients cocolonized by KPC-Kp and KPC-Ec and were included in the study. Of them, two were obtained at the beginning of the episode (including those isolated from the index patient), while the remaining were collected during the outbreak from patients admitted to those wards that were mainly involved in the episode. Isolates from rectal swabs were screened for carbapenemase production using chromogenic Brilliance CRE agar (Thermo Fisher Scientific). Bacterial isolates were identified to the species level using MALDI-TOF mass spectrometry (Vitek MS, bioMérieux), while susceptibility testing was routinely determined by the Vitek 2 system (bioMérieux). Isolates suspected of carbapenemase production (MIC values for ertapenem and/or meropenem >0.125 mg/L) were evaluated to assess the presence of specific carbapenem resistance determinants using the immunochromatographic technique (RESIST-4 O.K.N.V., Coris BioConcept) and/or a molecular dedicated assay (Xpert Carba-R, Cepheid).
To characterize epidemic isolates and better understand the dynamics of the outbreak, a total of 29 KPC-Ec strains (four of which were sporadically isolated in the post-outbreak period) were selected as representatives based on the site of infection or colonization, date and ward of admission (Table 2). MIC values of these isolates were determined by the MicroScan autoSCAN-4 system (NMDRM1 panel, Beckman Coulter). Selected antimicrobials (i.e., ceftazidime-avibactam, ceftolozane/tazobactam, and colistin) were evaluated by a broth microdilution Sensititre panel used for multidrug-resistant Gramnegative strains (DKMGN panel, Thermo Fisher Scientific). EUCAST criteria were used for determining susceptibility categories [22]. Finally, these isolates were analyzed by WGS-based typing and SNP-based phylogenetic reconstruction.

High Resolution Melting Assay
The selection of KPC-producing K. pneumoniae strains chosen for WGS investigation was made on the basis of the High Resolution Melting (HRM) assay results. The HRM was performed on the wzi hypervariable capsular gene as described by Perini et al. [23] using MeltingPlot software [24].

Whole-Genome Sequencing
A total of 35 strains (29 KPC-Ec and 6 KPC-Kp), representative of the epidemic event and of the post outbreak period, were processed for WGS analysis. In detail, inclusion criteria were: (i) isolates from ascertained infections (other than from screening rectal swabs) were chosen preferentially; (ii) KPC-Ec from all hospital wards involved in the epidemic event; (iii) KPC-Ec isolated in different periods of the outbreak (at the beginning, medium period, tail of the outbreak); (iv) In addition, 8 KPC-Ec from rectal swabs were included in order to evaluate the presence of the outbreak clone in the patients' intestinal microbiota. Genomic DNA was extracted using a QIAamp DNA minikit (Qiagen) following the manufacturer's instructions and sequenced using the Illumina Miseq platform with a 2 × 250 paired-end run after Nextera XT library preparation (Illumina Inc., San Diego, CA, USA).

CoreSNP Calling and Phylogenetic Analyses
For each of the 35 strains included in the study, the reads quality was assessed using FastqC software (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/, accessed on 29 June 2020), and the low quality terminal bases were trimmed using Trimmomatic software [25]. Reads were assembled using SPAdes software [26].
All the genome assemblies were submitted to the European Nucleotide Archive (ENA) with the project code PRJEB40388. All the ID codes are listed in Table S3.
The genome distance of Each KPC-Ec genome assembly was estimated using Mash software [27] against a collection of 3325 E. coli genomes retrieved from the PATRIC database [28] and the 50 most similar genomes were selected for subsequent analyses. All the selected genome assemblies (from this study and the PATRIC database) were aligned against the E. coli MG1655 reference genome using progressive Mauve and coreSNPs were called as described by Gona and colleagues [29]. Repeated regions in the reference genome assembly were detected using Blastn. Then coreSNPs localized within repeated regions were masked. From here, the obtained coreSNP alignment was called "Global coreSNP".
Global coreSNP alignments were subjected to phylogenetic analysis using RAxML software with a 100 pseudo-bootstrap, after best model selection using ModelTest-NG [30]. The strains were then clustered on the basis of the Global phylogenetic tree and SNP distance. At first, we identified on the ML phylogenetic tree the largest highly supported (>75 bootstrap) monophyletic groups including study strains only.

Whole-Genome Sequencing-Based Typing
Resistance genes of the 29 KPC-Ec strains were identified using the ResFinder online tool [31] and SRST2 software [32] with the ARG-ANNOT dataset [33]. Virulence genes were detected by the VirulenceFinder online tool [34]. Plasmid incompatibility groups were detected using PlasmidFinder [35]. The Multi Locus Sequence Typing profiles of the 29 E. coli and six K. pneumoniae strains were determined in silico according to the Achtman and Pasteur schemes, respectively, using an in-house Perl script.

KPC-Harboring Contigs Comparison
For each identified cluster (see above) one representative strain was selected and the genome assembly was analyzed as follows. The contig harboring bla KPC gene was identified by Blastn search (E-value threshold: 0.00001). The extracted contigs were oriented on the basis of bla KPC gene orientation, then annotated using Prokka [36] and aligned with progressive Mauve [37]. The Tn4401 transposon was annotated using TETyper [38]. Lastly, the gene composition and synteny of extracted contigs were graphically represented using the R library genoPlotR [39].

Conjugation Assay
To assess the possible transferability of resistance determinants identified, a conjugation assay was performed using the E. coli J53 Azide R as the recipient strain at temperatures of both 25 • C and 37 • C, and with MER 0.5mg/L for the selection of transconjugants.

Conclusions
Although KPC-producing Escherichia coli (KPC-Ec) remains uncommon, and mainly reported as the cause of sporadic episodes of infection rather than outbreaks, the present work shows that the acquisition of bla KPC gene by a high-risk successful clone, as the ST131, can lead to even large and potentially difficult to manage epidemic events. The attention on the presence and circulation of carbapenemase-producing enterobacteria (CPE) should be always kept high, especially in healthcare settings. In this context, the application of WGS could be useful to better understand the evolution and dynamic of outbreaks sustained by CPE in order to promptly address infection control and contrast interventions.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/antibiotics10060718/s1, Figure S1. Local prevalence of KPC-Kp and KPC-Ec in the period 2014-2016 in the hospital setting. Table S1. Escherichia coli (KPC-Ec) isolates selected for WGS analysis and co-isolation of Klebsiella pneumoniae (KPC-Kp) isolates from the same patients. Table S2. Klebsiella pneumoniae (KPC-Kp) isolates selected for WGS analysis and co-isolation of Escherichia coli (KPC-Ec) isolates from the same patients. Table S3. List of the ENA codes for the genome assemblies of the studied strains.
Author Contributions: A.P. conceived the study, performed microbiological experiments and sequencing, analyzed and interpreted the data and wrote the paper; L.P. conceived the study, performed microbiological experiments, analyzed and interpreted the data and drafted the paper; F.C. and M.P. performed bioinformatic analyses; E.M. performed microbiological experiments and provided the epidemiological data; V.M.M. performed microbiological experiments and the conjugation assay; R.M. conceived the study, supervised the activities and revised the manuscript; F.L. conceived the study, supervised the activities and revised the manuscript. All authors have read and agreed to the published version of the manuscript.
Funding: This research received no external funding.

Institutional Review Board Statement:
The study was conducted in the context of normal clinical routine. All information and metadata about patients had been anonymized. Samples were coded and analyses were performed with anonymized database.

Informed Consent Statement: Not applicable.
Data Availability Statement: The genome assemblies of the sequenced strains are available at the European Nucleotide Archive (ENA). The ID code for each genome submitted is reported in Supplementary Table S3.