The isolates in our study included two parts. Firstly, eleven non-repetitive CREC isolates collected from 6 hospitals during the period June-December 2015 in Nanjing city were analyzed for genomic characterization by whole genome sequencing. These strains were isolated from urine (n=4), sputum (n=3), blood (n=1), bile (n=1), secretion of uterus neck (n=1), and the source of one strain was unknown. The following hospitals were involved: Nanjing Mingji Hospital l (n=1), Nanjing Children’s Hospital (n=2), Nanjing Lishui Hospital (n=1), Nanjing Drum Tower Hospital (n=4), Nanjing jinyu Hospital (n=1), Nanjing Maternal and Child Health Hospital (n=1), and Nanjing First Hospital (n=1).
Secondly, considering the high prevalence of E. coli isolates co-producing KPC-2 and NDM among the 11 CREC strains from 6 hospitals, we tried to investigate the distribution of these strains among clinical CREC isolates. Therefore, a total of 43 consecutive non-duplicate isolates collected in Nanjing Drum Tower hospital during 2013-2017 were further analyzed for isolates co-producing KPC-2 and NDM by PCR and DNA sequencing. Among them, 4 strains were isolated in 2013, 10 in 2014, 2 in 2015, and 11 from 2016 and 16 in 2017. The source of the samples was as follows: urine (n=18), blood (n=9), sputum (n=6), secretion (n=3), bile (n=3), abdominal dropsy (n=2), and pus (n=1).
All the CREC strains were identified by Vitek 2.0 (BioMérieux. Marcy l'Etoile, France) or ATB 32E Semi-auto identification machine (Bio-Mérieux, France). Isolates resistant to at least one carbapenem (imipenem, meropenem, ertapenem) were included in the study.
Antimicrobial Susceptibility testing
Susceptibility of the 11 CREC isolates towards antimicrobial agents were tested by micro-broth method. The antimicrobial agents tested included ertapenem, imipenem, meropenem, cefepime, ceftazidime, cefotaxime, cefuroxime, cefazolin, piperacillin/tazobactam, amikacin, gentamicin, funantuoyin, trimethoprim and sulphame-thoxazole, aztreonam, piperacillin, ciprofloxacin, levofloxacin, aztreonam/avibactam, ceftazidime/avibactam, tigecycline, and colistin B. E. coli ATCC 25922 was used as quality control. The results were interpreted according to the CLSI 2019 guideline (16). Whereas, for tigecycline and colistin, the European Committee on Antimicrobial Susceptibility Testing breakpoints were referred (http://www.eucast.org/clinical_breakpoints).
The Ultraclean Microbial DNA Isolation Kit (MOBIO Laboratories, Carlsbad, CA, US) was used to extract genomic DNA. The NanoDrop 2000c spectrophotometer (Thermo Scientific, Waltham, MA, USA) was used for measuring the DNA concentration and purity for whole genome sequencing.
Whole Genome Sequencing, denovo Assembly, Scaffolding, and Annotation
The prepared pair-end DNA library was sequenced on the MiSeq (Illumina, SanDiego, CA, USA). Denovo assembly of the paired-end reads was performed by CLC Genomics Workbenchv7.0.4 (QIAGEN, Hilden, Germany) after quality trimming (Qs ≥ 20). Scaffolding was finished using SSPACE standard version 3.0 and the gaps within scaffolds were further closed by GapFiller (17, 18). Then genomes were then submitted to NCBI for annnotation.
Analysis of genomic epidemiology
The antimicrobial resistance determinants and virulence factors were identified using Resfinder v2.1 (http://cge.cbs.dtu.dk/services/ResFinder-2.1/) and Virulence Finder 2.0 (https://cge.cbs.dtu.dk/services/VirulenceFinder/); the 11 CREC were typed by multi-locus sequence typing (MLST) 2.0 (https://cge.cbs.dtu.dk/services/MLST/), Plasmid Finder 2.1 (https://cge.cbs.dtu.dk/services/PlasmidFinder/), Serotype Finder 2.0 (https://cge.cbs.dtu.dk/services/SerotypeFinder/), and FimTyper 1.0 (https://cge.cbs.dtu.dk/services/FimTyper/).
Phylogenetic relationship of 11 CREC
The core-genome phylogeny of the 11 CREC was constructed by using single-nucleotide polymorphisms (SNP)-sites detected from 1010 core genes (identity > 95%; coverage = 100% ) derived from 502 ST11 strains (19). A maximum-likelihood tree was calculated using RAxML version 8.2.8, with general time-reversible model and 100 bootstraps (20). Interactive Tree Of Life (https://itol.embl.de) was used to produce the phylogenetic tree (21).
The screening of Escherichia coli co-producing NDM and KPC-2 carbapenemases
In order to investigate the prevalence of E. coli isolates co-producing NDM-5 and KPC-2 in our hospital during 2013-2017, genes encoding carbapenemases (KPC and NDM) were detected by PCR and DNA sequencing (22). The positive products were sent to the Qingke Biotechnology Co., Ltd (Nanjing, China) for purification and sequencing. Sequences were further analyzed by using the Chromas-Pro application and BLAST (www.ncbi.nlm.nih.gov/BLAST).
Pulsed-field gel electrophoresis
Six E. coli isolates co-producing NDM-5 and KPC-2 including 3 ones from 11 CREC and 3 ones selected from the 43 CREC were further analyzed for genetic relatedness by PFGE, which was performed according to the protocol as previously described (23). Briefly, fresh colonies were mixed with proteinase K (Merck Sharp & Dohme Ltd, Germany) into plugs. After the plugs were digested by restriction endonuclease XbaI (Fermentas, ABI, Germany), the resultant DNA fragments were separated in a PFGE CHEF-DR III system (Bio-Rad Laboratories, Hercules, CA) in 0.5×Tris-borate-EDTA buffer at 120 V for 19 h. The pulse times ranged from 2.2 s to 54.2 s. Finally, the BioNumerics software (Applied Math, Sint-Maten-Latem, Belgium) was used to analyze the banding patterns.
For the 6 isolates co-carrying blaKPC and blaNDM, broth mating was performed in order to analyze the transferability of these genes according to the protocol prescribed previously(24). Azide resistant E. coli J53 was used as the recipient. Briefly, fresh colonies were inoculated into 5 ml LB broth and incubated at 37℃, 200 rpm. After 5 hours, 500 μl recipient cells and 100 ul donor were suspended in 5 ml LB broth for overnight culture at 37℃, 200 rpm, then 100 μl were plated onto the LB plates containing 30 mg/L cefoxitin and 100 mg/L sodium azide for E. coli J53. PCR (amplification for blaNDM and blaKPC) and Eric-PCR were used to verify conjugants.