Unravelling the Features of Success of VIM-Producing ST111 and ST235 Pseudomonas aeruginosa in a Greek Hospital

The objective of this study was to analyze the characteristics that contribute to the successful dissemination of VIM-producing Pseudomonas aeruginosa (P. aeruginosa), belonging to ST111 and ST235, in a Greek hospital. A total of 120 non-repetitive P. aeruginosa, which had meropenem minimal inhibitory concentrations (MICs) greater than 2 mg/L, were studied. VIM-encoding genes were amplified and sequenced within their integrons. Isolates were typed by multilocus sequence typing (MLST). Six VIM-producers, representative of different integron structures and sequence types (STs), were completely sequenced using Illumina platform. Sixty-one P. aeruginosa were confirmed to produce VIM-type carbapenemases. ST111 dominated (n = 34) among VIM-producers, while 15 VIM-producers belonged to ST235. The blaVIM-like genes were located in three integron types, including In59, In595 and In1760, which were integrated into P. aeruginosa chromosomes. Whole genome sequencing (WGS) data demonstrated that ST111 and ST235 MBL producers carried several resistance and virulence genes. Additionally, the presence of type I-C and type I-E clustered regularly interspaced short palindromic repeats (CRISPR)/Cas locus was observed in ST235 and ST395 isolates, respectively. In conclusion, our findings confirmed the clonal spread of ST111 P. aeruginosa, carrying the VIM-2-encoding integron In59, in the University Hospital of Larissa (UHL). In addition, they highlighted the important role of high-risk clones, ST111 and ST235, in the successful dissemination and establishment into hospital settings of clinically important pathogens carrying resistance determinants.


Introduction
Pseudomonas aeruginosa is an opportunistic pathogen causing a wide range of infections, including pulmonary infections, wound and sepsis. P. aeruginosa show a great tendency to form biofilms on medical devices [1]. Additionally, P. aeruginosa express virulence factors, like exotoxins, proteases and pigment. These factors help the bacterium to facilitate successful colonization and infection across a wide range of environments.
Furthermore, P. aeruginosa show a high prevalence of antimicrobial resistance. Approximately 13% of P. aeruginosa, causing infections, are multidrug-resistant (MDR) strains. Therefore, recently, the World Health Organization (WHO) has considered P. aeruginosa as a target of high priority for the development of new antibiotics [2].

Plasmid Analysis
Six VIM-producing P. aeruginosa (VPP) were selected as representatives of all different STs and integron types in order to define the genetic units harboring the bla VIM -like genes. The plasmid contents of the selected isolates were analyzed by pulsed-field gel electrophoresis (PFGE) of total DNA digested with S1 nuclease (Promega, Madison, WI, USA) [16], followed by bla VIM hybridization [13].

Nucleotide Sequence Accession Numbers
Nucleotide sequences representing the genetic context of VIM-encoding integrons In595 and In1760 were submitted to the GenBank under accession numbers MT428323 and MT437279, respectively.
In 70 P. aeruginosa isolates, the EDTA-meropenem test was positive, indicating MBL production. However, 61 out of 70 the isolates, being positive in the EDTA-meropenem test, were positive for bla VIM genes. Other types of carbapenemase-encoding genes were not detected. The population structure of P. aeruginosa isolates studied by MLST showed that the VPP isolates included four sequence types (STs) (Figure 1). ST111 was the most predominant comprising 34 VIM-producing isolates (Table 1). Twenty five VIM producers were distributed in STs 235 (n = 15) and 395 (n = 10). STs 111, 235 and 395 have been assumed as "high-risk clones" [20]. The remaining VIM-producing P. aeruginosa isolates belonged to ST773 (n = 2). However, higher genetic diversity was observed in the group of non-carbapenemase-producing isolates (Table 1). This group of isolates (n = 59) comprised 11 clones, with STs 235 (n = 14) and 111 (n = 11) accounting for 25 isolates. In 70 P. aeruginosa isolates, the EDTA-meropenem test was positive, indicating MBL production. However, 61 out of 70 the isolates, being positive in the EDTA-meropenem test, were positive for blaVIM genes. Other types of carbapenemase-encoding genes were not detected.

ST MBL No. of Isolates Integrons with bla VIM Genes
profiling did not observer any DNA bands that could be assigned to plasmids, and the bla VIM -like probe hybridized with only the largest DNA bands, indicating the chromosomal location of the VIM-encoding integrons. Thus, the genomes of these isolates ( Table 2) were sequenced by the Illumina platform. Sequence analysis showed that, in all cases, VIM-encoding integrons were inserted into P. aeruginosa chromosomes.
The In59-like integron was located in a Tn5060-like transposon integrated into the chromosome of ST111 P. aeruginosa isolates. However, insertion of the Tn3-like transposon Tn4661 [23] probably deleted the Tn5060 mer module. Tn4661 was found next to a partially deleted ICE1-like integrative conjugative element (∆ICE1), which was inserted into the tRNALys gene (PA4541.1 in GenBank accession no. AE004091). A similar structure has been observed in VIM-producing P. aeruginosa isolates belonging to ST111 isolated in Czech hospitals [13].
In595 integron was in integrative conjugative element ICEPae3483, identified in the P. aeruginosa chromosome ( Figure 2). ICEPae3483 was previously found in P. aeruginosa HSV3483 (GenBank accession no. MF168944) from Portugal. The In595 was embedded in a Tn1721-like transposon. The IRi of In595 was found in tnpM of the Tn1721-like, while the 3 CS of In595 was disrupted after the start codon of orf5 by an IS6100 element.
In ST395 VIM-2 producers, In1760 was associated with a composite transposon, bounded by two copies of the ISPsp7 insertion sequence in opposite orientation. The bla VIM-2 -carrying transposon was inserted in a novel genomic island (GI) (Figure 2), designated PAGI-8709. PAGI-8709 included open reading frames (ORFs) encoding hypothetical proteins of unknown function and proteins of various functions (e.g., regulation of gene expression, DNA recombination and metabolic activities). GIs closely related to PAGI-8709 have not been found previously in P. aeruginosa. PAGI-8709 was integrated in P. aeruginosa chromosome into gene encoding for tRNALys (PA0729 in GenBank accession no. AE004091). The 5 CS of the integron was intact, with the IRi of the integron being next to a 9685-bp segment (nucleotide 163493 to 166551 in GenBank accession no. MT437279), presenting 100% identity with α Tn3-like sequence from Pseudomonas sp. strain AG1 (GenBank accession no. CP045739), whereas an ISPa17 element was inserted into the 3 CS of the integron, downstream of orf5.
Additional genes conferring resistance to aminoglycosides, β-lactams and chloramphenicol were identified in the majority of the sequenced isolates (Table 2), as indicated by analysis of WGS data by the ResFinder 3.2 tool. Additionally, WGS data confirmed the absence of known carbapenemase-encoding genes in nine P. aeruginosa isolates being positive in the EDTA-meropenem test, but negative in PCR screening. In595 integron was in integrative conjugative element ICEPae3483, identified in the P. aeruginosa chromosome (Figure 2). ICEPae3483 was previously found in P. aeruginosa HSV3483 (GenBank accession no. MF168944) from Portugal. The In595 was embedded in a Tn1721-like transposon. The IRi of In595 was found in tnpM of the Tn1721-like, while the 3΄CS of In595 was disrupted after the start codon of orf5 by an IS6100 element. In ST395 VIM-2 producers, In1760 was associated with a composite transposon, bounded by two copies of the ISPsp7 insertion sequence in opposite orientation. The blaVIM-2-carrying transposon was inserted in a novel genomic island (GI) (Figure 2), designated PAGI-8709. PAGI-8709 included open reading frames (ORFs) encoding hypothetical proteins of unknown function and proteins of various functions (e.g., regulation of gene expression, DNA recombination and metabolic activities). GIs closely related to PAGI-8709 have not been found previously in P. aeruginosa. PAGI-8709 was integrated in P. aeruginosa chromosome into gene encoding for tRNALys (PA0729 in GenBank accession no. AE004091). Τhe 5΄CS of the integron was intact, with the IRi of the integron being next to a 9685-bp segment (nucleotide 163493 to 166551 in GenBank accession no. MT437279), presenting 100% identity with α Tn3-like sequence from Pseudomonas sp. strain AG1 (GenBank accession no. CP045739), whereas an ISPa17 element was inserted into the 3΄CS of the integron, downstream of orf5.
Additional genes conferring resistance to aminoglycosides, β-lactams and chloramphenicol were identified in the majority of the sequenced isolates (Table 2), as indicated by analysis of WGS data by the ResFinder 3.2 tool. Additionally, WGS data confirmed the absence of known carbapenemase-encoding genes in nine P. aeruginosa isolates being positive in the EDTAmeropenem test, but negative in PCR screening.
Examination of quinolone resistance-determining regions of gyrA, gyrB, parC, and parE and of the mexR, nfxB, and mexT genes, which regulate the MexAB-OprM, MexCD-OprJ, and MexEF-OprN multidrug efflux systems [24], revealed the presence of several point mutations Examination of quinolone resistance-determining regions of gyrA, gyrB, parC, and parE and of the mexR, nfxB, and mexT genes, which regulate the MexAB-OprM, MexCD-OprJ, and MexEF-OprN multidrug efflux systems [24], revealed the presence of several point mutations predicted to result in several amino acid substitutions ( Table 3). The majority of these amino acid substitutions have been identified previously from both ciprofloxacin-susceptible and ciprofloxacin-resistant P. aeruginosa isolates. However, in GyrA, the T83I amino acid substitution that previously has been associated with increased quinolone resistance [25] was found in all ciprofloxacin-resistant isolates. Furthermore, in the oprD gene, point mutations predicted to result in early termination of translation were identified in six P. aeruginosa isolates (Table 3). This finding is in agreement with increased resistance to carbapenems [26], even in non-carbapenemase-producing isolates.
Analysis of WGS data by VFDB showed that all sequenced isolates carried several virulence genes ( Table 2) that can be involved in colonization and cause bloodstream invasion, extensive tissue damage and dissemination [27]. These included secreted toxins, like exoenzyme S, exoenzyme T, exoenzyme U and exoenzyme Y, subverting host cell defense and signaling systems [28]. Our results confirmed that the clone ST111 correlated with the copresence of exoS, exoT, and exoY genes, whose expression correlates with a higher risk of mortality [29]. On the other hand, clone ST235 correlated with the copresence of exoT, exoU, and exoY. In particular, ExoU is associated with lung damage and acute cytotoxicity [30]. Additionally, sequenced isolates carried genes encoding phospholipases (PlcH and PlcN), elastase B, alkaline protease A, exotoxin A and phenazines (Table 2), which have been previously associated with increased virulence of P. aeruginosa isolates [31,32].
Finally, the presence of the CRISPR/Cas system in P. aeruginosa isolates was examined. CRISPR/Cas systems provide bacteria with an adaptive immunity system against invading genetic elements [33], like bacteriophages and plasmids. A CRISPR/Cas system was found in seven of the sequenced P. aeruginosa. Isolates belonging to STs 235 and 395 exhibited type I-C and type I-E CRISPR/cas locus, respectively, while a type I-F CRISPR/Cas locus was found in ST244 and ST253 isolates ( Table 2). Both ST395 isolates exhibited identical CRISPR/Cas systems, further confirming their clonal relationship. The type I-E CRISPR/Cas locus consisted of eight genes consecutively encoding the Cas2, Cas1, Cas6, Cas5, Cas7, Cse2, Cse1 and Cas3 proteins. The locus enclosed two arrays of 12 (CRISPR1) and 8 spacers (CRISPR2). The spacers, presenting variable sequences, had a common length of 32 bp. Each spacer was flanked by two direct repeats (DRs), being 29 bp long. The DRs in CRISPR1 had the 5 -CGGTTCATCCCCACGCATGTGGGGAACAC-3 consensus sequence, whereas the DRs in CRISPR2 had the 5 -CGGTTCATCCCCACACCCGTGGGGAACAC-3 consensus sequence.
The type I-C CRISPR/Cas locus consisted of seven genes consecutively encoding the Cas3, Cas5, Cas8, Cas7, Cas4, Cas1, and Cas2 proteins. The locus was followed by an array of 20 spacers, whose variable sequences were 33-36 bp long. DRs flanking the spacers had the 5 -GTCGCGCCCCGCACGGG CGCGTGGATTGAAAC-3 consensus sequence and a length of 32 bp. All ST235 isolates producing VIM-4 carbapenemase exhibited identical CRISPR/Cas systems. However, the type I-C CRISPR/Cas locus of the ST235 non-carbapenemase-producing isolate, Pae90-Lar, differed by an internal deletion erasing a spacer. This finding may indicate diverse evolution dynamics into those isolates. Type I-C locus was recently identified in ST235 isolates during a study analyzing the phylogenetic distribution of CRISPR/Cas systems in antibiotic-resistant P. aeruginosa [34].
In the ST244 isolate, Pae91-Lar, the type I-F CRISPR/Cas locus included six genes expressing the Cas1 endonuclease, Cas3/Cas2 helicase/RNAse, three Csy proteins (Csy1, Csy2, Csy3) and Cas6 endoribonuclease. The type I-F locus enclosed two arrays of 5 (CRISPR1) and 39 (CRISPR2) spacers. The spacers, presenting variable sequences, had a common length of 32 bp. Each spacer was flanked by two DRs 28 bp in length. The consensus sequence of the DRs found in CRIPSR1 was 5 -TTTTCTTAGCTGCCTATACGGCAGTGAAC-3 , whereas in CRISPR2, it was 5 -GTTCACTGCCGTG TAGGCAGCTAAGAAA-3 . Although a type I-F CRISPR/Cas locus was found in the ST253 isolate, Pae462-Lar, the locus exhibited 96.99% similarity to that identified in Pae91-Lar. Unlike Pae91-Lar, the locus included two arrays of 13 and 18 spacers. Type I-F locus was found previously in 38.5% of clinical P. aeruginosa isolates recovered from two Brazilian hospitals [35].
A BLASTn search with the identified spacers matched P. aeruginosa chromosomal sequences as well as different mobile elements, including bacteriophages and plasmids.

Discussion
In the current study, we analyzed the molecular characteristics of P. aeruginosa isolates belonging to the I and R susceptibility category according to meropenem MICs, recovered during surveillance program in UHL in 2018. A total of 50.8% of the isolates was confirmed to produce VIM-type carbapenemases. The majority (80.3%) of carbapenemase-producing isolates were assigned to "high-risk clones", ST111 and ST235, while higher genetic diversity was observed in the group of non-carbapenemase-producing isolates, which were assigned to 11 different clones. Three main types of class 1 integrons carrying bla VIM -like genes were found. The VIM-2-encoding integron In59, which was firstly described in France in 1998 [21], was the most common integron type identified in P. aeruginosa of STs 111, 235 and 773. An In59-like integron was also the predominant type among ST111 P. aeruginosa isolates collected from UK hospital laboratories [36]. Additionally, the VIM-4-encoding integron In595 was identified in 11 ST235 P. aeruginosa, while the integron In1760, expressing VIM-2 MBL, was found among ST395 isolates.
In P. aeruginosa isolates analyzed with Illumina sequencing, WGS data confirmed the chromosomal location of VIM-encoding integrons. This finding is in agreement with the clonal spread of VIM-producing pathogens in our setting. Of note, a novel genomic island (designated as PAGI-8709) associated with the bla VIM-2 -carrying integron In1760 was characterized from ST395 isolates. Additionally, WGS data confirmed the absence of carbapenemase-encoding genes in nine carbapenem-resistant P. aeruginosa that were positive in the EDTA-meropenem test. However, other traits corresponding to mutations in mexR, nfxB, and mexT genes, which regulate the MexAB-OprM, MexCD-OprJ, and MexEF-OprN multidrug efflux systems, and oprD were found in non-carbapenemase-producing isolates.
Furthermore, analysis of WGS data demonstrated that "high-risk clones", ST111 and ST235, contributing to the spread of VIM-type carbapenemases in UHL exhibited a vast variety in their armamentariums, including resistance genes, virulence genes and/or a CRISPR/Cas system. Those features could be involved in the establishment of ST111 and ST235 isolates into hospital environments. The presence of a CRISCPR/Cas system was identified in carbapenemase-producing isolates, belonging to STs 235 and 395. Interestingly, both ST395 isolates exhibited identical type I-E CRISPR/Cas systems, further indicating their clonal relationship. A similar observation was also made for all ST235 VIM-4-producing isolates, which carried type I-C CRISPR/Cas systems. These findings restated that CRISPR/Cas systems can be used as tools for typing of bacteria. Bacteria harboring an active CRISPR/Cas system present a selective advantage against the income of harmful DNA [37]. However, this inhibition may have a negative impact on cases where horizontal gene transfer would benefit bacteria by making them more resistant. Recently, an additional function has been attributed to the CRISPR/Cas system in the regulation of genes associated with virulence [38].

Conclusions
In conclusion, this is the first study analyzing the presence of CRISPR/Cas systems in carbapenemase-producing P. aeruginosa isolates of Greek origin. In line with previous studies [34,35], our data confirmed the presence of type I-C, I-E and I-F systems in P. aeruginosa of clinical origin, and their strong association with specific lineages. Isolates, which belonged to the same clone, exhibited identical CRISPR/Cas systems underlined the fact that CRISPR arrays is a simple, but precise, genotyping tool that can be used to track pathogenic bacterium. Additionally, our findings that are consistent with the results of previous studies [13,20,39] highlighted the important role of high-risk clones, namely, STs 111 and 235, in the successful dissemination of clinically important resistance determinants. Isolates, belonging to STs 111 and 235 harbored a vast variety in their armamentariums, including resistance genes, virulence genes and/or a CRISPR/Cas system, which could be involved in their successful dissemination. Therefore, recognition of carbapenemase-producing P. aeruginosa hyper-epidemic clones by molecular tools, and more specifically of WGS [40], represents an important step towards tracing transmission routes, developing targeted control and prevention strategies, and monitoring their effectiveness.
Author Contributions: C.C.P., A.V. and V.G. carried out experimental work. A.V., V.G., S.X., I.B. and J.H. assisted in the obtention of data. C.C.P. and E.P. played an important role in interpreting the results and writing the manuscript. C.C.P. supervised the experiments and revised the final manuscript. All authors have read and agreed to the published version of the manuscript.