Occurrence of bla_KPC-2, bla_CTX-M, and mcr-1 in Enterobacteriaceae from Well Water in Rural China

ABSTRACT

We report on the coexistence of mcr-1 and bla_CTX-M in multidrug-resistant, extended-spectrum β-lactamase-producing Escherichia coli belonging to the sequence type 10 complex isolated from well water in rural China. Raoultella ornithinolytica with bla_KPC-2 was also detected in well water from the same area. This study shows that genes coding for resistance to last-resort antibiotics are present in wells in rural China, indicating a potential source of antibiotic resistance.

TEXT

Extended-spectrum β-lactamase (ESBL)- and carbapenemase-producing members of the family Enterobacteriaceae such as Escherichia coli have spread worldwide both as commensal bacteria and as causative agents of infection in humans and animals (1–3). Carbapenems and polymyxins are regarded as last-resort antibiotics for the treatment of severe infections in humans caused by multidrug-resistant members of the family Enterobacteriaceae (4, 5). However, this paradigm was challenged when the plasmid-mediated colistin resistance gene mcr-1 was reported in members of the family Enterobacteriaceae isolated from humans and food animals (6). E. coli with both mcr-1 and the EBSL-encoding gene bla_CTX-M (7–9) and the carbapenemase-encoding gene bla_NDM (10, 11) have been isolated from both humans and animals, indicating that resistance to these last-resort antibiotics is emerging. The coexistence of mcr-1 and bla_CTX-M in E. coli has also been reported in surface waters such as rivers and ponds in Switzerland and Malaysia (9, 12).

In rural areas of China, well water is the primary source of irrigation water and drinking water for humans and farm animals. In this study, 71 wells were sampled in 12 rural villages of county A (n = 43) and 12 villages in county B (n = 28) in the Shandong province of China during July and September 2015 (Fig. 1). There was a long history of pig breeding in county A and a long history of chicken manure application for intensive vegetable cultivation in county B. About 1,000 ml of well water was collected in sterile bottles and filtered through sterile membrane filters with a pore size of 0.45 μm (Millipore, Billerica, MA). The filters were aseptically put into ESwab tubes (Copan, Brescia, Italy). Each sample was subsequently cultured on chromID ESBL, chromID CARBA, and chromID OXA-48 agar plates (bioMérieux, Marcy l'Etoile, France) for 18 to 24 h at 37°C to identify presumptive ESBL- and/or carbapenemase-producing members of the family Enterobacteriaceae. Identification to the species level was obtained by matrix-assisted laser desorption ionization-time of flight mass spectrometry. Antimicrobial susceptibility profiles were determined by the agar dilution method recommended by the Clinical and Laboratory Standards Institute (CLSI, 2016) and the EUCAST criteria (http://www.eucast.org/clinical_breakpoints/). The mcr-1, bla_CTX-M, bla_NDM, bla_KPC, bla_VIM, and bla_OXA-48 genes were screened for by PCR (6, 13, 14). PCR products were verified by sequencing (BGI Company, Beijing, China), and amplicons with at least 99% similarity to the mcr-1 amplicon were considered mcr-1. Multilocus sequence typing (MLST) was performed according to the scheme at http://mlst.warwick.ac.uk/mlst/mlst/dbs/Ecoli.

FIG 1.

The genotypes of E. coli and R. ornithinolytica isolates from agricultural counties A and B. DEM, digital elevation model.

Ten ESBL-producing E. coli isolates were detected in a total of 71 samples, all of which came from county A (10/43, 23.3%) (Table 1). The mcr-1 gene was identified in two isolates from two different samples (ECcz1 and ECdw2) that also carried the bla_CTX-M-14 and bla_CTX-M-65 genes, respectively. Antibiotic susceptibility testing revealed that both strains exhibited multidrug-resistant phenotypes (resistance to more than two antibiotic classes), but both were susceptible to meropenem, amikacin, and tigecycline (Table 1). No carbapenemase genes were detected in the E. coli isolates. MLST showed that isolate ECdw2 belonged to sequence type 10 (ST10) and isolate ECcz1 belonged to ST48 (Table 1). ST48 is part of the ST10 clonal complex and differs from ST10 by only a single nucleotide.

TABLE 1.

Characteristics of ESBL-producing E. coli with mcr-1 and carbapenemase-producing R. ornithinolytica isolated from well water

Strain	Deptha (m)	MIC (mg/liter)b															ESBL/carbapenemase type	mcr-1d	MLST
		CL	AMC	TZP	CTX	CAZ	MEMc	GEN	AMK	TE	TGC	CIP	SXT	FOF	F	FFC
ECaz1	5.0	1	16	32	>8	8	0.06	4	4	64	0.5	0.25	>16	>256	32	4	CTX-M-55	−	ST2973
ECaj4	4.5	2	16	32	>8	4	0.06	4	4	64	0.25	0.25	>16	>256	32	4	CTX-M-65	−	ST2526
ECaw1	4.5	1	16	256	>8	1	0.06	2	2	64	0.25	2	>16	128	>256	>32	CTX-M-65	−	ST2526
ECbw1	20.0	1	16	32	>8	32	0.06	2	4	64	0.5	0.25	>16	128	128	4	CTX-M-55	−	ST2973
ECbj3	5.0	1	16	32	>8	8	0.03	>32	>128	>128	2	>16	>16	256	32	>32	CTX-M-65	−	ST1642
ECcz1	7.0	16	16	8	>8	8	0.06	>32	4	64	0.5	16	>16	>256	128	>32	CTX-M-14	+	ST48
ECdw2	8.0	16	16	32	>8	16	0.06	>32	8	64	0.5	>16	>16	64	256	>32	CTX-M-65	+	ST10
ECdw3	30.0	1	16	8	>8	1	0.03	2	2	128	0.25	2	>16	8	256	>32	CTX-M-27	−	ST162
ECej3	5.0	2	8	≤2	>8	2	0.06	2	4	64	0.5	8	>16	>256	32	4	CTX-M-65	−	ST2309
ECiw1	20.0	2	4	64	>8	2	0.5	2	4	128	0.5	1	0.5	>256	>256	>32	CTX-M-55	−	ST4762
ROfj3	6.0	1	8	32	1	16	16	2	4	128	0.5	0.5	16	>256	32	8	KPC-2	−	NAe

^aDepth of the well from which the water sample was taken.

^bMICs were determined by agar dilution methods. Resistance is indicated by boldface type. Abbreviations: CL, colistin; AMC, amoxicillin-clavulanate; TZP, piperacillin-tazobactam; CTX, cefotaxime; CAZ, ceftazidime; MEM, meropenem; GEN, gentamicin; AMK, amikacin; TE, tetracycline; TGC, tigecycline; CIP, ciprofloxacin; SXT, trimethoprim-sulfamethoxazole; FOF, fosfomycin; F, nitrofurantoin; FFC, florfenicol. Antimicrobial susceptibility profiles were determined by the agar dilution method recommended by the EUCAST (http://www.eucast.org/clinical_breakpoints/), except for tetracycline and florfenicol, for which the Clinical and Laboratory Standards Institute criteria (CLSI, 2016) were applied.

^cThe meropenem epidemiological cutoff for E. coli is ≤0.125 mg/liter. According to EUCAST, a MIC of >0.125 mg/liter is used to detect suspected carbapenemase-producing isolates (http://www.eucast.org/mic_distributions_and_ecoffs/).

^dSymbols: +, present; −, absent.

^eNA, not available.

No ESBL-producing E. coli strains were detected in well water samples from county B. However, there are some essential differences between counties A and B that could have influenced the occurrence of resistant bacteria and ESBL-producing E. coli. The well depth in county B (40 to 80 m) was greater than that in county A (4.5 to 30 m), suggesting that depth and temperature can influence the growth or dissemination of resistant bacteria. Shallower wells are also more likely to be influenced by anthropogenic activity. There is also a higher intensity of livestock farming in county A, particularly pig farms, which could have increased the risk of contaminating the shallower wells. Furthermore, in county B, a larger part of the well water is used for irrigation; thus, the wells have a shorter hydraulic retention time, which might influence the dissemination of resistant bacteria.

The two E. coli strains that harbored mcr-1 and bla_CTX-M genes in our study belonged to the ST10 complex, which is the most common ST of E. coli detected in human feces in southeastern China (15). The CTX-M ESBL has been reported as the dominant ESBL family in China, with bla_CTX-M-14 being the major variant detected in community onset infections in county hospitals across China (16). bla_CTX-M-14 is also frequently detected in farm animals, including pigs, and in pets in China. Although bla_CTX-M-65 does occur in both humans and pets, it appears to be more common in pigs and broiler chickens (17–20). The other bla_CTX-M genes identified in this study (Table 1) are also frequently detected in humans and animals, especially bla_CTX-M-55 among clinical isolates. These findings demonstrate the wide spread of bla_CTX-M among E. coli isolates from humans, animals, and the environment in China.

Multidrug-resistant E. coli with carbapenemase genes and mcr-1 has previously been reported in clinical samples and constitutes a health risk (10, 21). While none of the E. coli isolates carrying mcr-1 in this study were carbapenemase producers, an isolate of Raoultella ornithinolytica carrying the carbapenemase gene bla_KPC-2 was also detected in a well water sample. R. ornithinolytica is a member of the family Enterobacteriaceae, and bla_KPC-2-carrying strains of this species have previously been reported in clinical samples in China (22). Although the isolate was not detected in the same wells as the E. coli isolates carrying mcr-1, its presence in the same area and its close relatedness to E. coli indicate a potential risk of co-occurrence and transfer of resistance genes. In rural areas, well water is essential in daily life for activities such as tooth brushing and bathing and as drinking water for humans and pigs. Such activities could potentially increase the risk of dissemination of antibiotic-resistant bacteria from the environment to humans and animals.

As only two mcr-1-positive isolates were detected and well water was the only environmental matrix investigated in this study, generalization of our findings should be done with caution. A limitation of this study is that phenotypic screening for ESBL- and/or carbapenemase-producing members of the family Enterobacteriaceae was performed and thus mcr-1-carrying isolates without either phenotype would have remained undetected. From a One Health perspective, further studies involving more samples from humans, animals, and different environmental matrices are urgently needed to study the dissemination dynamics of mcr-1 and other resistance genes.