Genomic Characterization of an Extensively Drug-Resistant Extra-Intestinal Pathogenic (ExPEC) Escherichia coli Clinical Isolate Co-Producing Two Carbapenemases and a 16S rRNA Methylase

An extensively drug-resistant Escherichia coli clinical isolate (N1606) belonging to Sequence Type 361 was recovered from the urine of a patient hospitalized in Switzerland. The strain showed resistance to virtually all β-lactams including the latest generation antibiotics cefiderocol and aztreonam–avibactam. Whole genome sequencing revealed that it possessed two carbapenemase-encoding genes, namely blaNDM-5 and blaKPC-3, and a series of additional β-lactamase genes, including blaCTX-M-15 and blaSHV-11 encoding extended-spectrum β-lactamases (ESBLs), blaCMY-145 encoding an AmpC-type cephalosporinase, and blaOXA-1 encoding a narrow-spectrum class D ß-lactamase. Most of these resistance genes were located on plasmids (IncFII-FIA, IncX3, IncIγ, IncFII). That strain exhibited also a four amino-acid insertion in its penicillin-binding protein 3 (PBP3) sequence, namely corresponding to YRIN. Complete genome analysis revealed that this E. coli isolate carried virulence factors (sitA, gad, hra, terC, traT, and cia) and many other non-β-lactam resistance determinants including rmtB, tet(A), dfrA17 (two copies), aadA1, aadA5 (two copies), sul1 (two copies), qacE (two copies), qepA, mdf(A), catA1, erm(B), mph(A), and qnrS1, being susceptible only to tigecycline, colistin and fosfomycin. In conclusion, we described here the phenotypic and genome characteristics of an extensively drug-resistant (XDR) E. coli ST361 being recognized as an emerging clone worldwide.


Introduction
Escherichia coli is considered as the most important pathogen for humans, receiving special attention in the microbiological world, since it causes a wide range of severe infections in humans and animals such as urinary tract infections and other communityacquired infections [1]. E. coli has a high capacity to accumulate resistance genes acquired mostly through horizontal gene transfer, leading to phenotypic resistance to antibiotics that are normally very effective against this bacterial species [2]. Difficult-to-treat infections caused by carbapenem-resistant and carbapenemase-producing E. coli has become an issue of utmost importance which is increasingly recognized in recent years, being associated with increased morbidity and mortality [3]. Five main carbapenemases are identified among enterobacterial species, including in E. coli clinical isolates, namely the Ambler class A KPC-type ß-lactamases, the class B ß-lactamases of the NDM, VIM, or IMP types, and the class D ß-lactamase OXA-48 and its derivatives [4]. During the last decade, clinical isolates co-producing multiple carbapenemases have been reported in different countries [5][6][7][8][9][10]. Considering that most NDM-type-producing Enterobacterales (including E. coli) are highly resistant both to most β-lactams and non-β-lactams antibiotics, the recently developed aztreonam/avibactam (ATM-AVI) combo offers an alternative therapeutic of choice [11]. However, the emergence of ATM-AVI resistance is being increasingly reported among NDM-producing E. coli clinical isolates [11,12]. On the other hand, cefiderocol (FDC), a novel siderophore cephalosporin, has been recently developed. FDC is one of the latest generations of commercialized antibiotics with excellent antibacterial activity against a large variety of Gram negatives including carbapenem-resistant Enterobacterales, using its so-called "Trojan horse" unique strategy [13,14]. In this study, we report an extensively drug-resistant E. coli clinical isolate carrying a large array of resistance genes leading to resistance to all β-lactams including FDC and most non-β-lactams.

Material and Methods
In November 2020, E. coli clinical strain N1606 was recovered from the human urine sample from a patient hospitalized in Switzerland. The isolate was first identified as E. coli using EnteroPluri-test (Liofilchem SRL, Roseto degli Abruzzi, Italy) that later was confirmed by whole genome sequencing (WGS). Two carbapenemase-encoding genes (bla NDM and bla KPC ) were detected using GeneXpert ® system (Cepheid, Sunnyvale, CA, USA), which is a molecular diagnostic platform. Antimicrobial susceptibility testing was then performed by using the disk diffusion method on Mueller-Hinton agar plates for selected antibiotics. Minimum inhibitory concentrations (MICs) were also determined using Etest strips (bioMérieux, La Balme-les-Grottes, France) on Mueller-Hinton agar plates at 37 • C except for aztreonam/avibactam (ATM-AVI), for which the MICs were determined using the broth microdilution (BMD) in cation-adjusted Mueller-Hinton broth (Bio-Rad, Marnesla-Coquette, France), and AVI was tested at a fixed concentration of 4 µg/mL. MICs of fosfomycin were determined using AD fosfomycin agar dilution test (Liofilchem, Italy) following the manufacturer's instructions. For FDC, MIC values were determined with the reference BMD method using iron-depleted cation-adjusted Mueller-Hinton (ID-CAMH) broth prepared following the protocol described by Hackel et al. [15]. The results were interpreted according to the latest EUCAST breakpoints (https://www.eucast.org/fileadmin/ src/media/PDFs/EUCAST_files/Breakpoint_tables/v_12.0_Breakpoint_Tables.pdf (accessed on 14 February 2022)) [16]. MICs of colistin were determined using broth microdilution in cation-adjusted Mueller-Hinton broth (Bio-Rad), and results were interpreted according to the EUCAST/CLSI joint guidelines (www.eucast.org, (accessed on 14 February 2022)). E. coli ATCC 25922 was used as control for all testing.
Mating-out assays using the filter-mating method were performed as described previously (6). The E. coli N1606 isolate was used as a donor, and the azide-resistant E. coli J53 strain was used as the recipient. Briefly, both donor and recipient strains were cultured separately in LB broth. After incubation, the donor and recipient strains were mixed at a ratio of 1:9 (donor/recipient) and centrifuged, the supernatant was removed, and the pellets were resuspended in 200 µL LB broth, which was plated on a conjugation filter on an LB agar plate. The plate was incubated for 5 h at 37 • C. Transconjugants were selected on LB agar supplemented with sodium azide (100 mg/L) and imipenem (4 mg/L) for bla NDM-5 carrying plasmid or cefoxitin 1 mg/L for bla KPC-3 carrying plasmid or gentamicin 50 mg/L and amikacin 50 mg/L for methylase. Successful transconjugants were confirmed by performing antimicrobial susceptibility testing and PCR targeting the genes of interest.
The entire genome of E. coli N1606 was sequenced using a combination of the MiSeq and MinION (Oxford Nanopore Technologies, Oxford, United Kingdom) platforms. Briefly, the total genomic DNA (gDNA) was extracted using a QIAamp DNA minikit and QIAcube (Qiagen) according to the manufacturer's instructions. For the MiSeq sequencing, a DNA library was constructed using the Nextera sample preparation with 2 × 150 bp pairedend reads (Illumina, San Diego, CA, USA) according to the manufacturer's instructions. MinION sequencing was performed using the MinION Mk1C (Oxford Nanopore Technologies, Oxford, UK), and sequencing libraries were prepared using a native barcoding kit (EXP-NBD104; Oxford Nanopore Technologies, UK) and 1D chemistry Ligation Sequencing Kit (SQK-LSK109; Oxford Nanopore Technologies) and performed on a R9.4.1 Flow Cell (FLO-MIN106; Oxford Nanopore Technologies). Assembly of both Illumina short reads and Nanopore long reads were performed using the CLC Genomic Workbench (version 20.0.4; CLC Bio, Aarhus, Denmark). The resulting assembled sequences were analyzed for antimicrobial resistance genes, multilocus sequence typing (MLST), Serotyping and fimH subtyping, plasmid incompatibility groups, and plasmid MLST using the Center for Genomic Epidemiology server (http://www.genomicepidemiology.org/, (accessed on 17 May 2022)). The complete genome sequence of E. coli N1606 strain and all plasmids had been deposited at GenBank under BioProject ID: PRJNA762038. The complete nucleotide sequences of chromosome (Chr1606) and plasmids (p1606A, p1606B, p1606C, p1606D, p1606E, p1606F, and p1606G) were deposited as GenBank accession numbers CP083701, CP083702, CP083703, CP083704, CP083705, CP083706, CP083707, and CP083708, respectively.
Mating-out assays were performed for the E. coli isolate N1606 with the aim to transfer the bla NDM-5 and bla KPC-3 genes. bla NDM-5 was transferred by conjugation, but bla KPC-3 was not, and the self-transmissible plasmid carrying bla NDM-5 was designated p1606A. By testing the bla NDM-5 positive transconjugant for antibiotic susceptibility, co-resistance to aminoglycosides (kanamycin, tobramycin, gentamicin, and amikacin) was observed, as was resistance to sulfonamides and tetracycline. Resistance to aminoglycosides was correlated with the acquisition of the 16S RNA methylase-encoding gene rmtB. PCR-based replicon typing (PBRT) analysis revealed that the bla NDM-5 gene was localized on an IncFII-FIA plasmid scaffold.
WGS analysis revealed that E. coli N1606 belonged to Sequence Type ST361 and to the O9:H30 serotype. The lineage in which E. coli ST361 belongs is not recognized so far as internationally widespread with very few reports either in human or animals, making it difficult to fully discuss and compare results from different epidemiological studies [22][23][24][25]. Very recently, an E. coli ST361 harboring the bla NDM-5 gene has been described from different human and animal sources in several European countries including Switzerland [26][27][28].
WGS-based analysis of the PBP3 sequence of E. coli isolate N1606 identified an insertion of four amino acids (YRIN) after residue 333 compared to the PBP3 sequence of the wildtype E. coli MG1655 reference strain (Genbank NC_000913.3). It was previously shown that four amino-acid insertions (such as YRIN or YRIK) are associated with increased resistance to ATM-AVI [12]. It is noteworthy that isogenic mutants possessing YIRNinserted PBP3 displayed a slight two-fold reduced susceptibility to cefiderocol (from 0.06 to 0.125 mg/L) [37], However, more studies are needed to investigate the real impact (if any) of such modifications in PBP3 protein, the primary target of FDC, on the antibacterial activity of this novel cephalosporin.
We investigated here several genes of specific interest considering they had been previously shown to be involved in reduced susceptibility to FDC, including the fiu and cirA iron-catecholate transporter encoding genes, and other iron transport-related genes (exbB, exbD, tonB3, baeS/R) [38,39]. The analyzed genes (fiu, cirA exbB, exbD, BaeS/R) showed a wild-type sequence. A single amino acid mutation (L133P) was identified in TonB3, a component of the TonB3-ExbB3/D3 complex, which is providing energy required for FDC transport and associated with iron acquisition [40]. Mutations in the tonB3 gene might impair the energy acquisition for FDC transport and iron availability, leading to reduced susceptibility to FDC as previously reported for other siderophore-conjugated antibiotics such as KP-736, BMS-180680, E-0702, pirazmonam, and U-78,608 [41].
To investigate whether the resistance to FDC observed in E. coli N1606 could be related to the production of acquired class A or C β-lactamases, MICs of FDC were determined in combination with avibactam (AVI) as an inhibitor of Ambler class A (including extended-spectrum β-lactamases (ESBLs), class C, and some class D ß-lactamases, including carbapenemases (e.g., KPC and OXA-48). Hence, MICs of FDC dropped from 64 to 4 mg/L when combined with AVI at a fixed concentration of 4 mg/L. This suggested that the high-level resistance to FDC in the tested isolate was likely due to the accumulation of a wide variety of different β-lactamases including carbapenemases, namely KPC-3, the ESBLs CTX-M-15 and SHV-11, and the AmpC enzyme CMY-145, as well as NDM-5 that had been previously shown to affect bacterial susceptibility toward FDC [42][43][44][45].

Conclusions
We described here the genome characteristics of an extensively drug-resistant (XDR) Escherichia coli ST361 isolate co-carrying the bla KPC-3 , bla NDM-5 and other resistance genes located on multiple plasmids. It is noteworthy that ST361 NDM-5-producing E. coli are increasingly identified worldwide, being either found in humans, the environment, or animals [28,[46][47][48]. The identification of such an XDR isolate may constitute a serious challenge for resistance control and the clinical treatment of related infections. The emergence of such an XDR phenotype in the E. coli strain due to the accumulation of many plasmid-associated MDR determinants can be explained by the bacterial potential to acquire additional resistance traits through mobile genetic elements such as plasmids by horizontal gene transfer not only among E. coli strains but also to other Enterobacterales. This case here further underlines that FOS-containing treatment may be an option for treating infections associated with such resistant strain [49].