Introduction

Urinary tract infections (UTIs) are the most common bacterial infection and have a significant economic impact on the health care system. About 70–95% of community-acquired, and 50% of nosocomial UTIs are caused by E. coli1. There is increasing evidence that epidemic clonal groups of E. coli are responsible for a rise in antimicrobial-resistance in UTIs. For example, the prevalence of extended spectrum beta lactamase (ESBL)-producing E. coli (ESBL-EC) sequence type (ST131) is increasing worldwide, accounting for 25% and 54% of all ESBLs in Mexico and the United States, respectively2,3. Moreover, the emergence of drug-resistant epidemic plasmids, that confer antimicrobial resistance among bacteria, also contributes to increased resistance, through the horizontal transmission of ESBL-encoding genes. IncFII, IncA/C, IncL/M, IncN and IncI1 plasmids, carrying blaCTX-M genes and acquired ampC genes, are frequently detected in Enterobacteriaceae of different origin and source4. Specifically, the IncFII-like type plasmids have been reported to be closely associated with E. coli ST131 carrying blaCTX-M-15, while the IncI1 plasmids carrying blaCTX-M-15 or blaCMY-2 are frequently isolated from non-ST131 E. coli strains, some of which are shared between humans and animals5,6.

In China, ST131 was the most abundant type (12.7%) in 256 ESBL-producer isolates based on a study on ESBL-EC isolates from patients with community-onset infections in Chinese country hospitals from 2010 to 20117, and its prevalence increased to 29.5% in another study on bloodstream infections patients in three hospitals in China from 2013 to 20148. However, E. coli isolates from UTIs show a significantly different drug resistant profile than those from Europe, with more than 50% of isolates from China estimated to be resistant to both ciprofloxacin and cefotaxime by the year 2000. Since 1998, the predominant ESBL type in E. coli in China has been CTX-M-14, which was detected in 71% of the ESBL-ECs in 2008http://pubmlst.org/plasmid/), with IncI1/ST16 the predominant subtype (23/34, 68%), with size of 97 kb. The remainder were ST15 and ST23 (one isolate each), ST136 (n = 2), and 7 were untypable. Of the seven untypable plasmids, a novel combination of alleles (repI1, 1; ardA, 5; trbA, 10; sogS, 8; pilL, 1) was detected in five plasmids, and another novel combination (repI1, 1; ardA, 2; trbA, 8; sogS,4; pilL, 10) was detected in one plasmid, while three alleles of the remaining one plasmid could not be typed. Further ty** of the lncF plasmids (10/45, 22%) revealed a variety of subgroups: F18:A-:B1(n = 4,145 kb); F-:A1:B1(n = 1); F-:A-:B20(n = 1); F51:A-:B10(n = 1); and three were untypable.

Characterization of the antibiotic resistance genes and genetic environment of blaCTX-M-55 revealed a variable genetic context of the IncI1 plasmid. Except for six IncI1 plasmids which contained only blaCTX-M-55 gene, the remaining plasmids (n = 17) contained at least two other antibiotic resistance genes. The blaTEM-1 gene and aac (3′)-II gene were detected in 65% and 44% of IncI1 plasmids, respectively(Table 1). Other antibiotic resistance genes were variably found (data not shown) while the blaSHV gene was not detected.

PCR identified the insertion sequence ISEcp1 or IS26 in its entirety or partially truncated upstream of the blaCTX-M-55 gene in 32 of 45 (71%) transconjugants. Nucleotide sequence analysis of the upstream region of blaCTX-M-55 genes revealed that the spacer region between the right inverted repeat (IRR) and the start codon of blaCTX-M-55 genes were 45 bp or 48 bp.

To understand if the IncI1/ST16 type plasmid was a possible epidemic plasmid, six of the IncI1/ST16 plasmids were selected randomly to determine their complete nucleotide sequences. The six plasmids consisted of a large backbone with considerable homology to the pEK204 plasmid isolated from E. coli in the UK, except for 8 regions missing in some of the plasmids compared to the reference. As shown in Fig. 1, BRIG output image of pseudo-plasmids compared to E. coli bla-CTX-M-15 plasmid pEK204, the innermost rings show GC content (black) and GC skew (purple/green). The remaining rings show BLAST comparisons of the 6 plasmids against the reference plasmid pEK204. The numbers 1–8 highlight the regions missing in some of the plasmids compared to the reference.

Figure 1
figure 1

BRIG output image of pseudo-plasmids compared to E.coli blaCTX-15 plasmid pEK204.

Full size image

Discussion

The emergence of ESBL-E.coli strains carrying blaCTX-M-55 gene across China poses a great threat to the health care systems, as these strains have been shown to cause UTIs more frequently than those carrying blaCTX-M-15 gene17. The IncI1 plasmids often carry a variety of β-lactamase antibiotic resistant genes such as blaCMY-2, blaCTX-M-1 and blaSHV-126,18. Our present findings also show that IncI1 plasmids can also carry resistance determining genes of other antibiotic classes, such as blaTEM-1, aac(3′)-II, aac-(6′)-Ib, qnrA and qnrB, which may contribute to their long-term existence in the intestinal flora of humans and animals, and thus become a reservoir for continual transmission of bacterial strains carrying these genes.

In order to understand the epidemiology of the IncI1 plasmids, Garcia and colleagues established an IncI1 plasmid MLST database (http://pubmlst.org/plasmid/), in which 158 subtypes have been registered. Among these subtypes, the blaCTX-M-1/IncI1/ST3, blaCTX-M-1/IncI1/ST7 and blaCMY-2/IncI1/ST2, are the frequently reported subtypes for ESBL-EC isolates of animal origin (such as poultry and pets) in Europe and America6,18,19. IncI1 plasmids carrying the blaCTX-M-55 were reported in ESBL-EC isolates of animal origin (dogs, chickens, ducks, etc.) from surveillance studies in Southern China, but unfortunately the MLST subtypes of these plasmids remain unknown13.

Contrary to these previous studies, the IncI1/ST16 subtype was the predominant subtype in the present study, constituting 68% of IncI1 carrying the blaCTX-M-55 gene.

We also identified another five IncI1/ST16 closely related plasmids that might be IncI1/ST16 variants due to one single allele gene mutation (pilL6 → 1), compared to the sequence of IncI1/ST16. The IncI1/ST16 subtype was detected first in the British bovine E. coli isolate (pEK204), and was subsequently reported in a French dog E. coli isolate18,20. Complete sequencing of the six IncI1/ST16 plasmids revealed that they consisted of a similar large backbone with considerable homology to the pEK204 plasmid, except for 8 regions missing in some of the plasmids compared to the reference, suggesting a possible similar origin.

Furthermore, five pairs of the IncI1/ST16 alleles in the present study(repI1, ardA, trbA, sogS and pill), are different with those of the IncI1/ST3 subtype, and four pairs of the IncI1/ST16 alleles (ardA, trbA, sogS and pill) are different to those of the IncI1/ST2 subtype, suggesting a lower genetic relationship between IncI1/ST16 and these two commonly described subtypes. Notably, the IncI2 (pHN1122–1) subtype plasmids that are commonly found in strains of animal origin in China and harbor blaCTX-M-55 were not detected in this study13. The IncI2 and IncI1 plasmids have an identical backbone gene encoding fimbriae that play an important role in engaging the plasmid and metastasis21. So far, the cause of high incidence of IncI1/ST16 carrying blaCTX-M-55 from the ESBL isolates remains unclear, and the relationship between IncI1/ST16 and other ESBL genotypes in clinical and animal-derived strains remains to be further studied.

Another significant plasmid subtype carrying the blaCTX-M-55 in our study was the IncF that accounted for 22%(10/45) of all blaCTX-M-55 carrying plasmids. IncF plasmids harboring blaCTX-M-15 are frequently detected among Enterobacteriaceae worldwide, and the acquisition of these plasmids is known to contribute to the dissemination of antimicrobial resistance and virulence genes22. There are currently 35 subtypes of IncF reported in plasmid MLST database. In the present study, we detected the IncF subtype F18: A-: B1, which is characterized by iron uptake (missing eitABCD), haemagglutinin and serum survival gene, and was firstly reported in the Avian Pathogenic E. coli (APEC)23. Other studies from Tunisia and China also reported high prevalence of the F18: A-: B1 subtype in E. coli strains of animal origin, and harbored the oqxAB or blaCTX-M-1 resistance genes24,25. These findings again suggest the possibility of horizontal transfer of the F18: A-: B1 plasmids between humans and animals.

An important feature of the ESBL-carrying plasmids is the existence of insertion fragments such as ISEcp1, IS26 and ISCR in these plasmids. For example, the ISEcp1 gene, often located at the 48 bp upstream of the blaCTX-M-55 gene, can be used to determine the homology of different CTX-M-ESBL genes26. Furthermore, the ISEcp1 gene was reported to be responsible for the mobility and expression of the blaCTX-M gene. Qu and colleagues reported that the insertion of ISEcp1 gene at 45 bp, 48 bp and 127 bp upstream of the blaCTX-M-55 gene, induces the expression of this gene in plasmids27. In this study, 94% of IncI1 plasmids contained the ISEcp1 or the IS26 insertion at 45 bp or 48 bp upstream of the blaCTX-M-55 gene, which suggests that acquisition of the blaCTX-M-55 gene by the Inc I1 plasmids might be mediated by the insertion fragments.

There are several limitations in our study. Firstly, only UTI ESBL-EC strains were analyzed; it is unclear whether the blaCTX-M-55 gene carrying ESBL-EC strains from other specimens such as blood belong to ST1193 and have IncI1 plasmids. Secondly, due to limited time and financial resources, a limited number (58 of 111, 52%) of isolates were used in conjugation studies and hence in plasmid analysis, which introduces some bias to our study. It’s arguable that the untested isolates would have yielded different results. Thirdly, to the best of our knowledge, this is the first report on the molecular epidemiology of the blaCTX-M-55-carrying major clone ST1193 and epidemic plasmid IncI1F/ST16 and F18: A-: B1. A whole genome analysis would greatly facilitate our understanding of the prevalence of the ST1193 and epidemic plasmid IncI1F/ST16 and F18: A-: B1 among the ESBL-EC strains. And finally, although blaCTX-M-55-carrying strains have been isolated from animal samples, there is insufficient evidence to conclude that there is dissemination of the blaCTX-M-55 gene between humans and animals.

In conclusion, the high prevalence of the blaCTX-M-55 gene among E. coli isolates from UTI patients was mainly attributed to the clonal dissemination of ST1193 strains, carrying the Inc I1/ST16 plasmids and mobile genetic element in China. This information can facilitate monitoring and controlling dissemination of the blaCTX-M-55 gene carrying strains so as to limit the rise in antibiotic resistance, which is a public health concern.

Materials and Methods

Ethics statement

This study protocol was approved by the Ethics Committee of The First Affiliated Hospital of Guangzhou Medical University. All subjects signed written informed consent prior to the study. Patient information was anonymized and de-identified prior to analysis. All methods were carried out in accordance with relevant guidelines and regulations that are stated in the methods below.

Bacterial isolate collection

In our previous study, 137 out of 660 clinical ESBL-EC isolates from cases of UTI were shown to harbor blaCTX-M-55 gene. Of these137 isolates, only those that harbored a single blaCTX-M-55 gene (n = 111), were included in the current study. The isolates were collected from 21 general hospitals from 19 cities in China. These hospitals are distributed in seven geographic regions across China: four in Eastern China, two in Southern China, five in Northern China, two in Central China, two in Northeast China, two in Southwest, and four in Northwest. Clinical information of patients from whom the isolates were obtained, was reviewed retrospectively, including age, gender, underlying diseases, site and type of infection. All the bacterial isolates were re-identified using a VITEK2 Compact automated microbial identification system (bioMérieux, France). E. coli ATCC 25922 was used as the control strain.

Antimicrobial susceptibility

The minimum inhibitory concentrations (MICs) of common antibiotics (National Institutes for Food and Drug Control, Bei**g, China) against ESBL-producing E. coli were determined using the broth microdilution method with E. coli ATCC 25922 as the control strain. Interpretation of results followed the criteria of CLSI 201328. The interpretation breakpoint of cefoperazone/Sulbactam (Pfizer Inc., USA) was extrapolated from that of cefoperazone, and those of tigecycline and colistin (Pfizer Inc.) were based on the criteria of the European Committee on Antimicrobial Susceptibility Testing (EUCAST-2012).

Pulsed-field gel electrophoresis (PFGE)

PFGE analysis was conducted on the 111 isolates following a standardized protocol29, with electrophoresis performed using a CHEF DRIII System (Bio-Rad). The results were analysed by BioNumerics software (Applied Maths, St-Martens-Latern, Belgium) using the Dice similarity coefficient.

Multilocus Sequence Ty** (MLST)

MLST was carried out as previously described30. Different sequences of a given locus were assigned an allele number based on E. coli MLST database (http://mlst.warwick.ac.uk/mlst/dbs/Ecoli). Sequence types (STs) were obtained via the electronic database at the above website. The allelic profiles and defined clonal complexes (CCs) of all STs were further analyzed with the Based upon Related Sequence Types (BURST) clustering algorithm (eburst.mlst.net).

Conjugation experiments and plasmid analysis

Conjugation experiments were performed on 16 of 20 ST1193 isolates, and a random selection (n = 42), of other isolates, making sure that at least one to three isolates from each hospital was included. Transferability of blaCTX-M-55 genes was determined by conjugation experiments using rifampin -resistant E. coli C600 as the recipient strain as described previously31. Transconjugants were selected on Mueller-Hinton agar plates containing rifampin (300 mg/L) and ceftazidime (2 mg/L). Incompatibility (Inc) groups, antibiotic resistance genes, and genetic environment were detected as described previously32,33. The sizes of plasmids of transconjugants were determined using S1-treated genomic DNA followed by PFGE (S1-PFGE). The epidemiological relatedness of IncI and IncF plasmids was determined by MLST and referred to the database (http://pubmlst.org/plasmid/)23,34.

Complete nucleotide sequencing of the epidemic plasmids

A random selection of the same ST Inc I1 plasmids (n = 6) was sequenced by next generation sequencing. Total DNA (includes genome and plasmid) from the transconjugants was sequenced using Illumina MiSeq 300 PE platform.The sequence reads were de novo assembled using CLC genomics workbench version 8.0 (Qiagen).

Additional Information

How to cite this article: **a, L. et al. Prevalence of ST1193 clone and IncI1/ST16 plasmid in E-coli isolates carrying blaCTX-M-55 gene from urinary tract infections patients in China. Sci. Rep. 7, 44866; doi: 10.1038/srep44866 (2017).

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.