INTRODUCTION
In China, extended-spectrum β-lactamase (ESBL) production has been increasingly prevalent in strains of Escherichia coli, the major aetiological agent of urinary tract infections (UTIs) [Reference Jean1]. Options for effective antibiotic treatment of infections, including UTIs, are limited owing to the frequent occurrence of expanded-spectrum cephalosporin-resistant and carbapenem-resistant, Gram-negative bacteria of the family Enterobacteriaceae [Reference Martirosov and Lodise2]. Use of older antibiotics such as fosfomycin has therefore been proposed as an alternative treatment of such infections [Reference Bergen3].
Fosfomycin is an organic phosphonate agent that inhibits cell wall synthesis by irreversibly inhibiting MurA, which is responsible for the initial step of peptidoglycan biosynthesis [Reference Falagas4]. Fosfomycin exhibits a broad spectrum of antimicrobial activity, including rapid bactericidal effects against several Gram-negative rods, particularly E. coli, and also has good activity against Staphylococcus aureus [Reference Falagas4]. Successful treatment of infections, especially UTIs, with fosfomycin has been documented in Japan [Reference Wachino5], and thus it is receiving renewed attention as an alternative agent for the treatment of UTIs caused by ESBL-producing E. coli [Reference Neuner6].
To date, fosfomycin resistance in E. coli has primarily involved either reduced uptake of the drug due to mutations in chromosomally encoded transporters [Reference Takahata7], or enzymatic inactivation by plasmid-mediated glutathione S-transferases (PMGST) such as FosA3, FosA4, and FosC2 [Reference Wachino5, Reference Chen8–Reference Ma11]. It has also been reported that fosfomycin-resistant isolates are more likely to be ESBL producers than fosfomycin-susceptible isolates [Reference Wachino5]. However, information on the prevalence of plasmid-mediated fosfomycin resistance genes in ESBL-producing urinary E. coli strains is lacking, with only limited knowledge of the molecular characteristics and prevalence of fosA3 and the ESBL genes bla CTX-M in strains in China.
The purpose of this study was to examine the occurrence of fosfomycin-resistant E. coli in ESBL-producing, urinary E. coli isolates, and to identify the distribution of PMGST and ESBL determinants. Furthermore, the genetic relatedness in fosA3-positive strains, transferability of fosA3, and replicon types of fosA3-carrying plasmids were analysed.
MATERIALS AND METHODS
Bacterial isolates
A total of 821 non-repetitive urinary E. coli isolates were streaked from the Strain Library of the Department of Laboratory Medicine, Nanjing Drum Tower Hospital. ESBL production was confirmed phenotypically, using both cefotaxime and ceftazidime alone or in combination with clavulanic acid. The susceptibility of strains to fosfomycin was tested by the disk diffusion method described previously [12], using Mueller–Hinton agar plates (Oxoid, UK) containing 25 mg/l glucose-6-phosphate (G6P). E. coli ATCC25922 was used as the quality control strain in antimicrobial susceptibility testing.
Detection of genes for fosfomycin resistance and ESBL production
Genes reported to be involved in fosfomycin resistance in Enterobacteriaceae, including fosA, fosB, fosC, and fosX, as well as the subtypes fosA1, fosA2, fosA3, fosA4, and fosC2, were detected by PCR and DNA sequencing analyses according to previously described protocols [Reference Wachino5, Reference Chen8–Reference Ma11]. The presence of bla genes for ESBL production (bla CTX-M, bla TEM, bla SHV) was assessed in each of the 57 ESBL-positive strains, following a previously described protocol [Reference Dallenne13].
Phylogenetic grouping
Phylogenetic grouping of fosfomycin-resistant E. coli isolates was conducted via triplex PCR, using six primers in a single reaction [Reference Clermont, Bonacorsi and Bingen14]. The amplification of three DNA markers (chuA, yjaA, TSPE4.C2) generated fragments of 279, 211, and 152 bp, respectively. This allowed E. coli isolates to be classified into the phylogenetic groups A, B1, B2, or D. E. coli strains ECOR 20 (yjaA positive), ECOR 48 (chuA positive), ECOR 58 (TSPE4.C2 positive), and ECOR 62 (chuA, yjaA, and TSPE4.C2 positive) were used as the positive controls, and E. coli strain ECOR 4 was used as the negative control. All controls were kindly provided by Statens Serum Institute, Denmark.
Genetic relatedness by pulsed-field gel electrophoresis (PFGE)
The fosA3-positive E. coli isolates were characterized by PFGE using the CHEF Mapper System (Bio-Rad Laboratories, USA) as described previously [Reference Ribot15]. Briefly, the chromosomal DNA of E. coli isolates was subjected to digestion with XbaI for 2 h at 37 °C. Electrophoresis was conducted at 6·0 V/cm and 14 °C for 19 h with an angle of 120°. The switch time was increased from 2·2 s to 54·2 s at a gradient of 6 V/cm. Salmonella enterica serovar Braenderup HP812 (kindly provided by the Centers for Disease Control and Prevention, USA) was used in parallel as a molecular weight standard. The results were analysed and interpreted using Bionumerics software v. 6.5 (Applied Maths, Belgium). The Dice similarity coefficient on the basis of the unweighted-pair group method using average linkages (UPGMA) with a 1·5% band tolerance was used. Furthermore, cut-off lines at 80% were used to analyse genetic relatedness.
Multi-locus sequence typing (MLST)
The fosA3-positive E. coli isolates were assessed for sequence types (STs) according to the MLST scheme developed for E. coli by the University College Cork (http://mlst.ucc.ie/mlst/dbs/Ecoli). Briefly, the housekeeping genes adk, fumC, gyrB, icd, mdh, purA, and recA were analysed using the primer sequences and amplification conditions available at http://mlst.warwick.ac.uk/mlst/.
Conjugation experiments
Conjugation experiments were performed using azide-resistant E. coli J53 as a recipient strain by the broth mating method. Trans-conjugants were selected on trypticase soy agar plates supplemented with 150 mg/l sodium azide, 40 mg/l fosfomycin, and 25 mg/l G6P. The presence of fos genes in phenotypically selected ESBL producers harbouring bla TEM, bla SHV, or bla CTX-M was assessed by PCR as described previously [Reference Wang16].
PCR-based replicon typing
DNA was extracted from 51 trans-conjugants, and main plasmid incompatibility groups, including F, FIA, FIB, FIC, HI1, HI2, I1-Ic, L/M, N, P, W, T, A/C, K, B/O, X, Y, and FII, were determined using the PCR-based replicon typing scheme, as described by Carattoli et al. [Reference Carattoli17].
Ethical statement
All procedures were performed in compliance with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008.
RESULTS
Susceptibility of ESBL-producing strains to fosfomycin
In total, 465 out of 821 E. coli isolates were found to be ESBL producers. Antimicrobial susceptibility testing revealed that fosfomycin exhibited good antibacterial activity towards ESBL-producing urinary E. coli strains, demonstrating effectiveness against 87·7% (408/465) of strains. The average fosfomycin resistance rate of ESBL-producing E. coli associated with UTIs was about 10% over the 5 years (2010–2014).
Prevalence of plasmid-mediated fosfomycin resistance genes and ESBL genes in fosfomycin-resistant E. coli
Molecular analysis showed that 89·5% (51/57) of the fosfomycin-resistant isolates were positive for fosA3, whereas only one was fosA1-positive; other fosfomycin resistance determinants were not identified. Fifty-five isolates were also bla CTX-M-positive, 26 harboured bla CTX-M-15, 22 bla CTX-M-14, 4 bla CTX-M-3, and three harboured bla CTX-M-123 (Fig. 1). In addition, 17 isolates carried bla TEM variants (14 bla TEM-104 and three bla TEM-1b) and 13 carried bla SHV variants (nine bla SHV-12 and four bla SHV-11) (Fig. 1).
Fig. 1. Dendrogram of PFGE profiles of 50 fosA3-carrying E. coli isolates, phylogentic groups (PGs), extended-spectrum β-lactamases (ESBLs), and sequence types (STs).
Phylogenetic groups
Of the 57 fosfomycin-resistant isolates in phylogenetic groups, 19 were classified as group D, 18 group A, 12 group B1, and eight group B2.
Genetic relatedness of fosA3-positive isolates determined by PFGE and MLST
The 50 strains harbouring fosA3 exhibited 44 different PFGE profiles and one strain was not typable. MLST revealed 37 STs and major STs were ST410 (n = 4 strains), ST10 (n = 4), ST405 (n = 3), ST156 (n = 3), and ST964 (n = 3), which together comprised 33·3% of the strains analysed. Similar or identical PFGE profiles were observed within ST10 clones, ST156, ST354, ST405, ST964, and ST2309. This level of genetic diversity indicates that most of the fosA3-carrying isolates were clonally unrelated (Fig. 1).
Transferability and replicon typing of fosA3 plasmids
Conjugation assays revealed that the fosA3 genes were transmissible. Moreover, bla CTX-M and bla TEM genes were able to be transferred simultaneously, indicating genetic linkage between fosA3 and bla CTX-M. Plasmids carrying fosA3 from 50 isolates were successfully transferred by conjugation. These 50 plasmids consisted of 39 that were replicon type IncFII, nine that were IncI1, four that were IncN, two that were IncA/C, and one that was IncP. In addition, plasmids from four isolates were fused, containing both the IncN and IncFII replication origins.
DISCUSSION
Fosfomycin has been extensively used in several European countries since 1988 for the treatment of uncomplicated UTIs [Reference Falagas4], but it was not approved for clinical use in China until recently. This is the first investigation of the prevalence of fosfomycin resistance (fos) genes in ESBL-producing urinary E. coli isolates in mainland China.
The strains in our study displayed a rate of resistance to fosfomycin of about 10%, which is higher than that previously reported [Reference Lai18]. However, our data indicate that fosfomycin should still be considered for the treatment of patients with infections due to ESBL-producing E. coli in China if they exhibit high resistance rates to other commonly used antimicrobial agents, including cephalosporins and fluoroquinolones [Reference Wang19]. This is because a previous study reported that fosfomycin retains its activity against both Gram-positive and Gram-negative multiple-drug-resistant (MDR) and extremely-drug-resistant (XDR) bacteria [Reference Falagas4]. To date, fosfomycin has not been used for clinical treatment in our hospital, and so we speculate that the observed resistance may be co-selected by antimicrobials other than fosfomycin.
Our study found that bla CTX-M-15 and bla CTX-M-14 were the main ESBL-encoding genes detected in fosfomycin-resistant urinary E. coli strains. This is in line with the results of previous studies investigating the global prevalence of bla genes in ESBL producers [Reference Ghafourian20]. It should be noted that we also detected bla CTX-M-123, which has been identified as a novel hybrid of the bla CTX-M-1 and bla CTX-M-9 β-lactamases recovered from E. coli isolates in China [Reference He21]. In parallel, there was a high prevalence of the bla TEM-104 variant in our study which to the best of our knowledge, has been identified in a MDR avian pathogenic E. coli strain isolated from septicaemic broilers in Egypt [Reference Ahmed, Shimamoto and Shimamoto22]. This is therefore the first report of TEM-104 variants in clinical urinary E. coli isolates in China.
It has been previously reported that fosA3 is the most prevalent PMGST in E. coli isolates of both clinical and non-clinical (healthy persons, companion and food animals) origins in several Asian countries (China, South Korea, Japan) [Reference Wachino5, Reference Sato23–Reference Cao27]. Thus, the high prevalence of fosA3 found here is consistent with these reports, and confirms that fosA3 is the primary mechanism of fosfomycin resistance in mainland China. Moreover, all but one of the 51 fosA3-positive isolates in our study were CTX-M producers, suggesting a high degree of association between the two resistance determinants. Indeed, the high transferability of these two genes via plasmids with identical replicon types further indicates that the two genes may be simultaneously disseminated by plasmids [Reference Sato23, Reference Hou26]. The implication of this is that there is a high risk for their widespread dissemination and suggests a critical need for close monitoring of such strains.
UTI-causing E. coli isolates have been closely associated with phylogroups D and B2 in China [Reference Cao27]. D was the main phylogroup in the ESBL-producing urinary E. coli isolates in this study, consistent with previous reports [Reference Cao27, Reference Navidinia28]. This indicates that group D may contribute more to MDR and UTI infections in China than other phylogroups. Phylogroups A and B1, however, were more common in our study than B2. Since phylogroups A and B1 have been reported in animal or human commensal E. coli strains [Reference Jakobsen, Hammerum and Frimodt-Møller29, Reference López-Cerero30], this provides evidence that animals may be the source of some UTI-causing E. coli isolates [Reference Osugui31].
Clonal diversity in fosA3- and bla CTX-M-harbouring E. coli from humans, as revealed by both PFGE and MLST, indicates that the spread of fosA3 in ESBL-producing E. coli is not attributable to clonal transfer of FosA3 producers in patients. In addition, MLST results suggest that several clonal strains involved in the dissemination of bla CTX-M-positive E. coli, such as ST450, also carry fosA3 [Reference Sato23]. IncFII, IncI1, and IncN plasmids carrying fosA3 as well as bla CTX-M β-lactamase genes have previously been reported in E. coli from chickens, pets, livestock, and other animals in China [Reference Hou26, Reference Yang32, Reference Ho33]. Furthermore, fosA3 and bla KPC-2 genes were found to be able to spread together worldwide through IncP plasmid transfer [Reference Jiang34, Reference Li35]. The high transferability of plasmids carrying fosA3 and multiple replicons found here provide further evidence of the high potential for transfer of fosfomycin resistance gene fosA3. Recently, fosA3 has been found on an epidemic plasmid carrying bla CTX-M-65 and rmtB [Reference He36], Of particular concern, the gene has also been identified on a novel IncR-F33:A-:B- plasmid harbouring bla KPC-2, bla CTX-M-65, bla SHV-12, and rmtB that was isolated from an epidemic Klebsiella pneumoniae ST11 strain in China [Reference Xiang37]. Therefore, close monitoring and continued surveillance of patterns of fosfomycin resistance are necessary in order to prevent further dissemination of fosA3 genes.
One limitation of this study stems from the fact that only ESBL producers in the 821 E. coli urinary isolates were screened for fosfomycin-resistance genes based on the strong association between the presence of fosA3 and that of bla CTX [Reference Sato23, Reference Hou26]. However, the high prevalence of fosA3 demonstrates the rapid spread of fosfomycin resistance in this region.
In summary, the high prevalence of fosfomycin resistance observed in ESBL-producing urinary E. coli isolates recovered during 2010–2014 is mainly attributed to the widespread occurrence of plasmid-mediated fosA3 genes. The dissemination of the fosA3 gene is closely associated with that of bla CTX-M. Rather than clonal expansion of fosA3-harbouring E. coli lineages, horizontal transfer of plasmid-mediated mobile elements carrying fosA3 played a central role in the spread of E. coli harbouring both fosA3 and bla CTX-M in our hospital.
ACKNOWLEDGEMENTS
The authors thank the EU Reference Laboratory for antimicrobial resistance and the National Food Institute of Technical University of Denmark for providing positive controls for our experiments. This study was supported by the Youth Fund of Jiangsu Province (grant no. BK20140099) and The Fundamental Research Funds for the Central Universities (grant no. 021414340283).
DECLARATION OF INTEREST
None.
